About this PDF file: This new digital representation of the original work has been recomposed from

XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
the authoritative version for attribution.

WASHINGTON, D.C. 1999

NATIONAL ACADEMY OF SCIENCES
COLLOQUIUM ON
NEUROBIOLOGY OF PAIN
i
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

COLLOQUIUM SERIES ii

NATIONAL ACADEMY OF SCIENCES

Colloquium Series

In 1991, the National Academy of Sciences inaugurated a series of scientific colloquia, five or six of which are scheduled each year
under the guidance of the NAS Councils Committee on Scientific Programs. Each colloquium addresses a scientific topic of broad and
topical interest, cutting across two or more of the traditional disciplines. Typically two days long, colloquia are international in scope and
bring together leading scientists in the field. Papers from colloquia are published in the Proceedings of the National Academy of Sciences
(PNAS).
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

COLLOQUIUM SERIES iii

National Academy of Sciences Colloquium Arnold and Mabel Beckman Center, Irvine
The Neurobiology of Pain
December 1113, 1998
PROGRAM
Friday, December 11, 1998
Introduction
Ronald Dubner, Colloquium Organizer
Session I: Channels
Chair and Discussion Leader John Hunter, Department of Analgesia, Roche Bioscience
Stephen G. Waxman, Department of Neurology, Yale University School of Medicine
Sodium Channels and the Pathophysiology of Pain
Michael Gold, Department of Oral and Craniofacial Biological Sciences, University of Maryland Dental School
TTX-R INa and Inflammatory Hyperalgesia
Daniel Weinreich, Department of Pharmacology and Experimental Therapeutics, University of Maryland, School of Medicine
Which Potassium Channels Keep Vagal Afferent Neurons Mellow?
Discussant Peter McNaughton, Department of Physiology, Kings College London
Heat-Activated Ion Currents in NociceptorsTransduction and Sensitization
Session II: Receptors
Chair and Discussion Leader Patrick Mantyh, Department of Preventive Sciences, School of Dentistry, University of Minnesota
Amy B. MacDermott, Department of Physiology and Cellular Biophysics and Center for Neurobiology and Behavior, Columbia
University
AMPA and Kainate Receptor Expression by DRG Neurons in Culture
Edwin W. McCleskey, Vollum Institute, Oregon Health Sciences University
The Role of Sensory Modality-Selective Gene Transcription in Opioid Analgesia
Michael W. Salter, Programmes in Brain and Behavior and Cell Biology, Hospital for Sick Children, and Department of
Physiology, University of Toronto
NMDA Receptors and Src in Synaptic Plasticity
Discussant Edward R. Perl, Department of Physiology, University of North Carolina School of Medicine, Chapel Hill
Receptor Expression and Regulation as Mechanisms Underlying Pain and Pain Pathology
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

COLLOQUIUM SERIES iv

Saturday, December 12, 1998

Introduction
Michael Gold, Colloquium Organizer
Tribute to John Liebeskind
Gregory Terman, Department of Anesthesiology, University of Washington, Seattle
Session III A: Systems and Imaging
Chair and Discussion Leader Donna Hammond, Anesthesia and Critical Care, University of Chicago
William D. Willis, Department of Anatomy and Neurosciences, University of Texas Medical Branch, Galveston, Texas
Evidence for a Visceral Pain Pathway in the Dorsal Column of the Spinal Cord
Tony L. Yaksh, Anesthesiology Research Laboratory, University of California
The Spinal Biology in Humans and Animals of Pain States Generated by Persistent Small Afferent Input
Discussant Gerald Gebhart, Department of Pharmacology, University of Iowa
Session III B: Systems and Imaging
Chair and Discussion Leader James Campbell, Department of Neurology, Johns Hopkins University Medical School
Howard Fields, University of California, San Francisco
Neuropeptides and Brainstem Pain Modulating Circuits
Kenneth L. Casey, Neurological Service, Veterans Administration Medical Center, Ann Arbor, Michigan
Forebrain Mechanisms of Nociception and Pain: Analysis through Imaging
Discussant M. Catherine Bushnell, McGill University, Montreal
What Are the Essential Cerebral Components?
Session IV: Growth Factors and Cytokines
Chair and Discussion Leader Kenneth Hargreaves, Department of Endodontics, University of Texas Health Sciences Center
William Snider, Department of Neurology, Washington University Medical Center
Trophic Factor Regulation of Nociceptor Development
Lorne M. Mendell, Department of Neurobiology and Behavior, State University of New York at Stony Brook
Neurotrophins and Pain
Linda R. Watkins, Department of Psychology, University of Colorado at Boulder
Immune-to-Brain Communication: Implications for Sickness and Pain
Discussant Stephen B. McMahon, Neuroscience Research Centre, Kings College London
Multiple Trophic Factor Influences on Nociceptive System
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

COLLOQUIUM SERIES v

Saturday, December 12, 1998

Session V: Development and Plasticity
Chair and Discussion Leader Ronald Dubner, Department of Oral and Craniofacial Biological Sciences, University of Maryland
Dental School
Maria Fitzgerald, Department of Anatomy and Developmental Biology, University College London
Postnatal Changes in Dorsal Horn Cell ActivityThe Development of Spinal Sensory Processing
Clifford J. Woolf, Department of Anesthesia and Critical Care, Massachusetts General Hospital and Harvard Medical School
Signal- and Activity-Dependent Plasticity in the Somatosensory SystemUnraveling the Cellular and Molecular
Mechanisms of Pain
David J. Mayer, Department of Anesthesiology, Medical College of Virginia
Cellular Mechanisms of Hyperalgesia and Morphine Tolerance and Their Interactions
Discussant Gary Bennett, Department of Neurology, Allegheny University, Philadelphia
The Revolution in Pain Research
Banquet for Colloquium Attendees
Banquet Speaker John Loeser, Department of Neurological Surgery and Anesthesiology, University of Washington Medical
Center
Sunday, December 13, 1998
Session VI: Molecular Genetics
Chair and Discussion Leader Charles Inturrisi, Department of Pharmacology, Cornell University Medical College
Allan Basbaum, Department of Anatomy, University of California, San Francisco
The Neurochemistry of Acute and Persistent Pain
Michael A. Moskowitz, Department of Neurosurgery and Neurology, Massachusetts General Hospital
Migraine Pathophysiology and Treatment Mechanisms
Jeffrey Mogil, Department of Psychology, University of Illinois at Urbana
The Genetics of Pain and Pain Inhibition: From Mice to Molecules
George Uhl, National Institute on Drug Abuse, National Institutes of Health
The Mu Opiate Receptor as a Model Gene for Individual Differences in Pain and Pain Modulation
Discussant Frank Porreca, Department of Pharmacology, University of Arizona Health Sciences Center
Antisense Oligodeoxynucleotides against the TTX-Resistant Sodium Channel, PN3, Prevent and Reverse Chronic,
Inflammatory and Neuropathic Pain in the Rat
the authoritative version for attribution.
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
the authoritative version for attribution.
COLLOQUIUM SERIES
vi
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

LIST OF ATTENDEES vii

Lisa Aimone
List of Attendees
George S. Aitken, Delta Medical Dental Centre
David J. Anderson, California Institute of Technology
David Andrew, Barrow Neurological Institute
K. Roger Aoki, Allergan
Dietrich Arndts, Boehringer Ingelheim
Ronald K. Ashley
Laura Audell
Victor Babenko, Aalborg University
Leif K. Bakland, Loma Linda University School of Dentistry
Allan Basbaum, University of California, San Francisco
Gary Bennett, MCP Hahnemann University
Jack M. Berger, University of Southern California School of Medicine
Scott Bowersox, Elan Pharmaceuticals
Walter Bowles, University of Minnesota
Robert Brennan, University of California, Los Angeles
Cathy Bushnell, McGill University
James Campbell, John Hopkins University
Yuqing Cao, University of California, San Francisco
Ken Casey, University of Michigan
Lin Chang, University of California, Los Angeles
Jenny Chen, University of California, Los Angeles
Zhou-Feng Chen, California Institute of Technology
Ken Chow, Allergan
Glenn Clark, University of California, Los Angeles
Patricia Claude, University of Texas, Health Science Center at San Antonio
Amy D. Clegg
Joseph R. Cohen, University of California, Los Angeles
Sean P. Cook, Oregon Health Sciences University-Vollum Institute, L-474
Santosh Coutinho, University of California, Los Angeles
Marie Csete, California Institute of Technology
Minglei Cui, Allergan
Bennet Davis
Xinzhong Dong, California Institute of Technology
Emma Dormand, California Institute of Technology
Ronald Dubner, University of Maryland Dental School
Helena Ennes, University of California, Los Angeles
Mark Erlander, RW Johnson Pharmalogical Research Institute
Stephen EspitiaJack L. Feldman, University of California,
Los Angeles
Howard Fields, University of California, San Francisco
Maria Fitzgerald, University College London
Christopher M. Flores, University of Texas
Nicholas Fuller, Cedars-Sinai Medical Center-The Pain Center
Gerald Gebhart, University of Iowa
Daniel W. Gil, Allergan, Inc.
Michael Gold, University of Maryland Dental School
Jay A. Goldstein
Robert Grimm
Paul Haerich, Loma Linda University
Aldric Hama
Donna Hammond, University of Chicago
Michael Hans
Kenneth Hargreaves, University of Texas, Health Science Center at San Antonio
Michael Hayward, Oregon Health Sciences University
Standiford Helm
David Helton, B. Braun Medical Inc.
Michael A. Henry, University of Colorado Health Science Center
H. Haydon Hill, Rehabilitation Medicine Associates
Dianne Hodges, CoCensys, Inc.
Lynda C. Honour
John Hunter, Roche
Victor Ilyin, CoCensys, Inc
Charles Inturrisi, Cornell University
Douglass L. Jackson, University of Washington-School of Dentistry
the authoritative version for attribution.

Stephen W. Jenkins, Allergan, Inc.

Russell Johnson, University of California, Los Angeles
Gerald J. Jonak, Dupont
Wade Kingery, Veterans Adminstration
Cheryl Kitt, National Institutes of Health
Beat Knusel, Allergan, Inc
Brad Kolls
Chris Konkoy
Lawrence Kruger, University of California, Los Angeles Medical Center
Nancy Lan, CoCensys Inc.
Letitia Lau
Jacqueline E. Lee, University of Colorado, Boulder
Julia Liebeskind
Wei-Jen Lin, Allergan
John Loeser
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

LIST OF ATTENDEES viii

Lin Luo, RW Johnson Pharmalogical Research Institute

Z. David Luo, University of California, San Diego
Kabirullah Lutfy, University of California, Los Angeles
Amy MacDermott, Columbia University
Nigel T. Maidment, University of California, Los Angeles
Celine Maillot, West Los Angeles V.A. Medical Center
Phil Malan, The University of Arizona
Mario G. Maldonado
Annika Malmberg, University of California, San Francisco
Corinne Manetto, Cedars-Sinai Comprehensive Cancer Center
Patrick Mantyh, University of Minnesota
Jianren Mao, VCU
Wendy Martin, Glaxo Wellcome, Inc.
William J. Martin, University of California, San Francisco
Juan Carlos Marvizon, University of California, Los Angeles
Emeran A. Mayer, University of California, Los Angeles, Division of Digestive Diseases
Ed McCleskey, Oregon Health Sciences University
Steve McMahon, Kings College London
Peter McNaughton, Kings College London
Lorne Mendell, State University of New York, Sunny Brook
Robert L. Merrill, University of California, Los Angeles School of Dentistry
Somsak Mitrirattanakul, University of California, Los Angeles
Jeffrey Mogil, University of Illinois at Urbana Champaign
Derek Molliver, Washington University of Saint Louis
Carlo V. Morpurgo
Mike Moskowitz, Harvard University
Million Mulugeta, West Los Angeles V.A. Medical Center
Alexander Nemirovsky
John K. Neubert
Alan Newman
Michael Ossipov, University of Arizona
A. Pace-Floridia
Pamela Pierce Palmer, University of California, San Francisco
Ed Perl, University of North Carolina
Karla Petersen, University of California, San Francisco
Frank Porreca, University of Arizona
Joshua P. Prager, California Pain Medical Center
Robert W. Presley, Pain Care Specialists
Steven Graff Radford, University of California, Los Angeles
John L. Reeves II, University of California, Los Angeles
Ke Ren, University of Maryland
Linda J. Rever
Steven Richeimer, University of California, Davis
Ruth Riley, University of California, San Francisco
Susan Roche, Orofacial Pain Management Clinic
Gary G. Rosengarten, The Pain Center, CSMC
Mike Salter, University of Toronto
Christine Sang
Harry Sernaker, University of Maryland
Seth Silbert, Oregon Health Sciences University-Vollum Inst.
William Snider, Washington University
Michael C. Son
Linda S. Sorkin, University of California, San Diego
Athena Spanoyannis, Allergan Inc.
Igor Spigelman, University of California, Los Angeles, School of Dentistry
Yvette Tache, University of California, Los Angeles
Brad Taylor, University of California, San Francisco
Edgar Tenorio, Elan Pharmaceuticals
Gregory Terman
Arnold Towe, University of Washington
Jodie Trafton, University of California, San Francisco
George Uhl, National Institute of Health
Clayton Varga, PRI
Douglas R. Wall
Wendy M. Walwyn, University of California
John Y.-X. Wang, Elan Pharmaceuticals
the authoritative version for attribution.

Yu Hua Wang, University of California, Los Angeles

Linda Watkins, University of Colorado
Steve Waxman, Yale University
Jen Yu Wei, University of California, Los Angeles, School of Medicine
Daniel Weinreich, University of Maryland
Ursula Wesselman, Johns Hopkins University
Larry A. Wheeler, Allergan
Joan Wilentz, National Institute of Dental and Cranofacial Research
William Willis, University of Texas Medical Branch
Steven P. Wilson, University of South Carolina School of Medicine
Richard Woodward, Cocensys Inc.
Clifford Woolf, Harvard University
Tony Yaksh, University of California, San Diego
Jen Yu, University of California, Irvine Medical Center
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TABLE OF CONTENTS ix

PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA

Table of Contents

Papers from a National Academy of Sciences Colloquium on The Neurobiology of Pain

The neurobiology of pain 76277630

Ronald Dubner and Michael Gold

John C. Liebeskind (19351997): A tribute 76317634

Gregory W. Terman

Sodium channels and pain 76357639

S. G. Waxman, S. Dib-Hajj, T. R. Cummins, and J. A. Black

A comparison of the potential role of the tetrodotoxin-insensitive sodium channels, PN3/SNS and NaN/SNS2, 76407644
in rat models of chronic pain
Frank Porreca, Josephine Lai, Di Bian, Sandra Wegert, Michael H. Ossipov, Richard M. Eglen, Laura Kas-
sotakis, Sanja Novakovic, Douglas K. Rabert, Lakshmi Sangameswaran, and John C. Hunter

Tetrodotoxin-resistant Na+ currents and inflammatory hyperalgesia 76457649

Michael S. Gold

Calcium regulation of a slow post-spike hyperpolarization in vagal afferent neurons 76507657

Ruth Cordoba-Rodriguez, Kimberly A. Moore, Joseph P. Y. Kao, and Daniel Weinreich

Ion channels gated by heat 76587663

P. Cesare, A. Moriondo, V. Vellani, and P. A. McNaughton

Causalgia, pathological pain, and adrenergic receptors 76647667

Edward R. Perl

Forebrain mechanisms of nociception and pain: Analysis through imaging 76687674

Kenneth L. Casey

A visceral pain pathway in the dorsal column of the spinal cord 76757679
William D. Willis, Elie D. Al-Chaer, Michael J. Quast, and Karin N. Westlund

The spinal biology in humans and animals of pain states generated by persistent small afferent input 76807686
Tony L. Yaksh, Xiao-Ying Hua, Iveta Kalcheva, Natsuko Nozaki-Taguchi, and Martin Marsala

Supraspinal contributions to hyperalgesia 76877692

M. O. Urban and G. F. Gebhart

Neurotrophins and hyperalgesia 76937696

X.-Q. Shu and L. M. Mendell

Src, a molecular switch governing gain control of synaptic transmission mediated by N-methyl-D- aspartate 76977704
receptors
Xian-Min Yu and Michael W. Salter
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TABLE OF CONTENTS x

Pain perception: Is there a role for primary somatosensory cortex? 77057709

M. C. Bushnell, G. H. Duncan, R. K. Hofbauer, B. Ha, J.-I. Chen, and B. Carrier

Implications of immune-to-brain communication for sickness and pain 77107713

Linda R. Watkins and Steven F. Maier

Brain-derived neurotrophic factor is an endogenous modulator of nociceptive responses in the spinal cord 77147718
S. W. N. Thompson, D. L. H. Bennett, B. J. Kerr, E. J. Bradbury, and S. B. McMahon

The postnatal development of spinal sensory processing 77197722

Maria Fitzgerald and Ernest Jennings

Transcriptional and posttranslational plasticity and the generation of inflammatory pain 77237730
Clifford J. Woolf and Michael Costigan

Cellular mechanisms of neuropathic pain, morphine tolerance, and their interactions 77317736
David J. Mayer, Jianren Mao, Jason Holt, and Donald D. Price

Does a neuroimmune interaction contribute to the genesis of painful peripheral neuropathies? 77377738
Gary J. Bennett

Distinct neurochemical features of acute and persistent pain 77397743

Allan I. Basbaum

The genetic mediation of individual differences in sensitivity to pain and its inhibition 77447751
Jeffrey S. Mogil

The opiate receptor as a candidate gene for pain: Polymorphisms, variations in expression, nociception, and 77527755
opiate responses
George R. Uhl, Ichiro Sora, and Zaijie Wang
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE NEUROBIOLOGY OF PAIN 7627

This paper is the introduction to the following papers, which were presented at the National Academy of Sciences colloquium
The Neurobiology of Pain, held December 1113, 1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

The neurobiology of pain

RONALD DUBNER * AND MICHAEL GOLD

Department of Oral and Craniofacial Biological Sciences, University of Maryland, School of Dentistry, Baltimore, MD 21201
This is a very exciting time in the field of pain research. Major advances are occurring at every level of analysis, from development to
neural plasticity in the adult and from the transduction of a noxious stimulus in a primary afferent neuron to the impact of this stimulus on
cortical circuitry. The molecular identity of nociceptors, their stimulus transduction processes, and the ion channels involved in the
generation, modulation, and propagation of action potentials along the axons in which these nociceptors are present are being vigorously
pursued. Similarly, tremendous progress has occurred in the identification of the receptors, transmitters, second messenger systems,
transcription factors, and signaling molecules underlying the neural plasticity observed in the spinal cord and brain stem after tissue or
nerve injury. With recent insight into the pharmacology of different neural circuits, the importance of descending modulatory systems in
the response of the nervous system to persistent pain after injury is being reevaluated. Finally, imaging studies have revealed that
information about tissue damage is distributed at multiple forebrain sites involved in attentional, motivational, and cognitive aspects of the
pain experience.
These major advances in pain research were the subject of a National Academy of Sciences colloquium entitled The Neurobiology
of Pain, held at the Beckman Center of the Academy in Irvine, California on December 1113, 1998. The meeting was organized by John
Liebeskind (deceased), Ronald Dubner, and Michael Gold. Its purpose was to bring together pain research scientists and those in related
fields who have made recent major advances in the development, cellular, and molecular biology and integrative neurosciences related to
the neurobiology of pain. The colloquium was organized into six sessions, each with a separate theme: channels, receptors, imaging and
systems neuroscience, growth factors and cytokines, development and plasticity, and molecular genetics. There was ample opportunity for
the discussion of the most fruitful and exciting lines of research and the identification of important future directions. One hundred and
sixty scientists attended the colloquium. We are indebted to Glaxo Wellcome, Inc. for its generous support that helped defray the expenses
of graduate students and the social events, as well as to Fran Addison for her invaluable assistance in the organization of the colloquium.
This colloquium was held because of John Liebeskinds commitment to the study of pain. Elected to the Academy after more than 20
years of pioneering research in the field, John always maintained that a critical component to progress in this or any field was a forum in
which leaders in the field could assemble to discuss recent advances and future directions. Given the tremendous advances that have
occurred in the field of pain research over the last decade, John felt that a colloquium held under the auspices of the Academy would be
both timely and appropriate. Over two years ago, he approached us and asked that we help him organize this colloquium. Soon after the
program was approved and the date was set, John learned that he had terminal cancer, and he died in September, 1997. He would have
been very pleased by the depth and breadth of research covered as well as the lively interactions of all the participants. While John was
remembered by many of the speakers, Greg Terman, one of his former students, delivered a moving and informative tribute ( 1 ).
The colloquium got underway with a spirited discussion of the role of ion channels in peripheral nerve, particularly their expression
in nociceptors. Researchers have long since appreciated that, in the presence of injury, nociceptors may become hyperexcitable. A change
in the expression of ion channels is one mechanism that may contribute to this hyperexcitability. Steve Waxman ( 2 ) summarized data
from an elegant series of experiments indicating that sodium channel expression in dorsal root ganglion neurons is dynamic, changing
markedly after tissue or nerve injury. Importantly, different forms of injury induce different changes in the expression of sodium channels.
For example, nerve injury in the form of axotomy results in a decrease in the expression of tetrodotoxin (TTX)-resistant currents and an
increase in a rapidly repriming TTX-sensitive sodium current. In contrast, inflammation results in an increase in the expression of TTX-
resistant sodium currents and a decrease in the expression of a TTX-sensitive current. Utilizing a different nerve injury model than that
employed by Waxman and colleagues, in combination with antisense oligodeoxynucleotides, Frank Porreca ( 3 ) presented evidence
indicating that a TTX-resistant sodium channel called SNS/PN3 is critical for the initiation and maintenance of nerve injury-induced
hyperalgesia and allodynia. In contrast, NaN, another TTX-resistant sodium channel recently identified by Waxman and colleagues ( 4 ),
does not appear to contribute to the maintenance of nerve injury-induced changes in nociceptive thresholds. Michael Gold ( 5 ) reported on
the role of the TTX-resistant sodium currents in inflammation and showed that the current is modulated by inflammatory mediators such
as prostaglandin E2, 5-HT, and adenosine, consistent with its role in peripheral sensitization. Gold provided additional data indicating that
TTX-resistant channels are not only present and functional in the peripheral terminals of nociceptors, but that modulation of these
channels contributes to prostaglandin-induced mechanical hyperalgesia. Daniel Weinreich ( 6 ) switched the focus of the discussion to
other channels by addressing the role of a calcium-dependent potassium current in controlling the excitability of vagal afferents. Through a
beautiful series of experiments, Weinreich was able to assess the relative contribution of various sources of calcium responsible for the
gating of the potassium currents.
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: TTX, tetrodotoxin; NGF, nerve growth factor.
* To whom reprint requests should be addressed at: Department of Oral and Craniofacial Biological Sciences, University of Maryland,
School of Dentistry, 666 West Baltimore Street, Room 5E-08, Baltimore, MD 21201.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE NEUROBIOLOGY OF PAIN 7628

Peter McNaughton ( 7 ) reviewed the data collected in his laboratory demonstrating for the first time the existence of an ion channel
specifically activated by heat. Pursuing observations indicating that bradykinin modulates the heat activated channel, McNaughton
presented evidence implicating activation of the epsilon isoform of protein kinase C in this process. McNaughtons findings are of even
greater interest because of the similarity of this ion channel to the properties of the recently cloned capsaicin/heat receptor by Julius and
colleagues ( 8 ).
Research on receptors involving the transduction, transmission, and modulation of nociceptive information is clearly one of the most
exciting and rapidly advancing areas in the field of pain research today. With the molecular characterization of many of the receptors
involved in the transmission of nociceptive stimuli as well as the cellular elements necessary for synaptic transmission, researchers have
begun to piece together the essential elements necessary for the first steps ultimately leading to the perception of pain. Amy MacDermott
started the session by describing results from recent experiments performed in her laboratory designed to investigate the role of
presynaptic non-N-methyl-D-aspartate receptors at the first synapse in the nociceptive pathway. Utilizing a dorsal root ganglion neuron/
dorsal horn neuron co-culture, MacDermott and her colleagues obtained evidence indicating dorsal root ganglion neurons express
functional -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid and kainate receptors. Importantly, activation of these receptors
appears to influence glutamate release from the sensory neuron and therefore the activation of the dorsal horn neurons. Edwin McCleskey
described an exciting series of experiments performed in his laboratory utilizing a combination of electrophysiology and single-cell PCR.
Through beautifully controlled reverse transcriptionPCR reactions, McCleskey and his colleagues were able to determine the number of
mRNA copies encoding the -opioid receptor in a single cell in which the presence of functional -opioid receptors had previously been
investigated. Their results provide a mechanistic explanation for some perplexing aspects of opioid analgesia. Michael Salter ( 9 ) provided
evidence supporting a revolutionary hypothesis concerning the cellular events underlying the development of long-term potentiation in the
hippocampus and, by analogy, central sensitization of spinal cord dorsal horn neurons after tissue injury. Salters compelling evidence
indicates that the initial steps underlying these two phenomena may involve an increase in the intracellular concentration of Na+, activation
of the nonreceptor protein tyrosine kinase, Src, and the subsequent phosphorylation of N-methyl-D-aspartate receptors. Edward Perl ( 10 )
pointed out that changes in the expression of receptors involved in the transmission of nociceptive stimuli may contribute to the
pathophysiology of pain. Perl cited evidence obtained in his laboratory supporting the hypothesis that an increase and/or change in the
expression of -adrenergic receptors present in sensory neurons is an underlying mechanism of adrenergic excitation of sensory neurons
often observed after nerve injury.
The session on imaging and systems neuroscience examined some of the most recent exciting findings on pain pathways and their
modulation. William Willis reported on a new visceral pain pathway that ascends in the dorsal column of the spinal cord ( 11 ).
Postsynaptic dorsal column neurons in the rat sacral spinal cord transmit visceral signals to the gracile nucleus, and this information is then
relayed to the ventral posterior lateral thalamic nucleus. Functional MRI studies have revealed that dorsal column lesions eliminate blood
volume changes in the thalamus produced by noxious pelvic visceral stimulation, suggesting the importance of this pathway. More studies
are needed to determine the functional significance of the spinothalamic and spinoparabrachial visceral pathways in comparison with this
newly discovered dorsal column pathway. Tony Yaksh ( 12 ) described studies with different animal models that support the importance of
spinal cord processing in pain states; he emphasized the functional and pharmacological comparability of symptoms across species and
pointed out that these models are an important source of information for the development of novel clinically relevant analgesics. Howard
Fields described his elegant findings on specific brain stem networks involved in potent pain modulation, -opioid receptor agonists
activate neurons in the periaqueductal gray and the rostral ventral medulla by inhibiting GABAergic inhibition. The behavioral
antinociception and inhibition of dorsal horn neurons is mediated by the release in rostral ventral medulla of an endogenous opioid peptide
acting at the -opioid receptor. Fields also presented exciting data indicating that -opioid receptor selective ligands have actions in rostral
ventral medulla that oppose those of the -opioid receptor-selective ligands and block the antinociceptive effect of periaqueductal gray-
administered morphine. Interestingly, this effect was only observed in male rats. Descending modulation systems were summarized by
Gerry Gebhart ( 13 ), who provided evidence that supraspinal structures make a significant contribution to the development and
maintenance of hyperalgesia associated with tissue injury. He suggested that persistent input engages spinobulbospinal facilitatory
mechanisms that contribute to secondary hyperalgesia that occurs outside the site of injury. The findings by others of descending
inhibitory systems contributing to hyperalgesia emphasizes the bimodal nature of these descending systems in the modulation of persistent
pain. Ken Casey ( 14 ) described the role of forebrain mechanisms of pain in imaging studies in humans and reviewed convincing evidence
that the perceived intensity of unilateral pain evoked by different inputs correlates with increases in regional cerebral blood flow in
primarily five structures: bilaterally in the thalamus, the contralateral insula, the bilateral premotor cortex, the contralateral anterior
cingulate, and the cerebellar vermis. In contrast, results on the role of primary somatosensory cortex are somewhat inconsistent. Cathy
Bushnell ( 15 ) reviewed the factors contributing to this inconsistency including cognitive modulation, average-related degradation of
signal due to anatomical variability in sulcal anatomy and differences in methodology. She provided behavioral evidence indicating that
manipulations that altered pain discrimination altered activity in primary somatosensory processing regions of the cerebral cortex. In
contrast, manipulation that preferentially altered the affective or motivational dimension of pain produced changes in the anterior cingulate
cortex. The combined use of psychophysical testing and brain imaging in humans should help reveal the functional role of these different
forebrain structures that have direct corticofugal projections to the thalamus, brain stem, and spinal cord and thereby modulate the pain
experience at those levels.
The role of trophic factors and cytokines in the development and maintenance of pain in response to various forms of tissue injury is
an area of research that has virtually exploded in the last several years. William Snider opened the session by describing recent
experiments performed in his laboratory designed to distinguish trophic influences of nerve growth factor (NGF) from its role in cell
survival. Through the use of knockout mice, Snider and his colleagues obtained striking results suggesting that, although activation of the
high affinity NGF receptor was necessary to establish proper innervation of peripheral targets, activation of this receptor was not necessary
for the growth and guidance of central terminals. In addition to its role in development, NGF and other growth factors and cytokines have
been shown to mediate pain and hyperalgesia associated with tissue injury. Lorne Mendell ( 16 ), the first to describe the
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE NEUROBIOLOGY OF PAIN 7629

link between NGF and pain, presented results obtained from experiments performed in his laboratory identifying the mechanisms
underlying the initial hyperalgesic response to NGF. The initial hyperalgesia in response to systemic or peripherally administered NGF
depends on indirect mechanisms, specifically mast cell degranulation. Mendell presented resent evidence indicating that NGF also is
capable of potentiating capsaicin-evoked currents in isolated sensory neurons. Utilizing this intriguing observation, Mendell presented a
model that would account for the initial NGFinduced thermal hyperalgesia. Focusing on the interaction between the immune system and
the nervous system, Linda Watkins ( 17 ) described additional pathways through which activation of the immune system results in changes
in multiple sites throughout the nervous system. Watkins described the molecules involved in the signaling pathways as well as how
activation of this system results in changes in behavior. Steve McMahon ( 18 ) brought trophic factors back to center stage with his
summary of a growing body of data implicating a critical role for brain derived neurotrophic factor in the altered nociceptive processing
observed in the presence of inflammation. Brain derived neurotrophic factor appears to function as a neurotransmitter/neuromodulator in
the dorsal horn of the spinal cord, where it is released from the central terminals of small-caliber afferents and increases the excitability of
dorsal horn neurons.
The session on development and plasticity explored plasticity that occurs in the central nervous system after tissue and nerve injury.
Maria Fitzgerald ( 19 ) reported on changes in the neonatal spinal cord that are not simply immature or incomplete versions of what occurs
in the adult. Central sensitization occurs in the normal immature spinal cord in response to electrical stimulation of A fibers whereas
activity-induced plasticity in the adult spinal cord takes place only in response to C fiber strength stimulation, unless the dorsal horn is
primed by previous peripheral injury. N-methyl-D-aspartate and -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors are
distributed in higher density in the neonatal cord, and the receptor subunit composition in the neonatal spinal cord maximizes non-N-
methyl-D-aspartate calcium influx. Clifford Woolf ( 20 ) provided an outstanding, concise, and up-to-date review of activity-induced and
signal-induced plasticity in sensory neurons after tissue and nerve injury. He showed the interaction of these mechanisms in the role of
brain derived neurotrophic factor in the generation of central sensitization. Jianren Mao ( 21 ) suggested that hyperalgesia and morphine
tolerance may be interrelated by common neural mechanisms involving excitatory amino acid receptor activation and subsequent
intracellular events, such as protein kinase C translocation and nitric oxide production. This hypothesis is supported by experiments
showing that hyperalgesia develops when animals are made tolerant to morphine and that both the hyperalgesia and morphine tolerance
develop as a consequence of peripheral nerve injury. Gary Bennett ( 22 ) reported on a new model of inflammation in which a focal
neuritis is produced in the rat sciatic nerve. The results suggest the presence of a neuroimmune interaction that occurs at the onset of nerve
injury and contributes to the development of neuropathic pain.
While the tools of molecular genetics were employed by many of the researchers who spoke throughout the colloquium, the issue was
addressed formally in the final session. Alan Basbaum ( 23 ) discussed the use of knockout mice to investigate the role of specific
receptors and second messengers in nociceptive processing. Basbaum eloquently illustrated how this powerful approach has shed new
light on our understanding of the mechanisms of action of molecules such as substance P and the isoform of protein kinase C. For
example, the contribution of substance P and neurokinin A to central sensitization may be considerably less than previously suspected
while, in contrast, activation of the isoform of protein kinase C appears to be vital to the development of nerve injury-induced
hyperexcitability of dorsal horn neurons. Michael Moskowitz focused the discussion somewhat by describing several approaches that have
been employed in the study of migraine. Moskowitz reviewed the data implicating the involvement of a specific class of serotonin
receptors in migraine headache. Based on an exciting series of functional MRI studies, Moskowitz demonstrated that brain metabolism
and blood flow may be uncoupled before the onset of headache. In contrast to the approach utilized by many researchers attempting to
identify a role for a specific protein in nociception (a bottom-up approach), Jeffrey Mogil ( 24 ) described a top-down approach in which
genetic mapping may be employed to identify genes responsible for specific behavioral phenotypes. Such an approach is readily applied to
pain research, where it provides a mechanism for the identification of unique molecules critically involved in nociceptive processing. To
illustrate a case in point, Mogil described the identification of a specific serotonin receptor subtype involved in the expression of morphine
analgesia. George Uhl ( 25 ) discussed the integration of the top-down and bottom-up approaches through experiments performed with the
-opioid receptor knockout mouse as well as populations of humans. His results suggest that polymorphisms in the gene encoding the -
opioid receptor may explain much of the variability observed among people with respect to their responsiveness to opiate analgesia.
Identification of the underlying mechanisms controlling opioid responsiveness may enable the development of individualized treatment
programs for ongoing pain.
The exciting new advances in pain research emphasize the importance of this field of neuroscience. The neural apparatus responsible
for the perception of pain includes mechanisms that clearly are prototypic components of all mammalian sensory systems. These
mechanisms include specialized receptors, stimulus transduction mechanisms, ion channel modulation, rapid and slow activity involving
excitatory and inhibitory transmitters and their receptors, amplification of relevant signals at peripheral and central nervous system sites
utilizing activity-dependent and signal-dependent mechanisms of neuronal plasticity, and, finally, distributed processing of environmental
signals and their interaction with learned memories at higher centers. The field of pain research has made giant steps in putting together
important segments of this puzzle. It is, of course, the hope of all of us that these advances will lead to an improvement in the quality of
life of acute and chronic pain sufferers.
1. Terman, G. W. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76317634 .
2. Waxman, S. G. , Dib-Hajj, S. , Cummins, T. R. & Black, J. A. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76357639 .
3. Porreca, F. Lai, J. , Bian, D. , Wegert, S. , Ossipov, M. H. , Eglen, R. M. , Kassotakis, L. , Novakovic, S. , Rabert, D. K. , Sangameswaran, L. &
Hunter, J. C. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76407644 .
4. Dib-Hajj, S. D. , Tyrell, L. , Black, J. A. & Waxman, S. G. ( 1998 ) Proc. Natl. Acad. Sci. USA 95 , 89638968 .
5. Gold, M. S. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76457649 .
6. Cordoba-Rodriguez, R. , Moore, K. A , Kao, J. P. Y. & Weinreich, D. , ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76507657 .
7. Cesare, P. , Moriondo, A. , Vellani, V. & McNaughton, P. A. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76587663 .
8. Caterina, M. J. , Schumacher, M. A. , Tominaga, M. , Rosen, T. A. , Levine, J. D. & Julius. D. ( 1997 ) Nature (London) 389 , 816824 .
9. Yu, X.-M. & Salter, M. W. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76977704 .
10. Perl, E. R. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76647667 .
11. Willis, W. D. , Al-Chaer, E. D. , Quast, M. J. & Westlund, K. N. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76757679 .
the authoritative version for attribution.

12. Yaksh, T. L. , Hua, X.-Y. , Kalcheva, I. , Nozaki-Taguchi, N. & Marsala, M. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76807686 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE NEUROBIOLOGY OF PAIN 7630

13. Urban, M. O. & Gebhart, G. F. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76877692 .
14. Casey, K. L. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76687674 .
15. Bushnell, M. C. , Duncan, G. H. , Hofbauer, R. K. , Ha, B. , Chen, J.-I. & Carrier, B. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77057709 .
16. Shu, X.-Q. & Mendell, L. M. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76937696 .
17. Watkins, L. R. & Maier, S. F. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77107713 .
18. Thompson, S. W. N. , Bennett, D. L. H. , Kerr, B. J. , Bradbury, E. J. & McMahon, S. B. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77147718 .
19. Fitzgerald, M. & Jennings, E. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77197722 .
20. Woolf, C. J. & Costigan, M. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77237730 .
21. Mayer, D. J. , Mao, J. , Holt, J. & Price, D. D. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77317736 .
22. Bennett, G. J. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77377738 .
23. Basbaum, A. I. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77397743 .
24. Mogil, J. S. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77447751 .
25. Uhl., G. R. , Sora, I. & Wang, Z. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 77527755 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

JOHN C. LIEBESKIND (19351997): A TRIBUTE 7631

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

John C. Liebeskind (19351997): A tribute

GREGORY W. TERMAN *
Department of Anesthesiology and the Graduate Program in Neurobiology and Behavior, University of Washington, Seattle, WA
98195
Id like to begin by thanking those who have helped me prepare this tributecontributing pictures, stories, and/or moral support. This
group includes many of Johns students and friends, his family, and especially those at the Louise M. Darling Biomedical Library at
UCLA where the John C. Liebeskind History of Pain Collection is housed. In particular, Marcia Meldrum, who worked closely with John
on the Pain Collection and, in 1995, took an oral history from him concerning his perspectives on his career, will be stolen from frequently
here. Also, Russell Johnson spent days helping me wade through many of Johns papers on a recent visit to the library.
My job is to address those of you who didnt know John Liebeskind; to give you a flavor for the importance of this man, not only for
the study of the neurobiology of painthe topic of this conferencebut also in positively influencing innumerable lives he came in
contact with and, literally, health care worldwidethough he never treated a patient.
I arrived at UCLA for graduate school in the fall of 1980. I had decided to attend UCLA because my psychology teacher, after
hearing John speak at a weekend seminar for undergraduate teachers on the East Coast, persuaded me that there was only one option for
pain research trainingLiebeskind. Fig. 1 is a picture of the John Liebeskind I met and got to know as a student in his laboratorynotice
the phone. John was never too busy for the phone or the people calling him on it. Although I actually enjoyed this characteristic once I left
UCLA, while I was there I hated it. The other imposing inanimate object in his office was the bulletin board. Now, the bulletin board may
have simply been a device John put up to give students he was supposed to be meeting with something to look at while he was on the
phone. Regardless, Johns bulletin board was such a fixture that when Tim Cannon, a former student, constructed his Unofficial
Liebeskind Lab Web Page a few years back, the whole bulletin board was lovingly reproduced. On it were pictures of his students (and
several teachers), friends, and family, as well as favorite sayings and over 100 misspellings of his name collected from various sources
over the years. Essentially, Johns bulletin board was a reflection of his two great loves, words and people.
John was born in Waterbury, Connecticut in 1935, the son of a clothing store owner. His family encouraged his education, sending
him to private school from 6th grade through college. It was in high school that he first developed his love of words, and in his oral history
he recounts specific teachers that he felt molded his academic interests. In my opinion, even more telling were his summers from age 8
until age 20, when he went off for several weeks each year to Camp Kennebec in Maine ( Fig. 2 ), first as a camper and then as a
counselor. Even after he began to attend college at Harvard he would head to Camp Kennebec in the summer. John described himself as a
good camp counselor; a teacher and mentor, helping younger kids; and psychologically aware. I wouldnt have understood in 1980 if
you asked me how I liked Camp Liebeskind, and I am only now beginning to realize, with trainees of my own, the effort it takes to be a
decent mentor. I remember fondly the group gatherings and his practice of inviting students not going home for holidays to his home for
backyard barbecues (the company was great and the food wasuhwell done). He treated his students/campers as if they were his
children, and the counselor made sure that once you visited Camp Liebeskind, you never left.
John majored in social relations at Harvard, taking the minimum science courses required (History of Science by I. B. Cohen and The
Science of Human Behavior by B. F. Skinner) before heading off to the University of Michigan to graduate school in clinical psychology
in 1957. At Michigan, he switched fairly quickly from clinical to physiological psychology, searching for what he called more precise
research. In his final 4 years of graduate school, he struggled with his inadequate science preparation and three different thesis advisors
before obtaining his Ph.D. in the fall of 1962. In his oral history, he says that these experiences helped him to find peace in studying and
to mold his attitudes toward mentoring. He cited his eventual thesis advisor, Steve Fox, as telling him, Whatever is good for you, John,
and your career, is going to reflect back on me and is going to end up being good for me. John learned well. One of his students, Hanan
Frenk, says of John, We were [each] the best student he ever had, until the next one needed a job [recommendation].
After graduation, John stayed on at Michigan, teaching and working with Steve Fox for another year. It was during this time that he
learned the basics of electrophysiology and decided to go to Paris to train with Madame Denise Albe-Fessard. There he studied muscle
spindle afferent inputs to the cortex in monkeys. When he returned to the United States in the summer of 1965, he came back as one of
few psychologists trained in electrophysiology and had several job offers to choose from before deciding to take a job at UCLA. He
arrived to work at UCLA in January of 1966, shortly after the Melzack/Wall Gate Control theory of pain was published in late 1965. John
was fascinated by this paper and began thinking of studying pain processes in his own new laboratory. His initial NIH grant proposal
concerned the modulation of cortical nociceptive responses by learning. This grant was funded and though I am unaware of him ever
having done any of the studies proposed in it, he kept that NIH grant, his only one, for the next 28 years before he closed his laboratory.
As John settled in, in Los Angeles, he was heavily influenced by his collaborators. Selected reviews of his early work include, offers
a valuable insight, a classic of its time, and [this] work blows my mind because it is so simple and so profound, and refer, of course,
to his uncredited walk-on part in Melvin Van Peebles movie Sweet Sweetbacks Baadasssss Song. Having sat through this movie to get a
glimpse of John, Im afraid I have to give it thumbs down. John may have agreed; it was the closest he would get to Hollywood stardom.
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

* To whom reprint requests should be addressed, e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

JOHN C. LIEBESKIND (19351997): A TRIBUTE 7632

FIG. 1. Dr. John Liebeskind at work in his office at UCLAon the phone, with his bulletin board in the background.
Meanwhile, in the laboratory, Dave Mayer and other early students found some interesting effects of lesions in the midbrain of rats on
their escape behaviors and began to investigate this relatively unstudied brain region in earnest. Among Johns papers, I found a 1971
letter from the journal Science rejecting his manuscript demonstrating an analgesic effect of midbrain stimulation in the rat. John and Dave
had seen a report by Reynolds suggesting an analgesic effect of midbrain stimulation, and while originally setting out to disprove the idea,
ended up supporting it. John was close to a tenure decision at UCLA at this time and whether this extra pressure prompted him or not, he
elected to call Science and ask for a second chance. He told the editor, Mr. Ringle, that he would rewrite the paper specifically for him,
emphasizing the high points, and if Ringle didnt agree that this was fascinating science he would not trouble him further. To Ringles
credit, he recognized the importance of the manuscript, now a Citation Classic, and published it. To Johns credit, in the rewrite, he argued
persuasivelywhat he said was Daves ideathat these data demonstrated the existence of endogenous pain inhibitory systems. The
articulation of this concept made this paper sing (as John would say) and almost certainly is what got it published in Science. For the rest
of his career, John would save his special magic for the discussion (his favorite part of any paper); taking methods and results and turning
out meaningadmittedly, sometimes, going just a bit beyond what the data actually demonstrated. Marcia Meldrum claims that, even as
[John] pursued his long and creative career in science, he was a historian because he could see ideas and events in context and knew how
to tell a story ). More cynically, one student joked that as a scientist, John would have made a wonderful used car salesman. Indeed, John
was always quick to deflect credit for his accomplishments to his excellent students and, truthfully, he did only rarely enter the lab the
last 25 years of his career. Nonetheless, his excellent students might only have been average elsewhere and, in fact, I am unaware of his
ever having turned away any student. He published nine times in Science with seven different first authors. Among his personal papers I
came upon a scribbled quotation attributed to T. S. Eliot, Where is the wisdom we have lost in knowledge? Where is the knowledge we
have lost in information? John found wisdom in the information we brought to him from the lab. Table 1 outlines his scientific findings
and the larger concepts they addressed. His early studies, for example, of stimulation-produced analgesia and his finding that naloxone
blocked this phenomenon probably facilitated the discovery of endogenous opioid peptides. His work on natural ways of activating these
pain-inhibitory systems, the influence of gender and other genetic factors on analgesic mechanisms, and pains inhibition of the immune
system, as well as much of his other work, at their core, all suggested that pain is not just unpleasant but dangerous in some circumstances.
His findings, themselves, were not always first in the literature, but the concepts were routinely ahead of their time, sometimes spawning
entire fields of study.

FIG. 2. John at Camp Kennebec in 1952.

the authoritative version for attribution.

John loved the word heuristic. I believe his research was heuristic. John was always sure of the right words. As a result, he was a
terrible pain to write with. An almost finished manuscript would come back marked beyond recognition. It was comforting then, looking
through his papers, to uncover a heavily edited version of his Pain Can Kill editorial on which he was the sole author. He couldnt even
write with himself, much less with others. He sure could schmooze with others, however. Doing important science and publishing it in
highprofile journals was only a start for John. He would hit the road, spreading the gospel, and from those first publications in the early
1970s he used his considerable political talents to encourage not just more pain research but better clinical pain management as well. He
was present in 1973 at the famous Issaquah meeting where the International Association for the Study of Pain (IASP) was formed and the
journal Pain was
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

JOHN C. LIEBESKIND (19351997): A TRIBUTE 7633

proposed. He started the Western Pain Society virtually by himself and then was instrumental in the formation of the American Pain
Society (APS). The pain-related organizational activities listed in Table 2 are selected from his C.V. and demonstrate the incredible time
and energy John must have spent putting forward the cause of pain research at the national and international levels. The foundation of the
pain societies by John, his good friend John Bonica, and others have fostered dialogue between basic and clinical scientists, setting the
standard for medical research. Moreover, working with the World Health Organization, these societies have succeeded in advancing the
cause of optimum pain management throughout the world. Such international educational and patient care successes have, of course,
raised the bar here at home for more and better pain care and research, benefiting industry, academia, and clinical practice alike. John
Liebeskind will not be able, as planned, to become president of the IASP next year. John Bonica is also dead. For those not already doing
so, I would encourage all who profit from the work of the pain societies every time a significance section of a grant proposal is written, to
consider contributing a portion of your efforts to these societies at other times of the year.

FIG. 3. John Liebeskind (1996).

In 1995, John was elected to the National Academy of Sciences, whose motto the furtherance of science and its use for the general
welfare he exemplified. As the only Academy member who was also a member of the APS or IASP, John began to think about how he
could use this honor to help the field of pain. This National Academy of Sciences-sponsored colloquium is one such effort.
From 1985, when John received his Jacob K. Javits Neuroscience Investigator award, his last NIH grant, until he closed his lab for
good 9 years later, he published 82 scientific articles and 15 book chapters. During that time, however, he became more and more
interested in telling the pain history story rather than his own research story. Wendy Sternberg, Johns last graduate student, recalls the
look of glee on his face as he showed his new [pain history] toys . . . his tape recorders, transcribers, and narratives of his interviews. I
remember John excitedly telling me of his plans to interview the pioneers of the pain field, to create a set of oral histories documenting
important people and events. I would nod numbly, clueless as to all I would learn looking at just one oral history, his, some years later. His
collection expanded to include personal papers of pain leaders, organizational records of the pain societies, and important historical works
on pain. Marcia Meldrum remembers that, John worked hard at learning to be a historian. He attended a training workshop given by the
Oral History Association and a graduate class in archival processing. He introduced himself to historians at UCLA and picked their brains.
But he took the greatest pleasure in having people tell him their stories, in finding rare books on pain, and in opening boxes of original
documents. John had found another way to meld his love of words and people for the benefit of the field of pain ( Fig. 3 ).
Unfortunately, Johns collection remains unfinished, although now bearing his name. In the fall of 1996, following a long bout of
laryngitis, a laryngeal tumor was diagnosed and resected. A total laryngectomy followed some weeks later, and his effortless speech was
replaced with an artificial voice box.
Table 1. John Liebeskinds studies and the concepts they addressed
Year Studies Concepts
1971 Stimulation produced analgesia Pain inhibitory systems
19721973 Dissecting analgesia/reward Analgesia vs. abuse
19721976 Naloxone blocks SPA Endorphins
1976 SPA for visceral pain Visceral vs. somatic pain
19771978 Enkephalin is epileptogenic Therapeutic window
19791985 Analgesia from stress and seizures Natural inhibition of pain
19831984 Stress inhibits immune function Psychoneuroimmunology
1991 Pain inhibits immune function Pain can kill
19791991 Effects of learning, pain, and NMDA receptors on opiate tolerance Tolerance is not simply receptor desensitization
19931995 Gender influences on analgesia Gender dependent effects
19851996 Genetic studies of analgesia Variance and mechanism

NMDA, N-methyl-D-aspartate; SPA, stimulation-produced analgesia.

the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

JOHN C. LIEBESKIND (19351997): A TRIBUTE 7634

Table 2. John Liebeskinds pain-related organizational activities

Year Organization Position
1973, 19751977 NIH Special pain study section
19751987; 1996 IASP Council member
19751978 IASP Scientific Program Chair
19751977 IASP Education chair
19751985 Pain Editorial board
19751981 IASP Publications committee
19781985 APS Board of directors
19781980 APS Scientific program committee
19781981 IASP Research and ethics committee
19811987 IASP Public information committee
19891991 APS President elect, President, Past President
1996 IASP President elect

IASP, International Association for the Study of Pain; APS, American Pain Society.

In the late spring of 1997, John again sought medical attention, this time for pain in his chest, and the tumor was found to have
metastasized to his lungs and beyond. Shortly after this diagnosis, he sent an e-mail to his studentsapologizing that, because he was
dying, he would not be able to attend a party we had scheduled for him in honor of his National Academy election. He went on to tell us of
our importance to him in a love letter I have no plans to ever delete.
John died on September 8th, 1997 at home with his family, in little painfortunately, he was spared that irony. As a student and
friend, it has been heartening to see the outpouring of remorse from around the world over the past year at the loss of this honorable
scholar, scientist, and statesman. It is fitting that we pay tribute to him at this conference today. Not long after learning of his poor
prognosis, John e-mailed the Academy, writing in part: Dear Edward. Ive had some bad news about my health . . . . This news calls into
serious question my ability to participate in the [National Academy of Sciences] colloquium Ron Dubner and I have been planning. I
havent said anything to Ron yet, but Im telling you now to see if you can help me in one matter. Ron really and truly is the brains behind
this colloquiumbelieve me. I initiated matters as an Academy member, but the first thing I did was bring Ron on board. My objective all
along has been to do what I could as an Academy member to help promote the field of pain. So I hope that if I am unable to participate
much longer in the planning and am not around for the colloquium itself that Ron . . . will be allowed to continue and actually hold it.
Sorry for the bad news. John. Before he died, John received reassurances that this conference would go on without him.
John Coleman Liebeskind, bringing together again, today, his two loves, words and people, for the benefit of the field of pain. Enjoy
the meeting.
I would like to thank Timothy Cannon, Deborah Colbern, Darryl Dearmore, Ronald Dubner, Michael Gold, Russell Johnson, James
Lewis, Julia Liebeskind, Marcia Meldrum, Michael Morgan, Wendy Sternberg, and the Louise M. Darling Biomedical Library at UCLA
(where the John C. Liebeskind History of Pain collection is housed) for their help in preparing this manuscript.
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SODIUM CHANNELS AND PAIN 7635

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Sodium channels and pain

(dorsal root ganglion neurons/hyperexcitability/ion channels/nerve injury/inflammation)

S. G. WAXMAN * , S. DIB-HAJJ, T. R. CUMMINS, AND J. A. BLACK
Department of Neurology, Yale University School of Medicine, New Haven, CT 06510; and Paralyzed Veterans of America/Eastern
Paralyzed Veterans Association Neuroscience Research Center, Veterans Administration Medical Center, West Haven, CT 06516
ABSTRACT Although it is well established that hyperexcitability and/or increased baseline sensitivity of primary sensory
neurons can lead to abnormal burst activity associated with pain, the underlying molecular mechanisms are not fully understood.
Early studies demonstrated that, after injury to their axons, neurons can display changes in excitability, suggesting increased
sodium channel expression, and, in fact, abnormal sodium channel accumulation has been observed at the tips of injured axons.
We have used an ensemble of molecular, electrophysiological, and pharmacological techniques to ask: what types of sodium
channels underlie hyperexcitability of primary sensory neurons after injury? Our studies demonstrate that multiple sodium
channels, with distinct electrophysiological properties, are encoded by distinct mRNAs within small dorsal root ganglion (DRG)
neurons, which include nociceptive cells. Moreover, several DRG neuron-specific sodium channels now have been cloned and
sequenced. After injury to the axons of DRG neurons, there is a dramatic change in sodium channel expression in these cells, with
down-regulation of some sodium channel genes and up-regulation of another, previously silent sodium channel gene. This
plasticity in sodium channel gene expression is accompanied by electrophysiological changes that poise these cells to fire
spontaneously or at inappropriate high frequencies. Changes in sodium channel gene expression also are observed in experimental
models of inflammatory pain. Thus, sodium channel expression in DRG neurons is dynamic, changing significantly after injury.
Sodium channels within primary sensory neurons may play an important role in the pathophysiology of pain.
Pain pathways begin with primary sensory neurons [dorsal root ganglion (DRG) neurons; trigeminal neurons]. It is now clear that, in
some pain syndromes, hyperexcitability and/or increased baseline sensitivity of these cells leads to abnormal bursting that can produce
chronic pain ( 1 3 ). The pivotal position of primary sensory neurons as distal sites of impulse generation along the nociceptive pathway,
and the experimental and clinical accessibility of these neurons, has resulted in intense interest in mechanisms underlying action potential
generation and transmission in them in disease states characterized by pain. Voltage-gated sodium channels, which produce the inward
membrane current necessary for regenerative action potential production within the mammalian nervous system, are, of course, expressed
in primary sensory neurons and have emerged as important targets in the study of the molecular pathophysiology of pain and in the search
for new pain therapies. In this paper we focus on the potential role of sodium channels in the molecular pathophysiology of pain. We will
emphasize, in particular, three motifs: first, that DRG neurons express a complex repertoire of multiple distinct sodium channels, encoded
by different genes; second, that some of these sodium channels are sensory neuron specific; and third, that sodium channel expression in
DRG neurons is highly dynamic, changing substantially not only during development, but also in various disease states, including some
that are accompanied by pain.
HYPEREXCITABILITY IN DRG CELLS AFTER INJURY
Early studies ( 4 , 5 ) demonstrated that, after injury to their axons, motor neurons display changes in excitability, suggesting
increased sodium channel expression over the cell body and the dendrites, and similar changes were subsequently observed in sensory
neurons ( 6 , 7 ). Abnormal sodium channel accumulation at the tips of injured axons also has been observed ( 8 10 ), and both
electrophysiological and computer simulation studies have suggested that abnormal increases in sodium conductance can lead to
inappropriate, repetitive firing ( 11 13 ). Indeed, there is substantial evidence indicating that the abnormal excitability of DRG neurons,
after axonal injury, is associated with an increased density of sodium channels ( 13 , 14 ). These observations, together with experimental
and clinical observations on partial efficacy of sodium channel-blocking agents in neuropathic pain ( 15 18 ), established a link between
sodium channel activity and sensory neuron hyperexcitability producing pain. However, these studies did not examine the crucial question:
what type(s) of sodium channels produce inappropriate sensory neuron discharge associated with pain?
MULTIPLE SODIUM CHANNELS IN PRIMARY SENSORY NEURONS
Over the past decade, it has become clear that nearly a dozen, molecularly distinct voltage-gated sodium channels are encoded within
mammals by different genes. DRG neurons, which had been known to display multiple, distinct sodium currents ( 19 22 ), express at
least six sodium channel transcripts ( 23 ), as illustrated by the in situ hybridizations and reverse transcriptionPCR shown in Figs. 1 and
2 . These include high levels of expression of the -I and Na6 channels, also expressed at high levels by other neuronal cell types within
the central nervous system, which are known to support tetrodotoxin (TTX)-sensitive sodium currents. In addition, DRG neurons are
unique in expressing four additional sodium channel transcripts that are not expressed at significant levels in other neuronal cell types: (i)
PN1/hNE, which is expressed preferentially in DRG neurons ( 24 ), produces a fast, transient TTX-sensitive sodium current in response to
sudden depolar
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: DRG, dorsal root ganglia; TTX; tetrodotoxin; NGF, nerve growth factor.
* To whom reprint requests should be addressed at: Department of Neurology, LCI 707, Yale Medical School, 333 Cedar Street, New
Haven, CT 06510. e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SODIUM CHANNELS AND PAIN 7636

izations and a persistent current elicited by slow depolarizations close to resting membrane potential ( 25 ); (ii) SNS/PN3, expressed
preferentially in small DRG and trigeminal neurons, encodes a TTX-resistent sodium current ( 26 , 27 ); (iii) NaN, expressed preferentially
in small and trigeminal neurons, exhibits an amino acid sequence that, although only 47% similar to SNS-PN3, predicts that it encodes a
TTX-resistant sodium channel ( 28 ); and (iv) NaG, another putative sodium channel that was originally cloned from astrocytes and at first
thought to be glial specific ( 29 ), is also preferentially expressed at high levels within DRG neurons ( 23 ) and at low levels within other
neurons of neural crest origin but not within other neuronal types ( 30 ).

FIG. 1. Sodium channel -subunit mRNAs visualized in sections from adult rat DRG by in situ hybridization with subtype-specific
antisense riboprobes. mRNAs for -I, Na6, hNE/PN1, SNS, NaN, and NaG are present at moderate to high levels in DRG neurons.
Hybridization signal is not present with sense riboprobes, e.g., for NaG (S). (Bar indicates 100 m.)

FIG. 2. Restriction enzyme profile analysis of Na channel domain 1 reverse transcriptionPCR products from DRG. M lanes contain
100-bp ladder marker. Lane 1 contains the amplification product from DRG cDNA. Lanes 29 show the result of cutting this DNA
with EcoRV, EcoN1, AvaI, SphI, BamHI, AflII, XbaI, and EcoRI, which are specific to subunits -I, -II, -III, Na6, PN1, SNS, NaG,
and NaN, respectively. Reproduced with permission from ref. 28 . (Copyright 1998, National Academy of Sciences, USA).
Preferential expression of SNS/PN3 and NaN within small DRG neurons provides a molecular correlate for the observation ( 19
22 , 32 , 33 ) that these cells express several distinct sodium currents, including TTX-resistant sodium currents. A role for TTX-resistant
sodium channels in action potential conduction along small diameter afferent fibers has been postulated ( 34 ), and TTX-resistant sodium
potentials have, in fact, been recorded from unmyelinated C-fibers ( 35 ).
Preferential expression of SNS/PN3 and NaN in small DRG neurons, which include nociceptive cells, and the demonstration of a role
of TTX-resistant sodium currents in conduction along their axons, have suggested that these channels may represent unique targets for the
pharmacologic treatment of pain. PN1 and NaG also may represent useful molecular targets for the pharmacologic manipulation of DRG
neurons because of their preferential expression in these cells.
SODIUM CHANNEL GENE EXPRESSION IS ALTERED AFTER INJURY TO DRG NEURONS
The first observations indicating that, in addition to production of excess channels, there is a switch in the type of channels produced
after axonal injury were provided by Waxman et al. ( 36 ), who found a significant up-regulation of expression of the previously silent -
III sodium channel gene in DRG neurons
the authoritative version for attribution.

FIG. 3. Transcripts for sodium channel -III (A) are up-regulated, neurons after transection of their axons within the sciatic nerve.
The and transcripts for SNS (B) and NaN (C) are down-regulated, in DRG micrographs (Right) show in situ hybridizations in control
DRG, and at 57 days postaxotomy. Reverse transcriptionPCR (Left) shows products of coamplification of -III (A) and SNS (B)
together with -actin transcripts in control (C) and axotomized (A) DRG (days postaxotomy indicated above gels in A and B), with
computerenhanced images of amplification products shown below gels. Coamplification of NaN (392 bp) and glyceraldehyde-3-
phosphate dehydrogenase (GAPDH) (606 bp) (C) shows decreased expression of NaN mRNA at 7 days postaxotomy (lanes 2, 4, and
6) compared with controls (lanes 1, 3, and 5). A and B modified from ref. 37 ; C modified from ref. 28 . (Copyright 1998, National
Academy of Sciences, USA).
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SODIUM CHANNELS AND PAIN 7637

after axotomy. This finding was followed by demonstration of down-regulation of the SNS/PN3 gene expression, which can persist as long
as 210 days after axotomy ( 37 ), and of down-regulation of the NaN gene ( 28 ). These changes are illustrated in Fig. 3 .

FIG. 4. TTX-resistant sodium currents in small DRG neurons are down-regulated after axotomy. (A and B, Left) Whole-cell patch-
clamp recordings from representative control (A) and axotomized (B, 6 days postaxotomy) DRG neurons. Note the loss of the TTX-
resistant slowly inactivating component of sodium current after axotomy. Steady-state inactivation curves (A and B, Right) show loss
of a component characteristic of TTX-resistant currents. (C) Attenuation of TTX-resistant current persists for at least 60 days
postaxotomy. (D) Cell capacitance, which provides a measure of cell size, does not change significantly after axotomy (modified
from ref. 39 ).
PHYSIOLOGIC CHANGES ACCOMPANY ALTERED SODIUM GENE EXPRESSION AFTER DRG NEURON
INJURY
On the basis of the down-regulation of SNS/PN3 and NaN genes in DRG neurons after axonal transection, it would be expected that
TTX-resistant sodium currents should be reduced in these cells after axotomy. Patch-clamp studies have demonstrated that, indeed, there is
a loss of TTX-resistant sodium currents in DRG neurons after axonal transection ( 38 ); this down-regulation persists in small DRG
neurons for at least 60 days ( 39 ), consistent with the long-lasting changes in gene expression that have been described ( 37 ) in these cells
( Fig. 4 ). In addition, as shown in Fig. 5 , there is a switch in the properties of the TTX-sensitive sodium currents in these cells after
axotomy, with the emergence of a rapidly repriming current (i.e., a current that recovers rapidly from inactivation) ( 39 ). Cummins and
Waxman ( 39 ) have suggested that the type III sodium channel is responsible for the rapidly repriming sodium current, but this conjecture
remains to be proven.
These changes may poise DRG neurons to fire spontaneously, or at inappropriately high frequencies, after injury. Increased sodium
channel densities, in themselves, will tend to lower threshold ( 12 ). In addition, Rizzo et al. ( 40 ) have pointed out that the overlap
between steady-state activation and inactivation curves, together with weak voltage dependence of TTX-resistant sodium channels may
confer instability on the neuronal membrane. Coexpression of abnormal combinations of several types of channels, whose window
currents can bracket each other, would be expected to permit subthreshold ocillations in voltage, supported by TTX-resistant channels, to
cross-activate other sodium channels, thereby producing spontaneous activity ( 40 ). Cummins and Waxman ( 39 ) noted that, because the
TTX-sensitive sodium current in DRG neurons after axotomy reprimes relatively rapidly, injured neurons would be expected to sustain
higher firing frequencies. Moreover, if persistent currents participate in setting the resting potential, as demonstrated in optic nerve axons (
41 ), loss of TTX-resistant currents in DRG neurons after axotomy could produce a hyperpolarizing shift in resting potential, which, by
relieving resting inactivation, might increase the amount of TTX-sensitive sodium current available for electrogenesis.
NEUROTROPHINS MODULATE SODIUM CHANNEL EXPRESSION IN DRG NEURONS
A number of studies have suggested that, in response to nerve or tissue injury, there are changes in synthesis or delivery of various
neurotrophins to neurons. Early studies in culture demonstrated that nerve growth factor (NGF) can affect sodium channel expression in
DRG neurons ( 42 , 43 ). Black et al. ( 44 ) showed that NGF, delivered directly to DRG cell bodies, acts to down-regulate -III mRNA
and maintain high levels of SNS/PN3 mRNA expression in small DRG neurons in an in vitro model that mimics axotomy. Following up
on these observations, Dib-Hajj et al. ( 45 ) studied small DRG neurons in vivo after axotomy and demonstrated that administration of
exogenous NGF to the proximal nerve stump results in an up-regulation of TTX-resistant sodium current and of SNS/PN3 mRNA levels in
small DRG neurons ( Fig. 6 ). These observations suggest that at least some of the changes
the authoritative version for attribution.

FIG. 5. The kinetics of recovery from inactivation in TTX-sensitive sodium currents are different in axotomized DRG neurons. The
graph shows recovery of TTX-sensitive sodium current from inactivation as a function of time in DRG neurons after axonal
transection (6 and 22 days postaxotomy, results pooled) compared with uninjured controls. Note the leftward shift in the recovery
curve. Modified from ref. 39 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SODIUM CHANNELS AND PAIN 7638

observed in DRG neurons after axotomy reflect loss of access to peripheral pools of neurotrophic factors.

FIG. 6. Reverse transcriptionPCR (A), in situ hybridization (B), and patch-clamp recordings (C), showing partial rescue of SNS
mRNA and TTX-resistant sodium currents in axotomized DRG neurons after delivery of NGF to the proximal nerve stump. (A)
Coamplification of SNS (479 bp) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (666 bp) products in Ringers solution-
treated axotomized DRG (lanes 1, 2, 5, and 6) and NGF-treated axotomized DRG (lanes 3, 4, 7, and 8). The graph shows the increase
in SNS amplification product in NGF-treated DRG. (B) In situ hybridization showing down-regulation of SNS mRNA in DRG after
axotomy (axotomy + Ringers solution compared with control), and the partial rescue of SNS mRNA by NGF. (C) Representative
patch-clamp recordings showing partial rescue of slowly inactivating TTX-resistant sodium currents in axotomized DRG neurons
after exposure to NGF. Corresponding steady-state inactivation curves are shown below the recordings. Modified from ref. 45 .
Brain-derived growth factor has been studied and has been found not to alter sodium currents in DRG neurons, although it affects the
expression of -aminobutyric acid receptormediated currents in these cells ( 46 ). Glial-derived growth factor has been found to modulate
the expression of NaN in a subpopulation of small DRG neurons, which are known to express the ret receptor ( 53 ). Multiple
neurotrophins and growth factors have effects on DRG neurons, and it is likely that sodium channel expression in these cells reflects
combinatorial effects of multiple factors.
SODIUM CHANNEL EXPRESSION IN INFLAMMATORY PAIN MODELS
Several studies have demonstrated that inflammatory molecules such as prostaglandins and serotonin can modulate TTX-resistant
sodium currents in DRG neurons ( 47 ), possibly acting through a cyclic AMP-protein kinase A cascade ( 48 ). However, the question, of
whether sodium channel gene expression is affected in inflammatory models of pain had not been addressed. To understand the role of
sodium channels in inflammatory pain, we have carried out studies in the carageenan inflammatory pain model in the rat ( 49 ). In these
studies, carried out before our cloning of NaN, we focused on SNS/PN3 because its expression was known to be labile. Based on our
previous observation in which we detected peak changes in SNS/PN3 mRNA 5 days after axotomy ( 37 ), we studied rats in the subacute
phase, 4 days after injection of carageenan into the hind paw. As shown in Fig. 7 , these experiments demonstrated significantly increased
SNS/PN3 mRNA expression in DRG neurons projecting to the inflamed limb, compared with DRG neurons from the contralateral side or
naive (uninjected) controls. Moreover, our patch-clamp recordings demonstrated that the amplitude of the TTXresistant sodium current in
small DRG neurons projecting to the inflamed limb was significantly larger than on the contralateral side 4 days postinjection (31.7 3.3
vs. 20.0 2.1 nA). The TTX-resistant current density was also significantly increased in the carageenan-challenged DRG neurons.
Consistent with these results, a persistent increase in sodium channel immunoreactivity is observed in DRG neurons within 24 hr of
injection of complete Freunds adjuvant into their projection field and persists for at least 2 months ( 50 ). The mechanism responsible for
this inflammation-associated change in sodium channel expression is not known. Interestingly, NGF normally is produced in peripheral
target tissues by supporting cells that include fibroblasts, Schwann cells, and keratinocytes; NGF production is stimulated in immune cells,
and increased NGF levels have been observed in the local area after treatment with inflammatory agents such as carageenan and Freunds
adjuvant ( 51 , 52 ), raising the possibility that inflammation may indirectly trigger changes in sodium channel gene expression via
changes in neurotrophin levels.
the authoritative version for attribution.

FIG. 7. SNS mRNA levels and TTX-resistant sodium currents are increased 4 days after injection of carrageenan into the projection
fields of DRG neurons. (Upper) In situ hybridization showing SNS mRNA in carrageenan-injected (A), contralateral control (B), and
naive (C) DRG. Patch-clamp recordings (DF) do not reveal any change in voltage dependence of activation or steady-state
inactivation of TTX-resistant sodium currents after carrageenan injection, but demonstrate an increase in TTX-resistant current
amplitude (D) and density. Modified from ref. 49 .
SODIUM CHANNELS AS MOLECULAR TARGETS IN PAIN RESEARCH
Given what we have learned about sodium channels, where do we go next in the search for better treatments for pain syndromes? The
answer to this question is not entirely clear at this time. We can, however, come to a number of conclusions. First, sodium channels are
important participants in electro
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SODIUM CHANNELS AND PAIN 7639

genesis within primary sensory neurons, including DRG neurons. Second, a multiplicity of sodium channels are present within DRG
neurons, where they probably subserve multiple functions (transduction, signal amplification, action potential electrogenesis, etc.) and
interact in a complex manner. Third, DRG neurons express a number of sodium channel genes (SNS/PN3, NaN, PN1, and NaG) in a
preferential manner, at levels much higher than in any other neuronal cell type. This observation may present a therapeutic opportunity for
the selective manipulation of primary sensory neurons in general, or nociceptive neurons in particular. Fourth, sodium channel expression
in DRG neurons is highly dynamic, with multiple sodium channel genes (including -III, SNS/PN3, and NaN) exhibiting up- or down-
regulation after various injuries to these cells. Importantly, different injuries may trigger opposing changes of certain sodium channel
genes (e.g., down-regulation of SNS/PN3 after axotomy vs. up-regulation in the carageenan inflammation model) in DRG neurons, so that
it may be difficult to extrapolate from one model system to another. Nevertheless, we have learned, at a minimum, that sodium channel
expression in DRG neurons is dynamic and can change significantly after injury, and that changes in sodium channel expression can
substantially alter excitability in these cells.
Delineation of the precise role(s) of each sodium channel subtype in the physiology of DRG neurons and the pathophysiology of pain
remains to be established, and the utility of selective blockade of each channel subtype as an approach to the treatment of pain will require
further careful study. However, the stage has been set for these investigations. It is quite likely, in our opinion, that sodium channel
blockade will emerge as a viable strategy for pharmacologic treatment of pain.
This work has been supported in part by grants from the National Multiple Sclerosis Society and the Paralyzed Veterans of America/
Eastern Paralyzed Veterans Association, and by the Medical Research Service, Department of Veterans Affairs. T.R.C. was supported in
part by a fellowship from the Spinal Cord Research Foundation.
1. Ochoa, J. & Torebjork, H. E. ( 1980 ) Brain 103 , 835854 .
2. Nordin, M. , Nystrom, B. , Wallin, U. & Hagbarth, K.-E. ( 1984 ) Pain 20 , 231245 .
3. Devor, M. ( 1994 ) in Textbook of Pain , eds. Wall, P. D. & Melzack, R. ( Churchill Livingstone , Edinburgh ), 2nd Ed. , pp. 79101 .
4. Eccles, J. C , Libet, B. & Young, R. R. ( 1958 ) J. Physiol. (London) 143 , 1140 .
5. Kuno, M. & Llinas, R. ( 1970 ) J. Physiol. (London) 210 , 807821 .
6. Gallego, R. , Ivorra, I. & Morales, A. ( 1987 ) J. Physiol. (London) 391 , 3956 .
7. Gurtu, S. & Smith, P. A. ( 1988 ) J. Neurophysiol. 59 , 408423 .
8. Devor, M. , Keller, C. H. , Deerinck, T. J. & Ellisman, M. H. ( 1989 ) Neurosci. Lett. 102 , 149154 .
9. England, J. D. , Gamboni, F. , Ferguson, M. A. & Levinson, S. R. ( 1994 ) Muscle Nerve 17 , 593598 .
10. England, J. D. , Happel, L. T. , Kline, D. G. , Gamboni, F. , Thouron, C. L. , Liu, Z. P. & Levinson, S. R. ( 1996 ) Neurology 47 , 272276 .
11. Waxman, S. G. & Brill, M. H. ( 1978 ) J. Neurol . Neurosurg. Psychiatr . 41 , 408417 .
12. Matzner, O. & Devor, M. ( 1992 ) Brain Res. 597 , 9298 .
13. Matzner, O. & Devor, M. ( 1994 ) J. Neurophysiol. 72 , 349359 .
14. Zhang, J.-M. , Donnelly, D. F. , Song, X.-J. & LaMotte, R. H. ( 1997 ) J. Neurophysiol. 78 , 27902794 .
15. Chabal, C. , Russell, L. C. & Burchiel, K. J. ( 1989 ) Pain 38 , 333338 .
16. Devor, M. , Wall, P. D. & Catalan, N. ( 1992 ) Pain 48 , 261268 .
17. Omana-Zapata, I. , Khabbaz, M. A. , Hunter, J. C. & Bley, K. R. ( 1997 ) Brain Res. 771 , 228237 .
18. Rizzo, M. A. ( 1997 ) J. Neurol. Sci. 152 , 103106 .
19. Kostyuk, P. G. , Veselovsky, N. S. & Tsyandryenko, A. Y. ( 1981 ) Neuroscience 6 , 24232430 .
20. Roy, M. L. & Narahashi, T. ( 1992 ) J. Neurosci. 12 , 21042111 .
21. Caffrey, J. M. , Eng, D. L. , Black, J. A. , Waxman, S. G. & Kocsis, J. D. ( 1992 ) Brain Res. 592 , 283297 .
22. Elliott, A. A. & Elliott, J. R. ( 1993 ) J. Physiol. (London) 463 , 3956 .
23. Black, J. A. , Dib-Hajj, S. , McNabola, K. , Jeste, S. , Rizzo, M. A. , Kocsis, J. D. & Waxman, S. G. ( 1996 ) Mol. Brain Res. 43 , 117132 .
24. Toledo-Aral, J. J. , Moss, B. L. , He, Z.J. , Koszowski, A. G. , Whisenand, T. , Levinson, S. R. , Wolff, J. J. , Silos-Santiago, I. , Halegoua, S. &
Mandel, G. ( 1997 ) Proc. Natl. Acad. Sci. USA 94 , 15271532 .
25. Cummins, T. R. , Howe, J. R. & Waxman, S. G. ( 1999 ) J. Neurosci. 18 , 96079619 .
26. Akopian, A. N. , Sivilotti, L. & Wood, J. N. ( 1996 ) J. Biol. Chem. 271 , 59535956 .
27. Sangameswaran, L. , Delgado, S. G. , Fish, L. M. , Koch, B. D. , Jakeman, L. B. , Stewart, G. R. , Sze, P. , Hunter, J. C. , Eglen, R. M. & Herman, R.
C. ( 1996 ) J. Biol. Chem. 271 , 59535956 .
28. Dib-Hajj, S. D. , Tyrrell, L. , Black, J. A. & Waxman, S. G. ( 1998 ) Proc. Natl. Acad. Sci. USA 95 , 89638969 .
29. Gautron, S. , Dos Santos, G. , Pinto-Henrique, D. , Koulkoff, A. , Gros, F. & Berwald-Netter, Y. ( 1992 ) Proc. Natl. Acad. Sci. USA 89 , 72727276 .
30. Felts, P. A. , Black, J. A. , Dib-Hajj, S. D. & Waxman, S. G. ( 1997 ) Glia 21 , 269277 .
31. Rizzo, M. A. , Kocsis, J. D. & Waxman, S. G. ( 1995 ) J. Neurophysiol. 72 , 27962816 .
32. Rush, A. M. , Brau, M. E. , Elliott, A. A. & Elliott, J. R. ( 1998 ) J. Physiol. (London) 511 , 771789 .
33. Scholz, A. , Appel, N. & Vogel, W. ( 1998 ) Eur. J. Neurosci. 10 , 25472556 .
34. Jeftinija, S. ( 1994 ) Brain Res. 639 , 125134 .
35. Quasthoff, S. , Grosskreutz, J. , Schroder, J. M. , Schneider, U. & Grafe, P. ( 1995 ) Neuroscience 69 , 955965 .
36. Waxman, S. G. , Kocsis, J. K. & Black, J. A. ( 1994 ) J. Neurophysiol. 72 , 466471 .
37. Dib-Hajj, S. , Black, J. A. , Felts, P. & Waxman, S. G. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 1495014954 .
38. Rizzo, M. A. , Kocsis, J. D. & Waxman, S. G. ( 1995 ) Neurobiol. Dis. 2 , 8796 .
39. Cummins, T. R. & Waxman, S. G. ( 1997 ) J. Neurosci. 17 , 35033514 .
40. Rizzo, M. A. , Kocsis, J. D. & Waxman, S. G. ( 1996 ) Eur. Neurol. 36 , 312 .
41. Stys, P. K. , Ransom, B. R. & Waxman, S. G. ( 1993 ) Proc. Natl. Acad. Sci. USA 90 , 69766980 .
42. Aguayo, L. G. & White, G. ( 1992 ) Brain Res. 570 , 6167 .
43. Zur, K. B. , Oh, Y. , Waxman, S. G. & Black, J. A. ( 1995 ) Mol. Brain Res. 30 , 97103 .
44. Black, J. A. , Langworthy, K. , Hinson, A. W. , Dib-Hajj, S. D. & Waxman, S. G. ( 1997 ) NeuroReport 8 , 23312335 .
45. Dib-Hajj, S. D. , Black, J. A. , Cummins, T. R. , Kenney, A. M. , Kocsis, J. D. & Waxman, S. G. ( 1998 ) J. Neurophysiol. 79 , 26682678 .
46. Oyelese, A. A. , Rizzo, M. A. , Waxman, S. G. & Kocsis, J. D. ( 1997 ) J. Neurophysiol. 78 , 3142 .
47. Gold, M. S. , Reichling, D. B. , Shuster, M. J. & Levine, J. D. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 11081112 .
48. England, S. , Bevan, S. & Docherty, R. J. ( 1996 ) J. Physiol. (London) 495 , 429440 .
49. Tanaka, M. , Cummins, T. R. , Ishikawa, K. , Dib-Hajj, S. D. , Black, J. A. & Waxman, S. G. ( 1998 ) NeuroReport 9 , 967972 .
50. Gould, H. J. III , England, J. D. , Liu, Z. P. & Levinson, S. R. ( 1998 ) Brain Res. 802 , 6974 .
the authoritative version for attribution.

51. Woolf, C. J. , Safieh-Garabedian, B. , Ma, Q.-P., Crilly, P. & Winter, J. ( 1994 ) Neuroscience 62 , 327331 .
52. Weskamp, G. & Otten, U. ( 1987 ) J. Neurochem. 48 , 17791786 .
53. Fzell, J. , Cummins, P. R. , Dib-Hajj, S. D. , Fried, K. , Black, J. A. & Waxman, S. G. ( 1999 ) Mol. Brain Res. 67 , 267282 .
 A COMPARISON OF THE POTENTIAL ROLE OF THE TETRODOTOXIN-INSENSITIVE SODIUM CHANNELS, PN3/SNS AND NAN/ 7640
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SNS2, IN RAT MODELS OF CHRONIC PAIN

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

A comparison of the potential role of the tetrodotoxin-insensitive

sodium channels, PN3/SNS and NaN/SNS2, in rat models of chronic
pain

FRANK PORRECA * , JOSEPHINE LAI *, DI BIAN *, SANDRA WEGERT *, MICHAEL H. OSSIPOV *, RICHARD M. EGLEN, LAURA KASSOTAKIS ,
SANJA NOVAKOVIC, DOUGLAS K. RABERT , LAKSHMI SANGAMESWARAN, AND JOHN C. HUNTER
* Department of Pharmacology, University of Arizona Health Sciences Center, Tucson, AZ 85724; and Center for Biological

Research, Roche Bioscience, Palo Alto, CA 94304

ABSTRACT Alterations in sodium channel expression and function have been suggested as a key molecular event underlying
the abnormal processing of pain after peripheral nerve or tissue injury. Although the relative contribution of individual sodium
channel subtypes to this process is unclear, the biophysical properties of the tetrodotoxin-resistant current, mediated, at least in
part, by the sodium channel PN3 (SNS), suggests that it may play a specialized, pathophysiological role in the sustained, repetitive
firing of the peripheral neuron after injury. Moreover, this hypothesis is supported by evidence demonstrating that selective
knock-down of PN3 protein in the dorsal root ganglion with specific antisense oligodeoxynucleotides prevents hyperalgesia and
allodynia caused by either chronic nerve or tissue injury. In contrast, knock-down of NaN/SNS2 protein, a sodium channel that
may be a second possible candidate for the tetrodotoxin-resistant current, appears to have no effect on nerve injury-induced
behavioral responses. These data suggest that relief from chronic inflammatory or neuropathic pain might be achieved by selective
blockade or inhibition of PN3 expression. In light of the restricted distribution of PN3 to sensory neurons, such an approach might
offer effective pain relief without a significant side-effect liability.
Spontaneous and/or evoked hyperexcitability of the peripheral nerve after injury is considered to be a principal feature of the
underlying pathophysiology associated with many chronic, in particular neuropathic, pain syndromes ( 1 , 2 ). A prominent molecular
basis for this abnormal, repetitive firing of injured primary afferents is an accumulation and increased membrane density of sodium
channels at focal sites of injury ( 3 , 4 ). The resultant membrane remodeling contributes to a lower threshold for action potential
generation at these sites and, consequently, precipitates ectopic impulse generation ( 5 , 6 ). Further, sodium channel blockade with
subanesthetic doses 5of a local anesthetic suppresses ectopic electrogenesis and may account for the analgesic effectiveness of these
agents ( 7 ). At present, the relative contribution of individual sodium channel subtypes toward this altered processing of sensory input
remains unclear. In the dorsal root ganglion (DRG), two main types of sodium currents, termed TTX-sensitive (TTX-S) and TTX-resistant
(TTX-R), have been identified on the basis of their kinetics and sensitivity to the neurotoxin, tetrodotoxin (TTX) ( 8 , 9 , 10 ). The fast-
inactivating, TTX-S current, found in all types of DRG cells, may be mediated by one or more of several -subunits known to be
expressed in these cells: brain types I, IIA, III ( 11 ), PN1 ( 12 , 13 ), and NaCh6 ( 14 ) [also known as either SCN8A ( 15 ) or PN4 ( 16 )].
In contrast, in normal adult DRG neurons, the more slowly inactivating TTX-R current appears to be preferentially expressed in a
subpopulation of small diameter, unmyelinated, capsaicin-sensitive neurons, otherwise referred to as nociceptors ( 8 , 9 , 17 ). Until
recently, only a single sodium channel -subunit, PN3 ( 18 ), also known as sensory neuron specific or SNS ( 19 ), had been identified that
displayed the biophysical properties, resistance to TTX, and anatomical distribution of the TTX-R current ( 18 , 19 , 20 ). However,
multiple types of TTX-R current, termed TTX-R1, R2, and R3, have now been suggested to be present in the small diameter neurons of
the adult rat DRG ( 21 ). Although the biophysical properties of PN3 make it a likely candidate for TTX-R1, the most abundant form of
TTX-R current, a second type of novel sodium channel was recently cloned from rat DRG, termed NaN ( 22 ). NaN, also referred to as
SNS2 ( 23 ), appears to be preferentially localized to an even more restricted subpopulation of small diameter sensory neurons within the
DRG ( 22 , 23 ). In comparison with PN3/SNS, it has intermediate resistance to TTX (1 M), and its biophysical properties ( 23 ) suggest
it may be a possible candidate for the TTX-R3 current ( 21 ). The physiological and/or pathophysiological role of NaN/SNS2 remains to
be elucidated, but it is possible that changes in the expression and function of this channel, in addition to PN3/SNS, may make an
important contribution to the establishment of certain chronic pain states. Further, implication of either channel has the additional
importance that selective blockade may produce pain relief in these states without many of the limiting central nervous system and other
side-effects associated with current therapies.
EVIDENCE FOR A ROLE FOR PN3 IN THE MEDIATION OF ABNORMAL PAIN BEHAVIORS AFTER NERVE
AND TISSUE INJURY
The selective expression of TTX-R INa, as well as PN3/SNS and NaN/SNS2, in a specific subpopulation of capsaicinsensitive,
primary afferent neurons suggests that these channels may play a crucial role in the regulation of sensory, nociceptive function. Moreover,
the rapid repriming properties, in addition to the higher threshold and slower rate of inactivation, of TTX-R INa (i.e., TTX-R1) and PN3/
SNS, further suggest that cells expressing a large proportion of TTX-R1 sodium channels should be ideally suited to sustain
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: TTX-R, tetrodotoxin-resistant; TTX-S; TTX-sensitive; DRG, dorsal root ganglion; SNL, spinal nerve ligation; CFA,
complete Freunds adjuvant; ODN, oligodeoxynucleotide; MM, mismatch; AS, antisense; CAR, carrageenan; PGE2, prostaglandin E2.
To whom reprint requests should be addressed, e-mail: [email protected] .
 A COMPARISON OF THE POTENTIAL ROLE OF THE TETRODOTOXIN-INSENSITIVE SODIUM CHANNELS, PN3/SNS AND NAN/ 7641
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SNS2, IN RAT MODELS OF CHRONIC PAIN

repetitive firing at the depolarized potentials characteristic of an injured peripheral nerve; that is, they will be slowly adapting in response
to a persistent, depolarizing stimulus ( 10 , 21 , 24 ).

FIG. 1. Immunohistochemical (peroxidase-diaminobenzidine) analysis of PN3 antibody labeling of L4 DRG cells from normal and
SNL animals, 7 days post-surgery. (A) Naive animals show a predominant labeling of the small diameter cells. In contrast, after SNL
injury (B), in animals that had received saline for 48 hr, a marked increase was observed in the number of large diameter cells
expressing PN3 in the L4 DRG ipsilateral to the side of injury. Animals treated with mismatch (MM) ODN (C) for 48 hr (day 7 post-
surgery) exhibited a similar pattern of labeling to the saline controls. In contrast, animals that had received antisense (AS) ODN (D)
for a similar time period to the MM ODN demonstrated a marked loss of PN3 immunolabeling in both small and large cells of the L4
DRG ipsilateral to the side of injury. (Bar = 50 m.)
Immunohistochemical Studies. Such a potentially specialized, pathophysiological role for TTX-R INa/PN3 has been supported by
alterations in channel distribution observed after sustained injury to the peripheral nerve. The effect, however, is complex and appears
dependent on the nature and degree of the injury. After axotomy, TTX-R INa is substantially reduced in the small cells of the DRG ( 25 ,
26 ) with the level of PN3/SNS mRNA ( 27 ) and protein ( 20 ) expression reduced in parallel. In contrast, in the chronic constriction
injury model of neuropathic pain in the rat, there was no significant change in the amplitude or IV relationship of either TTX-R or TTX-S
INa recorded from small-diameter DRG neurons at the maximum time of injury, i.e., 14 days post-surgery ( 20 ). This would suggest that
the chronic constriction injury type of injury has had little impact on the number of sodium channels present in the somal membrane at any
given time. However, an initial loss of PN3/SNS immunolabeling was observed from the DRG at all levels of the lumbar enlargement (L4-
L6), presumably from the predominantly intracellular pool, and correlated closely with a subsequent redistribution and accumulation of
channel protein within the peripheral nerve just proximal to the site of injury ( 20 ). The onset and subsequent reversal of PN3/SNS
channel redistribution appeared to correlate closely with temporal changes in behavioral thermal hyperalgesia and, morphologically, with
the damage and recovery of primary afferent fibers after this type of nerve ligation ( 28 ). A similar observation has been made in another
rat model of peripheral nerve injury in which the L5 and L6 dorsal spinal roots are tightly ligated, evoking behavioral signs of hyperalgesia
and allodynia ( 29 ). A loss of PN3/SNS protein was observed in the DRG at the level of L5 and L6 with a subsequent accumulation at the
injury site proximal to the ligatures (P. Mantyh, personal communication). However, although the expression of PN3 protein in the L5 and
L6 DRG decreases after spinal nerve ligation (SNL) injury, PN3 protein levels in the uninjured L4 DRG were preserved in small diameter
cells and significantly increased in the large diameter cells ( Fig. 1 ). It is possible to speculate that the increased number of large diameter
cells expressing PN3 protein after SNL may account, in part, for the observed tactile allodynia. Such observations provide additional
support for the involvement of uninjured primary afferents in adjacent segments of the sciatic nerve in mediating certain types of
neuropathic pain behaviors ( 30 ). The reasons for increased PN3 expression in large diameter cells in the non-injured L4 DRG after SNL
and decreases in PN3 expression in cell bodies of the injured L5/L6 DRG are unclear but may reflect the differential availability of factors
such as nerve growth factor, which is known to be released from peripheral tissues and to be retrogradely transported to the ganglion,
where it regulates mRNA expression of PN3 ( 31 ) as well as other sodium channels ( 32 , 33 ). However, it should be emphasized that the
extent to which the immunolabeling pattern observed in the L4 ganglion translates into axonal accumulation of the channel protein and/or

FIG. 2. Antisense (AS), but not mismatch (MM), ODN to PN3 prevents and reverses tactile allodynia (A) and thermal hyperalgesia
(B) after SNL. After determination of baseline (BL) responses, the rats received twice daily injections (45 g, intrathecal in a 5-l
volume) of either MM (open symbols) or AS (filled symbols) (arrow a). After 5 days of ODN pretreatment, rats were subjected to
either SNL (squares) or sham surgery (circles) (arrow b). The ODN injections were terminated after the afternoon injection on day
5 (arrow c), recommenced on the morning of day 12 (arrow d), and were terminated again after the afternoon injection on day 15
(arrow e). Tactile allodynia was indicated by a significant (P 0.05) reduction in paw withdrawal threshold to application of a
series of calibrated (0.415 g) von Frey filaments to the plantar surface of the hindpaw. Thermal hyperalgesia was indicated by a
significant (P 0.05) reduction in paw withdrawal latency to application of noxious radiant heat to the plantar surface of the affected
the authoritative version for attribution.

hindpaw of the nerve or sham-operated rats. A maximal cut-off of 40 sec was used to prevent tissue damage. n = 6 rats per group.
 A COMPARISON OF THE POTENTIAL ROLE OF THE TETRODOTOXIN-INSENSITIVE SODIUM CHANNELS, PN3/SNS AND NAN/ 7642
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SNS2, IN RAT MODELS OF CHRONIC PAIN

changes in the response thresholds to non-noxious sensory stimulation after nerve injury is unknown. In this regard, the possible
accumulation of the PN3 channel at the site of the injured nerve in the L5 or L6 fibers may be sufficient to generate sustained ectopic
discharge necessary to maintain the spinal cord in a sensitized state and to produce the well known changes in expression of central
proteins characteristic of the nerve-injured state. The importance of the contribution of the input from the injured fibers is well
documented as rhizotomy of the L5 and L6 dorsal roots blocks established SNL induced allodynia/ hyperalgesia ( 34 ).

FIG. 3. Immunohistochemical (peroxidase-diaminobenzidine) analysis of PN1 (AD) and PN4 (EH) antibody labeling of L4 DRG
cells from normal and SNL animals, 7 days post-surgery. (A) Naive animals show labeling of all cell types but with an increased
intensity observed in small diameter cells. (B) The pattern of PN1 immunolabeling remains unchanged by SNL injury in animals that
also received saline for 48 hr. SNL animals treated with either PN3 MM (C) or AS (D) ODNs exhibited a similar pattern of PN1
immunolabeling to the saline treated controls. Immunolabeling of L4 DRG cells with PN4 antibody followed a similar pattern to that
observed with the PN1 antibody. (E) PN4 immunolabeling of all cell types, in L4 DRGs taken from nave animals, but with an
increased intensity observed in small diameter cells. (F) Animals receiving a SNL injury and treated with saline showed a similar
PN4 immunolabeling pattern to the nave animals. SNL animals treated with either PN3 MM (G) or AS (H) ODNs exhibited a similar
pattern of PN4 immunolabeling to the saline treated controls. (Bar = 50 m.)
PN3 also has been implicated in the hyperesthesias that result from tissue injury. Thus, modulation of TTX-R INa, by a number of
naturally occurring hyperalgesic substances, e.g., prostaglandin E2 (PGE2), adenosine and serotonin, has been suggested to be a
mechanism that could underlie the subsequent increase in excitability and sensitization of sensory neurons mediated by these agents after a
peripheral nerve or tissue injury ( 35 , 36 , 37 ). Immunohistochemical studies have found that PN3 protein expression appears to increase
in the small-diameter DRG cells of the lumbar enlargement (L4-L6) after chronic inflammation induced by complete Freunds adjuvant
(CFA). In contrast, little evidence has been found for changes in PN3 protein levels after acute inflammation induced by either
carrageenan or formalin (P. Mantyh, personal communication).
Antisense Studies. In support of the immunohistochemical observations, evidence shows that spinal administration (45 g,
intrathecal, twice a day) of an antisense (5
-TCC-TCT-GTG-CTT-GGT-TCT-GGC-CT-3
), but not mismatch (5
-TCC-TTC-GTG-CTG-
TGT-TCG-TGC-CT-3
), oligodeoxynucleotide (ODN) to a unique sequence of PN3/SNS produces a selective and reversible block of
channel protein expression ( Fig. 1 ) in rats with spinal nerve (L5/L6) ligation and can prevent the behavioral thermal hyperalgesia and
tactile allodynia ( Fig. 2 ) evoked by this type of injury. Cessation of the antisense ODN treatment at any time after the SNL injury results
in the reoccurrence of tactile allodynia and thermal hyperalgesia within 48 hr, demonstrating the reversibility of the ODN effect. The
equivalent time course for the onset and reversibility of the antisense ODN effect is consistent with the reported 26-hr half life for the rate
of sodium channel turnover and biosynthesis ( 38 ). The lack of effect of the corresponding PN3 mismatch ODN, together with the lack of
effect on baseline (i.e., uninjured) responses to non-noxious or noxious stimuli, also suggests that this was not a nonspecific artifact caused
by repeated injections of an ODN. Intrathecal administration of ODNs previously have been demonstrated to alter the expression of
proteins in the spinal cord, the DRG, and peripheral nerve terminals ( 39 , 40 ). Further, the PN3 antisense, but not mismatch, ODN
affected the expression level of the PN3 channel protein in the DRG, but neither ODN had an effect ( Fig.3 ) on the TTX-S sodium
channels PN1 ( 12 , 13 ) and NaCh6/ SCN8A/PN4 ( 14 , 15 , 16 ). Pretreatment with PN3 antisense also appears to block the development
of both tactile allodynia and thermal hyperalgesia in rats treated with CFA but has no effect on the hyperesthesias evoked acutely by
carrageenan ( Fig. 4 ). In each case, the corresponding mismatch ODN was completely ineffective.

FIG. 4. Prevention of the development of tactile allodynia (A) or thermal hyperalgesia (B) after CFA, but not carrageenan (CAR)-
induced, inflammation by PN3 antisense (AS) ODN. Rats received twice-daily injections (45 g, intrathecal) of either mismatch
(MM) or AS for 2 days and again on the morning of day 3. On the afternoon of day 3, the rats received an injection of CFA (150 l)
or of CAR (200 l of 2%) to the hindpaw. Animals receiving CFA were tested 4 days afterward while receiving twice-daily
the authoritative version for attribution.

administration of AS and MM ODNs throughout this period. Rats receiving CAR were tested on the afternoon of day 3, 3 hr after
CAR injection. Tactile allodynia was indicated by a significant (*, P 0.05) reduction in paw withdrawal threshold in rats receiving
MM, but not AS, on the afternoon of the fourth day after CFA injection. Thermal hyperalgesia was indicated by a significant (*, P
0.05) reduction in paw withdrawal latency in rats receiving MM, but not AS, on the afternoon of the fourth day after CFA. All
treatment groups receiving CAR demonstrated significant (*, P 0.05) reductions in response thresholds to tactile stimuli (A) or
response latencies to thermal stimuli (B), and these responses were unaffected by either AS or MM ODNs. n = 6 rats per group.
 A COMPARISON OF THE POTENTIAL ROLE OF THE TETRODOTOXIN-INSENSITIVE SODIUM CHANNELS, PN3/SNS AND NAN/ 7643
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SNS2, IN RAT MODELS OF CHRONIC PAIN

FIG. 5. Antisense (AS) to NaN/SNS2 administered to rats with SNL injury produced no effect on either tactile allodynia (A) or
thermal hyperalgesia (B). Groups of six rats were used for each of the ODN or saline treatments and were monitored daily for tactile
(von Frey) and thermal nociceptive (radiant heat) responses of the ipsilateral hindpaw. Mismatch (MM) or AS ODN (45 g,
intrathecal) were given twice daily to sham-operated rats and rats with SNL injury. Neither MM nor AS reversed tactile allodynia or
thermal hyperalgesia in the ligated groups, and, likewise, neither treatment altered baseline values in the sham-operated groups. (),
sham-operated, AS; (), SNL, AS; () , sham-operated, MM; () , SNL, MM.
These data suggest that the behavioral consequences of SNL injury and chronic inflammation require de novo PN3 protein synthesis.
This view is borne out by noting that inhibition of PN3 protein expression by the antisense ODN does not affect normal noxious and non-
noxious sensory thresholds on the contralateral side to the nerve injury and fails to affect the allodynia and hyperalgesia seen after
carrageenan, an acute inflammation with a time-course that would not likely be associated with new channel protein synthesis. However,
prevention of protein synthesis by pretreatment with antisense ODN before, or during, chronic injuries, such as SNL or CFA, results in a
complete inhibition of the development of the behavioral consequences of the injury. It should be noted that the lack of effect of PN3
antisense ODN on the behavioral consequences of carrageenaninduced inflammation was studied only at the 3-hr time point after the
initiation of the inflammation. This was most probably insufficient time for PN3 gene transcription and translation to occur to a level that
would have had an impact on the inflammatory process. It is entirely possible that expression of PN3 might be regulated at a later time
point and that PN3 antisense ODN might be active under such circumstances, in a similar manner to the effect observed after CFA
treatment. In apparent contrast to these effects observed in vivo, it has been shown that the TTX-R current recorded from small diameter
cells of the DRG appears to be modulated rapidly in cell culture by inflammatory agents such as PGE2 ( 36 ). This may simply be related
to the much more rapid and sustainable local concentrations of the prostanoid achieved in vitro; that is, the time course remains to be
determined for such inflammatory-mediated increases in TTX-R current in vivo. Nevertheless, the PGE2 effect on the TTX-R INa in vitro
was partially reversed by application of an antisense ODN 21-mer, synthesized against a unique sequence of the PN3/SNS cDNA.
Moreover, the PN3 antisense ODN also reversed the behavioral mechanical hyperalgesia evoked by PGE2 intradermal administration into
the rat hindpaw ( 41 ).

FIG. 6. Immunohistochemical (peroxidase-diaminobenzidine) analysis of NaN/SNS2 antibody labeling of L4 DRG cells from normal
and SNL animals. (A) Naive animals show labeling of small diameter neurons with NaN/SNS2 antibody. No labeling is seen in the
preabsorbed control group (B). In contrast, sham-operated (C) or SNL (D) animals that had received AS ODN to NaN/SNS2 for 48 hr
(a total of four injections) demonstrated a marked loss of NaN/SNS2 immunolabeling in small cells of the L4 DRG ipsilateral to the
side of surgery. Labeling for NaN/SNS2 returned in both sham-operated (E) or SNL rats (F) that were perfused 4 days after the last
AS ODN to NaN/SNS2 injection. (Bar = 50 m.)
THE PATHOPHYSIOLOGICAL CONTRIBUTION OF NAN/SNS2 IN PERIPHERAL NERVE INJURY?
The novel sodium channel NaN/SNS2, recently cloned from rat DRG ( 22 , 23 ), has an even more restricted distribution within the
small diameter cell population of the DRG. This has led to its consideration as a second potential candidate for the TTX-R INa found in
these cells. However, in contrast to PN3, the low threshold for activation, fast rate of inactivation, and intermediate TTX sensitivity of
NaN/SNS2 resemble most closely the previously described properties of the cardiac channel and, more specifically, the TTX-R3 subtype
the authoritative version for attribution.

of INa ( 21 ). In chronic pain states, a potential role for TTX-R3 INa and, for that matter, NaN/SNS2 has not yet been elucidated, but, like
PN3/SNS, NaN/SNS2 mRNA levels in the DRG are markedly reduced after a peripheral nerve (sciatic) axotomy ( 23 ) and are elevated in
the small diameter DRG cells after persistent inflammation evoked by CFA ( 23 ).
In contrast to the findings with PN3 antisense, spinal administration of antisense (5
GCC TTG TCT TTG GAC TTC TTC 3
) and
mismatch (5
GCT CTG TTC TTG AGC TTT CTC 3
) ODN to NaN/SNS2 failed to produce any change in sensory thresholds after spinal
nerve (L5/L6) ligation injury ( Fig. 5 ) and did not alter gross behavior, as demonstrated by normal food intake, weight gain, and/or motor
performance, in spite of a significant knock-down of the NaN/SNS2 protein ( Fig. 6 ). This observation suggests, therefore, that the NaN/
SNS2 subtype of TTX-insensitive sodium channel appears unlikely to play a prominent role in the alterations in the sensory phenotype
that have been
 A COMPARISON OF THE POTENTIAL ROLE OF THE TETRODOTOXIN-INSENSITIVE SODIUM CHANNELS, PN3/SNS AND NAN/ 7644
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SNS2, IN RAT MODELS OF CHRONIC PAIN

proposed to contribute to the ongoing paresthesias and pain after peripheral nerve injury. NaN/SNS2 channels, like TTX-R3 INa, appear to
be activated, and possibly also inactivated, at much more hyperpolarized potentials than, for example, TTX-R1 INa/PN3. It is therefore
possible that the ineffectiveness of the NaN/SNS2 antisense may reflect the unavailability of NaN/SNS2 channels to contribute to
repetitive firing at the sustained level of membrane depolarization associated with injury to a peripheral nerve; that is, most will be in the
inactivated state ( 21 ).
The physiological or, indeed, pathophysiological role of NaN/ SNS2 therefore remains to be elucidated, particularly in persistent
inflammatory conditions, where an increased expression level of NaN/SNS2 protein in the small diameter cells of the DRG has been
recently reported ( 23 ). The lack of behavioral effects of NaN/SNS2 knock-down serves to validate that ODN administration does not
elicit changes in behaviorally determined thresholds to noxious and non-noxious sensory stimuli in normal or nerve-injured animals non-
specifically but, rather, that knockdown of a functionally important protein (i.e., PN3) is required for these effects to occur.
CONCLUSIONS
Collectively, therefore, the data suggest that the primary symptoms of neuropathic pain may be significantly attenuated by interfering
with the expression and, consequently, the function of PN3 but not of other sodium channels, which are mainly distributed within the DRG
such as NaN/SNS2. The PN3 antisense-mediated prevention of thermal hyperalgesia induced by CFA, but not by carrageenan, treatment
indicates that the clinical potential for a selective inhibitor of PN3 may extend to pain resulting from chronic tissue, as well as nerve,
injury. Such an effect is also consistent with recent evidence indicating that hyperalgesic substances associated with tissue injury can alter
the function of PN3 ( 35 , 36 ). The lack of an effect of the PN3 antisense on the noxious/non-noxious response thresholds of the
contralateral side to nerve and tissue injury and on the consequences of an acute inflammation (i.e., carrageenan), where the time-course of
the response makes new expression of the PN3 protein unlikely, suggests that this channel does not play a role in normal nociceptive
function. Consequently, these data strongly suggest that selective inhibition of PN3 will not result in changes in normal, nociceptive
function, although this clearly needs to be confirmed. The collective profile of the antisense ODNs against PN3 therefore implicates this
channel in the pathophysiology of pain after nerve and tissue injury. However, it will be important to determine whether PN3 antisense
administration will reverse an established injury in the same way that it has been found to prevent the injury. Moreover, although a role for
PN3 appears to be emerging in peripheral nerve injury, it will be intriguing to see whether this profile might expand to encompass some of
the other types of common neuropathies, e.g., metabolic and chemotherapy. The lack of any overt, particularly central nervous system,
adverse events with the antisense ODN was consistent with the previously described discrete localization of the channel to sensory nerve
fibers with an absence of staining in the central nervous system and cardiac tissue ( 20 ). It also reaffirms the potential that a selective
inhibitor of PN3 may be clinically analgesic, providing not only an improved therapeutic window over existing therapies, but offering
relief from neuropathic pain that is normally resistant to current therapies.
We thank Elda Tzoumaka, Dan Waligora, and Kimberly Wong for their expert technical assistance.
1. Devor, M. ( 1994 ) in Textbook of Pain , eds. Wall, P. D. & Melzack, R. ( Churchill Livingstone , Edinburgh ), pp. 79100 .
2. Woolf, C. J. ( 1994 ) in Textbook of Pain , eds. Wall, P. D. & Melzack, R. ( Churchill Livingstone , Edinburgh ), pp. 101112 .
3. Devor, M. , Govrin-Lippmann, R. & Angelides, K. ( 1993 ) J. Neurosci. 13 , 19761992 .
4. England, J. D. , Happel, L. T. , Kline, D. G. , Gamboni, F. Thouron, C. L. , Liu, Z. P. & Levinson, S. R. ( 1996 ) Neurology 47 , 272276 .
5. Wall, P. D. & Gutnick, M. ( 1974 ) Nature (London) 248 , 740743 .
6. Matzner, O. & Devor, M. J. ( 1994 ) J. Neurophysiol. 72 , 349359 .
7. Devor, M. , Wall, P. D. & Catalan, N. ( 1992 ) Pain 48 , 261268 .
8. Ogata, N. & Tatebayashi, H. ( 1993 ) J. Physiol. 466 , 937 .
9. Roy, M. L. & Narahashi, T. ( 1992 ) J. Neurosci. 12 , 21042111 .
10. Elliott, A. A. & Elliott, J. R. ( 1993 ) J. Physiol. 463 , 3956 .
11. Black, J. A. Dib-Hajj, S. , McNabola, K. , Jeste, S. , Rizzo, M. A. , Kocsis, J. D. & Waxman, S. G. ( 1996 ) Mol. Brain Res. 43 , 117132 .
12. Toledo-Aral, J. J. , Moss, B. L. , He, Z. J. , Koszowski, A. G. , Whisenand, T. , Levinson, S. R. , Wolf, J. J. , Silos-Santiago, I. , Halegoua, S. &
Mandel, G. ( 1997 ) Proc. Natl. Acad. Sci. USA 94 , 15271532 .
13. Sangameswaran, L. , Fish, L. M. , Koch, B. D. , Rabert, D. K. , Delgado, S. G. , Ilnicka, M. , Jakeman, L. B. , Novakovic, S. , Wong, K. , Sze, P. , et
al. ( 1997 ) J. Biol. Chem. 272 , 1480514809 .
14. Schaller, K. L. , Krzemien, D. M. , Yarowsky, P. J. , Krueger, B. K. & Caldwell, J. H. ( 1995 ) J. Neurosci. 15 , 32313242 .
15. Burgess, D. L. , Kohrman, D. C. , Galt, J. , Plummer, N. W. , Jones, J. M. , Spear, B. & Meisler, M. H. ( 1995 ) Nat. Genet . 10 , 461465 .
16. Dietrich, P. S. , McGivern, J. G. , Delgado, S. G. , Koch, B. D. , Eglen, R. M. , Hunter J. C. & Sangameswaran L. ( 1998 ) J. Neurochem. 70 , 2262
2272 .
17. Arbuckle, J. B. & Docherty, R. J. ( 1995 ) Neurosci. Lett. 185 , 7073 .
18. Sangameswaran, L. , Delgado, S. G. , Fish, L. M. , Koch, B. D. , Jakeman, L. B. , Stewart, G. R. , Sze, P. , Hunter, J. C. , Eglen, R. M. & Herman, R.
C. ( 1996 ) J. Biol. Chem. 271 , 59535956 .
19. Akopian, A. N. , Sivilotti, L. & Wood, J. N. ( 1996 ) Nature (London) 379 , 257262 .
20. Novakovic, S. D. , Tzoumaka, E. , McGivern, J. G. , Haraguchi, M. , Sangameswaran, L. , Gogas, K. R. , Eglen, R. M. & Hunter, J. C. ( 1998 ) J.
Neurosci. 18 , 21742187 .
21. Rush, A. M. , Brau, M. E. , Elliott, A. A. & Elliott, J. R. ( 1998 ) J. Physiol. 551 , 771789 .
22. Dib-Hajj, S. D. , Tyrrell, L. , Black, J. A. & Waxman, S. G. ( 1998 ) Proc. Natl. Acad. Sci. USA 95 , 89638968 .
23. Tate, S. , Benn, S. , Hick, C. , Trezise, D. , John, V. , Mannion, R. J. , Costigan, M. , Plumpton, C. , Grose, D. , Gladwell, Z. , et al. ( 1998 ) Nat.
Neurosci. 1 , 653655 .
24. Elliott, J. R. ( 1997 ) Brain Res 754 , 221226 .
25. Rizzo, M. A. , Kocsis, J. D. & Waxman, S. G. ( 1995 ) Neurobiol. Disease 2 , 8796 .
26. Cummins, T. R. & Waxman, S. G. ( 1997 ) J. Neurosci. 17 , 35033514 .
27. Dib-Hajj, S. , Black, J. A. , Felts, P. & Waxman, S. G. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 1495014954 .
28. Coggeshall, R. E. , Dougherty, P. M. , Pover, C. M. & Carlton, S. M. ( 1993 ) Pain 52 , 233242 .
29. Kim, S. H. & Chung, J. M. ( 1992 ) Pain 50 , 355363 .
30. Yoon, Y. W. , Na, H. S. & Chung, J. M. ( 1996 ) Pain 64 , 2736 .
31. Black, J. A. , Langworthy, K. , Hinson, A. W. , Dib-Hajj, S. D. & Waxman, S. G. ( 1997 ) NeuroReport 8 , 23312335 .
32. DArcangelo, G. , Paradiso, K. , Shepherd, D. , Brehm, P. , Halegoua, S. & Mandel, G. ( 1993 ) J. Cell Biol. 122 , 915921 .
the authoritative version for attribution.

33. Zur, K. B. , Oh, Y. , Waxman, S. G. & Black, J. A. ( 1995 ) Mol. Brain Res. 30 , 97103 .
34. Sheen, K. & Chung, J. M. ( 1993 ) Brain Res. 610 , 6268 .
35. England, S. , Bevan, S. & Docherty, R. J. ( 1996 ) J. Physiol. 495 , 429440 .
36. Gold, M. S. , Reichling, D. B. , Shuster, M. J. & Levine, J. D. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 11081112 .
37. Cardenas, C. G. , Del Mar, L. P. & Scroggs, R. S. ( 1997 ) J. Neurosci. 17 , 71817189 .
38. Waechter, C. J. , Schmidt, J. W. & Catterall, W. A. ( 1983 ) J. Biol. Chem. 258 , 51175123 .
39. Bilsky, E. J. , Wang, T. , Lai, J. & Porreca, F. ( 1996 ) Neurosci. Lett. 220 , 155158 .
40. Khasar, S. G. , Gold, M. S. , Dastmalchi, S. & Levine, J. D. ( 1996 ) Neurosci. Lett. 218 , 1720 .
41. Khasar, S. G. , Gold, M. S. & Levine, J. D. ( 1998 ) Neurosci. Lett. 256 , 1720 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TETRODOTOXIN-RESISTANT NA+ CURRENTS AND INFLAMMATORY HYPERALGESIA 7645

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Tetrodotoxin-resistant Na+ currents and inflammatory hyperalgesia

MICHAEL S. GOLD *
Department of Oral and Craniofacial Biological Sciences, University of Maryland Dental School, 666 West Baltimore Street,
Baltimore, MD 21201
ABSTRACT Several mechanisms have been identified that may underlie inflammation-induced sensitization of high-
threshold primary afferent neurons, including the modulation of voltage- and Ca2+-dependent ion channels and ion channels
responsible for the production of generator potentials. One such mechanism that has recently received a lot of attention is the
modulation of a tetrodotoxin (TTX)-resistant voltage-gated Na+ current. Evidence supporting a role for TTX-resistant Na+
currents in the sensitization of primary afferent neurons and inflammatory hyperalgesia is reviewed. Such evidence is derived
from studies on the distribution of TTX-resistant Na+ currents among primary afferent neurons and other tissues of the body that
suggest that these currents are expressed only in a subpopulation of primary afferent neurons that are likely to be involved in
nociception. Data from studies on the biophysical properties of these currents suggest that they are ideally suited to mediate the
repetitive discharge associated with prolonged membrane depolarizations. Data from studies on the effects of inflammatory
mediators and antinociceptive agents on TTX-resistant Na+ currents suggest that modulation of these currents is an underlying
mechanism of primary afferent neuron sensitization. In addition, the second-messenger pathways underlying inflammatory
mediator-induced modulation of these currents appear to underlie inflammatory mediator-induced hyperalgesia. Finally, recent
antisense studies have also yielded data supporting a role for TTX-resistant Na+ currents in inflammatory hyperalgesia. Although
data from these studies are compelling, data presented at the Neurobiology of Pain colloquium raised a number of interesting
questions regarding the role of TTX-resistant Na+ currents in inflammatory hyperalgesia; implications of three of these questions
are discussed.
Hyperalgesia that develops in the presence of tissue injury or inflammation reflects, at least in part, an increase in the excitability of
high-threshold primary afferent neurons innervating the site of injury. The increase in afferent excitability, or sensitization, develops
within minutes of an inflammatory stimulus and involves a leftward shift in neuronal stimulus response function and/or an increase in
spontaneous activity. The relatively rapid development of sensitization in response to inflammatory stimuli is likely to reflect the
modulation of proteins within or around the afferent terminal. In contrast, a change in the expression of protein(s) appears to be involved
in afferent sensitization observed in the presence of ongoing inflammation or nerve injury (see accompanying papers). At least three
underlying mechanisms have been identified that may contribute to the initial phase of inflammation-induced afferent sensitization: (i) a
change in the compliance of the tissue surrounding the afferent terminal ( 1 ); (ii) a change in efficacy of a transducer(s) within the afferent
terminal ( 2 4 ); and (iii) a change in a voltage- or Ca2+-dependent current within the afferent terminal ( 5 8 ). Because inflammatory
mediators may sensitize sensory neurons in vitro to stimuli that presumably bypass the afferent transduction apparati [i.e., high
extracellular potassium or current injection through a recording electrode ( 9 , 10 )], with no consistent changes either in resting membrane
conductance ( 7 ) or potential ( 10 ), modulation of a voltage- or Ca2+-dependent current must contribute to the sensitization of primary
afferent neurons. We have focused this review on the contribution of a particular class of voltage-gated Na+ currents (VGSCs), namely
tetrodotoxin (TTX)-resistant voltage-gated Na+ currents, to changes in afferent excitability.
WHY FOCUS ON NA+ CHANNELS?
Although a number of distinct voltage- and Ca2+-dependent currents have been identified that may underlie inflammationinduced
changes in afferent excitability (for review, see ref. 11 ), we have focused on the role of VGSCs for several reasons. First, VGSC
activation is critical for the generation and propagation of neuronal action potentials. Second, there is a growing body of evidence
indicating that modulation of these currents is an endogenous mechanism used to control neuronal excitability ( 8 , 12 15 ). Third,
evidence from injury in experimental animals ( 16 19 ) and humans ( 20 , 21 ) suggests that therapeutic interventions with compounds
known to block Na+ channels may be effective for the treatment of hyperalgesia and pain.
WHY TTX-RESISTANT CHANNELS?
Distribution. Evidence for the selective distribution of unique VGSC(s) among sensory neurons has been obtained in vivo and in
vitro. Intracellular recording from the cell bodies of sensory neurons in vivo indicated that the somal action potential of high-threshold
receptors is resistant to tetrodotoxin (TTX) at concentrations as high as 200 M applied to the surface of the ganglion ( 22 ). Similar
results were obtained with intracellular recording from intact ganglia in vitro, where it was observed that TTX-resistant action potentials
were present in neurons with slow-conducting axons (i.e., neurons likely to be associated with high-threshold receptors) ( 23 ).
Electrophysiological studies on dissociated sensory neurons have demonstrated that while TTX-sensitive Na+ currents are distributed
throughout the population of spinal sensory neurons, TTX-resistant Na+ currents are primarily restricted to a subpopulation of sensory
neurons likely to be involved in nociception ( 8 , 24 28 ). Specifically, TTX-resistant Na+ currents are present primarily in neurons that
have a small cell-body diameter (these are the neurons that tend to give rise to
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: PGE2, prostaglandin E2; NGF, nerve growth factor; ODN, oligodeoxynucleotide; PKA, protein kinase A; TTX,
tetrodotoxin; VGSC, voltage-gated sodium current.
* To whom reprint requests should be addressed, e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TETRODOTOXIN-RESISTANT NA+ CURRENTS AND INFLAMMATORY HYPERALGESIA 7646

small-diameter slow-conducting axons) and that are responsive to the algogenic compound capsaicin. Of note, a TTX-insensitive Na+
current (i.e., a current blocked by TTX at concentrations between 500 nM and 1 M) had been described in other tissues ( 29 31 ).
However, the biophysical properties of TTX-insensitive currents appears to differ from those of TTX-resistant currents ( 29 , 31 ).
Identification of a gene encoding a TTX-resistant Na+ channel confirmed the electrophysiological data indicating the existence of
unique Na+ currents in a subpopulation of primary afferent neurons. The first TTX-resistant Na+ channel cloned, referred to as SNS ( 32 ),
PN3 ( 33 ), and subsequently ScN10 ( 34 ), is only present in primary afferent neurons, in particular, a subpopulation of primary afferent
neurons with small-diameter cell bodies. Heterologous expression of SNS/ PN3 indicated that this clone encodes a voltage-gated Na+
channel with biophysical properties similar to those of the TTX-resistant channels present in sensory neurons ( 32 , 33 ). A second TTX-
resistant Na+ channel, referred to as NaN ( 35 ) or SNS2 ( 36 ), recently cloned from sensory neurons is also present in a subpopulation of
sensory neurons with a small-diameter cell body.
Biophysical Properties. At least three distinct TTX-resistant Na+ currents have been electrophysiologically isolated in rat primary
afferent neurons ( 37 , 38 ). The first TTX-resistant currents to be described ( 25 28 , 39 , 40 ) had several unique features. First, these
TTX-resistant currents have high thresholds for activation relative to TTX-sensitive currents ( 26 , 27 , 38 ). If the channels underlying
these currents are present in the peripheral terminals of primary afferent neurons and if spike initiation involves activation of these
channels, then the activation properties of these channels may explain why afferent neurons responsive to noxious stimuli have high
thresholds for activation. Second, TTX-resistant currents have high thresholds for steady-state inactivation relative to TTX-sensitive
currents ( 26 , 27 , 38 ). As a result, the majority of TTX-resistant channels are available for activation at membrane potentials as high as
40 mV ( 27 ). Consequently, it has been suggested that activation of these channels contributes to ongoing activity observed in the
presence of a sustained depolarization of primary afferent neurons ( 38 ). Third, TTX-resistant currents recover from inactivation rapidly
relative to TTX-sensitive currents (refs. 26 and 38 , but see ref. 27 ). Rapid recovery from inactivation is another factor that would enable
TTX-resistant currents to underlie sustained spiking in response to prolonged depolarizations ( 26 , 41 ). Fourth, the inactivation rates for
TTX-resistant Na+ currents are considerably slower than those of TTX-sensitive currents ( 38 ). This is particularly true at membrane
potentials close to the activation potential for TTX-resistant currents. Consequently, membrane depolarization may be facilitated after the
activation of a sustained inward current carried by TTX-resistant Na+ channels. The slow inactivation rate of TTX-resistant Na+ currents
also contributes to the broad action potential typically observed in high-threshold primary afferent neurons ( 41 ). Thus, the biophysical
properties of the first and most widely studied TTX-resistant Na+ currents are such that these currents may play a critical role in the
determination of the excitability of the afferent neurons in which they are present.
The biophysical properties of recently described TTX-resistant Na+ currents ( 37 , 38 ) more closely resemble TTX-sensitive Na+
currents than the TTX-resistant Na+ currents described above. For example, a second and third TTX-resistant Na+ current characterized by
Rush et al. ( 38 ) activated and inactivated at relatively hyperpolarized membrane potentials; the third current inactivated at membrane
potentials even more negative than those observed for TTX-sensitive Na+ currents. Of note, the inactivation rate of all three TTX-resistant
Na+ currents described by Rush et al. was relatively slow compared with that of TTX-sensitive Na+ currents. In contrast, a second TTX-
resistant Na+ current described by Sholz et al. ( 37 ) displayed both rapid kinetics for activation and inactivation and a voltage dependence
of activation and inactivation that occurred over relatively hyperpolarized membrane potentials. The role these additional TTX-resistant Na
+ currents play in regulating the excitability of high-threshold primary afferent neurons has yet to be determined.

Effects of Inflammatory Mediators. The distribution and biophysical properties of the classically described TTX-resistant Na+
currents suggests that these currents are involved in the control of the excitability of primary afferent neurons. Furthermore, several
inflammatory mediators released in response to injury are capable of directly sensitizing subpopulations of primary afferent neurons ( 9 ,
10 , 14 , 42 , 43 ). Therefore, we hypothesized that an inflammatory mediator-induced modulation of TTX-resistant Na+ currents is a
mechanism underlying the sensitization of primary afferent neurons. In support of this hypothesis, we observed that directly acting
hyperalgesic inflammatory mediators such as prostaglandin E2 (PGE2), serotonin, and adenosine decrease the activation threshold,
increase the rates of activation and inactivation, and increase the magnitude of TTX-resistant Na+ current ( 8 ). These changes could
contribute to both the decrease in threshold and increase in the number of action potentials evoked from a sensitized neuron. Further
support of our hypothesis is provided by the observation that the time course of inflammatory mediator-induced modulation of TTX-
resistant Na+ currents [developing within seconds and attaining a maximum within minutes ( 8 )] mirrors the time course of the
development of hyperalgesia in response to a peripheral injection of directly acting inflammatory mediators ( 44 ). Similar observations
have subsequently been reported by other investigators ( 14 , 45 ).
Effects of Antinociceptive Agents. Levine and Taiwo ( 46 ) previously observed that the peripheral administration of -opioid
receptor agonists blocked inflammatory mediatorinduced hyperalgesia. Therefore, we hypothesized that if inflammatory mediator-induced
modulation of TTX-resistant Na+ currents is an underlying mechanism of inflammatory hyperalgesia, then -opioid receptor agonists
should block inflammatory mediator-induced modulation of current. Consistent with this hypothesis, we observed that a -opioid receptor
agonist blocked PGE2-induced modulation of the current ( 47 ).
Second-Messenger Pathways. In a final series of experiments designed to test the hypothesis that inflammatory mediator-induced
modulation of TTX-resistant Na+ currents is an underlying mechanism of inflammatory hyperalgesia, we attempted to determine the role
of protein kinase A (PKA) in PGE2-induced modulation of the current. Previous studies performed in vivo suggested that direct acting
inflammatory mediators, including PGE2, serotonin, and adenosine, produced hyperalgesia ( 48 52 ) and afferent sensitization ( 53 ) via
the activation of a cAMP/PKA second-messenger cascade. The effects of these mediators were mimicked by compounds that increase the
intracellular concentration of cAMP, prolonged by agents that blocked the breakdown of cAMP, and blocked by agents that inhibit
adenylate cyclase and/or PKA.
Consistent with our hypothesis, England et al. ( 14 ) reported that PGE2-induced modulation of the TTX-resistant Na+ currents
involved activation of a cAMP/PKA second-messenger pathway. However, these experiments were performed on primary afferent
neurons from neonatal rats, and there are several lines of evidence suggesting primary afferent neurons from neonates may be qualitatively
and quantitatively different than neurons from adults ( 39 , 54 56 ). Furthermore, after failing to detect an effect with a membrane-
permeable analog of cAMP on TTX-resistant Na+ currents in primary afferent neurons from adult rats, Cardenas et al. ( 45 ) were
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TETRODOTOXIN-RESISTANT NA+ CURRENTS AND INFLAMMATORY HYPERALGESIA 7647

forced to conclude that modulation of the current must involve activation of another second-messenger pathway. We have recently
reported ( 57 ) that although an increase in the intracellular concentration of cAMP may result in the modulation of TTX-resistant Na+
current, the doseresponse relationship for such manipulations is bell-shaped. This observation may explain, at least in part, differences
between the observations of England et al. and those of Cardenas et al. It should be noted that a recent study involving heterologous
expression and site directed mutagenesis of the cloned TTX-resistant Na+ channel, SNS/PN3, indicates that the channel is phosphorylated
after activation of PKA ( 58 ). Furthermore, PKA-induced phosphorylation of the channel results in changes in gating properties similar to
those induced by inflammatory mediators ( 58 ), suggesting that inflammatory mediator-induced modulation of TTX-resistant currents
reflects a direct phosphorylation of the underlying channel(s). Importantly, and more to the point, our recent results ( 57 ) strongly support
the suggestion that PGE2-induced modulation of TTX-resistant Na+ currents in primary afferent neurons involves PKA activation.
The Function of TTX-Resistant Na+ Currents in Peripheral Terminals. Results obtained through the study of primary afferent
neuron cell bodies in vitro has provided compelling evidence in support of the hypothesis that modulation of TTX-resistant Na+ currents is
an underlying mechanism of inflammatory hyperalgesia. However, it is critical to determine whether these currents contribute to
inflammatory hyperalgesia in vivo. TTX-resistant Na+ currents are clearly present in the DRG cell body in vivo ( 22 ). Furthermore, results
from at least two studies suggest that these currents are present in the central terminals of primary afferent neurons ( 59 , 60 ). There also is
evidence that TTX-resistant Na+ currents are present in peripheral axons ( 61 ), but given that axonal conduction is blocked with TTX (
22 , 23 ), the function of TTX-resistant Na+ current in the axon has yet to be determined. Importantly, Brock et al. ( 62 ) have recently
obtained evidence suggesting that TTX-resistant Na+ channels play a role in action potential generation in the peripheral terminals of
corneal afferent neurons. Consistent with observations made while recording from cell bodies ( 22 , 23 ), recording from the peripheral
terminals revealed that electrical stimulation of the nerve trunk evoked action potentials that were blocked by TTX whereas spontaneous
or naturally evoked (with pressure or capsaicin) action potential persisted in the presence of TTX ( 62 ).
Although these recent results support the suggestion that TTX-resistant currents are present and functional in the peripheral terminals
of primary afferent neurons, they do not address the question of whether modulation of these currents contributes to inflammatory
hyperalgesia. Given the lack of specific pharmacological agents with which to manipulate TTX-resistant Na+ currents, it is not possible to
address this issue with traditional pharmacological approaches. However, through the use of antisense oligodeoxynucleotides (ODNs) to
selectively knock down expression of protein encoded by targeted mRNA ( 63 ), it has become possible to study the function of specific
proteins. Furthermore, we ( 64 ), and others ( 65 ) had previously demonstrated that the intrathecal administration of ODNs could be used
to knock down expression of proteins present in the peripheral terminals of primary afferent neurons. Therefore, we generated antisense
ODNs to a unique region of the cloned TTX-resistant Na+ channel, PN3/SNS, and assessed the effects of intrathecal ODN administration
on PGE2-induced hyperalgesia ( 66 ). Our results indicated that antisense, but not control, ODN sequences produced a small but significant
increase in baseline threshold to mechanical nociceptive stimuli, suggesting that activity in a TTX-resistant current contributes to the
determination of mechanical threshold. More importantly, antisense, but not control ODN sequences, resulted in a significant reduction in
PGE2-induced hyperalgesia. This observation is consistent with the electrophysiological data indicating that functional TTX-resistant Na+
channels are present in the peripheral terminals of primary afferent neurons. Furthermore, it supports the hypothesis that modulation of a
TTX-resistant Na+ current is an underlying mechanism of inflammatory hyperalgesia.
Questions Concerning the Role of TTX-Resistant Na+ Currents in Inflammatory Hyperalgesia. There are a number of questions
that remain to be answered concerning the role of TTX-resistant Na+ currents in inflammatory pain. At least three of these deserve
comment in light of data presented at the Neurobiology of Pain colloquium. First, what is the function of the second TTX-resistant Na+
current (NAN/ SNS2) cloned from spinal sensory neurons? Expression of the channel appears restricted to primary afferent neurons with
the smallest cell-body diameter in a subpopulation of neurons expressing PN3/SNS ( 36 ). The biophysical properties of the NAN/SNS2
expressed in HEK292 cells appear to more closely resemble TTX-sensitive Na+ currents (i.e., with faster activation and inactivation
kinetics), although these properties may not be reflective of the properties of the channel expressed in native tissue. Because there is no
homology between PN3/SNS and NaN/SNS2 in the region we targeted with our antisense ODN, it is unlikely that our results with
antisense ODNs reflect knock-down of both channels. Consequently, the residual PGE2-induced hyperalgesia we observed after PN3/SNS
antisense ODN administration may reflect an effect of PGE2 on NaN/SNS2. Porreca et al. ( 67 ) have recently obtained data suggesting
that NaN/SNS2 is not involved in either the establishment of nociceptive thresholds in control animals or in the maintenance of
hyperalgesia and allodynia in a neuropathy model. However, these investigators did not investigate the role of this channel in
inflammatory hyperalgesia. Thus, a role of NaN/SNS2 has yet to be determined.
The contribution of TTX-resistant Na+ currents to nociceptive thresholds in uninjured tissue is a second question concerning the role
of TTX-resistant Na+ currents in inflammatory pain. Results from our antisense study suggest that these currents do contribute, to a limited
extent, to the determination of mechanical nociceptive threshold ( 66 ). That the contribution of these currents to the determination of
nociceptive threshold is small is supported by observations made by Porreca et al. ( 67 ). These investigators were able to clearly
demonstrate a decrease in PN3/SNS protein in the cell bodies of primary afferent neurons and an attenuation of both inflammatory and
neuropathic hyperalgesia, by using an antisense strategy similar to the one we used. However, antisense ODNs had no effect on baseline
mechanical or thermal nociceptive thresholds. The small effect of antisense ODN treatment on baseline nociceptive threshold is striking in
light of the observation that in the nociceptor cell body, TTX-resistant Na+ current is the Na+ current primarily responsible for action
potential generation ( 14 ). The apparent difference between the role of TTX-resistant Na+ current in the cell body and in the peripheral
terminal suggests that either the current contributes little to the determination of baseline nociceptive threshold or activity in the population
of TTX-resistant Na+ current-containing afferent neurons contributes little to baseline nociceptive threshold. Although a single-unit
electrophysiological study is necessary to distinguish between these possibilities, observations obtained with the neurotoxin capsaicin
would suggest the latter. That is, rats treated neonatally with capsaicin to eliminate a vast majority of c-fiber afferent neurons have
baseline nociceptive thresholds that are only slightly elevated. However, PGE2-induced hyperalgesia is completely eliminated in these
animals ( 68 ).
Third, there is the question as to why the administration of antisense ODNs directed against SNS/PN3 had no effect on carrageenan-
induced hyperalgesia as observed by Porreca et al.
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TETRODOTOXIN-RESISTANT NA+ CURRENTS AND INFLAMMATORY HYPERALGESIA 7648

( 67 ). The peripheral administration of carrageenan is used as a model of acute inflammation associated with hyperalgesia that develops
within tens of minutes. In their carefully controlled study, Porreca et al. were able to demonstrate profound effects with antisense ODN
treatment on hyperalgesia resulting from the peripheral administration of Freunds adjuvant. Like carrageenan, the peripheral
administration of Freunds adjuvant is used as a model of inflammation, except the hyperalgesia associated with this model develops more
slowly, over many hours. Our results suggest that modulation of SNS/PN3 or a highly homologous species, contributes to inflammatory
mediator-induced hyperalgesia ( 66 ). Furthermore, carrageenan-induced hyperalgesia appears to involve the production and release of
hyperalgesic inflammatory mediators (i.e., see ref. 69 ). Consequently, we would predict that carrageenan-induced hyperalgesia should be
attenuated after SNS/PN3 antisense ODN administration. Identification of the reason(s) for the discrepancy between our predicted results
and the observations of Porreca et al. appears to require further experimentation.
CONCLUSIONS
Pain is clearly a complex process involving considerably more than the modulation of a single class of ion channels resulting in
changes in the excitability of a subpopulation of neurons. That the most effective analgesics available tend to have a wide spectrum of
action at a number of sites throughout the nervous system is largely reflective of this fact. Nevertheless, the study of a single class of ion
channels, TTX-resistant Na+ channels, has increased our understanding of the neurobiology of pain. Furthermore, because of the restricted
distribution of TTX-resistant Na+ currents and the observation that a decrease in the expression of these currents has little impact on low-
threshold mechanical transduction, targeting these currents may lead to the development of a therapeutic modality for the treatment of
hyperalgesia with fewer side effects than currently available modalities.
I would like to thank Normal Capra for helpful comments regarding the manuscript. Some of the work described in this article was
supported by grants from the National Institute of Health (Grant 1RO1NS3692901A1).
1. Cooper, B. ( 1993 ) J. Neurophysiol. 70 , 512521 .
2. Cesare, P. & McNaughton, P. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 1543515439 .
3. Reichling, D. B. & Levine, J. D. ( 1997 ) Proc. Natl. Acad. Sci. USA 94 , 70067011 .
4. Lopshire, J. C. & Nicol, G. D. ( 1997 ) J. Neurophysiol. 78 , 31543164 .
5. Nicol, G. D. , Vasko, M. R. & Evans, A. R. ( 1997 ) J. Neurophysiol. 77 , 167176 .
6. Weinreich, D. ( 1995 ) Pulm. Pharmacol. 8 , 173179 .
7. Gold, M. S. , Shuster, M. J. & Levine, J. D. ( 1996 ) Neurosci. Lett. 205 , 161164 .
8. Gold, M. S. , Reichling, D. B. , Shuster, M. J. & Levine, J. D. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 11081112 .
9. Baccaglini, P. I. & Hogan, P. G. ( 1983 ) Proc. Natl. Acad. Sci. USA 80 , 594598 .
10. Gold, M. S. , Dastmalchi, S. & Levine, J. D. ( 1996 ) Neuroscience 71 , 265275 .
11. Tanner, K. D. , Gold, M. S. , Reichling, D. B. & Levine, J. D. ( 1997 ) in Molecular Neurobiology of Pain , ed. Borsook, D. ( Int. Assoc. Study of
Pain , Seattle ), Vol. 9 , pp. 79105 .
12. Cantrell, A. R. , Ma, J. Y. , Scheuer, T. & Catterall, W. A. ( 1996 ) Neuron 16 , 10191026 .
13. Cantrell, A. R. , Smith, R. D. , Goldin, A. L , Scheuer, T. & Catterall, W. A. ( 1997 ) J. Neurosci. 17 , 73307338 .
14. England, S. , Bevan, S. & Docherty, R. J. ( 1996 ) J. Physiol. (London) 495 , 429440 .
15. Ono, K. , Fozzard, H. A. & Hanck, D. A. ( 1993 ) Circ. Res. 72 , 807815 .
16. Abram, S. E. & Yaksh, T. L. ( 1994 ) Anesthesiology 80 , 383391 .
17. Chabal, C. , Russell, L. C. & Burchiel, K. J. ( 1989 ) Pain 38 , 333338 .
18. Puig, S. & Sorkin, L. S. ( 1996 ) Pain 64 , 345355 .
19. Devor, M. , Wall, P. D. & Catalan, N. ( 1992 ) Pain 48 , 261268 .
20. Chabal, C. , Jacobson, L , Russell, L. C. & Burchiel, K. J. ( 1992 ) Pain 49 , 912 .
21. Rizzo, M. A. ( 1997 ) J. Neurol. Sci. 152 , 103106 .
22. Ritter, A. M. & Mendell, L. M. ( 1992 ) J. Neurophysiol. 68 , 20332041 .
23. Villiere, V. & McLachlan, E. M. ( 1996 ) J. Neurophysiol. 76 , 19241941 .
24. Arbuckle, J. B. & Docherty, R. J. ( 1995 ) Neurosci. Lett. 185 , 7073 .
25. Caffrey, J. M. , Eng, D. L. , Black, J. A. , Waxman, S. G. & Kocsis, J. D. ( 1992 ) Brain Res. 592 , 283297 .
26. Elliott, A. A. & Elliott, J. R. ( 1993 ) J. Physiol. (London) 463 , 3956 .
27. Ogata, N. & Tatebayashi, H. ( 1993 ) J. Physiol. (London) 466 , 937 .
28. Roy, M. L. & Narahashi, T. ( 1992 ) J. Neurosci. 12 , 21042111 .
29. Frohnwieser, B. , Weigl, L. & Schreibmayer, W. ( 1995 ) Pflugers Arch. 430 , 751753 .
30. Donahue, L. M. ( 1995 ) Neurochem. Res. 20 , 713717 .
31. Thio, C. L. & Sontheimer, H. ( 1993 ) J. Neurosci. 13 , 48894897 .
32. Akopian, A. N. , Sivilotti, L. & Wood, J. N. ( 1996 ) Nature (London) 379 , 257262 .
33. Sangameswaran, L , Delgado, S. G. , Fish, L. M. , Koch, B. D. , Jakeman, L. B. , Stewart, G. R. , Sze, P. , Hunter, J. C. , Eglen, R. M. & Herman, R.
C. ( 1996 ) J. Biol. Chem. 271 , 59535956 .
34. Oaklander, A. L. & Belzberg, A. J. ( 1997 ) Brain Res. Mol. Brain Res. 52 , 162165 .
35. Dib-Hajj, S. D. , Tyrrell, L , Black, J. A. & Waxman, S. G. ( 1998 ) Proc. Natl. Acad. Sci. USA 95 , 89638968 .
36. Tate, S. , Benn, S. , Hick, C. , Trezise, D. , John, V. , Mannion, R. J. , Costigan, M. , Plumpton, C. , Grose, D. , Gladwell, Z. , et al. ( 1998 ) Nat.
Neurosci. 1 , 653655 .
37. Scholz, A. , Appel, N. & Vogel, W. ( 1998 ) Eur. J. Neurosci. 10 Suppl. , 25472556 .
38. Rush, A. M. , Brau, M. E. , Elliott, A. A. & Elliott, J. R. ( 1998 ) J. Physiol. (London) 511 , 771789 .
39. Schwartz, A. , Palti, Y. & Meiri, H. ( 1990 ) J. Membr. Biol. 116 , 117128 .
40. Kostyuk, P. G. , Veselovsky, N. S. , Fedulova, S. A. & Tsyndrenko, A. Y. ( 1981 ) Neuroscience 6 , 24242430 .
41. Schild, J. H. & Kunze, D. L. ( 1997 ) J. Neurophysiol. 78 , 3198 3209 .
42. Levine, J. D. , Basbaum, A. I. & Fields, H. L. ( 1993 ) J. Neurosci. 13 , 22732286 .
43. Cui, M. & Nicol, G. D. ( 1995 ) Neuroscience 66 , 459466 .
44. Taiwo, Y. O. , Goetzl, E. J. & Levine, J. D. ( 1987 ) Brain Res. 423 , 333337 .
45. Cardenas, C. G. , Del Mar, L. P. , Cooper, B. Y. & Scroggs, R. S. ( 1997 ) J. Neurosci. 17 , 71817189 .
46. Levine, J. D. & Taiwo, Y. O. ( 1989 ) Neuroscience 32 , 571575 .
47. Gold, M. S. & Levine, J. D. ( 1996 ) Neurosci. Lett. 212 , 8386 .
the authoritative version for attribution.

48. Taiwo, Y. O. , Bjerknes, L K. , Goetzl, E. J. & Levine, J. D. ( 1989 ) Neuroscience 32 , 577580 .

49. Taiwo, Y. O. & Levine, J. D. ( 1989 ) Brain Res. 492 , 400403 .
50. Taiwo, Y. O. & Levine, J. D. ( 1991 ) Neuroscience 44 , 131135 .
51. Taiwo, Y. O. , Heller, P. H. & Levine, J. D. ( 1992 ) Neuroscience 48 , 479483 .
52. Taiwo, Y. O. & Levine, J. D. ( 1990 ) Neuroscience 38 , 752-762 .
53. Wang, J. F. , Khasar, S. G. , Ahlgren, S. C. & Levine, J. D. ( 1996 ) Neuroscience 71 , 259263 .
54. Bennett, D. L , Averill, S. , Clary, D. O. , Priestley, J. V. & McMahon, S. B. ( 1996 ) Eur. J. Neurosci. 8 , 22042208 .
55. Ogata, N. & Tatebayashi, H. ( 1992 ) Dev. Brain Res. 65 , 93100 .
56. Waxman, S. G. , Kocsis, J. D. & Black, J. A. ( 1994 ) J. Neurophysiol. 72 , 466470 .
57. Gold, M. S. , Levine, J. D. & Correa, A. M. ( 1998 ) J. Neurosci. 18 , 1034510355 .
58. Fitzgerald, E. M. , Okuse, K. , Wood, J. N. , Dolphin, A. C. & Moss, S. J. ( 1998 ) J. Physiol. (London) , in press .
59. Gu, J. G. & MacDermott, A. B. ( 1997 ) Nature (London) 389 , 749753 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TETRODOTOXIN-RESISTANT NA+ CURRENTS AND INFLAMMATORY HYPERALGESIA 7649

60. Jeftinija, S. ( 1994 ) Brain Res. 639 , 125134 .

61. Quasthoff, S. , Grosskreutz, J. , Schroder, J. M. , Schneider, U. & Grafe, P. ( 1995 ) Neuroscience 69 , 955965 .
62. Brock, J. A. , McLachlan, E. M. & Belmonte, C. ( 1998 ) J. Physiol. (London) 512 , 211217 .
63. Crooke, S. T. ( 1997 ) Ciba Found. Symp. 209 , 158164 .
64. Khasar, S. G. , Gold, M. S. , Dastmalchi, S. & Levine, J. D. ( 1996 ) Neurosci. Lett. 218 , 1720 .
65. Bilsky, E. J. , Wang, T. , Lai, J. & Porreca, F. ( 1996 ) Neurosci. Lett. 220 , 155158 .
66. Khasar, S. G. , Gold, M. S. & Levine, J. D. ( 1998 ) Neurosci. Lett. 256 , 1720 .
67. Porreca, F. , Lai, J. , Biau, D. , Wegert, S. , Ossipov, M. H. , Eglen, R. M. , Kassotakis, L. , Novakovic, S. , Robert, D. K. , Sangameswaran, L. &
Hunter, J. C. ( 1999 ) Proc. Natl. Acad. Sci. USA 96 , 76407644 .
68. Khasar, S. G. & Levine, J. D. ( 1996 ) Neurosci. Lett. 205 , 141 143 .
69. Dirig, D. M. , Isakson, P. C. & Yaksh, T. L. ( 1998 ) J. Pharmacol. Exp. Ther. 285 , 10311038 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CALCIUM REGULATION OF A SLOW POST-SPIKE HYPERPOLARIZATION IN VAGAL AFFERENT NEURONS 7650

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Calcium regulation of a slow post-spike hyperpolarization in vagal

afferent neurons

(spike frequency adaptation/ryanodine receptor/autacoids/allergic inflammation/mast cell)

RUTH CORDOBA-RODRIGUEZ *, KIMBERLY A. MOORE *, JOSEPH P. Y. KAO , AND DANIEL WEINREICH *
* Department of Pharmacology and Experimental Therapeutics and Medical Biotechnology Center and Department of Physiology,

University of Maryland, School of Medicine, Baltimore, MD 21201-1559

ABSTRACT Activation of distinct classes of potassium channels can dramatically affect the frequency and the pattern of
neuronal firing. In a subpopulation of vagal afferent neurons (nodose ganglion neurons), the pattern of impulse activity is
effectively modulated by a Ca2+-dependent K+ current. This current produces a post-spike hyperpolarization (AHPslow) that plays
a critical role in the regulation of membrane excitability and is responsible for spike-frequency accommodation in these neurons.
Inhibition of the AHPslow by a number of endogenous autacoids (e.g., histamine, serotonin, prostanoids, and bradykinin) results in
an increase in the firing frequency of vagal afferent neurons from <0.1 to >10 Hz. After a single action potential, the AHPslow in
nodose neurons displays a slow rise time to peak (0.30.5 s) and a long duration (315 s). The slow kinetics of the AHPslow are due,
in part, to Ca2+ discharge from an intracellular Ca2+-induced Ca2+ release (CICR) pool. Action potential-evoked Ca2+ influx via
either L or N type Ca2+ channels triggers CICR. Surprisingly, although L type channels generate 60% of action potential-induced
CICR, only Ca2+ influx through N type Ca2+ channels can trigger the CICR-dependent AHPslow. These observations suggest that a
close physical proximity exists between endoplasmic reticulum ryanodine receptors and plasma membrane N type Ca2+ channels
and AHPslow potassium channels. Such an anatomical relation might be particularly beneficial for modulation of spike-frequency
adaptation in vagal afferent neurons.
Activation and sensitization of primary afferent nerve fibers during allergic inflammation are orchestrated by inflammatory mediators
released from various cells, including tissue mast cells. Inflammatory mediators provoke excitability changes in sensory nerves through
diverse mechanisms, including (i) modification of the density and coupling efficacy of ligand-gated ionic channels; (ii) alteration in
voltage-gated sodium, potassium, and calcium channels; and (iii) manipulation of cellular mechanisms that control spike-frequency
adaptation.
After immunologic activation of mast cells in airway in vivo or in sensory ganglia in vitro, a wide range of electrophysiological
changes can be detected in peripheral sensory nerve terminals of the vagus ( 1 ) and in vagal primary afferent somata (located in the
nodose and jugular ganglia) ( 2 ). These changes range from transient (minutes) membrane depolarizations that sometimes reach action
potential (AP) threshold ( 3 ) to a sustained (days) unmasking of functional NK-2 tachykinin receptors ( 4 , 5 ). One electrical membrane
property that is particularly sensitive to inflammatory mediators is a slow post-spike afterhyperpolarization (AHPslow; see Fig. 1 ) ( 3 ).
This slow afterpotential influences neuronal excitability and determines the frequency and pattern of neuronal discharge. We have found
that the amplitude and duration of the AHPslow are exquisitely sensitive to known inflammatory mediators such as prostanoids, amines, and
kinins applied exogenously ( Table 1 ) or released endogenously (i.e., after immunologic activation of mast cells) ( 3 , 6 ). Inhibition of the
AHPslow is accompanied by a loss of spike-frequency adaptation. Thus, modulation of the AHPslow amplitude and duration provides a
mechanism for neuronal sensitization.
We are interested in identifying the ionic channels and second-messenger transduction pathways that participate in the initiation and
maintenance of the AHPslow in vagal primary afferent neurons. In this report, we describe the general properties of this slow afterpotential
and our progress in its characterization. Our working hypothesis is that a close functional proximity between three separate channels [N
type voltage-sensitive calcium channels, ryanodine (RY)-sensitive Ca2+-induced Ca2+ release (CICR) calcium channels, and AHPslow K+
(SK) channels that underlie the AHPslow] is essential for the initiation of the AHPslow.
RESULTS
General Properties of Vagal Afferent AHPslow. The AHPslow is observed in a wide variety of peripheral and central neurons (for
review, see ref. 7 ). In nodose neurons, AHPslow is always preceded by a fast post-spike afterhyperpolarization (AHPfast, 1050 ms) that
occurs at the end of the AP repolarization. In some neurons, the AHPfast is followed by a second afterpotential that lasts 50300 ms
(AHPmedium). The AHPmedium is voltage- and Ca2+- dependent and blocked by 10 mM tetraethylammonium in 50% of neurons, suggesting
that it is mediated by large-conductance Ca2+-activated K+ channels (BK channels) ( 8 ).
In vagal afferent somata, the AHPslow is particularly robust. After a single AP, the AHPslow displays a delayed onset (100500 ms), a
slow rise time to peak (0.35 s), and a long duration ( 2 15 s; see Fig. 1 ). The proportion of AHPslow neurons within nodose ganglia
varies among species: 20% in the guinea pig, 35% in rabbit, and 85% in ferret. Only nodose neurons classified as C fibers (conduction
velocity <1 m/s) possess AHPslow. To date, there have been few species differences in the pharmacological or physiological properties
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: AP, action potential; BK, large-conductance Ca2+-activated K+ channels; SK, small-conductance Ca2+ -activated K+
channels; CICR, Ca2+-induced Ca2+ release stores; RY, ryanodine; VDCC, voltage-dependent Ca2+ channels; L, N, R, L type, N type, and
R-type VDCC; AHP, afterhyperpolarization; DBHQ, 2,5,-di(t-butyl)hydroquinone.
To whom reprint requests should be addressed, e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CALCIUM REGULATION OF A SLOW POST-SPIKE HYPERPOLARIZATION IN VAGAL AFFERENT NEURONS 7651

of the AHPslow. An analogous slow AHP has also been recently described in 25% of C type dorsal root ganglion neurons of the rat (9, 10).

FIG. 1. A single AP can evoke three types of AHP in nodose neurons. (Top) A neuron with a single-component afterpotential lasting
30 ms. This AHP is designated AHPfast. All neurons have this short duration afterpotential. (Middle) Example of a neuron with two
afterpotentials, an AHPfast followed by a longer lasting afterpotential (300 ms), the AHPmedium. In approximately half of the neurons,
the AHPmedium is Ca2+-dependent. (Bottom) In a subset of C fiber type nodose neurons, a slowly developing (hundreds of ms) and
long-lasting (215 s) afterpotential is observed. This slow afterpotential (AHPslow) is always Ca2+-dependent. Intracellular recordings
were obtained at room temperature from adult neurons isolated from rabbit nodose ganglia. The values near the horizontal lines are
resting membrane potentials. The calibration in the Top also applies to the Middle. Similar results have been recorded in guinea pig
and ferret nodose neurons.
The AHPslow in vagal afferent neurons influences cellular excitability and controls AP frequency over the physiological range from
0.1 Hz to 10 Hz ( 11 , 12 ). One interesting property of the AHPslow is that its amplitude is tuned to both AP number and frequency. Over
the range of 1100 Hz, the amplitude of the AHPslow increases with the number of APs until it plateaus after 15 APs ( Fig. 2 ); similar
results were observed when the current underlying the AHPslow was monitored. For reasons still unresolved, 10 Hz (100-msec interspike
intervals) consistently evokes the largest responses.
Table 1. Inflammatory mediators that block AHPslow in vagal afferent neurons
Mediator Receptor type EC50, nM
Bradykinin B2 72
Histamine H1 2,000
Serotonin nd 300
PGD2, PGE2 nd 20
Leukotriene C4 nd 100

Bradykinin ( 26 ), histamine ( 27 ), serotonin ( 28 ), PGD2 and PGE2 ( 12 ), and leukotriene C4 ( 3 ) block the AHPslow. nd, not determined;
PG prostaglandin.

FIG. 2. Effects of varying numbers of APs and frequency on the amplitude of the AHPslow. All data points were recorded from a
single acutely dissociated adult rabbit nodose neuron at room temperature. Resting potential and membrane input resistance were 55
mV and 53 M, respectively. APs were evoked by transmembrane depolarizing current pulses (2 nA, 3 ms) at the frequencies
indicated. Similar results were obtained when measuring IAHP by using the hybrid voltage-clamp technique in rabbit, guinea pig, and
ferret nodose neurons.
The current generating the AHPslow (IAHP) is a voltage-insensitive Ca2+-dependent K+ current ( 13 , 14 ) that is unaffected by a wide
range of K+ channel antagonists: 100 nM apamin, 10 M d-tubocurarine, 5 mM Cs+, 30 mM tetraethylammonium, 10 mM Ba2+, 4 mM 4-
aminopyridine, and 10 nM charybdotoxin. The magnitude of the AHPslow (or the IAHP) is linearly related to the concentration of
extracellular Ca2+ ( Fig. 3 ) and requires a rise in cytosolic free Ca2+ ([Ca2+]i) for activation. Buffering intracellular Ca2+ with 1,2-bis(2-
aminophenoxy)ethane-N, N, N
, N
-tetraacetic acid (BAPTA) abolishes the AHPslow ( Fig. 4 ). Noise analysis of the IAHP suggests a single-
channel conductance of 10 pS (unpublished observations). These features are consistent with the properties of a small-conductance Ca2+-
activated K+ channel (SK channel; ref. 8 ). Of the several SK channels recently cloned from mammalian brain ( 15 ), the hSK1 channel
has a pharmacological and biophysical profile compatible with the K+ current underlying the AHPslow in nodose neurons.
the authoritative version for attribution.

Ca2+ Injection Evokes Two Temporally Distinct Outward Currents. To test whether the K+ channels associated with the
AHPmedium and the AHPslow are directly activated by Ca2+, we iontophoretically injected Ca2+ into nodose neurons. Independent of the
AHPslow, a large outward current with rapid activation and decay kinetics was elicited by Ca2+ injection. This current (IK-medium) was
evoked at holding potentials between 2 mV and 45 mV. It was completely blocked by 5 mM tetraethylammonium but unaffected by
inhibitors of the AHPslow (100 nM prostaglandins D2 or E2 or 1 M forskolin). IK-medium was strongly voltage-dependent, requiring
membrane holding potentials more positive than 55 mV. Assuming a reversal potential of 80 mV, IK-medium had an e-fold increase in
peak conductance for each 8.0 1.0 mV (mean SEM; n = 8) depolarization, as calculated from semilogarithmic plots of peak chord
conductance versus voltage-clamp holding potential. These properties are similar to those of large-conductance BK (AHPmedium) channels.
In neurons that exhibited AHPslow, Ca2+ injection provoked a slowly developing and protracted outward current (IK-slow).
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CALCIUM REGULATION OF A SLOW POST-SPIKE HYPERPOLARIZATION IN VAGAL AFFERENT NEURONS 7652

Fig. 5 shows an overlay of the outward current responses evoked by Ca2+ injection in a single nodose C type neuron at holding potentials
of 20 mV and 50 mV. The kinetic differences between IK-medium and IK-slow after Ca2+ injection are dramatic. In contrast to the rapid
activation of IK-medium, the onset of IK-slow is delayed, and the decay of IK-medium is nearly complete before the peak amplitude of the IK-
slow is reached. These two outward currents mirror the temporal and pharmacological differences between AHPmedium and AHPslow IK-slow,
like the AHPslow, was blocked by 100 nM prostaglandin D2. The data shown in Table 2 summarize quantitative differences between these
two Ca2+-induced outward currents.

FIG. 3. Effects of varying [Ca2+]o on the amplitude of the AHPslow recorded in isolated nodose neurons. (A) Sample traces of
AHPslow evoked by a train of four APs in the presence of different [Ca2+]o. APs are evoked by transmembrane depolarizing current
pulses (2 nA, 3 ms, 10 Hz) and are truncated. [Ca2+]o was varied from 2.0 to 0.0 mM in 0.5 mM decrements. The AHPslow is
completely blocked when [Ca2+]o is reduced to nominally zero. On returning to 2.0 mM [Ca2+]o, the AHPslow recovers to its original
amplitude. (B) Relation between [Ca2+]o and AHPslow amplitude recorded in several neurons. Values are means SEM of the number
of observations indicated near each data point. Data are normalized to the maximum response recorded in a given neuron. Linear
regression analysis yields the solid line (r = 0.993).
It is possible that the delayed onset of IK-slow compared with IK-medium results from unequal Ca2+ diffusion distances from the injection
site to the two types of K+ channels. This cause seems unlikely because the orientation of impalement was random, and the plasma
membranes of dissociated nodose neurons appear devoid of processes that would provide semiisolated regions where IK-slow might be
generated. An alternative possibility is that additional intermediate steps, such as the synthesis or release of a second messenger, are
required to activate IK-slow. The large Q10 (>3.0; ref. 14 ) supports the latter alternative. One candidate is mobilization of intracellularly
stored Ca2+.
Ca2+ Released by the CICR Pool Is Essential for the Generation of the AHPslow. Single APs produce transient increases in [Ca2+]i
(Cat) as measured by the fluorescent indicator fura-2. The magnitude of the Cat depends on both [Ca2+]o and the number of APs. Over
the range of one to eight APs, there is an approximately linear relation between the magnitude of the Cat and the number of APs (
Fig. 6 ). In the presence of drugs that block CICR but do not significantly affect AP-induced Ca2+ influx [(RY, 10 M), 2,5,-di(t-butyl)
hydroquinone (DBHQ, 10 M), or thapsigargin (TG, 100 nM)], we found that at least eight APs were required to evoke a detectable Cat (
Fig. 6 ). In the presence of RY, DBHQ, and TG, the CatAP relation exhibits slopes of 0.5, 1.1, and 0.8 nM per AP, respectively. When
compared with the slope of 9.6 nM per AP in control neurons, Ca2+ influx produced by a single nodose AP is amplified by 5- to 10-fold by
CICR ( 16 ). Nodose neurons demonstrate a relatively low stimulus threshold for eliciting CICR. For instance, a robust CICR response can
be observed after a single AP stimulus in nodose neurons, whereas many tens of APs are required in dorsal root ganglion neurons ( 17 ).
The greater CICR response in nodose neurons is not due to greater Ca2+ influx through voltage-dependent Ca2+ channels (VDCCs); a
single AP produces comparable Ca2+ influx in nodose and dorsal root ganglion neurons (39 vs. 49 pC, respectively; refs. 16 and 18 ).
Rather, the more responsive CICR pool in nodose neurons
the authoritative version for attribution.

FIG. 4. Effects of BAPTA on the AHPslow and on the excitability of an acutely dissociated rabbit nodose neuron. (A) Bath-applied
BAPTA/acetomethylester (10 M) blocks the AHPslow within 5 min without changing the resting membrane potential or membrane
input resistance. APs were evoked by transmembrane depolarizing current pulses (4 nA, 1.5 ms, 10 Hz) and are truncated. (B)
Responses recorded at a faster sweep speed to illustrate the kinetics of the AHPfast and AHPmedium, which precede the AHPslow. The
AHPfast is unaffected by 10 M BAPTA/acetomethylester (compare a with b). The Ca2+ dependence of the AHPmedium is illustrated in
c, where the neuron is superfused with 100 M CdCl2 for 30 s, which blocks most of the AHPmedium. The residual component of the
AHP recorded in CdCl2 is the AHPfast, which is mediated by delayed rectifier K+ channels. (C) Depression of the AHPslow markedly
increases neuronal excitability. The average AP firing frequency induced by a current ramp protocol (1 nA, 2 s) increased from 1 to
5.5 Hz when the AHPslow was blocked. Similar loss of spike-frequency adaptation was observed with bradykinin, prostaglandin D2,
histamine, and other inflammatory autacoids (see Table 2 ). The scale bar represents 3 mV, 2 s in A; 15 mV, 0.25 s in B; and 15 mV,
0.5 s in C. The dashed line represents the resting membrane potential (60 mV). Resting membrane input resistance was 70 M.
Data is from ref. 19 with permission from the American Physiological Society.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CALCIUM REGULATION OF A SLOW POST-SPIKE HYPERPOLARIZATION IN VAGAL AFFERENT NEURONS 7653

may reflect either a closer proximity between plasma membrane Ca2+ influx channels and endoplasmic reticulum RY receptors or a more
sensitive RY receptor.

FIG. 5. Comparison of two outward K+ currents evoked by intracellular Ca2+ injection. Recordings were made in a single acutely
isolated adult rabbit nodose neuron. A slow outward current (IK-slow) was activated by a 5-nA, 1-s iontophoretic Ca2+ injection at a
holding potential of 50 mV. A second outward current (IK-medium) was activated at 20 mV (5 nA, 0.5 sec). IK-medium activates and
decays blocked by 10 mM tetraethylammonium; IK-slow was blocked by 100 nM completely before IK-slow reaches peak amplitude. IK-
medium was prostaglandin D2. The iontophoretic pipette was filled with a 0.2 M CaCl2 solution. Voltage-clamp currents were recorded
with a second intracellular pipette. The discontinuous (switched) current injection mode of an Axoclamp II amplifier was used for
both currentand voltage-clamp applications. The larger calibration value is for IK-medium. Population data is shown in Table 2 .
By using physiological stimuli (APs) in conjunction with pharmacological manipulations of CICR, we have demonstrated that CICR
is essential for the development of the AHPslow. Over the range of 116 APs, the magnitudes of the AP-induced AHPslow and the Cat (a
monitor of CICR in these neurons) were highly correlated (r = 0.985). Simultaneous recordings of Cat and AHPslow before and during
bath application of CICR inhibitors (RY, TG, DBHQ, or 10 M cyclopiazonic acid) revealed that both responses were blocked in a
parallel fashion ( Fig. 7 ; see also Table 1 in ref. 19 ). These data indicate that a CICR pool is essential for the generation of the AHPslow.
They also provide a potential explanation for the slow kinetics of the AHPslow, namely Ca2+ mobilization from CICR.
Effects of Changing [Ca2+]o on the AHPslow, Cat, and Ca2+ influx. If the AHPslow depends on Ca2+ released from the CICR pool
triggered by AP-induced Ca2+ influx, it would follow that changes in [Ca2+]o should produce corresponding effects on both the AHPslow
and the Cat. The data shown in Fig. 3 A illustrate the effects of progressively lowering [Ca2+]o from 2.0 mM to nominally zero on the
amplitude of the AHPslow recorded in a single nodose neuron. As [Ca2+]o was decreased, the amplitude of the AHPslow was reduced
proportionally. When the results from this and five additional neurons were plotted ( Fig. 3 B), the relation between [Ca2+]o and the
amplitude of the AHPslow was linear (r = 0.993; n = 6, pooled data from three current-clamp and three hybrid voltage-clamp experiments).
Table 2. Comparison of IK-slow and IK-medium
Current Peak conductance, n Holding n Time-to- n Decay time n Duration, s n
nS potential, mV peak, ms constant, ms
IK-slow 27.9 6.5 14 55.4 2.7 14 6,570 1085 12 6,735 789 5 23 3.4 14
IK-medium 53.2 16.5 6 20 3.7 6 958 56 6 818 97 6 2.5 0.16 6
2+ into acutely isolated nodose neurons of the rabbit. The peak
K-slow and IK- medium are outward currents elicited by iontophoretic injection Ca
conductance is the largest conductance elicited, independent of membrane potential. The holding potential is the potential at which the peak conductance
was measured. The decay time constant was measured by fitting a line, by eye, to the log transform of the decay of the current. The duration was
calculated from the onset of Ca2+ injection to the time at which the current had decayed to 20% of its peak value. Data are summarized as the mean
SEM.

FIG. 6. (Upper) Effect of RY on AP-induced Ca2+ transients. Traces are Ca2+ transients evoked by varying numbers of APs, as
indicated below each trace. In control neurons, distinct Ca2+ transients can be elicited by very few APs. In contrast, in the presence of
10 M RY, a CICR inhibitor, at least eight APs are required to generate a discernible change in [Ca2+]i. Suppression of the Ca2+
transient by RY is due to its effect on CICR and not the result of nonspecific effects on Ca2+ channels; the kinetics and amplitude of
ICa elicited by APs are completely unaffected by RY. (Lower) Effect of RY on the relation between the amplitude of Ca2+ transients
the authoritative version for attribution.

and number of APs. and are mean amplitudes of Ca2+ transients evoked by varying numbers of action potentials for control (n =
10) and for RY-treated nodose neurons (n = 3), respectively. Linear regression of data from control (4 action potentials) and RY-
treated cells yielded slopes of 9.6 0.01 and 0.5 0.23 nM per AP, respectively. Comparison of the slopes illustrates that CICR is
capable of amplifying the trigger Ca2+ resulting from AP-induced Ca2+ influx by 20-fold. Data is modified from ref. 16 with
permission from Journal of Physiology (London).
Next, we examined the relation between [Ca2+]o and the magnitude of the AP-induced Cat. Fig. 8 A illustrates Cats elicited by
varying numbers of APs recorded from a single neuron in Locke solution containing 2.2 or 1.1 mM Ca2+. The population results relating
the normalized amplitude of the Cats recorded in four neurons to the number of APs is shown in Fig. 8 B. In 1.1 mM [Ca2+]o, the first
few APs did not elicit a measurable Cat. For the neuron shown in Fig. 8 A, at least eight APs were necessary to evoke a detectable Cat.
In three additional neurons, the minimum number of APs necessary to
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CALCIUM REGULATION OF A SLOW POST-SPIKE HYPERPOLARIZATION IN VAGAL AFFERENT NEURONS 7654

elicit a detectable Cat ranged from 4 to 32. The CatAP relation recorded in 1.1 mM [Ca2+]o, as in Locke solution containing normal
[Ca2+]o, followed a hyperbolic relation (2 = 6.75 and 0.31; r = 0.988 and 0.999 for 2.2 and 1.1 mM Ca2+,respectively; Fig. 8 B and see
also Fig. 1 in ref. 16 ). Given the hyperbolic nature of the CatAP relation, deducing the effects of altered [Ca2+]o on the magnitude of the
Cat clearly depends on where along this relation the comparison is made. At one extreme, there is a 2-fold change when comparing the
plateau phases of the curves in normal and one-half normal [Ca2+]o. It is also possible to calculate the limiting initial slopes for the rising
phase of the curves (dashed lines in Fig. 8 B). The limiting slopes, which represent the full Ca2+ release potential of the CICR pool before
any release has actually occurred, were 15 3.8 and 2 0.7 nM per AP in 2.2 and 1.1 mM [Ca2+]o, respectively. Thus, reducing [Ca2+]o by
a factor of 2 results in a reduction of the Cat by a factor of 7 2.8 when the rising phases of the two curves are compared. The 7-fold
reduction of the Cat associated with halving [Ca2+]o is much larger than the 2-fold reduction in the AHPslow amplitude ( Fig. 3 ),
suggesting that some, but not all, of the Ca2+ released from the CICR pool is required for the generation of the AHPslow.

FIG. 7. Effect of DBHQ, a functional CICR inhibitor, on the AP-induced Ca2+ transient and on the AHPslow recorded simultaneously
in an acutely isolated rabbit nodose neuron. Upper traces represent superimposed Ca2+ transients evoked by a train of four APs (10
Hz) recorded in control Locke solution and 7 min after switching to Locke solution containing 10 M DBHQ. The lower pair of
traces shows AHPslow. DBHQ treatment completely blocked both the Ca2+ transient and the AHPslow. Resting [Ca2+]i was 91 nM.
Fluorescence data were acquired at 10 Hz. Resting membrane potential was 67 mV. AP amplitudes are truncated. Data are from ref.
19 with permission from the American Physiological Society.
The disproportionate effect of reduced [Ca2+]o on the Cat versus the AHPslow could arise from a nonlinear reduction of Ca2+ influx
per AP and/or from a decreased Ca2+ release from CICR pool per unit Ca2+ influx. To study these possibilities, we examined the effect of
lowering [Ca2+]o on APinduced Ca2+ influx. The amount of Ca2+ entering a neuron with each AP in normal and in reduced [Ca2+]o was
determined by using a prerecorded AP as whole-cell voltage-clamp command under experimental conditions where the major inward
charge carrier is Ca2+ (for details, see Fig. 2 in ref. 16 ). When [Ca2+]o was decrementally reduced from 2 mM to nominally zero, the
magnitude of the ICa decreased proportionally. The charge movement caused by Ca2+ influx, normalized to cell membrane capacitance (pC/
pF), was plotted against varying [Ca2+]o for 12 neurons. Over the range of 02.0 mM [Ca2+]o, Ca2+ influx varied linearly with [Ca2+]o (r =
0.974). These results indicate that changes in Ca2+ influx alone cannot account for the disproportionate reduction in the Cat relative to the
AHPslow that is observed when [Ca2+]o is reduced.

FIG. 8. Effect of varying [Ca2+]o on the amplitude of AP-induced Ca2+ transients. (A) Representative traces of Ca2+ transients evoked
by varying numbers of APs in normal (2.2 mM) and reduced (1.1 mM) [Ca2+]o. APs were elicited by transmembrane depolarizing
current pulses (2 nA, 1.5 ms, 10 Hz). The number of APs is indicated below each trace. (B) The normalized (mean SEM) amplitude
of Ca2+ transients recorded in four neurons is plotted against varying numbers of APs. Data are normalized to the maximal response
recorded in a given neuron. represents Ca2+ transients recorded in 2.2 mM [Ca2+] o; represents Ca2+ transients recorded in the
same neurons in 1.1 mM [Ca2+]o. Continuous curves are rectangular hyperbolas fit to the data (2 = 6.75 and 0.31, r = 0.988 and 0.999
for 2.2 and 1.1 mM [Ca2+]o, respectively). The dashed lines represent the limiting initial slopes (15 3.8 and 2 0.7 nM per AP for
2.2 and 1.1 mM [Ca2+]o, respectively).
The disproportionate effect of reduced [Ca2+]o on the CatAHPslow relation could arise from a diminution in the amount of Ca2+
released from the CICR pool. Caffeine, a known agonist of CICR, is traditionally used to assess the releasable content of the CICR pool.
In 8 of the 13 neurons studied, halving [Ca2+]o reduced the caffeine-induced Cat by 2079% (100% vs. 47 7.2% in 2.2 and 1.1 mM [Ca2+]
2+
o, respectively; P = 0.0002). In other words, decreasing [Ca ]o by a factor of 2 caused a 1.25- to 5-fold reduction in the caffeine response.
the authoritative version for attribution.

On returning to normal Locke solution, the caffeine response was restored to near control values. In the remaining five neurons, the
caffeine-induced Cat was unaffected by reducing [Ca2+]o (100% vs. 112 8.4% in 2.2 and 1.1 mM [Ca2+]o, respectively; P = 0.690).
There was no significant difference in resting levels of [Ca2+]i between these two groups of neurons (93 29.5 nM vs. 111 29.7 nM; P =
0.530). Unfortunately, the wide variability in the effects of reduced [Ca2+]o on the caffeine responses prevents a meaningful interpretation
of the effect of [Ca2+]o on the releasable content of the CICR pool.
Ca2+ Influx Through N Type Calcium Channels Selectively Elicits AHPslow. Six types of VDCCs have been described in neurons:
L, N, P, Q, R, and T ( 20 ). Nodose neurons express several types of VDCCs. By using a panel of pharmacologic reagents that are selective
for different types of VDCCs, we tested the contribution of each to the total AP-induced Ca2+ current. Our results, summarized in Table 3 ,
reveal that 85%
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CALCIUM REGULATION OF A SLOW POST-SPIKE HYPERPOLARIZATION IN VAGAL AFFERENT NEURONS 7655

of the AP-induced inward Ca2+ current is shared by L and N type Ca2+ channels ( Fig. 9 ). P, Q, and T type Ca2+ channel antagonists were
ineffective, suggesting that the remaining Ca2+ current is associated with Ca2+ influx through R type channels. Nifedipine (10 M), an L
type Ca2+ channel blocker, produced no measurable effect on either the AHPfast, the AHPmedium, or the AHPslow. By contrast, -conotoxin-
GVIA (0.5 M), a selective N type Ca2+ channel blocker, always obliterated the AHPslow, and in 50% of the neurons abolished the
AHPmedium (about half of the AHPmedium are Ca2+sensitive, see above), while leaving the AHPfast unaffected ( Fig. 9 and Table 4 .).
These results indicate that the SK and BK type K+ channels are both regulated by Ca2+ influx through N type channels. BK channels are
gated by influx Ca2+ directly (8), whereas SK channels are affected by influx Ca2+ indirectly (i.e., Ca2+ entering through N type VDCC
triggers RY receptors to release Ca2+ from CICR pools). Such a sequence implies a functional coupling between N type Ca2+ channels

FIG. 9. Effects of VDCC antagonists on AP-induced calcium currents, AHPslow and AP-induced Ca2+ transients. (A) Inward calcium
currents recorded in isolated nodose neurons evoked by a prerecorded AP waveform from a holding potential of 60 mV. From Left
to Right, control inward current in the presence of 2 mM [Ca2+]o and in the presence of 10 M nifedipine. After reestablishing control
conditions, the neuron was exposed to 1 M -conotoxin-GVIA. The effects of 500 M cadmium were recorded in another neuron;
the control current for this cell was similar to the first trace. (B) AHPslow evoked by a train of four APs (10 Hz) recorded in another
nodose neuron. From Left to Right, AHPslow evoked in control conditions, in the presence of 100 M CdCl2, after washout, in the
presence of 500 nM -conotoxin-GVIA, and after washout. (C) AP-induced Ca2+ transients recorded in two nodose neurons. From
Left to Right, Ca2+ transients evoked by a train of eight APs in normal Locke solution, and in Locke solution containing 10 M
nifedipine. In another neuron, 1 M -conotoxin-GVIA reduced the Ca2+ transient 50% (see Table 4 ). APs were evoked by 2.5-ms,
10-Hz depolarizing current pulses.
Table 3. Effects of Ca2+ channel blockers on action potential-induced inward Ca2+ currents
Channel type Channel blocker Concentration M Reduction n
T Amiloride 500 00 18
L Nifedipine 10 44 5.6 9
P/Q -AGA IVA 0.2 00 2
Q -CTX MVIIC 0.25 00 6
N -CTX GVIA 1 40 4.0 15

The blocking effect of amiloride, nifedipine, -agatoxin (AGA) IVA, -conotoxin (CTX) MVIIC, and -conotoxin (CTX) GVIA is
expressed as percent reduction in the peak amplitude of the total calcium current SEM. n corresponds to the number of cells for each
condition.

Table 4. Actions of specific Ca2+ channel blockers on the action potential-induced Ca2+ transient and the AHP slow
Reduction, %
Channel type Channel blocker Ca2+ transient n AHPslow amplitude n
L Nifedipine 57 7.7 21 00 5
N -CTX GVIA 39 6.2 4 100 0 6
T, R Nickel nd 00 5
All Cadmium 100 0 2 100 0 6

The following concentrations of antagonists were used: nifedipine (10 M), -conotoxin GVIA (0.5 M or 1 M), nickel (50500 M),
the authoritative version for attribution.

and cadmium (100 M). nd, not determined.

lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CALCIUM REGULATION OF A SLOW POST-SPIKE HYPERPOLARIZATION IN VAGAL AFFERENT NEURONS 7656

and R Y channels in the endoplasmic reticulum. We tested this proposition by examining the effects of VDCC antagonists on the
magnitude of AP-induced Cat.
Ca2+ influx through both L and N type Ca2+ channels triggers CICR. The magnitude of the Cat is a sensitive indicator of Ca2+
release from the CICR pool. To determine the relative influence of Ca2+ influx through L and N type channels on release from the CICR
pool, we applied selective VDCC antagonists and monitored the amplitude of Cat. Nifedipine (10 M) and -conotoxin-GVIA (0.51.0
M) diminished the amplitude of the Cat by 57% and 39%, respectively ( Fig. 9 and Table 4 ). These results reveal that Ca2+ entering
through either L or N type Ca2+ channels provides trigger Ca2+ to stimulate CICR. Given that the amount of Ca2+ influx through L and N
type Ca2+ channels is comparable (44% and 40%, respectively, of total AP-induced Ca2+ influx; see Table 3 ), there must be a remarkable
spatial arrangement between plasma membrane N type Ca2+ channels, endoplasmic reticulum R Y receptors, and plasma membrane SK
channels. Our working hypothesis concerning the regulation of the AHPslow by Ca2+ is illustrated schematically in Fig. 10 .
DISCUSSION
Whether recorded in intact vagal sensory ganglia or in acutely isolated vagal afferent somata (nodose neurons), single APs can elicit
an AHPslow that exhibits a delayed onset (50300 ms), a slow time to peak amplitude (0.30.5 s), and a particularly long duration (215 s)
(14, 21). Inhibition of the AHPslow by numerous inflammatory mediators (e.g., bradykinin, prostanoids, histamine, serotonin, leukotriene
C4; see Table 1 ) results in an increased neuronal excitability and a loss of spike-frequency adaptation. Thus, modulation of the AHPslow by
these mediators provides a mechanism for peripheral nociceptor sensitization that may underlie allergic inflammation-induced hyperalgesia.
One unresolved but important mechanistic question revolves around the delayed onset and protracted duration of the AHPslow. Many
of our studies of nodose AHPslow were performed with acutely dissociated adult neurons, which are essentially spherical structures lacking
dendritic and axonal processes. Thus, the delayed onset of the AHPslow cannot be due to slow diffusion of Ca2+ from distal sites of influx
to somal SK channels. The high temperature coefficient (Q10 > 3.0) for the rising phase and the decay time constant of the nodose AHPslow
( 14 ) also argues against simple Ca2+ diffusion as an explanation for the slow kinetics of the AHPslow. The time course of the AHPslow
could arise from unusual channel kinetics of the SK channels. This also appears unlikely if SK channels in nodose neurons have activation
kinetics similar to those cloned from rat brain ( 22 ). Recombinant SK channels from rat brain have activation time constants that are
orders of magnitude shorter than the rise time of the AHPslow. It is more likely that the time course of the AHPslow is a consequence of the
Cat because of CICR.

FIG. 10. Schematic diagram of the relation between plasma membrane Ca2+ channels, BK, and SK potassium channels and
endoplasmic reticulum R Y receptors in primary vagal afferent neurons. Single APs evoke Ca2+ influx through L and N type VDCCs.
Ca2+ influx through either of these channels can trigger release of Ca2+ from the endoplasmic reticulum via RY receptors. Whereas
BK channels are activated directly by Ca2+ entering the neuron via N type VDCC, SK channels are activated indirectly. SK channels
require Ca2+ from CICR pools released after Ca2+ influx through N type channels.
If the AHPslow is directly dependent on Ca2+ released from the CICR pool, the AHPslow and the AP-induced rise in [Ca2+]i should
display similar kinetics. Quantitative kinetic comparisons between these two variables are subject to some uncertainty, because the time
course of the Cat reflects global changes in [Ca2+ ]I, whereas the kinetics of the AHPslow are determined by events at the plasma
membrane. Nonetheless, we determined the time-to-peak and 10-to-90% decay time for both the AHPslow and the Cat elicited by one to
eight APs ( 19 ). The time-to-peak for AHPslow was significantly slower than the Cat by nearly a factor of a two (1.0 s vs. 1.9 s); the Cat
also decayed more rapidly than the AHPslow (3 s vs. 7 s). Analogous temporal discrepancies have been reported between the Cat and
AHPslow in vagal motoneurons ( 23 ). Such temporal differences suggest that Ca2+ released from CICR pools does not act alone to gate
AHPslow K+ channels. Cloned SK channels contain many potential phosphorylation sites ( 15 ); Ca2+-dependent phosphorylation and/or
dephosphorylation may thus be additional processes in the signaltransduction pathway of AP-evoked AHPslow.
Unambiguous data now exist showing that Ca2+ can directly activate SK channels in hippocampal neurons ( 24 ) and in Xenopus
oocytes ( 22 ). In nodose neurons, it is less clear whether Ca2+ alone is sufficient to activate and sustain the AHPslow after an AP. In
hippocampal neurons, flash photolysis of a caged Ca2+ chelator immediately truncates AP-induced AHPslow, suggesting that elevated
intracellular Ca2+ is required to maintain the AHPslow ( 25 ). These results do not, however, distinguish between continuous Ca2+ gating of
SK channel and the involvement of other Ca2+-dependent factors sustaining the longevity of the AHPslow. It is also possible that Ca2+-
dependent factors act synergistically with Ca2+ to control SK channels ( 23 ). The nearly spherical morphology and large size of acutely
isolated adult nodose neurons provide a favorable preparation to determine the nature of second messengers required to activate and
sustain the AHPslow.
In conclusion, a subset of vagal primary afferent neurons possess a slowly developing and long-lasting spike afterhyperpolarization,
the AHPslow, that can profoundly affect the discharge frequency of these visceral afferent neurons. Although AP-evoked Ca2+ influx via
both L and N type Ca2+ channels triggers CICR, only Ca2+ flux through N type channels activates the CICR-dependent AHPslow. This type
of specificity suggests that spatial coupling between N type Ca2+ channels and SK channels may be critical for the generation of the
AHPslow in nodose neurons. The exact mechanism coupling Cat to the AHPslow current remains to be determined.
the authoritative version for attribution.

We thank our coworkers who participated in many of the experiments described in this manuscript: Drs. Akiva Cohen, Samir Jafri,
and Bill Wonderlin, and Mr. Glen Taylor. The authors also thank Dr. Liz Katz and Mr. Eric Lancaster for their constructive suggestions on
an earlier draft of this manuscript. This work was supported by National Institutes of Health Grants GM-46956 to J.P.Y.K., NS-22069 to
D.W. and Training Grant NS-07375 to K.A.M.
1. Undem, B. J. & Riccio, M. M. ( 1997 ) in Asthma , eds. Barnes, P. J. , Grunstein, M. M. , Leff, A. & Woolcock, A. J. ( Lippincott , Philadelphia ), pp.
10091026 .
2. Weinreich, D. ( 1995 ) Pulm. Pharmacol 8 , 173179 .
3. Undem, B. J. , Hubbard, W. & Weinreich, D. ( 1993 ) J. Auton. Nerv. Syst. 44 , 3544 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CALCIUM REGULATION OF A SLOW POST-SPIKE HYPERPOLARIZATION IN VAGAL AFFERENT NEURONS 7657

4. Weinreich, D. , Moore, K. A. & Taylor, G. E. ( 1997 ) J. Neurosci. 17 , 76837693 .

5. Moore, K. A. , Taylor, G. E. & Weinreich, D. ( 1999 ) J. Physiol. (London) 514.1 , 111124 .
6. Greene, R. , Fowler, J. C. , MacGlashlan, D., Jr. & Weinreich, D. ( 1988 ) J. Appl. Physiol. 64 , 22492253 .
7. Sah, P. ( 1996 ) Trends Neurosci. 19 , 150154 .
8. Blatz, A. L. & Magleby, K. L. ( 1987 ) Trends Neurosci. 10 , 463467 .
9. Gold, M. S. , Shuster, M. J. & Levine, J. D. ( 1996 ) Neurosci. Lett. 205 , 161164 .
10. Villire, V. & McLachlan, E. M. ( 1996 ) J. Physiol. (London) 76 , 19241941 .
11. Coleridge, J. C. G. & Coleridge, H. M. ( 1984 ) Rev. Physiol. Biochem. Pharmacol. 99 , 1110 .
12. Weinreich, D. & Wonderlin, W. F. ( 1987 ) J. Physiol. (London) 394 , 415427 .
13. Higashi, H. , Morita, K. & North, R. A. ( 1984 ) J. Physiol. (London) 355 , 479492 .
14. Fowler, J. C. , Greene, R. & Weinreich, D. ( 1985 ) J. Physiol. (London) 365 , 5975 .
15. Khler, M. , Hirschberg, B. , Bond, C. T. , Kinzie, J. M. , Marrion, N. V. & Adelman, J. P. ( 1996 ) Nature (London) 273 , 17091714 .
16. Cohen, A. S. , Moore, K. A. , Bangalore, R. , Jafri, M. S. , Weinreich, D. & Kao, J. P. Y. ( 1997 ) J. Physiol. (London) 499 , 315328 .
17. Shmigol, A. , Verkhratsky, A. & Isenberg, G. ( 1995 ) J. Physiol. (London) 489 , 627636 .
18. Scroggs, R. S. & Fox, A. P. ( 1992 ) J. Neurosci. 12 , 17891801 .
19. Moore, K. A. , Cohen, A. S. , Kao, J. P. Y. & Weinreich, D. ( 1998 ) J. Neurophysiol. 79 , 688694 .
20. Dunlap, K. , Luebke, J. I. & Turner, T. J. ( 1995 ) Trends Neurosci. 18 , 8998 .
21. Leal-Cardosa, H. , Koschorke, G. M. , Taylor, G. & Weinreich, D. ( 1993 ) J. Auton. Nerv. Syst. 45 , 2939 .
22. Hirschberg, B. , Maylie, J. , Adelman, J. P. & Marrion, N. V. ( 1998 ) J. Gen. Physiol. 111 , 565581 .
23. Lasser-Ross, B. , Ross, W. N. & Yarom, Y. ( 1997 ) J. Neurophysiol. 78 , 825834 .
24. Marrion, N. V. & Tavalin, S. J. ( 1998 ) Nature (London) 395 , 900905 .
25. Lancaster, B. & Zucker, R. S. ( 1994 ) J. Physiol. (London) 475 , 229239 .
26. Weinreich, D. , Koschorke, G. M. , Undem, B. J. & Taylor, G. E. ( 1995 ) J. Physiol. (London) 483.3 , 735746 .
27. Jafri, M. S. , Moore, K. A. , Taylor, G. E. & Weinreich, D. ( 1997 ) J. Physiol. (London) 503.3 , 533546 .
28. Christian, E. P. , Taylor, G. E. & Weinreich, D. ( 1989 ) J. Appl. Physiol. 67 , 584591 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ION CHANNELS GATED BY HEAT 7658

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Ion channels gated by heat

P. CESARE, A. MORIONDO, V. VELLANI, AND P. A. MCNAUGHTON *

Neuroscience Research Centre, Kings College London Strand, London WC2R 2LS, United Kingdom
ABSTRACT All animals need to sense temperature to avoid hostile environments and to regulate their internal homeostasis.
A particularly obvious example is that animals need to avoid damagingly hot stimuli. The mechanisms by which temperature is
sensed have until recently been mysterious, but in the last couple of years, we have begun to understand how noxious thermal
stimuli are detected by sensory neurons. Heat has been found to open a nonselective cation channel in primary sensory neurons,
probably by a direct action. In a separate study, an ion channel gated by capsaicin, the active ingredient of chili peppers, was
cloned from sensory neurons. This channel (vanilloid receptor subtype 1, VR1) is gated by heat in a manner similar to the native
heat-activated channel, and our current best guess is that this channel is the molecular substrate for the detection of painful heat.
Both the heat channel and VR1 are modulated in interesting ways. The response of the heat channel is potentiated by
phosphorylation by protein kinase C, whereas VR1 is potentiated by externally applied protons. Protein kinase C is known to be
activated by a variety of inflammatory medi ators, including bradykinin, whereas extracellular acidification is characteristically
produced by anoxia and inflammation. Both modulatory pathways are likely, therefore, to have important physiological correlates
in terms of the enhanced pain (hyperalgesia) produced by tissue damage and inflammation. Future work should focus on
establishing, in molecular terms, how a single ion channel can detect heat and how the detection threshold can be modulated by
hyperalgesic stimuli.
Organisms sense temperature for all sorts of reasons. Highly accurate thermosensation is required to set the body temperature of a
mammal. Simpler animals sense the external temperature to seek out favorable environments for feeding or for mating. Damaging
extremes of temperature must be avoided, of course, and for this purpose, pain-sensitive nerve terminals detect very low and very high
temperatures and induce an avoidance response. In all these instances, temperature must be detectedbut how? In many cases, the
detection mechanism involves a specialized temperature-sensitive nerve terminal, which, on application of a temperature change,
generates a depolarization and a resulting train of action potentials in the sensory nerve axon. So it is perhaps obvious to state that
temperature must gate an ion channel in the sensory nerve terminal. But how does it work? One could imagine a temperature-sensitive
biochemical pathway that modulates an internal transmitter and in turn gates the ion channel. There is some evidence for such a system in
at least one temperature-sensitive pathway in the nematode Caenorhabditis elegans (see below). But in the only other instance of which
we have any understanding, the action of temperature on the ion channel seems instead to be direct. To the existing voltage-gated, ligand-
gated, and mechanosensitive ion channels, we can therefore add a fourth major category of ion channels, namely, heat-sensitive ion
channels. This article reviews our understanding to date of this newly characterized class of ion channels.
Heat-Sensitive Ion Channels in Primary Sensory Neurons. The most direct way to study the detection of hot stimuli is in situ,
either by asking subjects at what temperature a sensation of warmth changes to a sensation of pain or alternatively by recording the
frequency of action potentials in the axon of a primary pain-sensitive neuron (a nociceptor) while a thermal stimulus is applied to the
receptive field. Experiments like these have shown that, as the temperature is raised, a sensation of warmth changes to pain at around 43
45C and that the intensity of the pain sensation increases steeply thereafter ( 1 ). Recordings of action potentials from nociceptive nerve
fibers show a similar picture, namely of a threshold for initiation of action potentials at 4345C and a steep increase in firing frequency as
the temperature is increased further ( 1 , 2 ).
To take things much further, for instance to study the pathways controlling ionic currents involved in the transduction process, a
preparation of isolated nociceptors is needed. Other sensory receptors can be isolated more or less intact, and the study of isolated
photoreceptors, auditory receptors, olfactory receptors, etc. has told us a great deal about their mechanisms of operation. Nociceptors are
unfortunately a much more difficult proposition. The sensory terminals are extremely fine and are embedded in a cellular matrix whose
disruption during dissection releases the very signaling molecules that the nociceptor nerve terminal is designed to detect. The difficulty of
isolating intact nociceptive nerve terminals has meant that studies on isolated nociceptors have all been on neuronal cell bodies. In a
typical procedure, the neuronal cell body is isolated by enzymatic treatment and is cultured for a few days before use ( 3 , 4 ). The sensory
terminals are, of course, completely removed during the isolation procedure, and we must hope that the properties of those terminals are
recreated in the cultured cell body and dendrites. When they are acutely isolated, nociceptive cell bodies often fail to respond to noxious
stimuli, but in a process that is poorly understood, nociceptive properties characteristic of the sensory terminal reappear after a few days in
culture in the presence of serum and nerve growth factor ( 3 6 ). This preparation of cultured nociceptors has been used for almost all
experiments investigating the cellular and molecular basis of detection of painful stimuli. The complexity of the procedure for isolating
nociceptors nonetheless makes it essential that we check very carefully that our nociceptors responses resemble those in vivo.
An example of the response of the membrane potential of a cultured nociceptor to application of a 49C heat stimulus is shown in
Fig. 1 ( 6 , 7 ). As is seen in nociceptive nerve terminals
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: PKC, protein kinase C; VR1, vanilloid receptor subtype 1.
* To whom reprint requests should be addressed. e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ION CHANNELS GATED BY HEAT 7659

in vivo ( 8 ), the heat stimulus causes a rapid depolarization and initiates a train of action potentials. The fact that this response is present in
isolated nociceptors shows that no other cells are necessary to produce the response to heat; there is no signal molecule released from
adjacent damaged cells to which the nociceptor responds. Nor is the response due to damage to the nociceptive neuron itselfas might
occur if heat were causing a breakdown in the plasma membrane and a consequent depolarizationbecause the depolarization and action-
potential firing ceases immediately when the stimulus is withdrawn and because similar responses can be elicited again and again on
repeated application of the heat stimulus.

FIG. 1. Depolarization and a train of action potentials initiated in a nociceptive neuron in culture by application of a brief heat
stimulus. Membrane potential recorded by using whole-cell patch clamp (see ref. 6 for details).
The reproducibility of the response in isolated nociceptors resembles that in other sensory receptors. Interestingly, though, the
behavior of nociceptors in vivo is different. Repeated application of a strong stimulus leads to a progressive increase in the response in
nociceptors in vivo but not in isolated nociceptors nor in other sensory receptors. This process, known as sensitization or hyperalgesia, is
characteristic of nociceptors in vivo and has obvious protective value for the organism as a whole, in that the pain caused by a damaging
stimulus becomes more urgent if the stimulus is repeated or maintained. The fact that sensitization is not observed in isolated nociceptors
suggests that the phenomenon is not intrinsic to the neuron but instead has its origin in extracellular signals released from nearby damaged
or inflamed tissue ( 9 ). Recent advances in our understanding of this process of sensitization are discussed below.
The membrane current induced by heat in a voltage-clamped nociceptor is shown in Fig. 2 A. The current is activated rapidly (but not
instantaneously) by heat, with a mean time to half activation of 35 ms at 50C ( 6 , 7 ). By contrast, neurons insensitive to heat show only
a small current change, which occurs as rapidly as the solution change and therefore probably has a simple physical origin such as a
temperature dependence of membrane leakage resistance. When the temperature dependence of the heat-sensitive current is examined (
Fig. 2 B), the current can be seen to be activated above 42C and to increase exponentially as the temperature is raised further, much as is
observed in nociceptors in vivo ( 1 , 2 ).
Experiments examining the ionic selectivity of the heat-activated current ( 6 , 7 ) have shown that the heat-activated channel
discriminates poorly amongst monovalent alkali cations, in common with many other ion channels such as those gated by glutamate,
acetylcholine, or cyclic nucleotides. Calcium ions can by themselves carry current through the channel but, in addition, have the effect of
partially blocking a current carried by monovalent ions. The channel must therefore possess a binding site in the pore region with a higher
affinity for Ca2+ than for monovalent cations. Contrary to early reports ( 10 ), the channel does not seem to be blocked by Cs+ ions. The
current-voltage relation shows outward rectification and a reversal potential of around 0 mV under physiological conditions ( 6 , 10 , 11 ).

FIG. 2. Responses of membrane current in isolated nociceptors to heat. (A) Application of a rapid step change in temperature (from
room temperature to 49C; time course shown by the top trace) elicits an inward current with a short delay (35 ms) in a nociceptive
neuron (lower of the two membrane current traces). In heat-insensitive neurons, a much smaller current change is elicited with no
delay (top current trace). Neurons were voltage-clamped by the whole-cell patch-clamp method at 70 mV. (B). Current as a function
the authoritative version for attribution.

of temperature in a heat-sensitive neuron (lower trace) and in a heat-insensitive neuron (upper trace). Modified from ref. 6.
Single heat-activated channels have a conductance of around 3040 pS ( 10 , 11 ). The single channel conductance itself is only
weakly temperature-dependent, in common with other ion channels, and the pronounced dependence of current on temperature is caused
by a strong temperature dependence of the probability of channel opening. The time constants of channel opening can be deduced from the
characteristics of the current noise produced when several channels are present simultaneously in a cell-attached membrane patch (see
Fig. 3 ). Application of heat activates inward current and current noise in heat-sensitive but not in heat-insensitive neurons ( Fig. 3 A). The
power spectrum of heat-induced noise ( Fig. 3 B) indicates the existence of two Lorentzian components with time constants 1 = 17 ms and
2 = 0.49 ms. A simple two-state model reproducing these features is
C1 C2 O,
1 2
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ION CHANNELS GATED BY HEAT 7660

where 1 is the time constant of transition from the closed state C1 to C2 and 2 is the time constant of transition from the closed state C2 to
the open state (O). The similarity of 1 to the half-time of channel opening after application of a heat jump ( Fig. 2 A) suggests that it is the
transition between C1 and C2 that is the main temperature-dependent event.

FIG. 3. Noise associated with opening of heat-activated ion channels. (A) Examples of cell-attached patch-clamp recordings of the
responses of a heat-insensitive neuron (top recording) and a heat-sensitive neuron (bottom recording) to temperature steps. Patch
pipette contained only 154 mM NaCl/10 mM Hepes to maximize current through heat-sensitive ion channels and was held at 0 mV.
Heat-sensitive currents through channels outside the patch were prevented by bathing the rest of the cell in solution free of permeant
ions (154 mM N-methyl glucamine/10 mM Hepes). (B) Power spectrum of cell-attached current activated by a temperature step to 49
C. For each experiment, 10 consecutive traces (16,384 samples per trace at 10 kHz; filtered at 4 kHz) were acquired, first at room
temperature then at 49C, and the power spectra were calculated. Heat-induced power spectrum (points) was calculated as the
difference between the two spectra. Power spectrum fitted with the sum of two Lorentzian functions (sum shown as a solid line;
component spectra as a dashed line) with half-power frequencies as shown, corresponding to time constants of 17 ms and 0.49 ms
(V.V., P.C., and P.A.M., unpublished data).
Electrophysiologists who work on cultured sensory neurons tend to think of their cells as a bimodal population, consisting of
nociceptive and nonnociceptive neurons. Whole-animal physiologists who work on nociceptors in situ know differently; nociceptors come
in many different varieties, with properties such as heat, mechanical, and chemical sensitivity present to variable extents in different single-
unit recordings. The main division, of course, is between slowly conducting, unmyelinated nociceptive nerve fibers, which commonly
respond to a wide range of stimuli (polymodal fibers) and more rapidly conducting myelinated nerve fibers, which frequently respond to a
smaller subset of noxious stimuli, but amongst which heat-sensitive units are also commonly encountered ( 12 , 13 ). A corresponding
division of nerve-cell bodies is seen, both in dorsal root ganglia and in cultured preparations, into small-diameter dark neurons and large-
diameter pale neurons ( 14 ). Heat sensitivity is, perhaps reassuringly, seen in both cell types in culture ( 10 , 15 ), but there is a
quantitative difference: the smaller cells have a threshold of around 45C, whereas larger cells form a different population with a threshold
of 51C ( 15 ). Interestingly, only the former population responds to capsaicin, suggesting that there is more than one heat-sensitive
channel at work in this diverse population of nociceptors (see further discussion below).
Thermosensation in C. elegans. The nematode worm C. elegans is capable of seeking out a preferred temperature at which to feed,
and mutants unable to detect temperature can therefore be selected by isolating individuals that stray from preferred-temperature areas.
These worms have a mutation either in a gene, tax-4, that codes for the -subunit of an ion channel gated by cyclic nucleotides ( 16 ) or in
a second gene, tax-2, that codes for a -subunit ( 17 ). The fact that these channels can be gated by cyclic nucleotides suggests (but does
not prove) that the mechanism of thermosensation is the modulation of the pathway that controls the level of cyclic nucleotides, rather than
the direct action of heat on the ion channel itself. In this respect, detection of nonnoxious temperatures in C. elegans is different from
mammalian noxious heat sensation. The latter depends only on expression of a heat-sensitive ion channel, which can be seen to function in
isolated membrane patches and therefore is not gated by diffusible messengers controlled by intracellular signaling pathways ( 11 , 18 ,
19 ), whereas thermosensation in C. elegans seems to depend on cyclic nucleotides as intracellular messengers, and the thermosensitive
element is therefore presumably some stage in the pathway modulating the level of cyclic nucleotides. There are many forms of
mammalian thermosensation, as outlined in the introduction, and it is quite possible that signaling pathways are involved in some of these,
even though they do not seem to be directly responsible for heat sensation in the nociceptors of higher vertebrates.
Sensitization of Nociceptors. The process of sensitization (or hyperalgesia) is familiar to us all: a stimulus strong enough to cause
tissue damage hurts more with time, and even after the stimulus has been removed, the damaged area is hypersensitive to touch and to
temperature. This phenomenon can be attributed partly to changes in pain transmission in the spinal cord or at higher levels, but an
important component results from processes occurring at the site of injury. A large number of molecules released by tissue damage are
known to act as mediators of hyperalgesia. Examples include neuropeptides, prostaglandins, histamine, platelet-activating factor, and
bradykinin ( 9 , 20 ).
the authoritative version for attribution.

With so many different factors able to cause hyperalgesia, it is perhaps no surprise that more than one cellular mechanism is involved.
One recently elucidated mechanism involves activation of protein kinase A. External inflammatory messengers such as prostaglandins,
serotonin, and adenosine activate adenylate cyclase and consequently increase the level of cAMP, leading to activation of protein kinase A
( 21 , 22 ). The principal physiologically important target of protein kinase A seems to be a recently identified voltage-sensitive Na
channel ( 23 ), which, unlike the more usual neuronal Na channel, is not blocked by tetrodotoxin. The effect of phosphorylation of the
tetrodotoxin-resistant Na channel is to lower its threshold, thereby making it more likely that an action potential will be elicited ( 21 , 22 ).
This membrane ionic current is probably not the only one modulated by cAMP, as actions on a K+ current and on a voltage and cyclic
nucleotide-gated conductance have also been identified ( 24 , 25 ). All of these cAMP-dependent mechanisms, however, operate in the
same direction, in that they sensitize the nociceptive nerve terminal to any stimulus that is capable of exciting it, because the effect is to
reduce the
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ION CHANNELS GATED BY HEAT 7661

threshold for action potential firing, rather than on the specific receptor current induced by the stimulus.
A second and more specific mechanism uses activation of protein kinase C (PKC) to sensitize the response to heat. Fig. 4 A shows
that the inflammatory mediator bradykinin potently increases the membrane current activated by a heat pulse ( 6 ). Bradykinin is known to
activate both phospholipase C and phospholipase A2, thereby releasing a number of potential intracellular signaling molecules. The
enhancement of the heat response can be shown to be due to activation of PKC, however, and not to other possible intracellular mediators,
because it can be mimicked by phorbol esters, which are specific activators of PKC ( Fig. 4 B). Fig. 4 C shows that activation of PKC
increases the current activated by a heat stimulus and shifts the relation between temperature and membrane current to lower temperatures.
These observations predict that normally innocuous temperatures, such as body warmth, will therefore become painful after sensitization,
an observation that corresponds well both with experiments on intact preparations and with our personal experience that even the warmth
of a hand can cause a sensation of pain when applied to an injured area of the body. Other evidence supporting the identity of PKC as an
intracellular mediator of sensitization includes the findings that sensitization can be reversed by PKC inhibitors and can be prolonged by
phosphatase inhibitors, which prevent dephosphorylation after a protein target has been phosphorylated by PKC ( 6 ). The possibility that
mediators other than bradykinin may also employ the PKC pathway to induce sensitization has not yet been investigated and certainly
deserves to be.

FIG. 4. Phosphorylation by PKC sensitizes the heat response of nociceptors. (A) Response of membrane current to a 49C heat pulse
before and after exposure to bradykinin (Bk). (B) Similar effect to that seen in A is observed after treatment by the specific PKC
activator phorbol myristate acetate (PMA). (C) Current vs. temperature relations before and after PMA treatment, showing
sensitization of the heat response. [Reproduced with permission from ref. 6 (Copyright 1996, Proceedings of the National Academy
of Sciences of the United States of America)].
Desensitization of Nociceptors. When a long pulse of moderate heat, insufficiently strong to cause cell damage and to release the
extracellular mediators responsible for sensitization, is applied to a heat-sensitive nociceptor in vivo, gradual adaptation or desensitization
in the firing frequency is observed ( 26 ). A similar phenomenon is seen in isolated nociceptors ( Fig. 5 ), showing that desensitization,
unlike sensitization, is intrinsic to the nociceptor. Recent experiments in our lab have shown that desensitization is triggered by an influx
of calcium ions from the external medium through the heat-sensitive ion channel. In this respect, desensitization of the heat response
resembles the desensitization in response to prolonged application of capsaicin ( 27 ), which is triggered by activation of the calcium-
dependent phosphatase calcineurin by an influx of calcium through the capsaicin-gated channel itself. These observations suggest that both
the heat-activated channel and the capsaicin-gated channel are desensitized when dephosphorylated by calcineurin, one of many
similarities between the two (see below). The molecular mechanisms of sensitization of the heat-activated channel (phosphorylation by
PKC; see above) and desensitization (dephosphorylation by calcineurin) may therefore be simply complementary aspects of the same
process, in which the heat sensitivity of the channel is regulated by phosphorylation (see Fig. 5 and discussion below).
The Capsaicin Receptor, Vanilloid Receptor Subtype 1 (VR1), and Its Relation to Heat Sensation. Capsaicin, the active
ingredient of chili peppers, has been known for some time to depolarize nociceptive nerve terminals by a direct action on an ion channel (
28 ). Capsaicin is not part of the normal environment of most animals. Therefore, it had always been supposed that the capsaicin receptor
was gated physiologically by an endogenous agonist, just as the morphine-receptor family is activated physiologically not by morphine but
by endogenous opiates. Capsaicin-responding neurons can be activated by low pH, and, as pH can drop considerably during inflammation,
hydrogen ions were a plausible candi
the authoritative version for attribution.

FIG. 5. The heat-sensitive current in a nociceptor undergoes desensitization in response to a maintained pulse of heat (P.C., A.M.,
V.V., and P.A.M., unpublished data).
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ION CHANNELS GATED BY HEAT 7662

date for the physiological agonist activating these nociceptors ( 28 ).

The capsaicin receptor VR1 has recently been cloned by an ingenious strategy by using imaging of the increase in internal calcium
caused by application of capsaicin to detect expression of capsaicin-receptor clones ( 18 ). The expressed receptor is indeed sensitive to
low pH, but, perhaps more interestingly, it responds to heat like the native heat receptor, as outlined above. The main points of
resemblance are as follows (see refs. 18 and 19). (i) The current passing through both channels is zero at room temperature and increases
sharply above about 42C. (ii) The capsaicin receptor is a cation channel with an ionic selectivity similar to that of the native heat receptor.
(iii) The single-channel conductance and the current-voltage relation are similar, (iv) The open time constant of VR1, 0.9 ms, is similar to
the fast open time constant of the heat-activated channel (0.5 ms; see above), (v) The actions of capsaicin and heat are synergistic on both
VR1 and the native heat receptor ( 19 , 29 ). (vi) VR1 is expressed exclusively in small neurons of primary sensory ganglia.
There may, however, be one crucial point of difference: the ion current through VR1 is blocked by the capsaicin-channel antagonists
capsazepine and ruthenium red, whether the current is elicited by capsaicin application or by heat ( 19 ). However, in cultured nociceptors,
the current induced by capsaicin is blocked by these agonists ( 28 ), but the response to heat does not seem to be ( 30 ).
One particularly interesting feature of VR1 is the interaction between its heat sensitivity and its pH sensitivity. At normal pH, VR1 is
activated only at temperatures above 42C. Low pH acts as a sensitizing agent, which reduces the threshold for activation by heat to 30
C at pH 6.3 ( 19 ). The sensitizing effect of pH explains the observation that capsaicin receptors are activated by low pH; at a sufficiently
low pH, room temperature is adequate to induce channel openings ( 19 ). In inflamed or anoxic tissue, the pH can drop to as low as 6.0,
and at this pH, body temperature would be sufficient to activate VR1. The pain of inflammation and anoxia may therefore be explained at
least partly by a combined effect of low pH and normally innocuous temperature on VR1.
Is VR1 the only heat-detecting mechanism in nociceptors? Probably not, in view of the observation by Nagy and Rang ( 15 ) that the
two properties of heat sensitivity and capsaicin sensitivity (and therefore presumably expression of VR1) are not absolutely colocalized in
sensory neurons, contrary to an earlier report based on a smaller number of experiments ( 31 ). A recent study ( 32 ) reports the cloning of
a vanilloid receptorlike channel (VRL-1) that is not sensitive to capsaicin but is gated by temperatures above 52C. Expression of this
channel may explain the responses to higher temperatures observed in some capsaicin-insensitive neurons ( 15 ).
How Does the Heat Receptor Work? How ligand-gated or voltage-gated ion channels might work is intuitively fairly obvious, at
least in terms of general principles. Ligand-gated channels operate like a lock and key; insertion of the key (the ligand) stabilizes the open
state of the channel. Voltage-gated channels possess a charged gating unit within the membrane field, such that changes in the membrane
potential move this unit and thereby induce a conformational change that gates the channel open or closed. How small elevations in
temperature might shift the heat-sensitive channel from the closed to the open state is less intuitively obvious but must depend on the well
known thermodynamic equation
G = H TS.
The change in the equilibrium between closed and open states of the channel, which depends on the Gibbs free energy change (G),
can be markedly temperature-dependent only if there is a large entropy difference (S) between the two states. Elevations in temperature
(T) must therefore cause the heat-sensitive ion channel to change from an ordered to a more disordered state, as occurs during melting of
ice or dissolving of a salt in water. It does not seem that any accessory protein or signaling pathway is needed to gate the channel, because
heat-sensitive ion channels can be seen in cell-free membrane patches from nociceptors ( 11 ) and because VR1 functions as a heat
receptor when heterologously expressed in HEK 293 cells or in Xenopus oocytes ( 18 , 19 ). The temperature-sensitive gating unit is likely,
therefore, to be intrinsic to the heat-sensitive channel protein.

FIG. 6. Possible model for sensitization of the heat response. Acidification of the external solution causes protonation of an external
site on the heat-sensitive ion channel and consequent sensitization ( 19 ). Phosphorylation at the internal surface has a similar effect (
6 ). Both effects are reversible. PP, phosphatase.
The process of sensitization, which shifts the relation between temperature and channel opening to lower temperatures (see Fig. 3 C),
must act by stabilizing the more disordered, higher-temperature state of the channel in such a way that lower temperatures are needed to
induce channel opening. How might this interesting and physiologically important process be operating? One possibility is that opposite
changes in the charge on either side of the membrane may be important. The work of Tominaga et al. ( 19 ) has shown that protonation of
an external site of VR1 induces sensitization, and work in our own lab has shown that phosphorylation of an internal site of the heat-
sensitive receptor induces an apparently identical sensitized state ( 6 ). If VR1 and the heat-sensitive receptors are one and the same, as is
suggested by most lines of evidence (see above), then we can put these two observations together in a simple (and speculative) model of
the sensitization process ( Fig. 6 ). In this model, addition of positive charge to the external face of the membrane or addition of negative
charge to the internal face have equivalent effects, with both manipulations leading to stabilization of a disordered state of the protein and
consequently to sensitization of the response to heat.
1. Treede, R. D. , Meyer, R. A. , Raja, S. N. & Campbell, J. N. ( 1992 ) Prog. Neurobiol. 38 , 397421 .
2. Belmonte, C. & Gallar, J. ( 1996 ) in Neurobiology of Nociceptors , eds. Belmonte, C. & Cervero, F. ( Oxford Univ. Press , Oxford ), Vol. 6 , pp. 146
183 .
the authoritative version for attribution.

3. Rang, H. P. , Bevan, S. & Dray, A. ( 1994 ) Textbook of Pain , eds. Melzack, R. & Wall, P. ( Churchill Livingstone , Edinburgh ), Vol. 3 , pp. 5778 .
4. Baccaglini, P. I. & Hogan, P. G. ( 1983 ) Proc. Natl. Acad. Sci. USA 80 , 594598 .
5. Gilabert, R. & McNaughton, P. A. ( 1997 ) J. Neurosci. Methods 71 , 191198 .
6. Cesare, P. & McNaughton, P. A ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 1543515439 .
7. Cesare, P. & McNaughton, P. A. ( 1997 ) Curr. Opin. Neurobiol. 7 , 493499 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ION CHANNELS GATED BY HEAT 7663

8. Treede, R. D. , Meyer, R. A. , Raja, S. N. & Campbell, J. N. ( 1995 ) J. Physiol. 483 , 747758 .

9. Kress, M. & Reeh, P. W. ( 1996 ) in Neurobiology of Nociceptors , eds. Belmonte, C. & Cervero, F. ( Oxford Univ. Press , Oxford ), Vol. 11 , pp.
258297 .
10. Reichling, D. B. & Levine, J. D. ( 1997 ) Proc. Natl. Acad. Sci. USA 94 , 70067011 .
11. Nagy, I. & Rang, H. P. ( 1998 ) J. Physiol. 507 , 29P .
12. Besson, J. M. & Chaouch, A. ( 1987 ) Physiol. Rev. 67 , 67155 .
13. Lynn, B. , Faulstroh, K. & Pierau, F. K. ( 1995 ) Eur. J. Neurosci. 7 , 431437 .
14. Lawson, S. N. , Perry, M. J. , Prabhakar, E. & McCarthy, P. W. ( 1993 ) Brain Res. Bull. 30 , 239243 .
15. Nagy, I. & Rang, H. P. ( 1999 ) Neuroscience , 88 , 995997 .
16. Komatsu, H. , Mori, I. , Rhee, J. , Akaike, N. & Ohshima, Y. ( 1996 ) Neuron 17 , 707718 .
17. Coburn, C. M. & Bargmann, C. I. ( 1996 ) Neuron 17 , 695706 .
18. Caterina, M. J. , Schumacher, M. A. , Tominaga, M. , Rosen, T. A. , Levine, J. D. & Julius, D. ( 1997 ) Nature (London) 389 , 816824 .
19. Tominaga, M. , Caterina, M. J. , Malmberg, A. B. , Rosen, T. A. , Gilbert, H. , Skinner, K. , Raumann, B. E. , Basbaum, A. I. & Julius, D. ( 1998 )
Neuron 21 , 531543 .
20. Levine, J. D. , Fields, H. L. & Basbaum, A. I. ( 1993 ) J. Neurosci. 13 , 22732286 .
21. Gold, M. S. , Reichling, D. B. , Shuster, M. J. & Levine, J. D. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 11081112 .
22. England, S. , Bevan, S. & Docherty, R. J. ( 1996 ) J. Physiol. 495 , 429440 .
23. Akopian, A. N. , Sivilotti, L. & Wood, J. N. ( 1996 ) Nature (London) 379 , 257262 .
24. Ingram, S. L. & Williams, J. T. ( 1996 ) J. Physiol. 492 , 97106 .
25. Nicol, G. D. , Vasko, M. R. & Evans, A. R. ( 1997 ) J. Neurophysiol. 77 , 167176 .
26. Treede, R. D. ( 1995 ) Ann. Med. 27 , 213216 .
27. Docherty, R. J. , Yeats, J. C , Bevan, S. & Boddeke, H. W. G. M. ( 1996 ) Pflgers Arch. Eur. J. Physiol. 431 , 828837 .
28. Bevan, S. & Geppetti, P. ( 1994 ) Trends Neurosci. 17 , 509512 .
29. Dittert, I. , Vlachova, V. , Knotkova, H. , Vitaskova, Z. , Vyklicky, L. , Kress, M. & Reeh, P. W. ( 1998 ) J. Neurosci . Methods 82 , 195201 .
30. Hepworth, M. B. & Pinnock, R. D. ( 1998 ) J. Physiol. , 513 , 133P.
31. Kirschstein, T. , Busselberg, D. & Treede, R. D. ( 1997 ) Neurosci. Lett. 231 , 3336 .
32. Caterina, M. J. , Rosen, T. A. , Tominaga, M. , Brake, A. J. & Julius, D. ( 1999 ) Nature (London) 398 , 436441 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CAUSALGIA, PATHOLOGICAL PAIN, AND ADRENERGIC RECEPTORS 7664

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Causalgia, pathological pain, and adrenergic receptors

EDWARD R. PERL*
Department of Cell and Molecular Physiology, CB 7545, University of North Carolina, Chapel Hill, NC 27599
ABSTRACT Control of expression of molecular receptors for chemical messengers and modulation of these receptors
activity are now established as ways to alter cellular reaction. This paper extends these mechanisms to the arena of pathological
pain by presenting the hypothesis that increased expression of -adrenergic receptors in primary afferent neurons is part of the
etiology of pain in classical causalgia. It is argued that partial denervation by lesion of peripheral nerve or by tissue destruction
induces a change in peripheral nociceptors, making them excitable by sympathetic activity and adrenergic substances. This
excitation is mediated by -adrenergic receptors and has a time course reminiscent of experimental denervation supersensitivity.
The change in neuronal phenotype is demonstrable after lesions of mixed nerves or of the sympathetic postganglionic supply.
Similar partial denervations also produce a substantial increase in the number of dorsal root ganglion neurons evidencing the
presence of -adrenergic receptors. The hypothesis proposes the increased presence of -adrenergic receptors in primary afferent
neurons to result from an altered gene expression triggered by cytokines/growth factors produced by disconnection of peripheral
nerve fibers from their cell bodies. These additional adrenergic receptors are suggested to make nociceptors and other primary
afferent neurons excitable by local or circulating norepinephrine and epinephrine. For central pathways, the adrenergic excitation
would be equivalent to that produced by noxious events and would consequently evoke pain. In support, evidence is cited for a
form of denervation supersensitivity in causalgia and for increased expression of human -adrenergic receptors after loss of
sympathetic activity.
This essay is an outgrowth of a Colloquium session in which new evidence was presented on how molecular receptors for chemical
synaptic mediators can specify and regulate neuronal responses in systems associated with pain mechanisms. These data build on the
concept that not only the nature, but also the magnitude, of the transfer of information between cells is at least as much a function of
receptive molecules as of the chemical messengers. Therefore, selective regulation of receptor expression and quantitative control of
receptor activity are factors defining or modulating synaptic function. Importantly, such concepts, when applied to disease, open novel
possibilities of pharmaceutical manipulation and treatment. My purpose is to extend such considerations of receptor regulation to a
pathological process involving pain.
There now is considerable agreement that in mammals, the detection and signaling of tissue damage or pathology, that is,
nociception, is a normal somatosensory function. In this view, pain, one of the reactions to tissue injury, represents the sensory
concomitant of nociception. By logical extension, pain in the absence of peripheral tissue damage is abnormal, in other words, pathological.
Causalgia and Postsympathectomy Pain. S. Weir Mitchell ( 1 ) named a syndrome causalgia after its most prominent symptom, a
burning pain referred to a particular body region appearing spontaneously or after innocuous stimulation. As classically described,
causalgia appears after partial disruption of the innervation to a limb, typically after injury to a large mixed nerve. It probably is relevant
that the full-blown syndrome is not usually reported after lesions of smaller, purely cutaneous nerves. Some years after the original
descriptions, Rene Leriche ( 2 ) pointed out that the syndrome of causalgia had features suggesting abnormal sympathetic nervous system
functioning and proposed sympathectomy as a treatment. Subsequently, the list of disorders in which pain was presumably related to
sympathetic nervous activity expanded beyond the original descriptions of classical causalgia and acquired other terminologies. It is not
clear that all of these later additions to the category of sympathetically related pain disorders share a common etiology and pathology to
the classical causalgic syndrome. For the purpose of focusing our consideration on a disorder with a common causative process, the
following starts from classical causalgia without implying extension to either more general or to more specific terminologies and
classifications: e.g., reflex sympathetic dystrophy ( 3 ), complex regional pain syndrome ( 4 ), and various others ( 5 ).
As outlined in Fig. 1 , the classical syndrome of causalgia includes the following features. (i) It follows partial denervation of a
region, usually by traumatic injury of a large mixed nerve. (ii) The partially denervated area is hypalgesic. (iii) Days to weeks after the
disturbance of innervation, spontaneous pain appears, typically burning in nature, referred to the partially denervated and nearby regions.
(iv) Pain is produced or increased by normally nonpainful stimuli, e.g., skin cooling or light touch (allodynia). (v) Abnormal sympathetic
function is evident in the region (e.g., vasomotion, perspiration). (vi) The pain is aggravated by emotional upset. (vii) Trophic changes
appear in the partially denervated tissues and nearby regions including abnormal coloration and turgor of the skin, unusual growth of hair,
and changes in bone and other subcutaneous tissues. Certain of the physical signs are suggestive of chronic inflammation.
After Leriches suggestion, regional sympathectomy or regional sympathetic block has been used as a therapy for this syndrome with,
in many cases, at least temporary success ( 6 , 7 ). In cases with successful outcome, the abnormal pain is reduced or abolished, and there
is amelioration of trophic changes ( 2 , 6 , 7 ). During remission of the signs and symptoms after sympathectomy or sympathetic block,
local injection of norepinephrine into the skin of the previously painful region has been reported to recreate the former causalgic pain (ref.
8 ; see also ref. 9 ). These observations suggest that sympathetic
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviation: DRG, dorsal root ganglion.
* To whom reprint requests should be addressed. e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CAUSALGIA, PATHOLOGICAL PAIN, AND ADRENERGIC RECEPTORS 7665

activity and adrenergic mediators have a part in the aberrant pain and other features of causalgia.
Classical Causalgia
Partial injury of a mixed nerve or of peripheral innervation

Decreased ability to detect and recognize stimuli (hypoesthesia and hypalgesia)

Varying burning pain beginning after days to weeks: intensified by emotion and temperature changes
Pain from normally nonpainful circumstances (allodynia)
Abnormal activity of sympathetic effectors (e.g.,blood flow, perspiration)
Trophic changes in skin and other tissues
Exaggerated subjective response to painful stimuli (hyperpathia)

FIG. 1. Features of causalgia as classically described.

The connection between adrenergic receptors and the pathological pain of causalgia proposed herein is circumstantial. The general
concepts have been addressed previously ( 10 , 11 ). Evidence is drawn from experimental studies in animal models and observations on
clinical cases. Admittedly, the choice of material from the voluminous literature on sympathetically related sensory phenomena is
selective. In part, the selection attempts to avoid mixing disparate material or possibly different clinical or experimental situations and
partially represents the writers view of relevance.
It is important to our argument that in addition to classical causalgia, there are clinical reports of a painful dystrophy after regional
sympathectomy, usually performed for vascular problems ( 12 14 ). The postsympathectomy painful dystrophies differ from classical
causalgia in that there often is a deep rather than cutaneous reference for the pain ( 15 ). Furthermore, postsympathetic pain is usually
described as spontaneously remitting, whereas the symptoms of classical causalgia persist in the absence of a remission produced by
sympathectomy.
Adrenergic Responsiveness of Primary Afferent Neurons. The logic for the proposal that an increased expression of adrenergic
receptors underlies the syndrome of causalgia begins with observations on the effects of peripherally applied adrenergic substances on
cutaneous nociceptors in vivo. Norepinephrine injected into the skin of normal human beings does not evoke pain. Prior work had shown
that some afferent fibers terminating in a neuroma at the central stump of a transected nerve, therefore injured, are excitable by
norepinephrine, an effect with -adrenergic features ( 16 , 17 ). In experimental animals, norepinephrine or epinephrine applied
peripherally does not excite nociceptors ( 18 20 ). On the other hand, after injury to part of a mixed peripheral nerve, some of the
nociceptors in the injured nerve become excitable by sympathetic stimulation and adrenergic substances ( 20 22 ). However, in these
experiments the primary afferent elements exhibiting the adrenergic excitation are not those whose peripheral fibers had been transected.
Furthermore, the adrenergically excitable nociceptors are otherwise functionally equivalent to those found in normal nerve. In rabbit, the
pharmacology of this adrenergic excitation proved consistent with mediation by 2-adrenergic receptors ( 20 , 21 ); however, in primate,
other -adrenergic receptors may be involved ( 23 ). The novel adrenergic excitation of nociceptors is manifest shortly after the time of
nerve injury and persists for months. Whether other classes of primary afferent neurons also participate in the changed response to
sympathetic stimuli and adrenergic substances has not been established. Importantly, though, regional sympathectomy also has been found
to induce an adrenergically mediated excitation of C-fiber nociceptors, although features of the adrenergically induced responses after
sympathectomy appear to differ from those seen after mixed peripheral nerve damage ( 24 ).
Changes Following Partial Denervations
Peripheral nerve injury induces an -AR mediated excitation of intact nociceptors.
Peripheral nerve injury causes increased numbers of intact DRG neurons to express -ARs.
Regional sympathectomy also induces adrenergic excitation of nociceptors.
Loss of sympathetic activity leads to increased -AR expression in human subjects.

FIG. 2. Prominent adrenergic consequences of partial denervations.

Adrenergic Receptors and Primary Afferent Neurons. Thus, the events unleashed by a partial denervation, consisting of
interruption of some peripheral sensory fibers and/or postganglionic sympathetic fibers, alter the phenotype of nociceptors that otherwise
remain functionally intact in the injured or another nerve supplying the region. A clue that this change is possibly related to an increase in
receptor population comes from the observations that the excitatory effect has a pharmacological profile of a specific adrenergic -
receptor. Moreover, the time course of the development of the adrenergic excitation suggested that although it was manifest within a few
days, the peak of excitatory effects is reached 23 weeks after the nerve injury ( 21 ). This time course is reminiscent of that for
denervation supersensitivity in which sympathetically denervated organs become much more responsive to adrenergic agents ( 25 ).
Sympathetic supersensitivity has been related to increased numbers of adrenergic receptors ( 26 28 ). Moreover, it is significant that the
adrenergic, sympathetic excitation of nociceptors appears to occur in the region of the peripheral receptive terminals ( 21 ).
We explored the possibility that the appearance of the excitatory response by cutaneous nociceptors to sympathetic stimulation and to
adrenergic agents is associated with alterations in adrenergic receptors in primary afferent neurons. The number of dorsal root ganglion
(DRG) cells labeled by an antibody putatively recognizing 2A- adrenergic receptors was found to markedly increase after both partial and
complete transections of the rat sciatic nerve ( 29 ). The principal increase in the population of DRG neurons expressing this
immunoreactivity appears in neurons with somata of medium-to-medium-large diameters (2040 m). Double labeling with markers for
injury and growth (c-jun protein) ( 30 , 31 ) or for transected fibers (fluorogold) ( 32 ) indicated that the increased immunoreactivity to the
2A-directed antibody occurs both in injured neurons and in those without evidence of damage. The latter predominate. The increase in the
number of DRG neurons expressing 2A-adrenergic receptor immunoreactivity following nerve injury is selective. Increased
the authoritative version for attribution.

immunoreactivity
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CAUSALGIA, PATHOLOGICAL PAIN, AND ADRENERGIC RECEPTORS 7666

to the 2A-adrenergic receptor antibody does not appear after localized artificial inflammation produced by injection of formalin or
Freunds complete adjuvant, and immunoreactivity to a 2C antibody is not increased after sciatic nerve injury ( 29 ). An earlier
autoradiographic study with the partial 2 agonist, p-iodoclonidine (125I-labeled) had indicated increased binding in ipsilateral DRG after
partial or complete nerve transection; however, the diameter spectrum of the p-iodoclonidine-labeled DRG neurons (mainly small-
diameter) partially differs from that with the 2A-adrenergic receptor antibody (refs. 10 and 29 ; K. Nishiyama and E.R.P., unpublished
data).
Linking -Adrenergic Receptors to Pain in Causalgia
Partial interruption of primary afferent and sympathetic innervation of a region
Reaction to injury by neurons and associated cells: production of cytokines for growth/regeneration
Activation of gene(s) producing -ARs in injured DRG neurons and in remaining intact nociceptors
Activation of -ARs in nociceptors by local and circulating adrenergic agents
Impulse traffic in peripheral and central pathways for nociception and pain
Peripheral and central changes leading to allodynia
Secondary changes in sensory and affective mechanisms from persisting pain-related input

FIG. 3. Factors in the hypothesis connecting increased expression of -adrenergic receptors to the pain of causalgia.
An Hypothesis. The effects of experimental nerve injury on the responsiveness of cutaneous nociceptors to adrenergic substances
and on -adrenergic receptor expression in dorsal root ganglia suggest a possible relationship to the etiology and symptoms of causalgia.
The most salient factors are outlined in Fig. 2 . This leads to the hypothesis summarized by Fig. 3 . It is proposed that injury to part of the
innervation to a bodily region, e.g., partial transection of a mixed nerve supplying part of a limb, induces production of cytokines and/or
growth factors by the injured neurons and associated cells (e.g., Schwann cells). These chemical factors, among other effects, mediate
responses to injury or are associated with regrowth and lead to altered gene expression in uninjured neurons of the nerve and region.
Disconnection of sympathetic postganglionic fibers from their targets, by itself and in conjunction with injury of primary afferent fibers,
contributes to these signals, leading to altered expression of -adrenergic receptors. The fact that classical causalgia usually follows injury
to large mixed nerves raises the possibility that the required lesion for the full syndrome is the combined interruption of primary afferent
fibers and sympathetic postganglionic fibers. Possibly the loss of the presence of sympathetic mediators in the vicinity of primary afferent
terminals because of interruption of sympathetic postganglionic fibers is part of the pathophysiological process. The net result after such
partial loss of innervation is that afferent neurons, normally expressing few -adrenergic receptors capable of producing an excitatory
response, develop them. As a result, sensory neurons, particularly nociceptors, become excitable through these newly formed receptors.
The afferent neuron excitation would occur by norepinephrine locally released by the remaining sympathetic supply to the vasculature or
norepinephrine and epinephrine circulating from other parts of the body. For central mechanisms, signals produced in nociceptors by
adrenergic agents are equivalent to those evoked by noxious events and lead to the sensation of pain. Signals interpreted as the result of
noxious stimuli can result from activation by -adrenergic receptors in DRG neurons by trace quantities of norepinephrine and
epinephrine. The outcome is spontaneous pain or sensitization of nociceptors.
Observations on human subjects offer circumstantial support for facets of this concept. Loss or decrease of sympathetic activity and
the consequent decrease in circulating sympathetic postganglionic mediators has been shown to increase expression of -adrenergic
receptors ( 27 , 28 , 33 ). Those observations were made on blood platelets. By inference, one can argue that the same process could affect
DRG neurons. Regional sympathectomy in experimental animals does increase binding of -adrenergic agonists in DRG neurons of the
innervated region (ref. 10 and K. Nishiyama and E.R.P., unpublished data). Thus, one manipulation that leads to induction of adrenergic
excitation of nociceptors, loss of sympathetic innervation, increases -adrenergic receptor expression in some tissues.
It is pertinent that the affected limb in persons suffering from pain disorders fitting the criteria of classical causalgia exhibits lower
concentrations of norepinephrine and a degradation product (3,4-dihydroxphenylethyleneglycol) in its venous return than contralaterally.
This could imply that the affected limb has less functioning sympathetic innervation. At the same time, the affected limb exhibits
increased activity by sympathetically innervated tissues ( 34 , 35 ). These human data suggest that a process akin to denervation
supersensitivity may operate in classical causalgia and possibly other varieties of sympathetically related pain disorders.
To repeat, my suggestion here is that primary afferent neuron excitation by adrenergic agents in classical causalgia results from novel
-adrenergic receptor production in dorsal root ganglia neurons evoked by direct and indirect effects of injury to peripheral innervation.
The increased adrenergic receptor expression, in part, involves primary afferent neurons, particularly nociceptors, with intact connections
to the periphery and the central nervous system. As a consequence of the novel -adrenergic receptor production, some of these afferent
neurons develop an excitatory response to trace amounts of adrenergic substances in peripheral tissues. Such excitation would be the start
of abnormal signals activating pain pathways and central pain-related mechanisms.
The concept just outlined, like most hypotheses, has difficulties. First, it cannot explain all parts of a complex syndrome. In
particular, it does not account for the trophic changes, allodynia, and psychological alterations. The hypothesis suggests only that nerve
injury and partial denervation unleash a set of circumstances leading to an abnormal production of -adrenergic receptors in sensory
neurons. Experimental studies suggest that the adrenergic receptor type may be of the 2 (possibly also 1) type ( 28 , 29 , 23 ). These
adrenergic receptors become part of a messenger system whereby adrenergic substances excite or sensitize peripheral sensory neurons
related to nociception and pain, which represents a step in the process leading to spontaneous pain and to activation of central pathways.
Subsequently, the abnormally initiated central activity can lead to sensitization and other plastic changes in central neuronal mechanisms.
Although not an explanation of all signs and symptoms of causalgia, this proposal provides a possible etiology of the pathological process
and some insight into factors that could operate to maintain the process. Second, there is the issue of the effects mediated by 2-adrenergic
receptors that usually are presumed in neurons to mediate
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CAUSALGIA, PATHOLOGICAL PAIN, AND ADRENERGIC RECEPTORS 7667

inhibitory actions. In this context, it should be remembered that 2-adrenergic receptors are intermediate arteriolar smooth muscle
constriction. Therefore, this class of receptors is capable of being part of an excitatory signal-transduction process ( 36 ); furthermore, the
signaling system induced by nerve injury may not be identical to that occurring in neurons normally.
Does the idea of a change in cellular phenotype by the enhanced production of membrane receptors possibly apply to other
situations? A similar process could operate in other versions of sympathetically related pain. It could also relate to Raynauds disease,
another pathological process which, in part, appears to represent overreaction to sympathetic mediators and could possibly result from an
increased expression of adrenergic receptors ( 37 ). Furthermore, enhanced reactions to adrenergic mediators by the vasculature have also
been postulated for certain forms of hypertension ( 38 , 39 ). To conclude, the concept of increased expression of molecular receptors as a
mechanism of disease, and in particular of pathological pain, deserves serious consideration and further exploration.
I thank Ms. S. Derr for her assistance. Preparation of this paper was aided by grants NS 10321 and NS 14899 of the National Institute
of Neurological Disorders and Stroke.
1. Mitchell, W. ( 1872 ) Injuries of Nerves and Their Consequences ( Lippincott , Philadelphia ).
2. Leriche, R. ( 1916 ) Presse Md. 24 , 178180 .
3. Evans, J. A. ( 1946 ) Surg. Clin. North Am. 26 , 435448 .
4. Merskey, H. & Bogduk, N. ( 1995 ) in Classification of Chronic Pain: Descriptions of Chronic Pain Syndromes and Definition of Terms ( IASP
Press , Seattle ).
5. Kozin, F. , McCarty, D. J. , Sims, J. & Genant, H. ( 1976 ) Am. J. Med. 60 , 321331 .
6. Schumacker, H. B. , Speigel, I. J. & Upjohn, R. H. ( 1948 ) Surg. Gynecol. Obstet. 86 , 7686 .
7. Richards, R. L. ( 1967 ) Arch. Neurol. 16 , 339350 .
8. Wallin, E. , Torebjrk, E. & Hallin, R. ( 1976 ) in Sensory Functions of the Skin in Primates, with Special Reference to Man , ed. Zotterman, Y. (
Pergamon , Oxford ), pp. 489502 .
9. Torebjrk, E. , Wahren, L. , Wallin, G. , Hallin, R. & Koltzenburg, M. ( 1995 ) Pain 63 , 1120 .
10. Perl, E. R. ( 1994 ) in Progress in Pain Research and Management , eds. Fields, H. L. & Liebeskind, J. C. ( IASP Press , Seattle ), pp. 129150 .
11. Perl, E. R. ( 1994 ) in Touch, Temperature, and Pain in Health and Disease: Mechanisms and Assessments, Progress in Pain Research and
Management eds. Boivie, J. , Hansson, P. & Lindblom, U. ( IASP Press , Seattle ), Vol. 3 , pp. 231248 .
12. Tracy, G. D. & Cockett, F. B. ( 1957 ) Lancet i (272) , 1214 .
13. Litwin, M. S. ( 1962 ) Arch. Surg. 84 , 591595 .
14. Raskin, N. H. , Levinson, S. A. , Hoffman, P. M. , Pickett, J. B. E. & Fields, H. L. ( 1974 ) Am. J. Surg. 128 , 7578 .
15. Churcher, M. D. ( 1984 ) Lancet ii (8395) , 131133 .
16. Devor, M. & Jnig, W. ( 1981 ) Neurosci. Lett. 24 , 4347 .
17. Devor, M. ( 1983 ) J. Auton. Nerv. Syst. 7 , 371384 .
18. Shea, V. & Perl, E. R. ( 1985 ) J. Neurophysiol. 54 , 491501 .
19. Barasi, S. & Lynn, B. ( 1986 ) Brain Res. 378 , 2127 .
20. OHalloran, K. D. & Perl, E. R. ( 1997 ) Brain Res. 759 , 233240 .
21. Sato, J. & Perl, E. R. ( 1991 ) Science 251 , 16081610 .
22. Bossut, D. F. & Perl, E. R. ( 1995 ) J. Neurophysiol. 73 , 17211723 .
23. Ali, Z. , Ringkamp, M. , Hartke, T. V. , Chien, H. F. , Flavahan, N. A. , Campbell, J. N. & Meyer, R. A. ( 1999 ) J. Neurophysiol. 81 , 455466 .
24. Bossut, D. F. , Shea, V. & Perl, E. R. ( 1995 ) J. Neurophysiol. 75 , 514517 .
25. Cannon, W. B. & Rosenblueth , ( 1949 ) The Supersensitivity of Denervated Structures: A Law of Denervation ( McMillan , New York ).
26. Arnett, C. O. & J. A. Davis ( 1979 ) J. Pharmacol. Exp. Ther. 211 , 394400 .
27. Davies, I. B. , Sudera, D. , Sagnella, G. , Marchesi-Saviotti, E. , Mathias, C. , Bannister, R. & Sever, P. S. ( 1982 ) J. Clin. Invest. 69 , 779784 .
28. Egan, B. , Neubig, R. & Julius, S. ( 1985 ) Clin. Pharmacol. Ther. 38 , 519524 .
29. Birder, L. A. & Perl, E. R. ( 1999 ) J. Physiol. (London) 515 , 533542 .
30. Jenkins, R. & Hunt, S. P. ( 1991 ) Neurosci. Lett. 129 , 107110 .
31. Jenkins, R. , McMahon, S. B. , Bond, A. B. & Hunt, S. P. ( 1993 ) Eur. J. Neurosci. 5 , 751759 .
32. Baranowski, A. P. , Anand, U. & McMahon, S. B. ( 1992 ) Neurosci. Lett. 141 , 5356 .
33. Davies, I. B. & Sever, P. S. ( 1988 ) in Autonomic Failure: A Textbook of Clinical Disorders of the Autonomic Nervous System , ed. Bannister, R. (
Oxford Univ. Press , Oxford ), pp. 348366 .
34. Drummond, P. D. , Finch, P. M. & Smythe, G. A. ( 1991 ) Brain 114 , 20252036 .
35. Drummond, P. D. , Finch, P. M. , Edvinsson, L. & Goadsby, P. J. ( 1994 ) Clin. Auto. Res. 4 , 113116 .
36. Nichols, A. J. & Ruffolo, R. R. , Jr. ( 1991 ) in Progress in Basic and Clinical Pharmacology , ed. Ruffolo, R. R., Jr. (Karger, Basel), Vol. 8 , pp.
115179 .
37. Edwards, J. M. , Phinney, E. S. , Taylor, L. M., Jr. , Keenan, E. J. & Porter, J. M. ( 1987 ) J. Vasc. Surg. 5 , 3845 .
38. Michel, M. C. , Insel, P. A. & Brodde, O. ( 1989 ) FASEB J. 3 , 139144 .
39. De Champlain, J. ( 1989 ) J. Hypertension 8, Suppl. 7 , S77S85 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

FOREBRAIN MECHANISMS OF NOCICEPTION AND PAIN: ANALYSIS THROUGH IMAGING 7668

This paper was presented at the National Academy of Sciences colloquium The Neurology of Pain, held December 1113, 1998,
at the Arnold and Mabel Beckman Center in Irvine, CA.

Forebrain mechanisms of nociception and pain: Analysis through

imaging

KENNETH L. CASEY
Neurology Service, Veterans Affairs Medical Center, University of Michigan, 2215 Fuller Road, Ann Arbor, MI 48105
ABSTRACT Pain is a unified experience composed of interacting discriminative, affective-motivational, and cognitive
components, each of which is mediated and modulated through forebrain mechanisms acting at spinal, brainstem, and cerebral
levels. The size of the human forebrain in relation to the spinal cord gives anatomical emphasis to forebrain control over
nociceptive processing. Human forebrain pathology can cause pain without the activation of nociceptors. Functional imaging of
the normal human brain with positron emission tomography (PET) shows synaptically induced increases in regional cerebral
blood flow (rCBF) in several regions specifically during pain. We have examined the variables of gender, type of noxious stimulus,
and the origin of nociceptive input as potential determinants of the pattern and intensity of rCBF responses. The structures most
consistently activated across genders and during contact heat pain, cold pain, cutaneous laser pain or intramuscular pain were the
contralateral insula and anterior cingulate cortex, the bilateral thalamus and premotor cortex, and the cerebellar vermis. These
regions are commonly activated in PET studies of pain conducted by other investigators, and the intensity of the brain rCBF
response correlates parametrically with perceived pain intensity. To complement the human studies, we developed an animal
model for investigating stimulus-induced rCBF responses in the rat. In accord with behavioral measures and the results of human
PET, there is a progressive and selective activation of somatosensory and limbic system structures in the brain and brainstem
following the subcutaneous injection of formalin. The animal model and human PET studies should be mutually reinforcing and
thus facilitate progress in understanding forebrain mechanisms of normal and pathological pain.
Forebrain Mediation of Pain. Pain is a conscious experience that includes discriminative, affective-motivational, and cognitive
components that produce the unified sensation of pain. These components are each mediated through separate, interactive forebrain
mechanisms ( 1 ). For example, the ability to localize somatic stimuli in time, space, and along a continuum of intensities is greatly
impaired following lesions limited to the primary somatosensory (S1) cortex or the ventral posterolateral thalamus. These lesions do not
produce analgesia, however, because the aversive nature of noxious stimuli, although poorly localized, is still evident in the behavior of
animals and the verbal reports of humans ( 2 ). Neurons in the S1 cortex and ventral posterolateral thalamus, including those responding
primarily to noxious stimuli, have small, contralateral receptive fields consistent with the mediation of spatial stimulus localization ( 3 ). In
contrast, lesions within the anterior cingulate cortex have no effect on innocuous or nociceptive somesthetic discriminative functions, but
impair the recognition of the noxious or aversive quality of the stimulus in animals and the perceived affective quality of pain in humans (
4 , 5 ). Anterior cingulate neurons that respond to noxious stimuli have large, often bilateral receptive fields, consistent with a limited role
in spatial discriminative capacity ( 6 ). There is no comparable information about the neuronal substrate for the cognitive dimension of
pain, but there are numerous studies and observations showing the profound influences of attention, suggestion, and emotional state on the
perception of pain ( 7 ). The broad range of environmental influences, such as attention, fear, and the placebo effect on the perception of
pain suggests that cortical association areas and their subcortical connections are critical participants in mediating the cognitive aspects of
pain.
The Forebrain Modulation of Pain. The processing of nociceptive stimuli is modulated by the forebrain at spinal, brainstem, and
diencephalic levels. Stimulation of the cerebral cortex or thalamus can facilitate or suppress the responses of spinothalamic or
trigeminothalamic tract neurons ( 8 , 9 ). In the awake monkey, the response of trigeminothalamic cells to noxious heat depends on
behavioral state ( 10 , 11 ). Corticobulbar and corticothalamic neurons have marked effects on the excitability of brainstem and thalamic
cells that receive nociceptive input ( 12 16 ).
Because of the large volume of the human forebrain in relation to that of the spinal cord (77% vs. 2% of central nervous system
volume), these descending modulatory influences may assume greater importance in humans than in other species, such as the laboratory
rat, where the forebrain is less anatomically dominant (31% vs. 35% of central nervous system volume) ( 17 ). The human spinothalamic
tract, for example, contains an estimated 2,000 to 5,000 fibers whereas the corticospinal tract, which includes fibers terminating in the
superficial layers of the dorsal horn ( 18 , 19 ), is estimated to contain from 5 105 to 1 106 fibers ( 20 , 21 ). Corticothalamic influences
are also likely to be dominant in the human; in the cat, approximately 50% of the estimated 5,000 to 9,000 synapses on thalamocortical
projection neurons are presumed to be of cortical origin, whereas only 15% are formed by ascending afferent fibers ( 22 ).
The Physiological Rationale of Positron Emission Tomography (PET). Synaptic activity generates increases in cerebral blood
flow (CBF). This physiological fact is the basis for both PET and functional magnetic resonance imaging (fMRI). The most commonly
used fMRI method relies on local shifts in the magnetic field that accompany the shift from deoxyhemoglobin to oxyhemoglobin within
activated perfused tissue ( 23 ). PET and fMRI are complementary methods of assessing brain activity. This article will be limited to a
discussion of PET.
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: CBF, cerebral blood flow; rCBF, regional CBF; fMRI, functional magnetic resonance imaging; PET, positron emission
tomography; ROI, region of interest; S1, primary somatosensory (cortex); S2, secondary somatosensory (cortex); VOI, volumes of interest.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

FOREBRAIN MECHANISMS OF NOCICEPTION AND PAIN: ANALYSIS THROUGH IMAGING 7669

The first indication that brain activity increases global CBF was reported by Roy and Sherrington over 100 years ago ( 24 ).
Subsequent radioactive tracer techniques revealed increases in regional CBF (rCBF) during sensory stimulation or the performance of
motor tasks ( 25 ). Recent studies that use the technique of optical imaging have demonstrated that cortical blood flow responses occur
within 3 sec of sensory stimulation and are initially restricted to the 300- to 500-m dimensions of cortical columns before spreading to
involve the surrounding 3 mm to 5 mm of cortical tissue ( 26 28 ). The biochemical coupling of rCBF and synaptic activity is unknown
and is still an area of active investigation. Studies of regional cerebral glucose utilization show that it and rCBF are normally tightly
coupled and that the coupling occurs within the synaptic neuropil. The degree of coupling may vary among regions and under special
experimental conditions ( 29 ), but it is reliably present in the normal mammal. Blocking the production of nitric oxide has no effect on
synaptically induced rCBF responses in the rat somatosensory system ( 30 ). There is evidence that adenosine may be a critical link in this
process, but it is likely that the action of several mediators may be important ( 31 ).
A reduction of rCBF should occur when synaptic activity is suppressed below some resting or background level. Indeed, reductions in
rCBF are observed in many PET studies of rCBF ( 32 ). However, the physiological significance of reduced rCBF is uncertain; it does not
necessarily indicate the presence of inhibitory synaptic activity, because both inhibitory and excitatory synaptic activity can contribute to
increases in synaptic metabolism ( 33 ). It is possible that some of the observed reductions in rCBF reflect autoregulatory mechanisms for
global CBF, these reductions may not affect neuronal function; others may signal the removal of synaptic excitation (disfacilitation) by an
inhibitory process located outside the area of rCBF decrease. In any event, it is not now possible to establish the valence of synaptic
activity by rCBF estimation methods.
PET Methodology. In most current studies, water (as H215O) or [15O]butanol is injected intravenously, or carbon dioxide (as C15O2)
is inhaled and converted in the lungs to H215O. The 15O has a half-life of 122 sec. This length of time is sufficient for CBF measurements,
because a bolus injection (e.g., 50 mCi) of this compound is nearly completely diffused into brain tissue on the first arterial pass ( 34 ).
The count of emissions from a given volume of brain tissue is therefore a good estimate of the perfusion of that brain region during the
counting period (approximately 60 sec for a typical scan).
With the analytical methods that we use, we find that there are approximately 95,000 voxels in the gray matter of the average human
brain. At our facility, a three-dimensional voxel is a cube 2.25 mm on each side. However, the spatial resolution of PET is limited by the
smoothing introduced by image reconstruction filters and by the ability of the radiation detectors to differentiate the radiation emitted from
two separate point sources. For PET, this distance, the full width at half maximum, is between 6 and 9 mm. However, the spatial accuracy
in the localization of an activation focus is improved (to less than half the full width at half maximum) when subtraction images are made.
Each image set is normalized to whole brain counts, and mean radioactivity concentration images are created by estimating rCBF
across all subjects with stereotactic anatomical standardization techniques. Image voxel intensities are normalized to global cerebral
activity with the use of a linear proportional model to remove baseline differences in global CBF between scans and subjects ( 35 ). In our
facility, CBF images are aligned onto the coordinates of a standard stereotactic atlas ( 36 ), by using anatomical landmarks identified
within the PET images of each individual so that the CBF differences are compared within the same brain regions ( 37 39 ). To
determine whether a task or a stimulus has produced an increase in rCBF, the rCBF computed during a control condition is subtracted
from that computed during the test condition. The resulting subtraction image, then, shows those brain regions with differences in CBF
between the two conditions.
A voxel-by-voxel statistical subtraction analysis (Z-score) with adjustment for multiple comparisons is performed by estimating the
smoothness of subtraction images ( 40 ) following three-dimensional Gaussian filtering to enhance signal-to-noise ratio and compensate
for residual anatomical variance. Typically, only those voxels with normalized CBF values larger than 60% of the global value are
analyzed, because these voxels represent the gray matter of the brain. Voxels showing a significantly increased CBF compared with the
average noise variance computed across all voxels (pooled variance) are identified ( 41 ). The critical level of significance is determined
by using this information to adjust P = 0.05 ( 42 ). With this method, the results of interest are revealed primarily through the data analysis.
However, it is also possible to perform correlations between the intensity of the rCBF responses throughout the brain and some behavioral
parameter of interest, such as the perceived intensity of stimulation ( 43 ).
In addition, volumes of interest (VOI) may be established within brain structures selected because of a priori hypotheses and the
results of previously published PET studies. The size and shape of each VOI may be standardized across studies or determined separately
according to functional criteria. We presently use a method similar to that described by Burton ( 44 ), in which voxels showing significant
peak increases in CBF between comparison conditions are identified within the brain structure of interest; the voxels are progressively
expanded in three dimensions to include contiguous voxels that meet the statistical criterion established by the voxel-by-voxel Z-score
analysis. To determine the statistical significance of rCBF increases, a paired t statistic is computed for each VOI from the average
percentage increase in CBF across all subjects. Levels of significance are established, based on the Bonferroni correction for multiple
comparisons among VOI.
Important Variables in the Conduct and Interpretation of PET Studies of Pain. There are now numerous studies from various
facilities that have used PET during the application of experimental pain. The results are difficult to compare because they are affected by
intersubject variability, type of stimulus, method of scanning, and data analysis methods. To assess the effect of some of these variables,
we have conducted several investigations in normal subjects with a variety of stimulation methods. The variables we have considered thus
far include gender, the physical characteristics of the stimulus, and the sources of nociceptive afferent input.
Gender. The prevailing evidence suggests that although there is no reliable gender difference in pain thresholds, pain tolerance is
generally higher in male than in female subjects ( 45 ). PET studies find gender differences in resting rCBF ( 46 ) or in the cerebral
metabolic rate of glucose utilization ( 47 49 ). These findings suggest that there may be underlying gender differences in the neural
mechanisms that mediate pain perception. Accordingly, we performed PET studies in normal right-handed male (n = 10) and female (n =
10) subjects (18 to 39 years old) as they discriminated differences in the intensity of innocuous and noxious heat stimuli applied to the left
forearm ( 50 ). Thermal stimuli were 40C or 50C heat, applied with a thermode as repetitive 5-sec contacts to the left volar forearm.
Both male and female subjects rated the 40C stimuli as warm but not painful and the 50C stimuli as painful, but females rated the 50C
stimuli as significantly more intense than did the males (P = 0.0052). Both genders showed a bilateral activation of premotor cortex during
heat pain in addition to the activation of a number of contralateral structures, including the posterior insula, anterior cingulate cortex, and
the authoritative version for attribution.

the cerebellar vermis ( Fig. 1 ). Overall, a nearly complete

lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

FOREBRAIN MECHANISMS OF NOCICEPTION AND PAIN: ANALYSIS THROUGH IMAGING 7670

overlap of the activation patterns occurred between genders. However, direct image subtraction showed that females had significantly
greater activation of the contralateral prefrontal cortex compared with males. A VOI comparison (t-statistic) also showed greater activation
of the contralateral insula and thalamus in females compared with males (P < 0.05). These pain-related differences in brain activation may
be attributed to gender, perceived pain intensity, or both factors. These results show that gender differences are important considerations in
the investigation of forebrain responses to noxious stimuli.

FIG. 1. Statistical map of rCBF responses of 10 males (M) and 10 females (F) to repetitive noxious heat stimulation (50C) of the left
volar forearm. Color coding of Z scores as indicated by flame bar at right. The right hemisphere of the MRI stereotactic template is
on the readers left. The numbers below columns of images indicate millimeters above a plane connecting the anterior and posterior
commissures. In both genders, there is significant activation of the contralateral anterior cingulate cortex (+41, +37), premotor, and
insular cortex (+15, +7), ipsilateral insula (+7, +15), and bilateral cerebellar vermis (12). Voxel-by-voxel analysis indicated that
some structures were significantly activated (Z > 4.0) only in males (contralateral sensorimotor cortex, +52; contralateral lenticular
nucleus, +2; ipsilateral prefrontal cortex, +15) and others only in females (contralateral prefrental cortex, +32; anterior insula, +2;
thalamus, +15; ipsilateral lenticular nucleus, +2; contralateral cerebellum, 25). Direct comparisons of percent increase in rCBF,
however, revealed that the only difference is that the contralateral thalamus, anterior insula, and prefrontal cortex show a greater
response in females compared with males. Reproduced from ref. 50 with permission from the International Association for the Study
of Pain.
Physical Characteristics of the Noxious Stimulus. Subtraction images have been interpreted as revealing those cerebral structures
that have increased synaptic activity related specifically to the central processing of the neuronal signals produced by noxious heat. In our
earlier report ( 51 ), we controlled for the cerebral processes mediating the discrimination of heat intensity within the innocuous range. By
cooling the skin, we were able to produce innocuous, easily perceptible differences in the degree of skin temperature increase and to
duplicate the intensity differences that were applied within the noxious range. However, it is possible that by cooling the skin, we reduced
the perceived difference between the innocuous stimuli below that which would exist normally. The resulting rCBF responses may have
been reduced below the sensitivity of our PET analysis. To test this possibility, we performed a series of PET studies on subjects who were
asked to discriminate the differences between innocuous warm stimuli delivered to the volar forearm at normal baseline skin temperature.
We wished to determine whether this procedure would lead to a demonstration of rCBF increases that could be compared with those
elicited by noxious heat stimulation. An equally important and related issue is whether other methods of producing pain result in the same
intensity and pattern of rCBF increases as the
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

FOREBRAIN MECHANISMS OF NOCICEPTION AND PAIN: ANALYSIS THROUGH IMAGING 7671

increases elicited by repetitive noxious heat applied to the skin. To examine this question, we performed PET studies on normal subjects as
they experienced the deep, aching pain produced by immersion of one hand in 6C water for 105 sec. We were then able to compare the
PET rCBF results obtained during warm discrimination and during tonic, noxious cold immersion with the results previously obtained
during repetitive cutaneous noxious heat ( 51 ).
We studied three groups of nine normal, healthy subjects, 18 to 39 years old, all of whom were given instruction and practice in the
use of the visual analog scale for the estimation of stimulus intensity and unpleasantness. One group was assigned a warm discrimination
task, another group rated innocuous and noxious heat intensity and unpleasantness, and the third group participated in the ice-water
immersion study. In the warm discrimination study, two intensities of innocuous heat (36C and 43C) were applied with a thermode as
repetitive 5-sec contacts to the volar forearm throughout the scan. Neither stimulus was rated as painful. All subjects discriminated the 43
C stimulus from the 36C stimulus (P < 0.0001). Significant increases in rCBF to the 43C stimuli were found in the contralateral ventral
posterior thalamus, lenticular nucleus, medial prefrontal cortex (Brodmanns areas 10 and 32), and the cerebellar vermis. In the study of
noxious and innocuous heat, all subjects rated the 50C stimuli as painful and the 40C stimuli as warm, but not painful. Significant rCBF
increases to 50C stimuli were found contralaterally in the lenticular nucleus, thalamus, anterior cingulate cortex, premotor cortex, and the
secondary somatosensory (S2) and insular cortices. The ipsilateral premotor cortex and thalamus as well as the medial dorsal midbrain and
cerebellar vermis showed significant rCBF increases. CBF increases just below the threshold for statistical significance were seen in the
contralateral sensorimotor cortex (S1/M1).
In the ice-water immersion study, the left hand was immersed to the wrist throughout each of six scans in water kept at an average
temperature of either 20.5C 1.15C or 6.02C 1.18C on alternate scans. All subjects rated the intensity of the stimuli on a scale in
which 0 = no pain and 10 = barely tolerable pain. Subjects rated the 20C water immersion as painless (average rating SD of 0.18
0.48), but gave ratings indicating intense pain during immersion in 6C water (7.89 1.45). All subjects expressed the perception of the
pain as very cold, steady, and deep. Highly significant increases in rCBF were found contralaterally in the sensorimotor cortex (M1/ S1),
premotor cortex, anterior cingulate cortex, and in the region of the anterior insula and lenticular nucleus. Ipsilateral increases in rCBF were
seen in the lateral prefrontal cortex (Brodmanns areas 10 and 46), anterior cingulate cortex, the region of the insular and opercular
precentral cortices, and the thalamus. The cerebellar vermis also showed a significant increase in rCBF. CBF increases just below the
threshold for statistical significance were seen in the contralateral thalamus.
Comparisons of rCBF response magnitude were made among the five stereotactically concordant brain regions that showed
significant responses in both the heat pain and cold pain conditions: the cerebellar vermis, ipsilateral thalamus, contralateral premotor
cortex, contralateral anterior cingulate cortex, and the region of the contralateral anterior insula and lenticular nucleus. Each region showed
a higher increase in rCBF during the cold pain study (3.26% 0.061%) than during the heat pain study (2.85% 0.124%; paired t4 = 3.60;
P < 0.022).
The results show that in conscious humans, two forms of noxious stimulation that are different in temporal pattern, afferent fiber
activation, and perceived spatiotemporal and qualitative characteristics produce similar, but not identical, patterns of brain rCBF increases.
These pain-related response patterns are each quite different from the brain responses observed during the discrimination between two
intensities of innocuous heat stimuli. The results suggest that the increased rCBF responses observed during noxious stimulation reflect
physiological differences in neuronal activity that are related to both nociceptive processing and the perception of pain. The overlap in the
spatial distribution of rCBF increases during noxious cutaneous heat and noxious deep-cold stimulation suggests that a reproducible
pattern of rCBF responses may occur that is common to the perceptions of pain produced by different stimuli. Differences in the intensity
and spatial patterns of these pain-related rCBF increases may reflect physiological differences in neuronal nociceptive processing that are
linked with these two forms of pain perception.
The Source of Nociceptive Input. In comparing the rCBF changes induced by cutaneous-contact heat pain with that induced by
deep-cold pain, we found that the brain activation patterns showed a considerable overlap; the contralateral anterior cingulate, anterior
insula/lenticular nucleus, premotor cortex, and the ipsilateral thalamus and cerebellar vermis were activated by both forms of noxious
stimulation ( 52 ). Because cold noxious stimulation activates cutaneous, subcutaneous, muscle, periosteal, and venous nociceptors ( 53 ),
we wished to compare two forms of noxious stimulation that more selectively activate nociceptive afferents from different sources.
For each of eight PET scans, 11 normal subjects rated the intensity of cutaneous and intramuscular stimuli delivered to the
nondominant (left) forearm on a visual analog scale; stimulus intensity was adjusted to approximate pain threshold levels. Cutaneous pain
was produced with a high-energy CO2 laser stimulator. Muscle pain was elicited with high-intensity intramuscular electrical stimulation.
The pain intensity ratings and the differences between noxious and innocuous ratings were similar for cutaneous and intramuscular stimuli
(P > 0.05). After stereotactic registration, statistical pixel-by-pixel summation (Z-score) and VOI analyses of subtraction images were
performed. Direct statistical comparisons between cutaneous and intramuscular stimulations showed no reliable differences between these
two forms of noxious stimulation, indicating that a substantial overlap occurred in brain activation patterns. These activated cerebral
structures may represent those recruited early in nociceptive processing, because both forms of stimuli were near pain threshold.
Increases in rCBF of 3.5% or more were seen in the contralateral S2, anterior insular, anterior cingulate, prefrontal, and inferior
parietal cortices; and in the contralateral thalamus, lenticular nucleus, ipsilateral premotor cortex, and cerebellum. Cutaneous laser
stimulation was relatively ineffective in evoking rCBF responses in the contralateral anterior cingulate or in the lenticular nucleus.
Intramuscular stimulation was similarly ineffective in activating the contralateral prefrontal and ipsilateral premotor cortex. However, each
form of stimulation evoked responses of sufficient magnitude in each structure, but a direct statistical comparison failed to differentiate
significantly between them. The similar cerebral activation patterns suggest that the perceived differences between acute skin and muscle
pain are mediated by differences in intensity and temporospatial patterns of neuronal activity within similar sets of forebrain structures.
An Emerging Pattern. In summarizing the data obtained from right-handed subjects in our facility, we found that certain structures
were activated by noxious stimuli across a wide range of conditions. The pattern of activation is best represented by inspection of the
results of our study of gender ( Fig. 1 ). Among the most prominent activation sites are the contralateral insular cortex, primarily the
anterior portion, and the cerebellar vermis. These structures responded to each form of noxious stimulation in all groups of subjects.
Bilateral insular activation is seen, but it often does not reach statistical significance in voxel-by-voxel analysis. The contralateral
thalamic responses to noxious stimuli are equally robust, but are
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

FOREBRAIN MECHANISMS OF NOCICEPTION AND PAIN: ANALYSIS THROUGH IMAGING 7672

more frequently bilateral. The contralateral anterior cingulate cortex responded to all noxious stimuli except cutaneous laser pulses; clear
bilateral responses were observed only during deep cold pain. Finally, the premotor cortex (Brodmanns area 6) responded bilaterally in all
studies, except the study performed with a cutaneous laser, where the response was ipsilateral only; and the study performed with
intramuscular stimulation, where the response was not significant. Other structures, including the S1 and S2, lateral prefrental
(Brodmanns areas 10 and 46), and inferior parietal (Brodmanns area 40) cortices were significantly activated in a minority of our studies.
The insular cortex has been considered a component of cerebral pain mechanisms based primarily on clinical information ( 54 57 ).
Because of its anatomical connections, this region is likely to mediate affective, mnemonic, and autonomic features of pain ( 58 ). The
insula receives input from the ventral medial posterior thalamus, a region that receives direct nociceptive input from the superficial dorsal
horn ( 59 , 60 ). The spinothalamic tract projects also to the medial and intralaminar thalamus, the origin of thalamocortical fibers to the
anterior cingulate gyrus; nociceptive neurons are found in both structures ( 6 , 61 63 ). Furthermore, clinical and experimental evidence
shows that the anterior cingulate cortex is critical for normal pain-related behaviors ( 64 66 ). The activation of these thalamocortical
pathways to the insular and cingulate cortices is therefore consistent with other information about nociceptive processing. How the
cerebellum and premotor cortices fit into this picture is currently unclear. Based on evidence from other PET studies, it is possible that
cortical and subcortical motor mechanisms become activated in anticipation of movements intended to escape the noxious stimulation (
67 ).
The contralateral S1 cortex responds significantly in a minority of our studies. Often there is a peak of activation within S1, but it
fails to reach statistical significance in most of our studies. The activation of S1 across other PET studies of pain also seems quite variable.
This observation raises again the question of the role of S1 in the cerebral processing of pain. Neurophysiological studies leave little doubt
that nociceptive information reaches the S1 cortex ( 3 , 68 , 69 ). Clinical observations and PET studies show that S1 is critical for
somesthetic discriminative performance, but that surgical extirpation of S1 does little to relieve pain ( 70 , 71 ). The conditions that require
S1 activity for nociceptive processing have yet to be determined and present an interesting challenge for the future.
Each of the five structures identified above (six, counting each hemithalmus separately) is known to participate in cerebral functions
other than nociceptive processing and pain. It is premature to consider that this particular pattern of activation is unique for pain.
Nonetheless, in nearly all other PET studies of experimentally applied pain, the structures named here have been activated by using a wide
variety of stimuli and different data processing methods. There are other important variables to investigate. The issue of cerebral
asymmetry in pain processing has yet to be addressed systematically; it is probably an important factor in pathological pain states in
humans ( 72 ). It is likely that further experience with wholebrain imaging methods will allow us to identify a pain-specific pattern of
cerebral activation.
Meanwhile we have the opportunity to test specific hypotheses about the participation of each of these regions in nociceptive
processing and pain. Introducing conditions that perturb the cerebral activation pattern can test hypotheses about the mechanisms of pain
and analgesia. For example, Rainville et al. ( 73 ), by using hypnotic suggestion, were able to uncouple perceived intensity from perceived
unpleasantness to demonstrate a strong correlation between the ratings of unpleasantness and the degree of rCBF increase within a portion
of the anterior cingulate gyrus. And Derbyshire et al. ( 43 ) have shown that several cerebral regions, including those named above, show a
significant correlation of rCBF response magnitude with perceived stimulus intensity. Information relevant to perceived stimulus intensity
thus appears to be distributed widely, not simply to structures such as the S1 cortex that are known to mediate discriminative function. The
ability of PET to provide information about nociceptive processing in the awake human brain offers an opportunity to study the effects of
neuropathic pain, central nervous system damage, and the unique effects of analgesics. For example, although opioid analgesia specifically
attenuates pain-activated, but not vibration-activated, cerebral responses, it strongly activates the anterior cingulate cortex (K.L.C., P.
Svenssen, T. J. Morrow, J. Raz, C. S. Jone, and S. Minoshima, unpublished data). Such a result suggests the involvement of both spinal
and supraspinal sites of analgesic action, including the participation of descending inhibitory modulation.
Future progress in the analysis of the physiology of this pain-related network will require the development of an animal model for
invasive studies that cannot be performed in humans. We have recently developed a model for studies of nocifensive behaviors in the rat.
An Animal Model for Future rCBF Studies. We used rCBF in an animal model to identify the patterns of forebrain nociceptive
processing that occur during early and late phases of the well-established formalin test of inflammatory pain in the rat ( 74 ). During the
early phase, immediately after the injection of formalin into the dorsal hindpaw, pain behaviors are frequently elicited that are most
intense. This phase continues for approximately 5 min, after which nociception is considerably reduced. The late phase is marked by the
return of moderate to high levels of pain-related behaviors, beginning 10 to 15 minutes after formalin injection and continuing for 1 h.
The early phase is thought to be caused by the direct activation of peripheral nociceptors by formalin, whereas the late phase is believed to
be related to the development of inflammation and sensitization of central nociceptive neurons ( 75 78 ).
We measured normalized rCBF increases by an autoradiographic method that uses the radiotracer [99m]Tc-exametazime. Rats were
restraint-adapted to a soft towel for 2 to 3 weeks. To examine changes in rCBF during the early acute pain phase of the formalin test, we
injected the left hindfoot of the restrained rat with 0.05 ml of a 2.5% solution of formalin. After 2 min, we injected each animal with an
intravenous bolus of radiotracer. The same procedure was followed in the late phase of the test, but we injected the radiotracer 20 minutes
later. After the first injection, these adapted rats showed little or no movement while in the restraint. Two to five minutes after the
radiotracer injection, the rat was overdosed with anesthetic (chloral hydrate, 300 mg/kg i.v.) and decapitated; slides of the frozen brain
were prepared for routine histological staining and quantitative autoradiography. Eighteen regions of interest (ROIs) were selected,
representing various structures within the limbic and somatosensory systems. Densitometric analysis of autoradiograms was performed
with microcomputer-assisted video imaging. Anatomic location of selected ROIs was accomplished by overlaying matching transparencies
from a standard stereotactic atlas. We converted the film densities to apparent tissue radioactivity concentrations (nCi/mg) by comparing
them with the film optical densities of 14C-labeled standards, allowing ROI comparisons across different films and animals. An index of
activation was then calculated from individual ROI activities as a percentage of the average total activity of the entire brain. Significant
differences in activation for each ROI were detected between experimental groups by using ANOVA with post-hoc t tests (P 0.05).
During the early phase of the formalin test, a highly significant (31%) increase in rCBF occurred in the contralateral hindlimb cortex.
At the same time the retrosplenial portion of the cingulate cortex and the midbrain periaqueductal gray
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

FOREBRAIN MECHANISMS OF NOCICEPTION AND PAIN: ANALYSIS THROUGH IMAGING 7673

were activated bilaterally (31% and 7.8%, respectively). During the late phase, these structures remained active but the hindlimb activation
became bilateral. In addition, the intensity of periaqueductal gray activation increased to 20% and was joined by significant rCBF
increases in the interpeduncular and paraventricular nuclei (66% and 30%, respectively), in the habenular complex (58%), anterior dorsal
nucleus of the thalamus (30%), and the parietal cortex (30%) adjacent to the hindlimb cortex. The somatotopic organization of the
somatosensory thalamus and the small number of neurons excited by hindlimb stimulation probably resulted in an underestimation of
specific thalamic nuclei activity. Nonetheless, we detected blood flow increases in the ventral posterolateral thalamus (8.7%) and in the
medial thalamus (9.0%) that did not reach statistical significance but did tend to be greater in the late phase compared with the early phase
of the formalin test.
These results show that specific structures known to be important in nociceptive processing and modulation are selectively activated
in the awake rat during the formalin test. Activation of a structure may be related to nociception, antinociception, or both. The contralateral
hindlimb cortex and midbrain periaqueductal gray received nociceptive input and were active during the early phase. In the late phase,
bilateral activity was seen throughout the forebrain, with the recruitment of limbic system components, each of which has been shown to
participate in mediating or modulating nocifensive behaviors. In addition to the well-known analgesic effect of periaqueductal gray
stimulation, interpeduncular nucleus stimulation modulates antinociceptive circuitry in the medullary raphe nuclei ( 79 ), and stimulation
of the paraventricular nucleus produces analgesia ( 80 ). Analgesia also follows the microinjection of morphine and electrical stimulation
within the habenular complex ( 81 , 82 ). Activation of the cingulate cortex is consistent with the activation of one of its major inputs, the
anterior dorsal thalamic nuclei, and is in accord with the limbic cortical activation seen in human PET studies. Overall, the bilateral
activation of somatosensory and limbic structures agrees with 2-deoxyglucose studies of glucose uptake in rats with chronic constriction
injures of the sciatic nerve ( 83 ). Here we show that rCBF analysis is useful in studying central responses to acute and chronic stimuli.
The Future of Pain Imaging. This developing technology may undergo significant improvements in both spatial and temporal
resolution. Currently, PET provides a quantitative, statistically reliable method for assessing the activity of large brain and brainstem
regions. Hypotheses can then focus on the conditions necessary and sufficient to activate one or more regions in a group of subjects.
Although it is now possible to obtain reliable and quantitative information from single subjects with PET, fMRI has the ability to focus
with great precision on rCBF responses in specific regions. Working together in a complementary manner, the two procedures should help
develop a more precise understanding of the functional organization of pain and nociceptive processing. This progress will be facilitated
by the parallel use of animal models, allowing questions about dynamics and functional connectivity to be addressed by selective
stimulation, lesion, and drug microinjection studies.
The clinical impact of this effort will be apparent as we develop an understanding of how the central nervous system adapts to
chronic nociceptive input and injury. The changes in nociceptive processing demonstrated at the spinal cord level in experimental animals
are likely to affect nociceptive processing and hence pain at higher levels. Such studies may have an important impact on descending
modulatory influences, especially in forebrain-dominated animals such as humans. Evidence has accumulated showing that peripheral
injury can profoundly affect thalamic and cortical sensory processes over long periods of time ( 84 86 ). In some cases, these plastic
changes can be correlated with pain ( 87 ). A significant minority of patients with injury or disease of the central nervous system also
suffer chronic, often unremitting pain as a consequence of the central lesion(s) ( 88 ). The pathophysiology of this condition is unknown,
but the methods discussed here hold the promise for better solutions to the treatment and prevention of these chronic pain conditions.
1. Melzack, R. & Casey, K. L. ( 1968 ) in The Skin Senses , eds. Kenshalo, D. R. & Thomas, C. C. ( Thomas , Springfield, IL ), pp. 423439 .
2. Head, H. & Holmes, G. ( 1911 ) Brain 34 , 102254 .
3. Kenshalo, D. R., Jr. , & Isensee, O. ( 1983 ) J. Neurophysiol. 50 , 14791496 .
4. Foltz, E. L. & White, L. E. ( 1962 ) J. Neurosurg. 19 , 89100 .
5. Hurt, R. W. & Ballantine, H. T. ( 1973 ) Clin. Neurosurg. 21 , 334351 .
6. Sikes, R. W. & Vogt, B. A. ( 1992 ) J. Neurophysiol. 68 , 17201732 .
7. Price, D. D. ( 1988 ) Psychological and Neural Mechanisms of Pain ( Raven , New York ), pp. 1241 .
8. Gerhart, K. D. , Yezierski, R. P. , Fang, Z. R. & Willis, W. D. ( 1983 ) J. Neurophysiol. 49 , 406423 .
9. Yezierski, R. P. , Gerhart, K. D. , Schrock, B. J. & Willis, W. D. ( 1983 ) J. Neurophysiol. 49 , 424441 .
10. Dubner, R. , Hoffman, D. S. & Hayes, R. L. ( 1981 ) J. Neurophysiol. 46 , 444464 .
11. Hayes, R. L. , Dubner, R. & Hoffman, D. S. ( 1981 ) J. Neurophysiol. 46 , 428443 .
12. Basbaum, A. I. & Fields, H. L. ( 1984 ) Annu. Rev. Neurosci. 7 , 309338 .
13. Yuan, B. , Morrow, T. J. & Casey, K. L. ( 1986 ) J. Neurosci. 6 , 36113617 .
14. Shin, H. C. & Chapin, J. K. ( 1990 ) Somatosens. Mot. Res. 7 , 421434 .
15. Alloway, K. D. , Johnson, M. J. & Wallace, M. B. ( 1993 ) J. Neurophysiol. 70 , 892908 .
16. Liu, X. B. , Honda, C. N. & Jones, E. G. ( 1995 ) J. Comp. Neurol. 352 , 6991 .
17. Swanson, L. W. ( 1995 ) Trends Neurosci. 18 , 471474 .
18. Cheema, S. S. , Rustioni, A. & Whitsel, B. L. ( 1984 ) J. Comp. Neurol. 225 , 276290 .
19. Ralston, D. D. & Ralston, H. J. , III ( 1985 ) J. Comp. Neurol. 242 , 325337 .
20. Blinkov, S. M. & Glezer, I. I. ( 1968 ) The Human Brain in Figures and Tables ( Plenum , New York ).
21. Towe, A. L. ( 1995 ) J. Brain Res. 36 , 393398 .
22. Liu, X.-B. , Honda, C. N. & Jones, E. G. ( 1995 ) J. Comp. Neurol. 352 , 6991 .
23. Ogawa, S. , Lee, T.-M., Kay, A. R. & Tank, D. ( 1990 ) Proc. Natl. Acad. Sci. USA 87 , 98689872 .
24. Roy, C. S. & Sherrington, C. S. ( 1890 ) J. Physiol. (London) 11 , 85108 .
25. Roland, P. E. ( 1993 ) Brain Activation ( Wiley-Liss , New York ).
26. Malonek, D. & Grinvald, A. ( 1996 ) Science 272 , 551554 .
27. Wang, G. , Tanaka, K. & Tanifuji, M. ( 1996 ) Science 272 , 1665 1668 .
28. MacVicar, B. A. ( 1997 ) Neuroscientist 3 , 381388 .
29. Mraovitch, S. , Calando, Y. , Pinard, E. , Pearce, W. J. & Seylaz, J. ( 1992 ) Neuroscience 49 , 451466 .
30. Adachi, K. , Takahashi, S. , Melzer, P. , Campos, K. L. , Nelson, T. , Kennedy, C. & Sokoloff, L. ( 1994 ) Am. J. Physiol. 267 , H2155 H2162 .
31. Dirnagl, U. , Niwa, K. , Lindauer, U. & Villringer, A. ( 1994 ) Am. J. Physiol. 267 , H296H301 .
32. Seitz, R. J. & Roland, P. E. ( 1992 ) Acta Neurol. Scand. 86 , 6067 .
33. Auker, C. R. , Meszler, R. M. & Carpenter, D. O. ( 1983 ) J. Neurophysiol. 49 , 15041516 .
the authoritative version for attribution.

34. Ter-Pogossian, M. M. , Eichling, J. O. & Davis, D. O. ( 1969 ) Radiology 93 , 3140 .

35. Fox, P. T. & Raichle, M. E. ( 1984 ) J. Neurophysiol. 51 , 11091120 .
36. Talairach, J. & Tournoux, A. ( 1988 ) A Coplanar Stereotaxic Atlas of the Human Brain ( Thieme , New York ).
37. Minoshima, S. , Berger, K. L. , Lee, K. S. & Mintun, M. A. ( 1992 ) J. Nucl. Med. 33 , 15791585 .
38. Minoshima, S. , Koeppe, R. A. , Mintun, M. A. , Berger, K. L. , Taylor, S. F. , Frey, K. A. & Kuhl, D. E. ( 1993 ) J. Nucl. Med. 34 , 322329 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

FOREBRAIN MECHANISMS OF NOCICEPTION AND PAIN: ANALYSIS THROUGH IMAGING 7674

39. Minoshima, S. , Koeppe, R. A. , Frey, K. A. & Kuhl, D. E. ( 1994 ) J. Nucl. Med. 35 , 15281536 .
40. Friston, K. J. , Frith, C. D. , Liddle, P. F. & Frackowiak, R. S. ( 1991 ) J. Cereb. Blood Flow Metab. 11 , 690699 .
41. Worsley, K. J. , Evans, A. C. , Marrett, S. & Neelin, P. ( 1992 ) J. Cereb. Blood Flow Metab. 12 , 900918 .
42. Adler, R. J. & Hasofer, A. M. ( 1976 ) Ann. Probabil. 4 , 112 .
43. Derbyshire, S. W. , Jones, A. K. , Gyulai, F. , Clark, S. , Townsend, D. & Firestone, L. L. ( 1997 ) Pain 73 , 431445 .
44. Burton, H. , Videen, T. O. & Raichle, M. E. ( 1993 ) Somatosens. Mot. Res. 10 , 297308 .
45. Goolkasian, P. ( 1985 ) Psychol. Women Q. 9 , 1528 .
46. Gur, R. C. , Gur, R. E. , Obrist, W. D. , Hungerbuhler, J. P. , Younkin, D. , Rosen, A. D. , Skolnick, B. E. & Reivich, M. ( 1982 ) Science 217 , 659
661 .
47. Gur, R. C. , Mozley, L. H. , Mozley, P. D. , Resnick, S. M. , Karp, J. S. , Alavi, A. , Arnold, S. E. & Gur, R. E. ( 1995 ) Science 267 , 528531 .
48. Andreason, P. J. , Zametkin, A. J. , Guo, A. C. , Baldwin, P. & Cohen, R. M. ( 1994 ) Psychiatry Res. 51 , 175183 .
49. Azari, N. P. , Rapoport, S. I. , Grady, C. L. , DeCarli, C. , Haxby, J. V. , Schapiro, M. B. & Horwitz, B. ( 1992 ) Brain Res. 574 , 198208 .
50. Paulson, P. E. , Minoshima, S. , Morrow, T. J. & Casey, K. L. ( 1998 ) Pain 76 , 223229 .
51. Casey, K. L. , Minoshima, S. , Berger, K. L. , Koeppe, R. A. , Morrow, T. J. & Frey, K. A. ( 1994 ) J. Neurophysiol. 71 , 802807 .
52. Casey, K. L. , Minoshima, S. , Morrow, T. J. & Koeppe, R. A. ( 1996 ) J. Neurophysiol. 76 , 571581 .
53. Klement, W. & Arndt, J. O. ( 1992 ) J. Physiol. (London) 449 , 7383 .
54. Biemond, A. ( 1956 ) A. M. A. Arch. Neurol. Psychiatry 75 , 231244 .
55. Schmahmann, J. D. & Leifer, D. ( 1992 ) Arch. Neurol. 49 , 10321037 .
56. Bassetti, C. , Bogousslavsky, J. & Regli, F. ( 1993 ) Neurology 43 , 19421949 .
57. Greenspan, J. D. & Winfield, J. A. ( 1992 ) Pain 50 , 2939 .
58. Saper, C. B. ( 1982 ) J. Comp. Neurol. 210 , 163173 .
59. Craig, A. D. , Bushnell, M. C. , Zhang, E.-T. & Blomqvist, A. ( 1994 ) Nature (London) 372 , 770773 .
60. Craig, A. D. & Bushnell, M. C. ( 1994 ) Science 265 , 252255 .
61. Casey, K. L. ( 1966 ) J. Neurophysiol. 29 , 727750 .
62. Dong, W. K. , Ryu, H. & Wagman, I. H. ( 1978 ) J. Neurophysiol. 41 , 15921613 .
63. Bushnell, M. C. & Duncan, G. H. ( 1989 ) Exp. Brain Res. 78 , 415418 .
64. Vogt, B. A. , Sikes, R. W. & Vogt, L. J. ( 1993 ) in Neurobiology of Cingulate Cortex and Limbic Thalamus: A Comprehensive Handbook , eds.
Vogt, B. A. & Gabriel, M. ( Birkhauser , Boston ).
65. Devinsky, O. , Morrell, M. J. & Vogt, B. A. ( 1995 ) Brain 118 , 279306 .
66. Gabriel, M. & Poremba, A. ( 1995 ) in The Role of Pain in Cingulate Cortical and Limbic Thalamic Mediation of Avoidance Learning , eds. Besson,
J.-M. , Guilbaud, G. & Ollat, H. ( John Libbey, Eurotext , Paris ) pp. 197212 .
67. Hsieh, J.-C. , Hagermark, O. , Stahle-Backdahl, M. , Ericson, K. , Eriksson, L. , Stone-Elander, S. & Ingvar, M. ( 1994 ) J. Neurophysiol. 72 , 3004
3008 .
68. Kenshalo, D. R. , Chudler, E. H. , Anton, F. & Dubner, R. ( 1988 ) Brain Res. 454 , 378382 .
69. Chudler, E. H. , Anton, F. , Dubner, R. & Kenshalo, D. R. , Jr. ( 1990 ) J. Neurophysiol. 63 , 559569 .
70. White, J. C. & Sweet, W. H. ( 1969 ) Pain and the Neurosurgeon. A Forty-Year Experience ( Thomas , Springfield, IL ).
71. Sweet, W. H. ( 1982 ) in New Perspectives in Cerebral Localization , eds. Thompson, R. A. & Green, J. R. ( Raven , New York ), pp. 205242 .
72. Nasreddine, Z. S. & Saver, J. L. ( 1997 ) Neurology 48 , 11961199 .
73. Rainville, P. , Duncan, G. H. , Price, D. D. , Carrier, M. & Bushnell, M. C. ( 1997 ) Science 277 , 968971 .
74. Morrow, T. J. , Paulson, P. E. , Danneman, P. J. & Casey, K. L. ( 1998 ) Pain 75 , 355365 .
75. Coderre, T. J. , Vaccarino, A. L. & Melzack, R. ( 1990 ) Brain Res. 535 , 155158 .
76. Coderre, T. J , Katz, J. , Vaccarino, A. L. & Melzack, R. ( 1993 ) Pain 52 , 259285 .
77. Wiebe, J. P. & Kavaliers, M. ( 1988 ) Brain Res. 461 , 150157 .
78. Brown, A. G. ( 1981 ) Organization in the Spinal Cord ( Springer , New York ).
79. Hentall, I. D. & Budhrani, V. M. ( 1990 ) Brain Res. 522 , 322324 .
80. Yirmiya, R. , Ben-Eliyahu, S. , Shavit, Y. , Marek, P. & Liebeskind, J. C. ( 1990 ) Brain Res. 537 , 169174 .
81. Cohen, S. R. & Melzack, R. ( 1985 ) Brain Res. 359 , 131139 .
82. Cohen, S. R. & Melzack, R. ( 1986 ) Neurosci. Lett. 70 , 165169 .
83. Mao, J. , Mayer, D. J. & Price, D. D. ( 1993 ) J. Neurosci. 13 , 26892702 .
84. Kaas, J. H. ( 1991 ) Annu. Rev. Neurosci. 14 , 137167 .
85. Dubner, R. & Ruda, M. A. ( 1992 ) Trends Neurosci. 15 , 96103 .
86. Kaas, J. H. , Florence, S. L. & Neeraj, J. ( 1997 ) Neuroscientist 3 , 123130 .
87. Flor, H. , Elbert, T. , Knecht, S. , Wienbruch, C. , Pantev, C. , Birbaumer, N. , Larbig, W. & Taub, E. ( 1995 ) Nature (London) 375 , 482484 .
88. Casey, K. L. ( 1991 ) Pain and Central Nervous System Disease: The Central Pain Syndromes ( Raven , New York ), pp. 1280 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASCENDING PATHWAYS THAT MEDIATE VISCERAL NOCICEPTIVE TRANSMISSION 7675

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

A visceral pain pathway in the dorsal column of the spinal cord

WILLIAM D. WILLIS* , ELIE D. AL-CHAER*, MICHAEL J. QUAST*, AND KARIN N. WESTLUND*

Departments of *Anatomy and Neurosciences, Radiology, and Internal Medicine, University of Texas Medical Branch, Galveston,
TX 77555
ABSTRACT A limited midline myelotomy at T10 can relieve pelvic cancer pain in patients. This observation is explainable in
light of strong evidence in support of the existence of a visceral pain pathway that ascends in the dorsal column (DC) of the spinal
cord. In rats and monkeys, responses of neurons in the ventral posterolateral thalamic nucleus to noxious colorectal distention are
dramatically reduced after a lesion of the DC at T10, but not by interruption of the spinothalamic tract. Blockade of transmission
of visceral nociceptive signals through the rat sacral cord by microdialysis administration of morphine or 6-cyano-7-
nitroquinoxaline-2,3-dione shows that postsynaptic DC neurons in the sacral cord transmit visceral nociceptive signals to the
gracile nucleus. Retrograde tracing studies in rats demonstrate a concentration of postsynaptic DC neurons in the central gray
matter of the L6-S1 spinal segments, and anterograde tracing studies show that labeled axons ascend from this region to the
gracile nucleus. A similar projection from the midthoracic spinal cord ends in the gracile and cuneate nuclei. Behavioral
experiments demonstrate that DC lesions reduce the nocifensive responses produced by noxious stimulation of the pancreas and
duodenum, as well as the electrophysiological responses of ventral posterolateral neurons to these stimuli. Repeated regional blood
volume measurements were made in the thalamus and other brain structures in anesthetized monkeys in response to colorectal
distention by functional MRI. Sham surgery did not reduce the regional blood volume changes, whereas the changes were
eliminated by a DC lesion at T10.
Visceral pain caused by cancer of pelvic organs can be difficult to manage even with high doses of morphine. Neurosurgical
approaches to this problem have included anterolateral cordotomy to interrupt the spinothalamic tract and related associated pathways (see
ref. 1 for review). However, visceral pain is often bilateral in origin, and bilateral cordotomies can produce undesirable complications.
Commissural myelotomy was introduced as a means to interrupt the crossing axons of the spinothalamic tract over several segments
without damaging long tracts in the lateral funiculus. However, commissural myelotomy also can result in unwanted side effects ( 1 ).
Limited midline myelotomy at C1 was found to be surprisingly effective in relieving pain in distant parts of the body ( 2 , 3 ). The
proposed rationale for this procedure was interruption of a hypothetical multisynaptic pain transmission system in the central gray region
of the spinal cord ( 3 ). For pelvic cancer pain, a limited midline myelotomy at T810 has been reported to be effective with minimal side
effects ( 4 6 ).
ASCENDING PATHWAYS THAT MEDIATE VISCERAL NOCICEPTIVE TRANSMISSION
In one particularly well-documented clinical case, a patient who was treated successfully by a limited midline myelotomy had colon
cancer pain that was not relieved by large doses of morphine delivered i.v. by using a patient-controlled analgesia pump ( 5 ). After the
myelotomy was done at T10, the cancer pain was totally relieved, and the morphine was discontinued over 3 days. The patient died in 3
months without recurrence of the visceral pain and without need for pain medication. His spinal cord became available for postmortem
study and it was
the authoritative version for attribution.

FIG. 1. Transverse section of the spinal cord of a patient whose colon cancer pain was relieved by a limited midline myelotomy at
T10. The section was taken at a level just rostral to the site of the lesion and was stained for myelin. A bilateral demyelinated area in
the fasciculus gracilis is seen below the arrow. The drawing shows the demyelinated area in black. [Figure reproduced with
permission from ref. 5 (Copyright 1996, International Association for the Study of Pain).]

PNAS is available online at www.pnas.org .

Abbreviations: CRD, colorectal distention; DC, dorsal column; VLC, ventral lateral column; VPL, ventral posterolateral nucleus.
To whom reprint requests should be addressed at: University of Texas Medical Branch, 301 University Boulevard, Galveston, TX

77555-1069. e-mail: [email protected] .

lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASCENDING PATHWAYS THAT MEDIATE VISCERAL NOCICEPTIVE TRANSMISSION 7676

determined that the surgical lesion interrupted the medial fasciculus gracilis bilaterally, as shown in Fig. 1 . The lesion did not appear to
intrude into the central gray matter.

FIG. 2. Effects of DC and VLC lesions at T10 in rats (n = 20) on responses to mechanical stimulation of the skin (Brush, Press, and
Pinch) and to graded intensities of CRD (20, 40, 60 and 80 mm Hg). The responses were normalized. * indicates a significant change.
[Figure reproduced with permission from ref. 7 (Copyright 1996, The American Physiological Society).]

FIG. 3. Effects of microdialysis administration of drugs that block nociceptive synaptic transmission into the sacral spinal cord on
responses of gracile neurons (n = 10) to graded intensities of CRD. The drugs included morphine and CNQX (6-cyano-7-
nitroquinoxaline-2,3-dione). Naloxone was given systemically to reverse the action of morphine. In some animals (n = 3), a lesion
was placed in the DC. * indicates a significant change. [Figure reproduced with permission from ref. 11 (Copyright 1996, The
American Physiological Society).]
Experimental studies therefore were begun to determine whether there is a visceral nociceptive pathway in the dorsal column (DC).
Recordings were made in the ventral posterolateral (VPL) nucleus of the rat thalamus from neurons that responded to colorectal distention
(CRD) or to inflammation of the colon ( 5 , 7 ). The neurons also responded to mechanical stimuli applied to the skin. Sequential lesions of
the DC and the spinothalamic tract in the ventrolateral column (VLC) were made to see whether these affected the responses to noxious
stimulation of the colon or to stimulation of the cutaneous receptive field. A DC lesion was found to reduce the responses to CRD by 60
80%, whereas a VLC lesion only reduced such responses by 20% ( Fig. 2 B). The DC lesion also
the authoritative version for attribution.

FIG. 4. Morphological studies of the origins of axons in the medial DC that might mediate the transmission of visceral nociceptive
information. The retrograde tracer, WGA-HRP, was injected into the DCs at the cervical level, taken up by axons, and transported
caudally to label numerous cell bodies in the central region of the spinal cord adjacent to the central canal (CC). [Figure reproduced
with permission from ref. 5 (Copyright 1996, International Association for the Study of pain).]
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

MORPHOLOGICAL STUDIES OF THE VISCERAL POSTSYNAPTIC DC SYSTEM 7677

profoundly reduced the responses to weak stimulation of the skin but had little effect on those to noxious pinch ( Fig. 2 A). A VLC lesion
nearly eliminated responses to pinch but had only limited effects on responses to weak stimulation of the skin. A DC lesion also reduced
the heightened activity of VPL neurons after colon inflammation; a VLC lesion had much less effect. A small electrolytic or kainic acid
lesion in the gracile nucleus dramatically reduced the responses of VPL neurons to CRD ( 8 ). It was concluded that the DC is more
important than the VLC for transmitting nociceptive signals from the colon to the VPL nucleus ( 7 ). Similar results with respect to CRD
were obtained in monkeys ( 9 ).

FIG. 5. Summary diagram depicting the course of axons arising from postsynaptic DC neurons in the visceral processing region
adjacent to the central canal. Axons arising from neurons near the central canal at sacral levels of the spinal cord ascend in the
midline of the DC adjacent to the medial septum to innervate the medial gracile nucleus. Axons arising from neurons near the central
canal at thoracic levels of the spinal cord ascend in the DC adjacent to the dorsal intermediate septum and innervate both the gracile
and cuneate nuclei.
The visceral nociceptive signals that reach the gracile nucleus could be transmitted by the collaterals of primary afferent neurons that
ascend directly to the gracile nucleus or by the axons of postsynaptic DC neurons (see review in ref. 10 ). Which of the routes was the
more effective was tested by blocking nociceptive transmission from the colon in the sacral spinal cord by using microdialysis
administration of morphine or of the non-N-methyl-D-aspartic acid receptor antagonist CNQX (6-cyano-7-nitroquinoxaline-2,3-dione) (
11 ). Access to the spinal cord by colon afferents was restricted to the pelvic nerve distribution by sectioning the hypogastric nerves
bilaterally. Administration of either drug blocked the responses of gracile neurons to CRD ( Fig. 3 ). Naloxone reversed the action of
morphine. Similar effects were observed when recordings were made from identified postsynaptic DC neurons in the central gray region
of the sacral cord. Neurons in this region are known to have a strong visceral input, and it proved to be the case for postsynaptic DC
neurons located in this region. It was concluded that most of the responses of gracile neurons to CRD distention (or colon inflammation)
depended on activation of postsynaptic DC neurons, although it could not be ruled out that some of the responses were the result of
conduction in directly projecting primary afferents.

FIG. 6. Reduction in the writhing reflex (WR) evoked by balloon distention of the duodenum after a DC lesion, (a) The intensities of
the WR evoked by graded distention (by 0.1 to 0.7 ml) of the duodenum are shown. No differences in the behavioral response were
seen between rats tested before the sham surgery (pre-SH) and rats tested after the sham surgery (SH). The scale used was from 0
(normal) to 4 (stretching of the body, extension of the hind limbs). (b) The effect of the DC lesion on the behavioral response to
duodenal distention is shown (prelesion vs. after lesion). (c) A comparison of the behavior in the sham surgery and lesion groups. The
numbers of animals used are shown. * indicate significant changes. [Figure reproduced with permission from ref. 13 (Copyright
1998, Lippincot Williams & Wilkins, http://lww.com ).]
the authoritative version for attribution.

MORPHOLOGICAL STUDIES OF THE VISCERAL POSTSYNAPTIC DC SYSTEM

The cells of origin of the midline DC pathway and their brainstem terminations have been identified anatomically (Figs. 4 and 5 ).
The injection of a retrograde tracer, WGA-HRP, into the DC at an upper cervical level labeled a large number of postsynaptic DC neurons
in the central gray and surrounding area at all levels of the spinal cord examined ( 5 ). Unpublished studies have shown that retrograde
label microinjected into the medial aspect of the gracile nucleus also labels a large population of postsynaptic DC neurons in the region of
the central gray at sacral levels.
Injection of an anterograde tracer, biotin dextran, into the central gray region of the sacral cord labeled axonal projections from this
region that ascend in the fasciculus gracilis near the midline and have terminal arbors in the medial part of the gracile nucleus ( 5 ). These
observations were confirmed by small injections of Phaseolus vulgaris leucoagglutinin as the
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

FUNCTIONAL MRI STUDIES IN MONKEYS 7678

anterograde tracer made into the central canal region of the sacral cord (C.-C. Wang, W.D.W., and K.N.W., unpublished work). The
visceral postsynaptic DC pathway arising from the sacral spinal cord ascends in the DC midline adjacent to the medial septum. Similar
injections into the central gray region at a midthoracic level label axons that ascend adjacent to the dorsal intermediate septum. These
axonal projections also have terminations in the gracile and cuneate nuclei. A summary diagram representing these pathways is shown in
Fig. 5 .

FIG. 7. Functional MRI (fMRI) study of the effects of a DC lesion on blood volume changes evoked in the brain after noxious
distention of the colon in monkeys. (Left) The brain slices shown were taken from a monkey that was subjected to sham surgery at
T10. The fMRI images were from a comparable location before (Upper) and then 4 months after the sham surgery (Lower). (Right)
The images were made before and 4 months after a DC lesion at T10 in a different monkey. Colored regions indicate locations where
regional cerebral blood volume increased relative to the nonstimulated state.
BEHAVIORAL EVIDENCE FOR A DC VISCERAL PAIN PATHWAY
To determine whether the DC helps mediate pain originating in abdominal viscera, experiments were done in awake, behaving rats on
the effects of a DC lesion on the behavioral responses to pancreatitis and noxious chemical stimulation of the pancreas and to duodenal
distention.
A pancreatitis model induced by infusion of glycodeoxycholic acid into the pancreas and i.p. caerulian was shown to reduce home-
cage exploratory behaviors (rearing) in rats ( 12 ). Decreases in normal activity did not occur in animals that had received lesions of their
DC at the C1 spinal level 1 week before induction of the pancreatitis, suggesting that the DC lesion provided a protective effect. More
recent unpublished data confirmed that responses of VPL neurons to stimulation of the pancreas with bradykinin also are reduced by a DC
lesion and also by spinal administration of morphine at a midthoracic level. The effect of morphine is naloxone reversible.
In another series of experiments, a balloon catheter was chronically implanted in the duodenum in rats through the stomach wall (
13 ). Graded distention of the duodenum in awake, behaving animals resulted in graded intensities of the writhing reflex. A lesion of the
DC at C2 produced a dramatic reduction in the intensity of the writhing reflex ( Fig. 6 ). However, to be effective, the lesion had to include
the region of the dorsal intermediate septum bilaterally. Parallel experiments were done in anesthetized rats to determine the effects of a
DC lesion at C2 on the responses of VPL neurons to duodenal distention. Again, a lesion of the DC that included the region of the dorsal
intermediate septum bilaterally resulted in a profound reduction in the responses of the VPL neurons. The requirement for the placement
of the DC lesions laterally to the midline is explained by the morphological study described in the previous section and illustrated in
Fig. 5 . Axons originating from postsynaptic dorsal horn neurons in the central gray region of the midthoracic spinal cord travel toward the
DC nuclei near the dorsal intermediate septum.
FUNCTIONAL MRI STUDIES IN MONKEYS
In a recent unpublished study, the brain structures involved in mediating visceral nociceptive responses have been investigated by
using functional MRI. Monkeys were anesthetized with isoflurane and placed in a 4.7-Tesla magnet for imaging regional cerebral blood
volume. The blood was labeled with a superparamagentic iron oxide compound to enhance the contrast in the image in proportion to
increases in regional blood volume. Relative cerebral blood flow was estimated by gradient echo bolus tracking, and changes in cerebral
the authoritative version for attribution.

blood volume were estimated by steady-state spin echo imaging. Noxious

lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

FUNCTIONAL MRI STUDIES IN MONKEYS 7679

CRD consistently caused regional blood volume increases in a number of brain structures, including the VPL nucleus. Blood volume
changes were not reduced after sham surgery, but they were essentially completely eliminated after the DC lesion. The effects of the DC
lesion persisted for at least 4 months ( Fig. 7 ).
We thank Kelli Gondesen and Jingna Wei for their expert technical assistance. The work was supported by National Institutes of
Health Grant NS 11255.
1. Gybels, J. M. & Sweet, W. H. ( 1989 ) Neurosurgical Treatment of Persistent Pain (Karger, Basel).
2. Hitchcock, E. R. ( 1970 ) J. Neurol. Neurosurg. Psychiatry 33 , 224230 .
3. Schvarcz, J. R. ( 1984 ) Acta Neurochir. 33, Suppl. , 431435 .
4. Gildenberg, P. L. & Hirshberg, R. M. ( 1984 ) J. Neurol. Neurosurg. Psychiatry 47 , 9496 .
5. Hirshberg, R. M. , Al-Chaer, E. D. , Lawand, N. B. , Westlund, K. N. & Willis, W. D. ( 1996 ) Pain 67 , 291305 .
6. Nauta, H. J. W. , Hewitt, E. , Westlund, K. N. & Willis, W. D. ( 1997 ) J. Neurosurg. 86 , 538542 .
7. Al-Chaer, E. D. , Lawand, N. B. , Westlund, K. N. & Willis, W. D. ( 1996 ) J. Neurophysiol 76 , 26612674 .
8. Al-Chaer, E. D. , Westlund, K. N. & Willis, W. D. ( 1997 ) J. Neurophysiol. 78 , 521527 .
9. Al-Chaer, E. D. , Feng, Y. & Willis, W. D. ( 1998 ) J. Neurophysiol. 79 , 31433150 .
10. Willis, W. D. & Coggeshall, R. E. ( 1991 ) Sensory Mechanisms of the Spinal Cord ( Plenum , New York ).
11. Al-Chaer, E. D , Lawand, N. B. , Westlund, K. N. & Willis, W. D. ( 1996 ) J. Neurophysiol. 76 , 26752690 .
12. Houghton, A. K. , Kadura, S. & Westlund, K. N. ( 1997 ) NeuroReport 8 , 37953800 .
13. Feng, Y. , Cui, M. , Al-Chaer, E. D. & Willis, W. D. ( 1998 ) Anesthesiology 89 , 411420 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

BEHAVIORAL EFFECTS OF CUTANEOUS STIMULI AFTER INJURY 7680

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

The spinal biology in humans and animals of pain states generated

by persistent small afferent input

TONY L. YAKSH * , XIAO-YING HUA* , IVETA KALCHEVA*, NATSUKO NOZAKI-TAGUCHI * , AND MARTIN MARSALA*
* Department of Anesthesiology, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0818; and
Department of Anesthesiology, Chiba University School of Medicine, 18-1 Inohana Chuo-ku Chiba-shi, 260 Japan

ABSTRACT Behavioral models indicate that persistent small afferent input, as generated by tissue injury, results in a
hyperalgesia at the site of injury and a tactile allodynia in areas adjacent to the injury site. Hyperalgesia reflects a sensitization of
the peripheral terminal and a central facilitation evoked by the persistent small afferent input. The allodynia reflects a central
sensitization. The spinal pharmacology of these pain states has been defined in the unanesthetized rat prepared with spinal
catheters for injection and dialysis. After tissue injury, excitatory transmitters (e.g., glutamate and substance P) acting though N-
methyl-D-aspartate (NMDA) and neurokinin 1 receptors initiate a cascade that evokes release of (i) NO, (ii) cyclooxygenase
products, and (iii) activation of several kinases. Spinal dialysis show amino acid and prostanoid release after cutaneous injury.
Spinal neurokinin 1, NMDA, and non-NMDA receptors enhance spinal prostaglandin E2 release. Spinal prostaglandins facilitate
release of spinal amino acids and peptides. Activation by intrathecal injection of receptors on spinal C fiber terminals (,/ opiate,
2 adrenergic, neuropeptide Y) prevents release of primary afferent peptides and spinal amino acids and blocks acute and
facilitated pain states. Conversely, consistent with their role in facilitated processing, NMDA, cyclooxygenase 2, and NO synthase
inhibitors act to diminish only hyperalgesia. Importantly, spinal delivery of several of these agents diminishes human injury pain
states. This efficacy emphasizes (i) the role of facilitated states in humans, (ii) shows the importance of spinal systems in human
pain processing, and (iii) indicates that these preclinical mechanisms reflect processes that regulate the human pain experience.
Local tissue injury and inflammation yields well-defined escape behaviors in animals and pain reports in humans. Examination of the
histochemistry and electrophysiology of spinal systems has revealed considerable detail regarding the elements of systems that are
activated by these stimuli. Nevertheless, the functional contribution of different spinal systems in pain processing ultimately must be
defined in terms of the systems in which such end points can be measured, e.g., the behavior of the intact organism. We will consider
below how certain spinal systems contribute to the observed behavioral states.
BEHAVIORAL EFFECTS OF CUTANEOUS STIMULI AFTER INJURY
An acute, unconditioned, thermal, or mechanical stimulus sufficient to activate polymodel nociceptive afferents (C fibers) depolarizes
populations of dorsal horn wide dynamic range (WDR) neurons that project supraspinally. This output in turn evokes a supraspinally
organized escape behavior. The hot plate test (thermal stimulus to the paw) or the local injection of an irritant such as formalin or
capsaicin where the unconditioned stimulus evokes a somatotopically directed behavior (e.g., withdrawal or licking) are behavioral
paradigms believed to reflect this underlying mechanism ( 1 ). The more intense the stimulus, the more robust will be the afferent volley
and the more vigorous or shorter latencied is the escape behavior ( 2 ).
An acute stimulus of intensity and duration that leads to tissue injury also produces an acute discharge. In addition, the injury leads to
the local release of active factors that evoke and sustain persistent activity in the sensory afferents innervating the injured or inflamed
tissue ( 3 ). Thus, in contrast to the acute response, injury leads to persistent activity in populations of small afferents and also may activate
afferent populations that are excited only in the presence of local factors generated by the injury (e.g., silent nociceptors) ( 4 ).
Electrophysiological studies have shown that the persistent activation of spinal WDR neurons by small, but not large, afferents, will lead:
(i) a progressive enhancement of the WDR response to each subsequent input, and (ii) an increase in the dimensions of the peripheral
receptive field to which the spinal neuron will respond ( 5 ). This electrophysiological observation parallels behavioral changes in which
the animal displays an enhanced response to a given stimulus or a reduction in the intensity of the stimulus required to evoke an escape
response. Thus, the injection of an irritant (formalin) into one hind paw evokes a high frequency barrage during the first 1020 min
followed by a modest ongoing discharge over the next hour ( 6 ). Coincident with the initial afferent barrage, WDR neurons display an
initial burst of activity followed by a period of quiescence and then a progressive enhanced barrage ( 7 ). In rats, injection of formalin
results in a prominent licking and flinching of the injected paw with the incidence of flinching showing a biphasic time course that
parallels that reported for the discharge of spinal WDR neurons (see Fig. 1 ). The first-phase behavior is the result of an initial intense
afferent barrage. The secondphase behavior is believed to represent the induction of a state of spinal facilitation in which the diminished
formalin-initiated afferent input yields a prominent response.
Alternately, after a mild local burn, there is a decreased nociceptive threshold to heat within the burn area (a 1 thermal hyperalgesia)
and a pain response generated by light touch applied to uninjured skin regions adjacent to the area of injury (a 2 tactile allodynia) (see
Fig. 2 ) ( 8 ). Importantly, at least the initiating component of this hyperalgesia reflects on
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: WDR, wide dynamic range; NMDA, N-methyl-D-aspartate; COX, cyclooxygenase; NK, neurokinin; AMPA, -amino3-
hydroxy-5-methylisoxazole-4-propionic acid; PK, protein kinase; NOS, NO synthase; PG, prostaglandin; sP, substance P; CGRP,
calcitonin gene-related peptide; DRG, dorsal root ganglia.
To whom reprint requests should be addressed. e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CHARACTERIZATION OF SEVERAL SPINAL COMPONENTS LEADING TO POSTINJURY PAIN STATES 7681

small afferent input. The intradermal injection of capsaicin, an agent known to selectively activate C fibers, can induce a 2 allodynia in
humans and animals ( 9 ). This altered sensory condition persists after the termination of the pain produced by the capsaicin injection and
extends anatomically beyond the local site in which the capsaicin was shown to exert an effect.

FIG. 1. (Upper) C fiber activity recorded in situ in rats from single sural nerve fibers, identified by their conduction velocity and
modality as C fibers. Immediately after formalin injection (as indicated by the dashed line) into their receptive fields, high activity
was observed in high-threshold C nociceptive afferent fibers (as well as in A beta and A delta fibers, data not shown). At later
intervals, activity was observed in all mechanically sensitive C fibers, at rates that were less (1/22/3) than those achieved initially
(adapted from ref. 6 ). (Lower) Frequency of flinching as measured by an automated motion detector is plotted at 5-min intervals after
the injection of formalin into the paw at the time indicated by the vertical dashed line. As indicated, the flinching behavior displays a
biphasic occurrence (phase 1 and phase 2). The data represent the mean SEM of eight rats.
ROLE OF SPINAL AND PERIPHERAL SYSTEMS IN THE POST-TISSUE INJURY PAIN STATE
The behavioral sequelae outlined above, showing a hyperalgesic/allodynic state after tissue injury, may result from a peripheral
sensitization secondary to the injury and/or to a change in central processing initiated by the persistent small afferent input generated by
the injury. Blockade of spinal activation by the spinal delivery of a local anesthetic ( 11 ) or a selective blockade of small afferent input by
the intrathecal infusion of a short-lasting opiate during the initial period of injury ( 12 ) will attenuate the second phase of the formalin
response and abolish the 2 tactile allodynia, but not the 1 hyperalgesia observed secondary to a mild thermal injury (N.N.-T. and T.L.Y.,
unpublished observations). Importantly, as the pain behavior observed during the second phase after formalin injection is blocked by the
injection of local anesthetic into the paw ( 13 ), it is clear that the exaggerated responding indeed depends on the concurrent low-level
afferent input observed during the second phase of the formalin test (see Fig. 1 ). These findings thus support the hypothesis that (i) the
initial injury-induced afferent barrage generated in an opiatesensitive spinal system initiates a cascade that supports the 2 allodynia
observed after injury, and (ii) the 1 hyperalgesia is mediated in part by a peripheral sensitization of small opiatesensitive C fibers.

FIG. 2. Time course of change in mechanical threshold necessary to evoke acute withdrawal (Upper) or the thermal escape latency to
evoke withdrawal (Lower) in the normal (noninjured) and injured paw. The injury was induced with the exposure of the shaded area
indicated in the paw diagram (Left) to a 52C thermal stimulus applied for 45 sec at the time indicated in the graphs by the vertical
dashed line. As indicated, the test stimuli were applied to the sites as indicated. Only a modest change in tactile thresholds were
observed at the injury site, and no change in thermal escape thresholds were noted in the off injury site (data not shown). Hence the
lower response latency corresponds to a 1 thermal hyperalgesia and 2 tactile allodynia. Contralateral paws showed no systematic
change. Mechanical thresholds were determined with Von Frey hairs, and the thermal escape thresholds were assessed with
Hargreaves apparatus ( 10 ). All points represent mean SEM of five animals. B, baseline threshold.
the authoritative version for attribution.

CHARACTERIZATION OF SEVERAL SPINAL COMPONENTS LEADING TO POSTINJURY PAIN STATES

Based on immunohistochemistry and electrophysiology, several points are evident regarding the biology of several spinal systems
that may mediate the consequences of small afferent activation. (i) Populations of C fibers jointly contain peptides such as substance P
(sP) and calcitonin gene-related peptide (CGRP), as well as amino acids such as glutamate. (ii) Small afferent activation will evoke the Ca2
+-dependent spinal release of these products. (iii) Focusing on sP and glutamate, these agents evoke excitation of second-order neurons

through an effect mediated by the tachykinin neurokinin 1 (NK-1) and the glutamatergic -amino-3-hydroxy-5-
methylisoxazole-4propionic acid (AMPA)/N-methyl-D-aspartate (NMDA) receptors, respectively. In situ hybridization shows labeling for
NK-1 and NMDA receptor units in the dorsal gray matter, particularly in the substantia gelatinosa where small afferents are known to
terminate, (iv) Electrophysiologically, NK-1 and AMPA receptor antagonists will diminish small afferentevoked excitation. NMDA
antagonists do not appear to reduce monosynaptically mediated afferent-evoked excitation and thus are not believed to be immediately
postsynaptic to the primary afferent terminal, though some binding may be on the C-fiber terminal itself (see refs. 5 and 14 for references).
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CHARACTERIZATION OF SEVERAL SPINAL COMPONENTS LEADING TO POSTINJURY PAIN STATES 7682

Spinal Pharmacology of Facilitated Processing

The preceding section emphasizes that after tissue injury there is an excitation of small sensory afferents and the production of a
behaviorally defined state of hyperalgesia and allodynia. The overview of connectivity suggests elements that define a portion of the
organization of spinal systems that encode activity generated by small afferent input. The contribution of these several spinal systems in
nociceptive processing can be determined by considering the effects of systematically altering spinal pharmacology on pain behavior
generated by acute high intensity and tissue injurious stimuli. Modifications of spinal pharmacology can be accomplished by the spinal
delivery of pharmacological agents in animal behavior models by using chronically implanted catheter systems as noted above ( 15 ).
Regulation of Spinal Terminal Excitability. Based on the described role of small afferents activated by tissue injury, it is
reasonable to hypothesize that regulation of small afferent terminal excitability will diminish afferent-evoked pain behavior. Such
regulation should be achieved by receptors having a presynaptic inhibitory effect on spinal C-fiber terminals as defined by: (i) presence of
receptor binding on terminals of C fibers [e.g., receptor mRNA in the dorsal root ganglia (DRG), particularly in small DRG cell bodies,
binding, or receptor protein in the spinal substantia gelatinosa]; (ii) negative coupling of receptor to the opening of voltage-sensitive Ca2+
channels, and (iii) their ability to block release of small afferent transmitters (sP or CGRP). Mu and delta opioid ( 16 ), alpha 2 agonist (
17 ), and neuropeptide Y ( 18 ) receptor systems possess such a presynaptic distribution and effect (see Fig. 3 ). In addition to the
presynaptic action of these agents, binding, receptor protein, and/or mRNA is in dorsal horn neurons. These postsynaptic receptors are
coupled to Gi/o-protein and increase potassium conductance, serving to hyperpolarize those membranes and directly block depolarization
of that neuron ( 1 ). This joint action, reducing small afferent excitatory input and diminishing postsynaptic excitability, maximizes the
likelihood of a selective effect on acute nociceptive processing. The functional importance to pain processing of this concurrent spinal
action is demonstrated by the dose-dependent and pharmacologically specific blockade of the acute response to an acute high intensity or
injurious thermal (hot plate and tail flick), mechanical (paw pressure), or chemical (intradermal formalin) stimuli produced when these
agents are delivered intrathecally in a variety of animal models (see ref. 1 ). Consistent with the electrophysiology, at doses that alter pain
behavior there is no effect on the response to proprioceptive stimuli or on motor function.

FIG. 3. Schematic summarizes the organization of several dorsal horn systems that contribute to the processing of nociceptive
information. Primary afferent C fibers release peptide (e.g., sP/CGRP, etc.) and excitatory amino acid (glutamate) products. Small
DRG as well as some postsynaptic elements contain NOS) and are able on depolarization to release NO. Peptides and excitatory
amino acids evoke excitation in second-order neurons. For glutamate, direct monosynaptic excitation is mediated by non-NMDA
receptors (i.e., acute primary afferent excitation of WDR neurons is not mediated by the NMDA or NK-1 receptor). Interneurons
excited by afferent barrage induce excitation in second-order neuron via a NMDA receptor, which leads to an increase in intracellular
Ca2+, activation of phospholipase A2, NOS, and phosphorylating enzymes. COX products (PG) and NO are formed and released.
These agents diffuse extracellularly and facilitate transmitter release (retrograde transmission) from primary and nonprimary afferent
terminals by either a direct cellular action (e.g., NO) or by an interaction with a specific class of receptors [e.g., PG type E (EP)
receptors for prostanoids]. Phosphorylation of intracellular protein (e.g., enzymes and receptors such as NMDA) leads to additional
enhanced sensitivity. See text for other details.
sP. The spinal delivery of NK-1 receptor agonists results in a mild acute pain behavior and a subsequent reduced response latency
to thermal stimuli (thermal hyperalgesia). Blockade of the NK-1 receptor by intrathecal antagonists ( 19 ) or down-regulation of NK-1
receptor expression by intrathecal treatment with NK-1 receptor mRNA antisense ( 20 ) has no effect on acute nociceptive thresholds, but
reduces the second phase of the formalin response. Intrathecal injection of NK-1 antagonists after phase 1 reduces their effect on the
second phase ( 19 ).
Glutamate Receptors. Repetitive small afferent input (as that which occurs after tissue injury) will evoke spinal glutamate release
(see Fig. 4 ) ( 21 , 22 ). The spinal delivery of agonists for the ionotrophic glutamate receptors (NMDA/ AMPA) will evoke a potent
spontaneous pain behavior and a subsequent thermal hyperalgesia and tactile allodynia ( 23 ).
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SYSTEM INTERACTIONS 7683

Blockade of spinal AMPA receptors by intrathecal antagonists ( 24 ) will elevate acute nociceptive thresholds, as well as the first and
second phase of the formalin test. In contrast, intrathecal NMDA antagonists have little effect on acute nociception, but diminish the
second phase of the formalin test ( 25 ). As with the NK-1 antagonists, NMDA antagonists in the formalin model show a diminished effect
on phase two when delivered after phase 1 ( 26 ), which reflects the fact that after an acute injurious stimulus, such as with formalin
injection, there is an initiating barrage of activity leading to transmitter release (see Fig. 4 ). This release, for example of glutamate, sP, and
prostanoids, leads to biochemical changes within the spinal cord that must persist after the initial occupancy of the NMDA (or NK-1
receptor) receptor during phase 1 has passed.
Prostaglandins (PGs). PGs are released in vivo from the spinal cord by a peripherally injurious stimulus that is associated with small
afferent activation (see Fig. 4 ) ( 28 ) and by the direct spinal delivery of NK-1 and glutamate receptor agonists ( 22 , 29 ). Prostanoid
receptors are present in the dorsal horn and on DRG ( 30 ). Activation of prostanoid receptors has been shown to increase the opening of
voltage-sensitive Ca channels and to enhance primary afferent peptide release ( 31 ). Consistent with these events, intrathecal PG receptor
agonists will evoke a hyperalgesia and allodynia ( 32 ). The spinal delivery of cyclooxygenase (COX) inhibitors or antagonists has no
effect on acute nociceptive processing, but will reduce the phase 2 of the formalin test at doses that are inactive when given systemically (
33 , 34 ). Consistent with the observation that NMDA/ NK-1 antagonists can block certain hyperalgesic states, spinal NMDA agonists can
evoke a hyperalgesic state and this hyperalgesia also can be blocked by spinal COX inhibitors ( 23 ). Several COX isozymes have been
identified ( 35 ). Current evidence indicates that COX-2 is constitutive in the spinal cord (but only in a few peripheral organs). The
intrathecal delivery of selective COX inhibitors suggests that the hyperalgesia induced by intrathecal NK1/NMDA is mediated in part by
the local release of PGs. In recent work, we have shown that spinal COX2 inhibitors are also effective in models of inflammatory
hyperalgesia (ref. 36 and D. M. Dirig and T.L.Y., unpublished observations).

FIG. 4. Time course of flinching behavior (Top) and concurrent assessment of lumbar spinal glutamate (Middle), and PGs E2
(Bottom) release measured in unanesthetized rats before and after the injection of formalin (5%/50 l) into the left hind paw of the rat
at time = 0. At 10 min, the rats received an intrathecal (IT) injection of saline (vehicle: n = 4) or morphine (10 g; n = 7). Release
was assessed by using a chronically implanted loop dialysis probe ( 27 ), and the intrathecal injection was through a chronic
intrathecal catheter. Each line presents the mean of four and seven rats, respectively. Error bars deleted for clarity. *, P < 0.05 versus
formalin; #, P < 0.05 vs. IT saline.
NO. NO is synthesized by NO synthase (NOS). Evidence of in vivo NO release from cord secondary to repetitive afferent stimulation
and by intrathecal NMDA has been presented ( 21 ). The hyperalgesia induced by intrathecal NMDA and the second phase of the formalin
test has been shown to be reduced by intrathecal competitive NOS inhibitors ( 37 ).
Kinases and Phosphorylation. Increasing intracellular Ca2+I, though the inosital triphosphate pathway by activation of NK receptor
and/or by the influx of Ca2+ through voltagegated Ca2+ channels or ionophores (NMDA receptor) ( 38 ), activates kinases that
phosphorylate and phosphatases, which dephosphorylate local proteins. Phosphorylating enzyme systems consist of several classes of
kinases that are distinguished by structure and the pharmacology of their inhibitors. In the spinal dorsal horn, cAMP-dependent kinase (
39 ) and camkinase II ( 40 ) have been observed in the spinal dorsal horn and DRG. Protein kinase (PK) C consists of a large family of
isozymes. In the spinal cord PKC ( 41 ) and PKC ( 42 ) are limited to the spinal gray, with PKC being reported to be largely in cells in
lamina I and II inner of the dorsal horn. Although it is possible that any or all of the phosphorylating enzymes noted above may play a
role, the use of inhibitors for PKA and PKC have shown the particular importance of this family of kinases in regulating spinal facilitation.
the authoritative version for attribution.

Many hyperalgesic states are mediated by a spinal NMDA receptor. The NMDA receptor is multiply phosphorylated by PKA and PKC (
43 ). Intrathecal delivery of PKC inhibitors has been shown to stereospecifically diminish the hyperalgesic effects of intrathecal NMDA.
In addition, the augmented activity in dorsal horn neurons after intradermal mustard oil ( 44 , 45 ) or spinal NMDA ( 46 ) is reduced by the
local spinal delivery of PKC inhibitors. PKA and PKC inhibitors, but not inactive isomers, will diminish capsaicin-evoked hyperpathia (
45 ) and the second phase of the formalin test ( 47 , 48 ).
SYSTEM INTERACTIONS
As reviewed above, after local injury, the behaviorally defined components of post-tissue injury pain states reflect an increased
receptive field and a left shift in the stimulus response curve for spinal dorsal horn neurons, which is evoked initially and then sustained in
part by persistent small afferent input. The contribution of these changes in spinal function to the behaviorally relevant nociceptive state is
substantiated by comparing the pharmacology associated with the effects on the behavior of the unanesthetized animal with the effects of
the drugs on the underlying electrophysiology. Table 1 provides a summary of the effects of several classes of agonists and
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SYSTEM INTERACTIONS 7684

antagonists given intrathecally in rats on acute pain behavior (as measured by thermal escape) and facilitated processing (as defined by
their effects on the second phase of the formalin test). Based on such observations, it is possible to formulate a heuristic picture of the
organization of several pharmacologically defined spinal systems that mediate the response of the animal to a strong and injurious
stimulus. Thus, repetitive afferent input increases excitatory amino acid and peptide release from primary afferents that serve to initially
depolarize dorsal horn neurons. Persistent depolarization serves to increase intracellular calcium, activating a variety of intracellular
enzymes (COX-2 and NOS) and various kinases (PKC). PGs and NO are released spinally and serve to acutely enhance the subsequent
release of afferent peptides and glutamate. Activation of local kinases serves to phosphorylate membrane receptors and channels. As an
example, the NMDA receptor when phosphorylated displays an enhanced calcium flux (see Fig. 3 ). The role of these system-level
changes in spinal nociceptive processing in pain behavior is supported by the analgesic effects of spinally delivered agents known to
reduce small afferent transmitter release (, opioid, and 2 adrenergic agonists) and the antihyperalgesic actions of spinally delivered
NK-1 and NMDA receptor antagonists, as well as inhibitors of spinal COX-2, NOS, and PKC.
Table 1. Spinal drug action in nociceptive processing in animals models and human pain states
Agonists Rat acute thermal escape Rat phase 2 formalin Human pain states Reference
agonist + + Morphine (16)
agonist + + DADL (16, 49)
2 agonist + + Clonidine (17, 50)
Aden A-1 agonist +/ + R-PIA/adenosine (5153)
GABA-A/B agonist +/ + Baclofen* (54, 55)
GABAPENTIN 0 + (56)
NMDA antagonist 0 + Ketamine/CPP (25, 57, 58)
AMPA antagonist + + (24, 59)
Metab Glu-antagonist +/ + (60)
NK-1 agonist 0 + (20)
COX inhibitor 0 + Lysine acetylsalicylate (23, 61)
EP-antagonist 0 + (34)
NOS inhibitor 0 + (37)
AChase inhibitor (mus) + + Neostigmine (6266)
N-Ca Ch blocker 0 + Ziconotide (SNX-111) (67, 68)

GABA, -aminobutyric acid; DADL, d-Ala2-d-Leu5-encephalin; R-PIA, R(-)N6-(2-phenylisopropyl) adenosine; CPP, ()-3-(2-
carboxypiperazin-4-yl)-propyl-1-phosphonic acid; EP, PG type E.
*Used intrathecally for spasticity.

Several additional points should be considered in interpreting these data. The above comments are limited to several specific
components of dorsal horn biology. Other systems that no doubt play an important role in spinal nociceptive processing, such as the
purinergic receptors ( 69 ) and the metabotrophic glutamate receptors ( 70 ) are not considered. In each case, the effects of manipulations
associated with a single system, e.g., NK-1, glutamate, or AMPA, are considered. In the case

FIG. 5. Graph plots the relative intrathecal potency of several opioids as defined in rats on the hot plate test versus the relative
potency when given epidurally (EP) or intrathecally (IT) in human postoperative or cancer pain. The axis plot the potency of the
agent (in g for rat or mg in human) relative to the potency of morphine given in that model. These calculations are based on a
standard analgesic dose for intrathecal morphine in rats on the hot plate (10 g) and after intrathecal (0.5 mg) or epidural (3 mg)
delivery for pain in humans. Human data are based on reported doses necessary to produce an adequate clinical analgesia (data
derived from ref. 16 ). DADL, d-Ala2-d-Leu5-encephalin.
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

HUMAN SPINAL PROCESSING 7685

of the primary afferent, terminals are known to routinely contain and likely release combinations of amino acids (glutamate), peptides
(CGRP, sP, vasoactive intestinal polypeptide; ref. 71 ) and peptides (such as growth factors, ref. 72 ). The net combination of these drug
effects are poorly studied. In the case of sP and NMDA, both contribute to the postsynaptic action ( 73 ). This observed excitation is
consistent with (i) the ability of either agent to activate the second-order neurons, (ii) a depolarization by sP serving to remove the Mg2+
block, and (iii) sP activating by local kinases to phosphorylate the NMDA channel. Agents that block the opening of Ca2+ channels in
primary afferents likely block all transmitter release from that terminal, which accounts for the potent antinociception that is produced by
these agents in contrast to that produced by antagonists for specific receptors (e.g., NK-1 or glutamate). The present comments focus
primarily on the events that occur in the interval around the injury period. Over extended intervals of hours to days there is an up-
regulation of receptors (NK-1) ( 74 ) and enzymes [COX ( 75 ) and NOS ( 76 )], leading to additional changes in system function. The
evidence presented here clearly reflects the functional complexity of the events that occur secondary to a focal injury, leading to a
persistent small afferent barrage. The fact that such stimuli will lead to a local 1 hyperalgesia and 2 tactile allodynia raises the likelihood
that specific components of the post-tissue injury pain state may have distinct components. Thus, as noted, after a mild, local tissue injury
a 1 hyperalgesia and a 2 tactile allodynia are noted. Treatment with spinal opiates during the injury phase will prevent the appearance of
the allodynia, but not the hyperalgesia. This finding suggests that the allodynia after an acute injury depends on a cascade that is initiated,
but not sustained, by the injury stimulus. In contrast, the hyperalgesia does not appear to depend on that cascade to be made manifest. Such
differences may reflect on the clinical phenomena of preemptive analgesia ( 77 ). In pre-emptive analgesia the patient receiving opiates
during the surgery is hypothesized to require less analgesic postoperatively. To the extent that the postoperative pain state reflects the
allodynic component noted here, that would indeed be true. To the extent that the postoperative state involves a hyperalgesic mechanisms,
the differences produced by intraoperative opiates might be slight. Finally, the early discussions on the events that occur during the periods
immediately after injury focused on the phenomena as if it were a unitary phenomena. The initial observations, for example, that
demonstrated that dorsal horn wind-up and several inflammatory models all were diminished by spinal NMDA receptor antagonists
supported such homogeneity. It is now clear that variations in mechanism can be defined even with acute injury stimulus conditions. Thus,
on examining the allodynia observed after the healing of a skin incision ( 59 ) or the hyperalgesia induced by a local burn (N.N.-T. and
T.L.Y., unpublished observations), the hyperpathia was noted to be poorly diminished by NMDA receptor antagonists.
HUMAN SPINAL PROCESSING
Although the above work is of importance in defining the biology of spinal processing in the mechanistic sense, such preclinical
insights also appear to be relevant to our understanding of spinal system function in humans. Two points can be made: (i) comparability of
the behavioral components and (ii) parallels in pharmacological activity.
Comparability of Behavioral Pain Components in Humans and Animal Models. As in the preclinical models, after focal tissue
injury (whether experimental or pathological) in humans, there is a clearly defined 1 hyperalgesia and 2 off-site tactile allodynia ( 78 )
Though as yet poorly studied, it is clear that the postoperative or postinjury pain state in humans possess the same complexity (see refs. 79
and 80 ). Still, the typical postoperative pain evaluation typically is limited to a univariate assessment (e.g., visual analogue score or
postoperative narcotic consumption). Although clinically practical, such limited surveys may well obscure the benefits or actions of a drug
that influences one of the components of the pain state.
Comparability of Spinal Pharmacology in Humans and Animal Models. The pharmacology and activity of drug effects at the
spinal level as defined in rodent systems have been shown to be extraordinarily predictive of the activity in human pain states. The best
evaluated pharmacology is that of the opiates that have been widely examined in both humans and animals. As presented in Fig. 5 ,
plotting the spinal potency of such agents relative to morphine in rats (intrathecal) and humans (epidural or intrathecal) reveals a high
correlation. More importantly, a variety of nonopioid agents have been delivered intrathecally or epidurally in animal models and then in
humans. Table 1 summarizes such work in which humans have received the respective novel class of agent. Importantly, it should be
noted that agents, which, unlike opiates, have little effect on acute pain behavior (e.g., thermal escape) are indeed active in human clinical
pain states.
These results jointly (i) support the importance of spinal processing in human pain states, (ii) emphasize the functional and
pharmacological comparability of these systems across species, and (iii) provide an important source of targets for the development of
novel clinically relevant analgesics processing can be expected to yield even greater rewards.
1. Yaksh, T. L. , Lynch III, C , Zapol, W. M. , Maze, M. , Biebuyck, J.F. & Saidman, L. J. , eds. ( 1997 ) Anesthesia: Biologic Foundations ( Lippincott ,
Philadelphia ), pp. 685718 .
2. Dirig, D. M. & Yaksh, T. L. ( 1995 ) Pain 62 , 321328 .
3. Handwerker, H. O. ( 1991 ) Agents Actions 32 , 9199 .
4. Messlinger, K. ( 1997 ) Anaesthesist 46 , 142153 .
5. Dickenson, A. H. , Stanfa, L. C. , Chapman, V. & Yaksh, T. L. ( 1997 ) in Anesthesia: Biologic Foundations , eds. Yaksh, T. L. , Lynch III, C. ,
Zapol, W. M. , Maze, M. , Biebuyck, J. F. & Saidman, L. J. ( Lippincott , Philadelphia ), pp. 611624 .
6. Puig, S. & Sorkin L. S. ( 1996 ) Pain 64 , 345355 .
7. Dickenson, A. H. & Sullivan, A. F. ( 1987 ) Pain 30 , 349360 .
8. Nozaki-Taguchi, N. & Yaksh, T. L. ( 1998 ) Neurosci. Lett. 254 , 2528 .
9. Simone, D. A. , Sorkin, L. S. , Oh, U. , Chung, J. M. , Owens, C. , LaMotte, R. H. , Willis, W. D. ( 1991 ) J. Neurophysiol. 66 , 228246 .
10. Dirig, D. M. , Salami, A. , Rathbun, M. L. , Ozaki, G. T. & Yaksh, T. L. ( 1997 ) J. Neurosci. Methods 76 , 183191 .
11. Yashpal, K. , Katz, J. & Coderre, T. J. ( 1996 ) Anesthesiology 84 , 11191128 .
12. Buerkle, H. , Marsala, M. & Yaksh, T. L. ( 1998 ) Br. J. Anaesth . 80 , 348353 .
13. Taylor, B. K. , Peterson, M. A. & Basbaum, A. I. ( 1995 ) J. Neurosci. 15 , 75757584 .
14. Wilcox, G. L. & Seybold, V. ( 1997 ) in Anesthesia: Biologic Foundations , eds. Yaksh, T. L. , Lynch III, C. , Zapol, W. M. , Maze, M. , Biebuyck,
J. F. & Saidman, L. J. ( Lippincott , Philadelphia ), pp. 557576 .
15. Yaksh, T. L. & Rudy, T. A. ( 1976 ) Physiol. Behav. 17 , 10311036 .
16. Yaksh, T. L. ( 1993 ) in Handbook of Experimental Pharmacology , ed. Herz A. ( Springer , Berlin ), Vol. 104/II , pp. 5390 .
17. Yaksh, T. L. , Jage, J. & Takano, Y. ( 1993 ) Baillires Clin. Anaesthesiol. 7 , 597614 .
18. Hua, X.-Y , Boublik, J. H. , Spicer, M. A. , Rivier, J. E. , Brown, M. R. & Yaksh, T. L. ( 1991 ) J. Pharmacol. Exp. Ther. 258 , 243258 .
the authoritative version for attribution.

19. Yamamoto, T. & Yaksh, T. L. ( 1991 ) Life Sci. 49 , 19551963 .

20. Hua, X.-Y. , Chen, P. , Polgar, E. , Nagy, L , Marsala, M. , Phillips, E. , Wollaston, L. , Urban, L. , Yaksh, T. L. & Webb, M. ( 1998 ) J. Neurochem.
70 , 688698 .
21. Yang, L. C. , Marsala, M. & Yaksh, T. L. ( 1996 ) Pain 67 , 345354 .
22. Hua, X.-Y. , Chen, P. , Marsala, M. & Yaksh, T. L. ( 1999 ) Neuroscience 89 , 525534 .
23. Malmberg, A. B. & Yaksh, T. L. ( 1992 ) Science 257 , 12761279 .
24. Nishiyama, T. , Yaksh, T. L. & Weber, E. ( 1998 ) Anesthesiology 89 , 715722 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

HUMAN SPINAL PROCESSING 7686

25. Chaplan, S. R. , Malmberg, A. B. & Yaksh, T. L. ( 1997 ) J. Pharmacol. Exp. Ther. 280 , 829838 .
26. Yamamoto, T. & Yaksh, T. L. ( 1992 ) Anesthesiology 77 , 757763 .
27. Marsala, M. , Malmberg, A. B. & Yaksh, T. L. ( 1995 ) J. Neurosci. Methods 62 , 4353 .
28. Malmberg, A. B. & Yaksh, T. L. ( 1995 ) J. Neurosci. 15 , 768776 .
29. Yang, L. C , Marsala, M. & Yaksh, T. L. ( 1996 ) Neuroscience 75 , 453461 .
30. Coleman, R. A. , Smith, W. L. & Narumiya, S. ( 1994 ) Pharmacol. Rev. 46 , 205229 .
31. Hingtgen, C. M. & Vasko, M. R. ( 1994 ) Brain Res. 655 , 5160 .
32. Minami, T. , Uda, R. , Horiguchi, S. , Ito, S. , Hyodo, M. & Hayaishi, O. ( 1994 ) Pain 57 , 217223 .
33. Malmberg, A. B. & Yaksh, T. L. ( 1992 ) J. Pharmacol. Exp. Ther. 263 , 264275 .
34. Malmberg, A. B. , Rafferty, M. F. & Yaksh, T. L. ( 1994 ) Neurosci. Lett. 173 , 193196 .
35. Seibert, K. , Zhang, Y. , Leahy, K. , Hauser, S. , Masferrer, J. , Perkins, W. , Lee, L. & Isakson, P. ( 1994 ) Proc. Natl. Acad. Sci. USA 91 , 12013
12017 .
36. Yamamoto, T. & Nozaki-Taguchi, N. ( 1997 ) NeuroReport 8 , 21792182 .
37. Malmberg, A. B. & Yaksh, T. L. ( 1993 ) Pain 54 , 291300 .
38. Ghosh, A. & Greenberg, M. E. ( 1995 ) Science 268 , 239247 .
39. Malmberg, A. B. , Brandon, E. P. , Idzerda, R. L. , Liu, H. , McKnight, G. S. & Basbaum, A. I. ( 1997 ) J. Neurosci. 17 , 74627470 .
40. Kocsis, J. D. , Rand, M. N. , Lankford, K. L. & Waxman, S. G. ( 1994 ) J. Neurobiol. 25 , 252264 .
41. Roberts, R. E. & McLean, W. G. ( 1997 ) Brain Res. 754 , 147156 .
42. Malmberg, A. B. , Chen, C. , Tonegawa, S. & Basbaum, A. I. ( 1997 ) Science 278 , 279283 .
43. Leonard, A. S. & Hell, J. W. ( 1997 ) J. Biol. Chem. 272 , 12107 12115 .
44. Munro, F. E. , Fleetwood-Walker, S. M. & Mitchell, R. ( 1994 ) Neurosci. Lett. 170 , 199202 .
45. Sluka, K. A. & Willis, W. D. ( 1997 ) Pain 71 , 165178 .
46. Cerne, R. , Rusin, K. I. & Randic, M. ( 1993 ) Neurosci. Lett. 161 , 124128 .
47. Coderre, T. J. & Yashpal, K. ( 1994 ) Eur. J. Neurosci. 6 , 1328 1334 .
48. Yashpal, K. , Pitcher, G. M. , Parent, A. , Quirion, R. & Coderre, T. J. ( 1995 ) J. Neurosci. 15 , 32633272 .
49. Onofrio, B. M. & Yaksh, T. L. ( 1983 ) Lancet 1 , 13861387 .
50. Eisenach, J. C. , De Kock, M. & Klimscha, W. ( 1996 ) Anesthesiology 85 , 655674 .
51. Poon, A & Sawynok, J. ( 1998 ) Pain 74 , 235245 .
52. Karlsten, R. & Gordh, T. , Jr. ( 1995 ) Anesth. Analg. 80 , 844847 .
53. Rane, K. , Segerdahl, M. , Goiny, M. & Sollevi, A. ( 1998 ) Anesthesiology 5 , 11081115 .
54. Dirig, D. M. & Yaksh, T. L. ( 1995 ) Pharmacol. Exp. Ther. 275 , 219227 .
55. Porter, B. ( 1997 ) Br. J. Nursing 6 , 253260 .
56. Partridge, B. J. , Chaplan, S. R. , Sakamoto, E. & Yaksh, T. L. ( 1998 ) Anesthesiology 88 , 196205 .
57. Abdel-Ghaffar, M. E. , Abdulatif, M. A. , al-Ghamdi, A. , Mowafi, H. & Anwar, A. ( 1998 ) Can. J. Anaes. 45 , 103109 .
58. Kristensen, J. D. , Svensson, B. & Gordh, T. , Jr. ( 1992 ) Pain 51 , 249253 .
59. Brennan, T. J. ( 1998 ) Anesthesiology 89 , 10491051 .
60. Fisher, K. & Coderre, T. J. ( 1996 ) Pain 68 , 255263 .
61. Devoghel, J. C. ( 1983 ) J. Int. Med. Res. 11 , 9091 .
62. Prado, W. A. & Goncalves, A. S. ( 1997 ) Braz. J. Med. Biol. Res. 30 , 12251231 .
63. Naguib, M. & Yaksh, T. L. ( 1994 ) Anesthesiology 80 , 13381348 .
64. Naguib, M. & Yaksh, T. L. ( 1997 ) Anesth. Analg. 85 , 847853 .
65. Hood, D. D. , Eisenach, J. C. & Tuttle, R. ( 1995 ) Anesthesiology 82 , 331343 .
66. Lauretti, G. R. & Lima, I. C. ( 1996 ) Anest. Analg. 82 , 617620 .
67. Malmberg, A. B. & Yaksh, T. L. ( 1994 ) J. Neurosci. 14 , 4882 4890 .
68. Brose, W. G. , Gutlove, D. P. , Luther, R. R. , Bowersox, S. S. & McGuire, D. ( 1997 ) Clin. J. Pain 13 , 256259 .
69. Driessen, B. , Reimann, W. , Selve, N. , Friderichs, E. & Bultmann, R. ( 1994 ) Brain Res. 666 , 182188 .
70. Young, M. R. , Blackburn-Munro, G. , Dickinson, T. , Johnson, M. J. , Anderson, H. , Nakalembe, I. & Fleetwood-Walker, S. M. ( 1998 ) J.
Neurosci. 18 , 1018010188 .
71. Levine, J. D. , Fields, H. L. & Basbaum, A. I. ( 1993 ) J. Neurosci. 13 , 22732286 .
72. Michael, G. J. , Averill, S. , Nitkunan, A. , Rattray, M. , Bennett, D. L. , Yan, Q. & Priestley, J. V. ( 1997 ) J. Neurosci. 17 , 84768490 .
73. Chapman, V. , Buritova, J. , Honore, P. & Besson, J. M. ( 1996 ) J. Neurophysiol. 76 , 18171827 .
74. Kar, S. , Rees, R. G. & Quirion, R. ( 1994 ) Eur. J. Neurosci. 6 , 345354 .
75. Beiche, F. , Scheuerer, S. , Brune, K. , Geisslinger, G. & Goppelt-Struebe, M. ( 1996 ) FEBS Lett. 390 , 165169 .
76. Wu, J. , Lin, Q , Lu, Y. , Willis, W. D. & Westlund, K. N. ( 1998 ) Exp. Brain Res. 118 , 457465 .
77. McQuay, H. J. ( 1995 ) Ann. Med. 27 , 249256 .
78. Handwerker, H. O. & Kobal, G. ( 1993 ) Physiol. Rev. 73 , 639671 .
79. OConnor, T. C. & Abram, S. E. ( 1997 ) in Anesthesia: Biologic Foundations, eds. Yaksh, T. L. , Lynch III, C. , Zapol, W. M. , Maze, M. ,
Biebuyck, J. F. & Saidman, L. J. ( Lippincott , Philadelphia ), pp. 747758 .
80. Silbert, B. S. , Osgoodm P. F. & Carr, D. B. ( 1997 ) in Anesthesia: Biologic Foundations , eds. Yaksh, T. L. , Lynch III, C. , Zapol, W. M. , Maze,
M. , Biebuyck, J. F. & Saidman, L. J. ( Lippincott , Philadelphia ), pp. 759773 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SUPRASPINAL CONTRIBUTIONS TO HYPERALGESIA 7687

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Supraspinal contributions to hyperalgesia

M. O. URBAN AND G. F. GEBHART *

Department of Pharmacology, College of Medicine, University of Iowa, Iowa City, IA 52242
ABSTRACT Tissue injury is associated with sensitization of nociceptors and subsequent changes in the excitability of central
(spinal) neurons, termed central sensitization. Nociceptor sensitization and central sensitization are considered to underlie,
respectively, development of primary hyperalgesia and secondary hyperalgesia. Because central sensitization is considered to
reflect plasticity at spinal synapses, the spinal cord has been the principal focus of studies of mechanisms of hyperalgesia. Not
surprisingly, glutamate, acting at a spinal N-methyl-D-aspartate (NMDA) receptor, has been implicated in development of
secondary hyperalgesia associated with somatic, neural, and visceral structures. Downstream of NMDA receptor activation, spinal
nitric oxide (NO ), protein kinase C, and other mediators have been implicated in maintaining such hyperalgesia. Accumulating
evidence, however, reveals a significant contribution of supraspinal influences to development and maintenance of hyperalgesia.
Spinal cord transection prevents development of secondary, but not primary, mechanical and/or thermal hyperalgesia after
topical mustard oil application, carrageenan inflammation, or nerveroot ligation. Similarly, inactivation of the rostral
ventromedial medulla (RVM) attenuates hyperalgesia and central sensitization in several models of persistent pain. Inhibition of
medullary NMDA receptors or NO generation attenuates somatic and visceral hyperalgesia. In support, topical mustard oil
application or colonic inflammation increases expression of NO synthase in the RVM. These data suggest a prominent role for
the RVM in mediating the sensitization of spinal neurons and development of secondary hyperalgesia. Results to date suggest that
peripheral injury and persistent input engage spinobulbospinal mechanisms that may be the prepotent contributors to central
sensitization and development of secondary hyperalgesia.
Hardy et al. ( 1 ) investigated two types of experimentally produced cutaneous hyperalgesia, primary and secondary. Primary
hyperalgesia occurs at the site of injury; secondary hyperalgesia is associated with the injury, but occurs in undamaged tissues adjacent to
and at some distance from the site of an injury. They proposed a new formulation to explain the spread of hyperalgesia away from the
site of injury, namely that a central (spinal) excitatory state, and not a peripheral mechanism as advanced by Lewis ( 2 ), was responsible
for secondary hyperalgesia. Subsequent intensive study of the altered sensations that arise from and adjacent to injured tissues has
supported this formulation and it is now widely accepted that mechanisms of primary and secondary hyperalgesia are, respectively,
peripheral and central (e.g., see refs. 3 , 4 ).
The increase in excitability of spinal neurons after peripheral injury, termed central sensitization, has been extensively studied by
Woolf and colleagues (see ref. 5 for overview). They documented that the enhanced reflex excitability after peripheral tissue damage did
not require ongoing peripheral input, and that spinal dorsal horn neuron receptive fields expanded, responsiveness to suprathreshold
stimuli increased, response thresholds decreased, and sensitivity to novel stimuli was acquired after peripheral injury. The focus of
investigation has remained the spinal cord, and many investigators have since documented the importance of the spinal N-methyl-
Daspartate (NMDA) receptor to the induction and maintenance of central sensitization (see ref. 6 for recent overview). A growing body of
evidence, however, reveals a significant contribution of descending influences from supraspinal sites in the development and maintenance
of central sensitization/ secondary hyperalgesia. We review here and discuss evidence that peripheral tissue injury engages
spinobulbospinal circuitry that may be important to the development and maintenance of central sensitization and secondary hyperalgesia.
Descending Facilitation. Although the potency of descending inhibitory influences has long been appreciated, the study and
characterization of descending facilitatory influences have been more recent developments. Interestingly, inhibitory and facilitatory
influences can be produced at many of the same sites in the brainstem, particularly in the rostral ventromedial medulla (RVM). Generally,
low intensities of electrical stimulation or low concentrations of chemical (e.g., glutamate, neurotensin) facilitate spinal nociception,
whereas greater intensities of stimulation or concentrations of chemical at the same sites typically inhibit spinal nociception ( 7 10 ).
These dual influences appear to involve anatomically distinct independent spinal pathways and are mediated by different lumbar spinal
receptors. For example, high-intensity electrical stimulation or high-dose glutamate or neurotensin injection into the RVM inhibits spinal
nociceptive transmission via descending projections in the dorsolateral funiculi and activation of spinal cholinergic and monoaminergic
receptors. In contrast, facilitatory influences from the RVM produced by electrical stimulation, glutamate injection, or neurotensin
injection involve descending projections in the ventrolateral funiculi and are mediated by spinal serotonin and cholecystokinin receptors. (
7 , 9 , 11 13 ).
In addition to the RVM, adjacent medullary sites also have been implicated in descending facilitation of spinal nociceptive
transmission. Electrical and/or selective chemical stimulation in these areas have been shown to enhance spinal behavioral and dorsal horn
neuron responses to noxious stimulation ( 14 ).
Fields et al. ( 15 ) have characterized cells in the RVM that may constitute the physiological basis for generation of bidirectional
modulation of spinal nociceptive transmission. They have operationally defined three classes of neurons in the RVM: on-cells, off-cells,
and neutral cells, which are intermixed in the RVM and not anatomically separable. Off-cells
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: NMDA, N-methyl-D-aspartate; RVM, rostral ventromedial medulla; NO , nitric oxide; LPS, lipopolysaccharide; NTS,
nucleus tractus solitarius; APV, 2-amino-5-phosphonovaleric acid.
* To whom reprint requests should be addressed, e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SUPRASPINAL CONTRIBUTIONS TO HYPERALGESIA 7688

display an abrupt pause in ongoing activity immediately before nociceptive reflexes and are proposed to contribute to inhibitory influences
that descend from the RVM. On-cells display a burst of activity immediately before nociceptive reflexes and are proposed to contribute to
facilitatory influences that descend from the RVM. Neutral cells show no nociception-related change in activity. Off-cells, on-cells, and
neutral cells all project to the spinal dorsal horn ( 16 ), placing on-cell and off-cell terminals in appropriate laminae (I, II, and V) to
modulate nociceptive transmission. That on- and off-cells mediate descending facilitatory and inhibitory influences from the RVM is
supported by several reports demonstrating enhanced on- or off-cell activity during facilitation or inhibition of spinal nociceptive
transmission, respectively ( 17 19 ).
We hypothesize that there exists a spinobulbospinal circuit that contributes significantly to central sensitization and secondary
hyperalgesia. Anatomically, this circuit is in place. Both the RVM and adjacent areas receive direct afferent input from the superficial
spinal dorsal horn and in turn send descending projections through spinal funiculi that terminate in the superficial dorsal horn, completing
a spinobulbospinal loop ( 20 23 ). We review below recent studies that document that spinal transection, or inactivation of supraspinal
sites, prevents the expression of secondary hyperalgesia in a variety of animal models of persistent inflammatory, neurogenic, or
neuropathic pain, thus providing the functional context in support of the anatomy (see Table 1 ).
Inflammatory/Neurogenic Models of Hyperalgesia. Mustard oil. Mustard oil (allyl isothiocyanate) is a chemical irritant that
produces a neurogenic inflammation and excites chemosensitive C-fibers, resulting in behavioral hyperalgesia and central sensitization (
24 , 25 ). An involvement of supraspinal sites in mustard oil-induced sensitization was reported by Mansikka and Pertovaara ( 26 ), who
found that tactile allodynia of the glabrous skin of the foot after topical application of mustard oil to the ankle was prevented in animals
that had received spinal transection. Additionally, in spinally intact rats, the tactile allodynia was blocked after inactivation of the medial
RVM by local lidocaine microinjection. The authors concluded that persistent nociceptor stimulation by topical mustard oil activates a
positive feedback loop involving descending facilitatory influences from the RVM. In an electrophysiological study of spinal cord
neurons, Pertovaara ( 27 ) subsequently reported that midthoracic spinal transection or lidocaine inactivation of the RVM blocked mustard
oil-induced enhanced excitability of wide dynamic range neurons to mechanical stimulation. In these experiments, mustard oil was applied
12 cm outside the border of the receptive field of the spinal neuron. Thus, in both studies, the allodynia/ hyperalgesia was tested at a site
distant from the site of application of mustard oil (i.e., it was secondary in nature).
Table 1. Summary of supraspinal contributions to hyperalgesia
Model of hyperalgesia Nociceptive response Manipulation Effect Ref.
Inflammation/neurogenic
Mustard oil (ankle) Tactile allodynia, foot Spinal transection Intra-RVM Block 26
lidocaine
Mustard oil (foot, outside receptive Enhanced excitability of WDR Spinal transection Intra-RVM Block 27
field) dorsal horn neurons lidocaine
Mustard oil (leg) Facilitation of the thermal tail- Spinal transection Electrolytic Block 28, 29
flick reflex RVM lesion Ibotenic acid RVM
lesion
Carrageenan (knee joint) Enhanced C-fiber-mediated Spinal transection Block 37
flexor motoneuron wind-up
Carrageenan (knee joint) Facilitation of the thermal paw- Intra-RVM lidocaine Ibotenic Block 29
withdrawal response acid RVM lesion
Carrageenan (plantar foot) Facilitation of the thermal paw- Intra-RVM lidocaine Ibotenic No effect 29
withdrawal response acid RVM lesion
Formalin (foot) Facilitation of the thermal tail- Spinal transection Electrolytic block 41, 42
flick reflex RVM lesion
Neuropathic
Spinal nerve ligation Tactile allodynia, foot Intra-RVM lidocaine Block 44
Spinal nerve ligation Tactile allodynia, foot Spinal transection Block 45
Facilitation of the thermal paw-
withdrawal response
Spinal nerve cut Tactile allodynia, foot Spinal transection Block 46
Illness
LPS (intraperitoneal) Facilitation of the thermal tail- Electrolytic RVM lesion Block 42, 50
flick reflex Electrolytic NTS lesion
In related studies, we documented a significant contribution of descending facilitatory influences in a model of thermal hyperalgesia
involving topical application of mustard oil to the hind leg and measurement of the spinal nociceptive tail-flick reflex ( 28 ). It was found
that midthoracic spinal transection or electrolytic lesion of the RVM prevented facilitation of the tail-flick reflex produced by mustard oil.
To confirm an involvement of cells in the RVM in modulating this secondary thermal hyperalgesia, we found that RVM lesion using the
soma-selective neurotoxin ibotenic acid resulted in a similar block of mustard oil-induced hyperalgesia ( 29 ).
Active participation of descending facilitatory influences from the RVM in modulating mustard oil-induced hyperalgesia is supported
further by evidence that NMDA and neurotensin receptors in the RVM modulate this secondary thermal hyperalgesia. As indicated above,
neurotensin receptors ( 7 , 8 ) and NMDA receptors ( 30 , 31 ) in the RVM have been implicated in descending facilitation of spinal
nociception. Selective blockage of these receptors should then modulate hyperalgesia. Indeed, intra-RVM injection of a selective
neurotensin receptor antagonist (SR48692) or NMDA receptor antagonist [2-amino-5-phosphonovaleric acid (APV)] fully and dose
dependently prevented mustard oil-induced facilitation of the tail-flick reflex ( 28 , 30 ). It is known that generation of nitric oxide (NO )
is one downstream consequence of NMDA receptor activation ( 32 ). In complementary studies, we showed that intra-RVM administration
the authoritative version for attribution.

of the NO -synthase inhibitor

lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SUPRASPINAL CONTRIBUTIONS TO HYPERALGESIA 7689

N-nitro-L-arginine methyl ester (L-NAME), like the NMDA receptor antagonist APV, attenuated mustard oil-induced hyperalgesia (
30 ). Conversely, microinjection of the NO donor GEA 5024 (or of NMDA itself) dose dependently facilitated the tail-flick reflex in
nave rats. The involvement of NO in the RVM was further supported by a significant increase in the number of NADPHdiaphorase-
labeled cells at the time of maximal mustard oil-induced hyperalgesia. Finally, in a model of visceral hyperalgesia in which the
inflammogen zymosan is instilled into the colon, both APV and L-NAME given into the RVM 3 hr after colonic inflammation reversed
the hyperalgesia for the duration of drug action, suggesting that the RVM plays a role in maintenance of the hyperalgesia ( 31 ). Similar to
what was seen in the model of mustard oil hyperalgesia, both NADPHdiaphorase-labeled cell numbers and the number of cells
immunostained for the neuronal isoform of NO -synthase were significantly increased in the RVM 3 hr after colonic inflammation. These
results support a role for descending facilitatory influences in the maintenance of mustard oil-induced and visceral hyperalgesia involving
activation of NMDA and neurotensin receptors in the RVM.
Carrageenan. Several models of hyperalgesia involving subcutaneous injection of carrageenan have been characterized. Carrageenan
is a water-extractable polysaccharide obtained from various seaweeds. Injection of lambda carrageenan (a hydrocolloid that does not form
a gel) into the plantar foot, or intraarticular injection into the knee joint, results in a localized inflammation, decreased weight bearing,
guarding of the affected limb, and hyperalgesia (e.g., refs. 33 and 34 ). Carrageenan-induced hyperalgesia is believed to occur as a
consequence of sensitization of primary afferent nociceptors and neuron plasticity intrinsic to the spinal cord ( 35 , 36 ).
Herrero and Cervero ( 37 ) first reported that the A- and C-fiber mediated wind-up of flexor motoneurons after intraarticular (knee)
carrageenan injection was prevented by spinal transection. They concluded that supraspinal modulatory systems, either direct excitatory
influences on spinal neurons or release of local inhibitory controls, are essential for wind-up. We examined a potential contribution of
descending facilitatory influences from the RVM to enhanced behavioral nociceptive responses after intraplantar or intraarticular (knee)
injection of carrageenan ( 29 ). Intraplantar injection of carrageenan and subsequent thermal stimulation of the plantar surface of the
hindpaw is a model of primary hyperalgesia; intraarticular injection of carrageenan and subsequent thermal stimulation of the plantar
surface of the hindpaw is a model of secondary hyperalgesia. Inactivation of the RVM by lidocaine microinjection reversed, and prior
permanent inactivation of the RVM by ibotenic acid lesion completely blocked, facilitation of the thermal paw-withdrawal response after
intraarticular carrageenan injection. RVM inactivation by either lidocaine or ibotenic acid was ineffective, however, in preventing thermal
hyperalgesia after intraplantar carrageenan injection (i.e., model of primary hyperalgesia). These results suggest that these two models of
carrageenan-induced thermal hyperalgesia are differentially modulated in the central nervous system. Additionally, similar to mustard oil-
induced secondary hyperalgesia, intra-RVM injection of a selective neurotensin receptor antagonist (SR48692) or NMDA receptor
antagonist (APV) was found to block facilitation of the thermal pawwithdrawal response after intraarticular, but not intraplantar,
carrageenan injection ( Fig. 1 ). These results further support a contribution of descending facilitatory influences to secondary hyperalgesia
that is mediated by neurotensin and NMDA receptors in the RVM.
Formalin. Subcutaneous injection of formalin into the dorsum of the rodent hindpaw is a well characterized model in which animals
exhibit spontaneous pain behaviors (shaking, licking of the injected hindpaw) as well as hyperalgesia ( 38 , 39 ). Additionally, formalin
has been shown to produce secondary hyperalgesia after subcutaneous injection into either the hindpaw or tail ( 40 , 41 ). A significant
contribution of supraspinal sites to formalin-produced secondary hyperalgesia was reported by Wiertelak et al. ( 41 ), who found that
spinal transection prevented facilitation of the tail-flick reflex after formalin injection into the hindpaw. That activation of descending
facilitatory influences from the RVM modulates this hyperalgesia was subsequently supported by the finding that electrolytic lesion of the
RVM prevented facilitation of the tail-flick reflex after formalin injection ( 42 ).
Neuropathic Models of Hyperalgesia. Animal models of neuropathic pain generally involve loose ligation of peripheral nerves,
which results in spontaneous pain behaviors, enhanced responses of spinal-dorsal horn nociceptive neurons, and hyperalgesia (for review,
see ref. 43 ). A contribution of supraspinal sites to neuropathic pain after spinal nerve ligation was initially reported by Pertovaara et al. (
44 ). In that study, the tactile allodynia that develops after unilateral ligation of the L5 and L6 spinal nerves was found to be attenuated by
inactivation of the RVM by lidocaine injection. The lidocaine effect was determined to be localized within the RVM and independent of
an opioid mechanism, suggesting an inactivation of a descending facilitatory influence from the RVM. These results were supported in a
subsequent study ( 45 ), in which spinal transection was found to abolish the tactile allodynia as well as thermal hyperalgesia produced by
ligation of the L5 and L6 spinal nerves. Additionally, Kauppila ( 46 ) found spinal transection to block mechanical hyperalgesia observed
after a chronic sciatic nerve cut. Thus, neuropathic pain after peripheral nerve injury appears to involve, at least in part, activation of
descending facilitatory influences from supraspinal sites, including the RVM. ropcapli9
Illness-Induced Models of Hyperalgesia. The systemic administration of lipopolysaccharide (LPS) has been shown to produce a
number of symptoms associated with illness, such as fever, lethargy, decreased food and water intake, and increased sleep (for review, see
ref. 47 ). Additionally, administration of LPS produces hyperalgesia through the release of peripheral cytokines (e.g., IL-1) from immune
cells ( 48 , 49 ). In a series of experiments, Watkins et al. ( 49 , 50 ) determined that facilitation of the tail-flick reflex after intraperitoneal
injection of LPS does not involve primary afferent nociceptor input to the spinal dorsal horn. Instead, a novel circuit was proposed
involving IL-1 activation of hepatic vagal afferent fibers that terminate in the nucleus tractus solitarius (NTS). Consistent with this
proposal, electrolytic lesion of the NTS or RVM was found to block facilitation of the tail-flick reflex produced by intraperitoneal LPS.
Because the NTS and RVM are reciprocally connected, direct afferent input to the RVM may mediate this effect, although Watkins et al. (
50 ) implicated an unidentified site rostral to the midmesencephalon as an important relay. This interpretation is consistent with earlier
studies of biphasic effects of electrical stimulation of vagal afferent fibers (see ref. 51 for review). In those experiments, low-intensity
stimulation of vagal afferent fibers was documented to facilitate spinal nociceptive reflexes (tail-flick reflex) and spinal dorsal horn neuron
responses to noxious stimuli. The facilitatory effect of vagal stimulation was abolished after midcollicular decerebration, implicating an
NTSforebrain circuit in descending influences that ultimately exit the brainstem via the RVM. Although the tail-flick reflex is a spinally
organized response, facilitation of this reflex after intraperitoneal LPS similarly appears to involve activation of descending facilitatory
influences from the RVM.
Primary vs. Secondary Hyperalgesia. We and others have studied the effects of spinal cord transection and of reversible (lidocaine)
or permanent (ibotenic acid) inactivation of the RVM in models of primary and secondary hyperalgesia after peripheral tissue insult. The
results reviewed above uniformly support the hypothesis that facilitatory influences from the
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SUPRASPINAL CONTRIBUTIONS TO HYPERALGESIA 7690

brainstem significantly contribute to secondary, but not primary, hyperalgesia. What has not yet been addressed specifically is whether the
RVM is necessary and sufficient for development or for maintenance of secondary hyperalgesia. Intra-RVM injection of lidocaine
reverses, in a time-limited fashion, already established secondary hyperalgesia, suggesting a clear role for the RVM in maintenance of
secondary hyperalgesia. Other studies reveal that spinal-cord transection or soma-selective lesion of the RVM prevents development of
secondary, but not primary, hyperalgesia. Accordingly, available evidence suggests that the RVM is important to both the development
and maintenance of secondary hyperalgesia. The studies reviewed here all have examined behavioral consequences of peripheral tissue
insult, and there are limited data available yet with respect to the direct influence of the RVM on spinal neuron plasticity (central
sensitization).

FIG. 1. Involvement of descending facilitatory influences from the RVM in models of secondary, but not primary, thermal
hyperalgesia after peripheral inflammation. (A) RVM lesion produced by ibotenic acid prevented facilitation of the thermal paw-
withdrawal response after intraarticular carrageenan/kaolin injection into the knee (t test, P < 0.05), but was ineffective in preventing
facilitation of the thermal paw-withdrawal response after intraplantar carrageenan injection into the foot (model of primary
hyperalgesia). (B) Intra-RVM microinjection of the NMDA receptor antagonist APV (1 pmol/1 l), or (C) intra-RVM microinjection
of the neurotensin receptor antagonist SR48692 (3 nmol/1 l) attenuated secondary, but not primary, hyperalgesia (t test, P < 0.05).
All data are represented as mean SEM of the percent change in thermal paw-withdrawal latency (%) from the control response for
the ipsilateral (inflamed) hindlimb. In experiments involving ibotenic acid RVM lesion, responses are represented at the time of
maximal hyperalgesia (3 hr after carrageenan injection). Intra-RVM microinjection of APV or SR48692 was performed at the time of
maximal hyperalgesia (3 hr), and responses are represented at the time of maximal drug effect after intra-RVM injection (10 min).
Two studies have examined changes in spinal neuron behavior associated with peripheral tissue insult. Schaible et al. ( 52 )
examined, in the cat, the effect of acute inflammation of the knee joint with a mixture of kaolin and carrageenan on spinal dorsal horn
neurons. They documented that spontaneous activity and responses to both innocuous and noxious stimulation of the joint were increased
progressively as the inflammation progressed. Neuron activity and responses to stimulation were increased further when spinal cord
transmission was interrupted temporarily by cold block of the lower thoracic spinal cord. They concluded that spinal neuron
hyperexcitability associated with a peripheral inflammation was counteracted by enhancement of descending inhibitory influences. Ren
and Dubner ( 53 ) studied, in the rat, the effect of lidocaine injection into the midline RVM on spinal neuron responses to stimulation of a
hindpaw inflamed with complete Freunds adjuvant. During the action of lidocaine, neuron spontaneous activity and responses to
mechanical and thermal stimulation applied to the hindpaw were significantly increased, which was interpreted to indicate that peripheral
inflammation leads to an enhanced descending inhibition. Both of these studies used models of primary hyperalgesia (stimuli were applied
to the injured tissue). Both also noted, however, an increase in the size of neuron receptive fields, usually taken as an indication of
secondary hyperalgesia. Although neither report directly addresses the hypothesis advanced here, both contribute relevant information.
Both document an active modulation by the brainstem of spinal neuron excitability in the presence of tissue injury, confirming activation
by peripheral noxious inputs of descending inhibition that can modulate further spinal nociceptive transmission.
The generality of the present hypothesis remains to be established. Most of the studies done to date have examined secondary thermal
the authoritative version for attribution.

hyperalgesia. Thermal hyperalgesia is widely used in studies with nonhuman animals, but secondary thermal hyperalgesia is not of
significant consequence in most instances of tissue injury in humans. The extent to which secondary mechanical hyperalgesia is modulated
by the RVM is unclear. The limited data available to date relate to tactile allodynia and mechanical hyperalgesia in models of neuropathic
pain. Additional studies that use other models of hyperalgesia are necessary. Models of chemically produced hyperalgesia, which may
involve more selective actions on different types of nociceptors, have not been studied extensively. Secondary thermal hyperalgesia
produced by topical application of the C-fiber excitant mustard oil has been documented by several investigators to be influenced by the
RVM. Whether secondary hyperalgesia produced by intradermal injection of capsaicin, which acts at the vanilloid-1 receptor, is similarly
modulated by the RVM has not been reported.
It is also unknown how blockage of central sensitization at the level of the spinal cord (by antagonism of the NMDA receptor, for
example) influences the RVM. It may be that the spinal cord and RVM are both necessary and sufficient to development and maintenance
of secondary hyperalgesia. Results reviewed here clearly indicate that central sensitization at the level of the spinal cord can be modulated
by the RVM, even if the spinal cord is the portal of first entry of the relevant
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SUPRASPINAL CONTRIBUTIONS TO HYPERALGESIA 7691

input. Temporally, input to the spinal cord likely precedes receipt of similar input in the brain stem, but it may be that other avenues of
input (e.g., via the vagus) provide an important (more important?) trigger for the RVM.

FIG. 2. Summary diagram illustrating a significant supraspinal contribution to secondary, but not primary, thermal hyperalgesia after
peripheral inflammation. Peripheral injury results in activation and sensitization of peripheral nociceptors and subsequent enhanced
excitability of dorsal horn nociceptive neurons (central sensitization) that contributes to primary hyperalgesia (at site of injury) and
secondary hyperalgesia (adjacent/distant from site of injury). Additionally, it is proposed that stimulation of nociceptors activates a
spinobulbospinal loop, engaging a centrifugal descending nociceptive facilitatory influence from the RVM. Facilitatory influences
are activated by NMDA receptors and NO , and neurotensin (NT) receptors in the RVM and descend to multiple spinal segments to
contribute significantly to secondary hyperalgesia. In contrast, primary hyperalgesia does not involve descending facilitatory
influences from supraspinal sites and is likely the direct result of peripheral nociceptor sensitization and neuroplasticity intrinsic to
the spinal cord. For clarity, the afferent input to the spinal dorsal horn from the site of injury is illustrated as not entering the spinal
cord (which it certainly does).
Returning to the formulation advanced almost 50 years ago by Hardy et al. (1), we believe that a dominant active influence from the
brainstem is necessary for the expression of secondary hyperalgesia (see Fig. 2 ). We acknowledge that there are likely multiple
supraspinal sites involved in responding to peripheral tissue insult. Indeed, the limited data available suggest that forebrain sites can play
an important role, even if the RVM is the final common pathway of facilitatory influences that mediate spinal neuron excitability.
This work was supported by National Institutes of Health awards DA11431 (M.O.U.), NS19912 (G.F.G.), and DA02879 (G.F.G.).
1. Hardy, J. D. , Wolff, H. G. & Goodell, H. ( 1950 ) J. Clin. Invest 29 , 115140 .
2. Lewis, T. ( 1936 ) Clin. Sci. 2 , 373421 .
3. Woolf, C. J. ( 1983 ) Nature (London) 306 , 686688 .
4. LaMotte, R. H. , Shain, C. N. , Simone, D. A. & Tsai, E. F. P. ( 1991 ) J. Neurophysiol. 66 , 190211 .
5. Woolf, C. J. ( 1992 ) in Hyperalgesia and Allodynia , ed. Willis, W. ( Raven , New York ) pp. 221243 .
6. Urban, M. O. & Gebhart, G. F. ( 1998 ) Prog. Brain Res. 116 , 407420 .
7. Urban, M. O. & Gebhart, G. F. ( 1997 ) J. Neurophysiol. 78 , 15501562 .
8. Urban, M. O. & Smith, D. J. ( 1993 ) J. Pharmacol. Exp. Ther. 265 , 580586 .
9. Zhuo, M. & Gebhart, G. F. ( 1992 ) J. Neurophysiol. 67 , 15991614 .
10. Zhuo, M. & Gebhart, G. F. ( 1997 ) J. Neurophysiol. 78 , 746758 .
11. Urban, M. O. , Smith, D. J. & Gebhart, G. F. ( 1996 ) J. Pharmacol. Exp. Ther. 278 , 9096 .
12. Zhuo, M. & Gebhart, G. F. ( 1990 ) Brain Res. 535 , 6778 .
13. Zhuo, M. & Gebhart, G. F. ( 1991 ) Brain Res. 550 , 3548 .
14. Almeida, A. , Tjolsen, A. , Lima, D. , Coimbra, A. & Hole, K. ( 1996 ) Brain Res. Bull. 39 , 715 .
15. Fields, H. L. , Bry, J. , Hentall, I. & Zorman, G. ( 1983 ) J. Neurosci. 3 , 25452552 .
16. Fields, H. L. , Malick, A. & Burstein, R. ( 1995 ) J. Neurophysiol. 74 , 17421759 .
17. Fields, H. L. , Vanegas, H. , Hentall, I. D. & Zorman, G. ( 1983 ) Nature (London) 306 , 684686 .
18. Bederson, J. B. , Fields, H. L. & Barbaro, N. M. ( 1990 ) Somatosens. Res. 7 , 185203 .
19. Morgan, M. M. & Fields, H. L. ( 1994 ) J. Neurophysiol. 72 , 11611170 .
20. Almeida, A. , Tavares, L , Lima, D. & Coimbra, A. ( 1993 ) Neuroscience 55 , 10931106 .
21. Basbaum, A. I. , Clanton, C. H. & Fields, H. L. ( 1978 ) J. Comp. Neurol. 178 , 209224 .
22. Craig, A. D. ( 1995 ) J. Comp. Neurol. 361 , 225248 .
23. Martin, G. F. , Vertes, R. P. & Waltzer, R. ( 1985 ) Exp. Brain Res. 58 , 154162 .
24. Woolf, C. J. & Wall, P. D. ( 1986 ) J. Neurosci. 6 , 14331442 .
25. Woolf, C. J. , Shortland, P. & Sivilotti, L. G. ( 1994 ) Pain 58 , 141155 .
the authoritative version for attribution.

26. Mansikka, H. & Pertovaara, A. ( 1997 ) Brain Res. Bull. 42 , 359365 .

27. Pertovaara, A. ( 1998 ) Exp. Neurol. 149 , 193202 .
28. Urban, M. O. , Jiang, M. C. & Gebhart, G. F. ( 1996 ) Brain Res. 737 , 8391 .
29. Urban, M. O. , Zahn, P. K. & Gebhart, G. F. ( 1999 ) Neuroscience 90 , 349352 .
30. Urban, M. O. , Coutinho, S. V. & Gebhart, G. F. ( 1999 ) Pain 81 , 4555 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SUPRASPINAL CONTRIBUTIONS TO HYPERALGESIA 7692

31. Coutinho, S. V. , Urban, M. O. & Gebhart, G. F. ( 1998 ) Pain 78 , 5969 .

32. Meller, S. T. & Gebhart, G. F. ( 1993 ) Pain 52 , 127136 .
33. Hargreaves, K. , Dubner, R. , Brown, F. , Flores, C. & Joris, J. ( 1998 ) Pain 32 , 7788 .
34. Sluka, K. A. & Westlund, K. N. ( 1993 ) Pain 55 , 367377 .
35. Schaible, H.-G. & Schmidt, R. F. ( 1985 ) J. Neurophysiol. 54 , 11091122 .
36. Schaible, H.-G. , Schmidt, R. F. & Willis, W. D. ( 1987 ) Exp. Brain Res. 66 , 466489 .
37. Herrero, J. F. & Cervero, F. ( 1996 ) Neurosci. Lett. 209 , 2124 .
38. Coderre, T. J. , Vaccarino, A. L. & Melzack, R. ( 1990 ) Brain Res. 535 , 155158 .
39. Dubuisson, D. & Dennis, S. G. ( 1977 ) Pain 4 , 161174 .
40. Bianchi, M. & Panerai, A. E. ( 1997 ) Neurosci. Lett. 237 , 8992 .
41. Wiertelak, E. P. , Furness, L. E. , Horan, R. , Martinez, J. , Maier, S. F. & Watkins, L. R. ( 1994 ) Brain Res. 649 , 1926 .
42. Wiertelak, E. P. , Roemer, B. , Maier, S. F. & Watkins, L. R. ( 1997 ) Brain Res. 748 , 143150 .
43. Bennett, G. J. ( 1993 ) Muscle Nerve 16 , 10401048 .
44. Pertovaara, A. , Wei, H. & Hamalainen, M. M. ( 1996 ) Neurosci. Lett. 218 , 127130 .
45. Bian, D. , Ossipov, M. H. , Zhong, C. M. , Malan, T. P & Porreca, F. ( 1998 ) Neurosci. Lett. 241 , 7982 .
46. Kauppila, T. ( 1997 ) Brain Res. 770 , 310312 .
47. Watkins, L. R. , Maier, S. F. & Goehler, L. E. ( 1995 ) Pain 63 , 289302 .
48. Maier, S. F. , Wiertelak, E. P. , Martin, D. & Watkins, L. R. ( 1993 ) Brain Res. 623 , 321324 .
49. Watkins, L. R. , Wiertelak, E. P. , Goehler, L. E. , Smith, K. P. , Martin, D. & Maier, S. F. ( 1994 ) Brain Res. 654 , 1526 .
50. Watkins, L. R. , Wiertelak, E. P. , Goehler, L. E. , Mooney-Heiberger, K. , Martinez, J. , Furness, L. , Smith, K. P. & Maier, S. F. ( 1994 ) Brain Res.
639 , 283299 .
51. Randich, A. & Gebhart, G. F. ( 1992 ) Brain Res. Rev. 17 , 7799 .
52. Schaible, H.-G. , Neugebauer, V. , Cervero, F. & Schmidt, R. F. ( 1991 ) J. Neurophysiol. 66 , 10211032 .
53. Ren, K. & Dubner, R. ( 1996 ) J. Neurophysiol. 76 , 30253037 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NEUROTROPHINS AND HYPERALGESIA 7693

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Neurotrophins and hyperalgesia

X.-Q. SHU AND L. M. MENDELL*

Department of Neurobiology and Behavior, State University of New York, Stony Brook, NY 11794
ABSTRACT Nerve growth factor (NGF), a member of the neurotrophin family, is crucial for survival of nociceptive neurons
during development. Recently, it has been shown to play an important role in nociceptive function in adults. NGF is up-regulated
after inflammatory injury of the skin. Administration of exogenous NGF either systemically or in the skin causes thermal
hyperalgesia within minutes. Mast cells are considered important components in the action of NGF, because prior degranulation
abolishes the early NGF-induced component of hyperalgesia. Substances degranulated by mast cells include serotonin, histamine,
and NGF. Blockade of histamine receptors does not prevent NGF-induced hyperalgesia. The effects of blocking serotonin
receptors are complex and cannot be interpretable uniquely as NGF losing its ability to induce hyperalgesia. To determine
whether NGF has a direct effect on dorsal root ganglion neurons, we have begun to investigate the acute effects of NGF on
capsaicin responses of small-diameter dorsal root ganglion cells in culture. NGF acutely conditions the response to capsaicin,
suggesting that NGF may be important in sensitizing the response of sensory neurons to heat (a process that is thought to operate
via the capsaicin receptor VR1). We also have found that ligands for the trkB receptor (brain-derived neurotrophic factor and
neurotrophin-4/5) acutely sensitize nociceptive afferents and elicit hyperalgesia. Because brain-derived neurotrophic factor is up-
regulated in trkA positive cells after inflammatory injury and is transported anterogradely, we consider it to be a potentially
important peripheral component involved in neurotrophin-induced hyperalgesia.
It is now well established that, in late embryonic life, sensory neurons depend on the availability of peripherally derived factors for
survival; this dependence is referred to as the neurotrophic hypothesis ( 1 ). Nociceptive neurons require nerve growth factor (NGF),
which is a member of the neurotrophin family; the other members are brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3),
and NT-4/5 ( 2 ). Neurotrophins signal via two types of receptors, the high-affinity trk receptor and the low-affinity p75 receptor ( 3 ). The
high-affinity receptor for NGF is trkA; for BDNF and NT-4/5, it is trkB; for NT-3, it is trkC. The p75 receptor can be activated by all
members of the neurotrophin family ( 2 , 3 ).
Although neurotrophins have been generally considered to function during embryonic life, it is now clear that their importance
continues well beyond this period. For example, nociceptors do not depend on NGF for survival beyond postnatal day 2, but they require
the availability of NGF to maintain their phenotype during a postnatal critical period ( 4 ). trk receptors continue to be expressed on
sensory neurons in adults ( 5 ), and neurotrophins also continue to be synthesized by numerous cell types ( 6 ).
NGF and Thermal Hyperalgesia. In recent years, it has become clear that NGF plays an important role in the function of
nociceptive afferents in the adult ( 7 ). Specifically, the continued presence of trkA receptors on nociceptive afferent fibers ( 8 ) and the up-
regulation of its ligand NGF in the skin during inflammation ( 9 ) indicate a potential role for NGF in inflammatory pain. Confirmation of
a hyperalgesic action for NGF has been obtained by demonstrating that administration of NGF produces thermal and mechanical
hyperalgesia ( 10 ). The thermal hyperalgesia has its onset within 1 h of NGF application whether it is systemic ( 10 ) or local ( 11 ). This
short latency and the effectiveness of local peripheral administration suggest a peripheral mechanism underlying the hyperalgesia.
Mechanical hyperalgesia typically has a latency of several hours, indicating a more complex mechanism, probably involving central
processes ( 12 ). There is also evidence for independent, longer-latency, peripheral ( 13 ) and central ( 12 ) mechanisms underlying thermal
hyperalgesia. The following discussion is restricted to peripheral mechanisms of neurotrophin-induced hyperalgesia.
Two types of evidence have been obtained in support of a peripheral locus for the short-latency thermal hyperalgesia initiated by
NGF. Pharmacological studies indicate a role for peripherally located nonneural cells in NGF-induced hyperalgesia. The cell most
centrally implicated in this action is the mast cell. These immunocompetent cells express trkA receptors ( 14 ) and degranulate their
contents in response to NGF stimulation ( 15 ). These contents include serotonin (5-HT), histamine, and NGF itself ( 16 ). Lewin et al. (
12 ) showed that prior degranulation of mast cells with compound 48/80 prevented the short-latency hyperalgesic effects of NGF without
affecting long-latency ones. Systemic application of receptor blockers for 5-HT blocked the hyperalgesic effect of 5-HT. Because 5-HT
has been shown, under some conditions, to sensitize the response of polymodal nociceptors to noxious heat ( 17 , 18 ), the conclusion was
that NGF was acting to sensitize primary afferent fibers via the release of 5-HT from mast cells. However, a more careful consideration of
this effect indicates that the 5-HT receptor blockers had a paradoxical effect of converting the action of NGF from hyperalgesia to
hypoalgesia ( 12 ). This effect raises the possibility that these blockers are acting to disturb some balance of effects initiated by NGF rather
than simply blocking the sensitizing action of mast-cell 5-HT. Recently, blockers of the 5-HT2 receptor and the histamine H1 receptor
were shown to be ineffective in blocking the effect of NGF-induced hyperalgesia ( 19 ). Together, these findings suggest that other mast-
cell contents (e.g., NGF) may play a more important role than 5-HT or histamine in NGF-induced hyperalgesia (see below).
A more direct approach toward demonstrating a peripheral component to NGF-induced hyperalgesia was adopted by Rueff and
Mendell ( 20 ), who examined the response of
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: NGF, nerve growth factor; BDNF, brain-derived neurotrophic factor; NT-n, neurotrophin-n; 5-HT, serotonin; DRG,
dorsal root ganglion.
* To whom reprint requests should be addressed, e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NEUROTROPHINS AND HYPERALGESIA 7694

small-diameter nociceptive afferents in an isolated skin-nerve preparation to NGF applied directly to the receptive field. Individual small-
diameter afferent fibers whose conduction velocity was established by the latency of their response to electrical stimulation in the
receptive field were selected if they responded to high-intensity mechanical stimulation. The response to thermal stimulation was
established. NGF then was applied directly to the receptive field for 20 min, and the mechanical and thermal responses were determined
again. NGF was found to lower the threshold to thermal stimulation by about 2C, a change that was statistically significant. However, no
change was noted in the mechanical threshold. Application of saline was found to elicit no effect. In later experiments, NT-3 applied to the
receptive field also was determined not to alter the threshold to noxious heat ( 11 ). If the animals were pretreated with compound 48/80,
NGF had no effect on the noxious heat threshold of individual nociceptive afferents. These experiments established that NGF changes the
threshold of nociceptive afferents and that mast cells are involved. They also confirmed a peripheral locus for the sensitizing action of NGF.
The experiments described so far show that NGF is sufficient to elicit hyperalgesia. A crucial question is whether it is a necessary
intermediate. This issue has been explored by preventing the increase in NGF after an inflammatory injury induced experimentally by
agents such as complete Freunds adjuvant. Antibodies to NGF ( 12 , 21 ) or the immunoadhesin molecule trkA-IgG ( 22 ) were used to
prevent the increase of NGF levels. A uniform finding in these experiments was that hyperalgesia was abolished, suggesting that NGF is a
necessary intermediate in inflammatory hyperalgesia induced by molecules such as complete Freunds adjuvant. It is presently believed
that injury leads to release of cytokines, such as tumor necrosis factor- and IL-1, which cause the release of NGF from cells such as
keratinocytes and fibroblasts ( 23 ). Such a release would initiate the degranulation of mast cells as illustrated in Fig. 1 .
Direct Effects of NGF on Nociceptors. Although the experiments described thus far indicate an indirect role for NGF in peripheral
sensitization via mast cells, the exact locus of action for NGF is not fixed by these findings. The reason for this uncertainty is that mast
cells contain NGF, which would be released on degranulation. Thus, activation of mast cells by NGF might lead to the release of more
NGF, and the presence of trkA receptors on sensory afferents could provide the ability for a direct effect on the peripheral threshold.
Furthermore, repeated daily administration of NGF can eventually produce hyperalgesia despite maintained mast-cell degranulation ( 24 ),
suggesting that, under some circumstances, mast cells can be bypassed. It is already known that NGF can affect the function of sensory
neurons directly, because, both in culture ( 25 ) and in vivo ( 21 ), exogenous NGF administration leads to up-regulation of peptides such
as substance P and calcitonin gene-related peptide in the cell body. However, this up-regulation is a relatively slow effect (hours to days)
involving transcriptional mechanisms and would be much too slow to account for the rapid effect of NGF on sensory thresholds.

FIG. 1. Schematic diagram outlining the relationship of mast cells, nociceptors, and NGF as well as how this system is activated as a
consequence of peripheral injury. Skin injury leads to release of cytokines, such as tumor necrosis factor- and IL-1, which activate
cells, such as keratinocytes and fibroblasts, to release NGF. The NGF can activate nociceptors directly but, in addition, can cause
mast cells to degranulate their products, including 5-HT, histamine, and NGF. This endogenous source of NGF seems to be more
potent than exogenous NGF in sensitizing nociceptors (see text for further details).
We have initiated experiments to examine whether NGF can rapidly increase the threshold of sensory neurons directly. To
accomplish this increase, it is necessary to provide a stimulus that excites these neurons and to determine whether NGF sensitizes the
response. To determine whether this effect is direct, these experiments must be done in culture to avoid the potential actions of other cells
such as mast cells.
Both behavioral evidence and electrophysiological evidence indicate that peripheral nociceptors are very sensitive to capsaicin ( 26 ,
27 ), an ingredient in hot peppers. When capsaicin is applied directly to a cell, it is depolarized as a consequence of a nonspecific increase
in permeability to cations including Ca. Recently, the capsaicin receptor was cloned ( 28 ). This receptor, named VR1, when expressed in
oocytes, is also sensitive to noxious heat, suggesting that the noxious-heat response of primary afferents is mediated via the VR1 receptor.
However, the component of the VR1 receptor sensitive to heat and to capsaicin may be different. NGF has been shown to play some role
in the expression of these receptors, because dorsal root ganglion (DRG) cells cultured for several days in the absence of NGF fail to
display capsaicin sensitivity ( 29 ).
Recently, we have shown that capsaicin responses are influenced acutely by NGF (X.-Q.S. and L.M.M., unpublished work).
Capsaicin was pressure ejected on the somata of small dissociated (<30 m) DRG neurons that were recorded in whole-cell or perforated
patch clamp. With the cell voltage clamped at 60 mV, an inward current was observed. A second identical capsaicin pulse 10 min later
resulted uniformly in a substantially smaller current (Figs. 2 and 3 ). We found that bath-applying NGF during the 10-min interval (
Fig. 2 ) often resulted in elimination of the tachyphylaxis and that the second response was often larger than the first ( Fig. 3 ), as much as
twice as large. In some cells, tachyphylaxis after NGF
the authoritative version for attribution.

FIG. 2. Protocol for testing NGF effects on capsaicin currents on DRG cells recorded in perforated patch clamp. Cells chosen for this
analysis were <30 m in diameter. A microelectrode filled with 1 M capsaicin was placed close to the patched cell, and the
capsaicin was ejected via a brief (400-ms) pressure pulse. The cell was clamped at 60 mV, and the inward capsaicin current lasting
10 s is shown at the top. A second pressure pulse was delivered 10 min later. During the interval, the cell was exposed either to
control saline solution or to 100 ng/ml of NGF. In controls, the response to the second capsaicin pulse was always smaller than the
initial one.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NEUROTROPHINS AND HYPERALGESIA 7695

treatment was similar to that observed after saline treatment. We assume that these cells do not express trkA.

FIG. 3. Examples of capsaicin responses in two DRG cells, the first conditioned by a 10-min exposure to saline (Upper) and the
second conditioned by 100 ng/ml NGF (Lower). Note the smaller response to the second capsaicin pulse (tachyphylaxis) after saline
treatment and the larger response when NGF is placed in the medium.
If we assume that the heat response of nociceptors is mediated by peripheral capsaicin receptors ( 28 ), it is apparent that the
sensitizing effect of NGF may be direct on the capsaicin receptor. It might involve phosphorylation of the capsaicin receptor that is
dephosphorylated as a consequence of the initial capsaicin stimulus ( 30 ). However, the assumption that the heat response is sensitized in
the same way as the response to capsaicin requires direct proof.
These experiments were carried out on DRG cell bodies maintained in culture for up to 1 day. The question that inevitably arises is
whether these conditions adequately test what might be occurring in the terminals. One problem is the presence of the nucleus in the
cultured DRG cells, which would not be present in the terminals in vivo. However, the NGF effect on the capsaicin response in the present
experiments was very rapid, within 10 minutes, which is too fast to require transcription. It follows that the influence of NGF on the
capsaicin response does not involve transcription and probably occurs at a membrane and/or cytoplasmic level.
The finding that NGF directly affects the capsaicin response of DRG cells suggests that NGF should be able to directly sensitize the
response to noxious heat. If so, why should prior degranulation of mast cells be able to abolish the hyperalgesic effects of NGF? In other
words, why should mast cells be necessary for NGF-induced hyperalgesia if NGF can directly sensitize nociceptive neurons as shown in
Fig. 3 ? One possibility is that exogenous NGF is insufficient in amount or does not reach the terminal in large enough concentrations in
the intact skin to influence nociceptive endings in vivo. However, if exogenous NGF degranulates mast cells, and these, in turn, release
NGF, then the resulting positive feedback cycle would amplify the amount of NGF activating the sensory ending. In effect, this hypothesis
suggests that exogenous NGF acts as a trigger that liberates NGF from mast cells. This latter source of NGF is postulated to be necessary
because of its intimate relationship with nociceptive afferent terminals, which allows it to condition the nociceptive terminal. The second
possibility is that other substances released from the mast cell, such as 5-HT or histamine, are necessary to enable NGF to have its full
effect on nociceptive terminals.
trkB Agonists and Hyperalgesia. A recent finding of interest in the context of neurotrophin involvement in hyperalgesia is that
administration of exogenous NGF leads to up-regulation of the trkB agonist BDNF in trkA-expressing sensory neurons ( 31 , 32 ). There is
evidence that BDNF can be transported anterogradely in intact axons to the periphery ( 33 ), raising the possibility that this neurotrophin
plays some role in inflammation-triggered events. In support of this possibility, both trkB agonists (BDNF and NT-4) were found to evoke
heat hyperalgesia when injected locally into the skin, and both of these agents were shown to elicit sensitization of individual nociceptive
afferent fibers to noxious heat ( 11 , 20 ). Both the hyperalgesia ( 11 ) and the acute fiber sensitization ( 20 ) elicited by NT-4 were blocked
in mast-cell-depleted preparations, suggesting the operation of a mechanism similar to the one mediating the hyperalgesic effects of NGF.
The interpretation of these findings is still uncertain, because it is not known whether skin mast cells express functional trkB receptors.
Intraperitoneal mast cells are known to express only trkA receptors ( 14 ), but blood mast cells, for example, express trkA, trkB, and trkC
receptors ( 34 ).
NT-3 did not elicit peripheral sensitization to noxious heat ( 11 ). This result might be interpreted as indicating that mast cells
responsible for this effect do not express trkC receptors. However, the selectivity probably resides with the afferent fibers, because it is
known that trkC receptors are expressed only on large-diameter afferents, whereas trkA is expressed only on small-diameter afferents ( 5 ).
trkB is expressed on cells in the DRG with a wide range of sizes, with considerable numbers of cells coexpressing trkA and trkB.
However, there is virtually no coexpression of trkA and trkC on sensory neurons ( 5 ).
CONCLUSIONS
If we consider capsaicin as a surrogate for noxious heat in activating the VR1 receptor ( 28 ), our findings suggest that NGF acts as a
peripheral sensitizing agent, at least in part by sensitizing the response of the nociceptor to noxious heat directly. NGF is not the first such
sensitizing agent to be described. It has been known that other sensitizing agents exist, including prostaglandins and bradykinin.
Previously, prostaglandin E2 has been shown to sensitize sensory neurons to capsaicin ( 35 ). Acute exposure of cultured neonatal DRG
cells to bradykinin can enhance their sensitivity to capsaicin and to low pH ( 36 ).
These agents do not seem to function in isolation. For example, bradykinin activates postganglionic efferents in the skin, and these
release prostaglandin E2 ( 37 ). Inactivation of postganglionic efferents reduces NGF-induced hyperalgesia (38), indicating interaction
between NGF and prostaglandins in hyperalgesia. NGF also interacts with bradykinin, in part by stimulating the delayed up-regulation of
BK1 receptors via release of kallikreins from mast cells (see ref. 13 for review). Blockade of BK1 receptors can transiently diminish NGF-
induced hyperalgesia ( 13 ), further indication of the interrelationship of these agents in causing peripheral sensitization.
Thus, NGF seems to be one of a number of sensitizing agents present in peripheral tissues. What is not clear at present is what the
integrative function of these individual agents might be in causing pain associated with inflammatory injury. An answer might be found by
the authoritative version for attribution.

considering the principal cellular events responsible for production of these different agents: cell breakdown for prostaglandin E2, clotting
for bradykinin, and the immune reaction (mast cells) for NGF (see ref. 39 for review). Evidently, the hyperalgesia accompanying
inflammation is sufficiently adaptive in terms of protection by immobilization of the affected body part that redundant mechanisms
involving mast cells ( 12 ), sympathetic efferents ( 38 ), and neutrophils ( 19 ) have evolved. However, in some cases hyperalgesia
becomes maladaptive, particularly if it elicits central sensitization ( 40 ) that causes the hyperalgesia to outlast the peripheral damage. It
then becomes essential to minimize peripheral sensitization to reduce both the immediate nociceptive effects as well as the long-lasting
ones produced by central sensitization ( 40 ).
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NEUROTROPHINS AND HYPERALGESIA 7696

This work was supported by National Institutes of Health Grants NS-14899 and NS-32264. Partial support was furnished by the
Javits Neuroscience Award NS-16996 (to L.M.M.). We thank Genentech for the NGF and NT-4/5 and Regeneron Pharmaceuticals
(Tarrytown, NY) for the BDNF and NT-3 used in our studies reviewed in this paper.
1. Davies, A. M. ( 1996 ) Philos. Trans. R. Soc. London B 351 , 389394 .
2. Thoenen, H. ( 1991 ) Trends Neurosci. 14 , 165170 .
3. Barbacid, M. ( 1994 ) J. Neurobiol. 2 , 13861403 .
4. Lewin, G. R. , Ritter, A. M. & Mendell, L. M. ( 1992 ) J. Neurosci. 12 , 18961905 .
5. McMahon, S. B. , Armanini, M. P. , Ling, L. H. & Phillips, H. S. ( 1994 ) Neuron 12 , 11611171 .
6. Lindsay, R. M. ( 1996 ) Philos. Trans. R. Soc. London B 351 , 365373 .
7. Lewin, G. R. & Mendell, L. M. ( 1993 ) Trends Neurosci. 16 , 353359 .
8. Averill, S. , McMahon, S. B. , Clary, D. O. , Reichardt, L. F. & Priestley, J. V. ( 1995 ) Eur. J. Neurosci. 7 , 14841494 .
9. Weskamp, G. & Otten, U. ( 1987 ) J. Neurochem. 48 , 17791786 .
10. Lewin, G. R. , Ritter, A. M. & Mendell, L. M. ( 1993 ) J. Neurosci. 13 , 21362148 .
11. Shu, X. Q. , Llinas, A. & Mendell, L. M. ( 1999 ) Pain 80 , 463470 .
12. Lewin, G. M. , Rueff, A. & Mendell, L. M. ( 1994 ) Eur. J. Neurosci. 6 , 19031912 .
13. Rueff, A. , Dawson, A. J. & Mendell L. M. ( 1996 ) Pain 66 , 359372 .
14. Horigome, K. , Pryor, J. C , Bullock, E. D. & Johnson, E. M. , Jr. ( 1993 ) J. Biol. Chem. 268 , 1488114887 .
15. Mazurek, N. , Weskamp, G. , Erne, P. & Otten, U. ( 1986 ) FEBS Lett. 198 , 315320 .
16. Leon, A. , Buriani, A. , Dal Toso, R. , Fabris, M. , Romanello, S. , Aloe, L. & Levi-Montalcini, R. ( 1994 ) Proc. Natl. Acad. Sci. USA 91 , 3739
3743 .
17. Beck, P. W. & Handwerker, H. O. ( 1974 ) Pflgers Arch. 347 , 209222 .
18. Rueff, A. & Dray, A. ( 1993 ) Agents Actions 38 , 1315 .
19. Bennett, G. , al-Rasheed, S. , Hoult, J. R. S. & Brain, S. D. ( 1998 ) Pain 77 , 315322 .
20. Rueff, A. & Mendell, L. M. ( 1996 ) J. Neurophysiol. 76 , 35933596 .
21. Woolf, C. J. , Safieh-Garabedian, B. , Ma, Q.-P. , Crilly, P. & Winter, J. ( 1994 ) Neuroscience 62 , 327331 .
22. McMahon, S. B. , Bennett, D. L. , Priestley, J. V. & Shelton, D. L. ( 1995 ) Nat. Med. 1 , 774780 .
23. Woolf, C. J. , Allchorne, A. , Safieh-Garabedian, B. & Poole, S. ( 1997 ) Br. J. Pharmacol. 121 , 417424 .
24. Amann, R. , Schuligoi, R. , Herzeg, G. & Donnerer, J. ( 1995 ) Pain 64 , 323329 .
25. Lindsay, R. M. & Harmar, A. J. , ( 1989 ) Nature (London) 337 , 362364 .
26. LaMotte, R. , Lundberg, L. E. & Torebjork, H. E. ( 1992 ) J. Physiol. (London) 448 , 749764 .
27. Szolcsanyi, J. , Anton, R , Reeh, P. W. & Handwerker, H. O. ( 1988 ) Brain Res. 446 , 262268 .
28. Caterina, M. J. , Schumacher, M. A. , Tominaga, M. , Rosen, T. A. , Levine, J. D. & Julius D. ( 1997 ) Nature (London) 389 , 816824 .
29. Winter, J. , Forbes, C. A , Sternberg, J. & Lindsay, R. M. ( 1988 ) Neuron 1 , 973981 .
30. Koplas, P. A. , Rosenberg, R. L. & Oxford , G. S. ( 1997 ) J. Neurosci. 17 , 35253537 .
31. Apfel, S. C. , Wright, D. E. , Wiideman, A. M. , Dormia, C. , Snider, W. D. & Kessler, J. A. ( 1996 ) Mol. Cell. Neurosci. 7 , 134142 .
32. Michael, G. J. , Averill, S. , Nitkunan, A. , Rattray, M. , Bennett, D. L. H. , Yan, Q. & Priestley, J. V. ( 1997 ) J. Neurosci. 17 , 84768490 .
33. Tonra, J. R. , Curtis, R. , Wong, V. , differ, K. D. , Park, J. S. , Timmes, A. , Nguyen, T. , Lindsay, R. M. , Acheson, A. & DiStefano, P. S. ( 1998 ) J.
Neurosci. 18 , 43744383 .
34. Tam, S. Y. , Tsai, M. , Yamaguchi, M. , Yano, K. , Butterfield, J. H. & Galli, S. J. ( 1997 ) Blood 90 , 18071820 .
35. Pitchford, S. & Levine, J. D. ( 1991 ) Neurosci. Lett. 132 , 105108 .
36. Stucky, C. L. , Abrahams, L. G. & Seybold, V. S. ( 1998 ) Neuroscience 84 , 12571265 .
37. Taiwo, Y. O. , Heller, P. H. & Levine, J. D. ( 1990 ) Neuroscience 39 , 523531 .
38. Andreev, N. Y. , Dimitrieva, N. , Koltzenburg, M. & McMahon, S. B. ( 1995 ) Pain 63 , 109115 .
39. Levine, J. & Taiwo, Y. ( 1994 ) in Textbook of Pain , eds. Wall, P. D. & Melzack, R. ( Churchill Livingstone , Edinburgh ), pp. 4556 .
40. McMahon, S. B. , Lewin, G. R. & Wall, P. D. ( 1993 ) Curr. Opin. Neurobiol. 3 , 602610 .
the authoritative version for attribution.
 SRC, A MOLECULAR SWITCH GOVERNING GAIN CONTROL OF SYNAPTIC TRANSMISSION MEDIATED BY N-METHYL-D- 7697
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASPARTATE RECEPTORS

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Src, a molecular switch governing gain control of synaptic

transmission mediated by N-methyl-D-aspartate receptors

XIAN-MIN YU * AND MICHAEL W. SALTER

* Molecular Neurobiology Section, Centre for Addiction and Mental Health and Programmes in Brain and Behavior and Cell
Biology, Hospital for Sick Children, Toronto, Ontario M5G 1X8, Canada; and Department of Physiology and Faculty of Dentistry,
University of Toronto, Toronto, M5G 1G6 Ontario, Canada
ABSTRACT The N-methyl-D-aspartate (NMDA) receptor is a principal subtype of glutamate receptor mediating fast
excitatory transmission at synapses in the dorsal horn of the spinal cord and other regions of the central nervous system. NMDA
receptors are crucial for the lasting enhancement of synaptic transmission that occurs both physiologically and in pathological
conditions such as chronic pain. Over the past several years, evidence has accumulated indicating that the activity of NMDA
receptors is regulated by the protein tyrosine kinase, Src. Recently it has been discovered that, by means of up-regulating NMDA
receptor function, activation of Src mediates the induction of the lasting enhancement of excitatory transmission known as long-
term potentiation in the CA1 region of the hippocampus. Also, Src has been found to amplify the up-regulation of NMDA receptor
function that is produced by raising the intracellular concentration of sodium. Sodium concentration increases in neuronal
dendrites during high levels of firing activity, which is precisely when Src becomes activated. Therefore, we propose that the boost
in NMDA receptor function produced by the coincidence of activating Src and raising intracellular sodium may be important in
physiological and pathophysiological enhancement of excitatory transmission in the dorsal horn of the spinal cord and elsewhere
in the central nervous system.
Appropriate modification of the transmission of information at synapses in the central nervous system (CNS) is essential for
physiological processes such as development, learning, and memory. On the other hand, inappropriate alteration of synaptic transmission
is a fundamental underpinning of various pathological conditions, including epilepsy and chronic pain. In the case of chronic pain,
enhanced transmission in nociceptive pathways, i.e., pathways conveying pain-related information, is known from animal experiments to
occur at various levels of the neuraxis including the dorsal horn of the spinal cord ( 1 3 ) and the trigeminal nucleus caudalis, the
homologous region in the brainstem ( 4 , 5 ). Synaptic transmission at fast excitatory synapses in the dorsal horn, as in most regions of the
CNS, is mediated by glutamate receptors, and there is a growing body of evidence indicating that these receptors are crucial in conditions
of enhanced nociceptive transmission ( 6 8 ). Activation of the N-methyl-D-aspartate (NMDA) subtype of glutamate receptor, in
particular, appears critical for the initiation and maintenance of the enhanced responsiveness of dorsal horn nociceptive neurons that
occurs in experimental pain models ( 4 , 5 , 9 11 ). The function of NMDA receptors, rather than being fixed at one level, is modulated
over a wide range, and thus understanding the processes by which this modulation occurs has the potential to shed new light on our
understanding of pathological alterations of synaptic transmission in chronic pain and other conditions in the CNS.
Over the past several years, it has become apparent that a fundamental process for regulating the function of NMDA receptors and
other ion channels in neurons is tyrosine phosphorylation ( 12 16 ). The protein tyrosine kinase (PTK) Src has been identified as an
endogenous PTK regulating NMDA receptor function ( 17 ). Src is one of the most well studied of the PTKs (for review, see ref. 18 ) and
is highly expressed in the CNS ( 19 , 20 ). Paradoxically, the functions of Src in the nervous system had been enigmatic. Recent
observations indicate that the regulation of NMDA receptors by Src may mediate the induction of a form of synaptic plasticity known as
long-term potentiation (LTP) in the hippocampus. Below, we outline the evidence for Src regulation of NMDA receptors ( 17 ) and the
role of this kinase in LTP induction ( 21 ). We also describe novel observations showing that Src kinase governs the regulation of NMDA
receptors by intracellular sodium ( 22 ).
NMDA Receptors and Regulation by Tyrosine Phosphorylation. NMDA receptors, as a main subtype of glutamate receptor,
participate in rapid excitatory synaptic transmission in the spinal cord and throughout the CNS ( 23 ). NMDA receptors are members of the
superfamily of ligand-gated ion channels, and a variety of NMDA receptor subunit proteins (NR1, NR2A-D, NR3) ( 24 , 25 ) have been
identified by using molecular cloning. Native NMDA receptors appear to be heterooligomeric complexes with the second membrane
region of the subunits coming together to form a conductance pathway that is selectively permeable to cations ( 24 , 26 ). NMDA receptors
are activated by the binding of two molecules of glutamate ( 27 ) and two molecules of glycine, which acts as a coagonist at an
extracellular site on the channel complex ( 28 ). Native NMDA receptors likely consist of one or more NR1 subunits, which may bind
glycine ( 29 , 30 ), a glutamatebinding subunit, NR2 ( 31 ), and possibly the more recently identified subunit, NR3 ( 32 , 33 ).
Activated NMDA channels are permeable to monovalent cations, such as Na+ and K+, and also divalent cations, the most important of
which is Ca2+ ( 34 , 35 ). NMDA receptors are known to be regulated ( 23 ) at diverse extracellular ( 36 40 ) as well as intracellular sites
( 41 ), and the key intracellular process regulating NMDA receptor function is phosphorylation ( 42 ). Both serine/threonine ( 43 52 ) and
tyrosine ( 12 ) phosphorylation have been shown to regulate NMDA receptor function. In terms of NMDA receptor regulation by tyrosine
phosphorylation, it has been found that when recombinant purified
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: CNS, central nervous system; LTP, long-term potentiation; mEPSCs, miniature excitatory postsynaptic currents;
NMDA, N-methyl-D-aspartate; NR, NMDA receptor; PTK, protein tyrosine kinase; SH2, Src-homology 2; SH3, Src-homology 3; AMPA,
-amino3-hydroxy-5-methyl-4-isoxazolepropionic acid.
To whom reprint requests should be addressed, e-mail: [email protected] .
 SRC, A MOLECULAR SWITCH GOVERNING GAIN CONTROL OF SYNAPTIC TRANSMISSION MEDIATED BY N-METHYL-D- 7698
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASPARTATE RECEPTORS

protein tyrosine kinase (PTK) or protein tyrosine phosphatase (PTP) enzymes are applied into neurons, the whole-cell currents through
native NMDA receptors are increased by PTK and are decreased by PTP (e.g., Fig. 1 ). Conversely, applying PTK inhibitors has been
found to decrease NMDA currents, whereas PTP inhibitors potentiate these currents ( 12 ), indicating that native NMDA receptors are
controlled by the balance of PTK and PTP activity. The increase in the ensemble NMDA currents that are measured by using the whole-
cell recording method was found to be caused by increased activity of individual NMDA channels and there is no change in the single-
channel conductance ( 53 ). This increase in NMDA channel activity is produced through enhancing the gating of already active receptors
rather than through recruiting previously inactive NMDA receptors ( 54 ).
As it is known that NMDA receptor subunit proteins, in particular NR2A ( 55 ) and NR2B ( 56 ), are phosphorylated on tyrosine it is
logical to ask whether the up-regulation of NMDA receptor function is due to phosphorylation of the subunit proteins themselves.
Investigating this would appear to be technologically feasible, because in addition to increasing the function of native NMDA receptors,
PTKs have been found to potentiate the function of recombinant NMDA receptors expressed heterologously ( 57 , 58 ). The number of
tyrosine residues on presumed intracellular domains of NR2A and -2B is large ( 54 ), but some residues are better candidates than others,
depending on the sequence of the surrounding amino acids, which confers selectively for particular PTKs or groups of PTKs ( 59 ).
Recently, through mutagenesis of residues in the C-terminal region of NR2A, three tyrosine residues in this region were found to be
necessary for the up-regulation of recombinant NR1/NR2A receptors expressed in HEK293 cells ( 60 ). This combination of receptor
subunits was known to be especially sensitive to inhibition by Zn2+ ( 61 ), and evidence was found indicating that the enhancement of
currents by tyrosine phosphorylation was caused by removal of this inhibition for NR1/NR2A, and also for NR1/NR2B, receptors. These
findings are surprising and intriguing ( 62 ) because the site for inhibition by Zn2+ is on the extracellular region of the receptor, whereas
tyrosine phosphorylation is presumed to occur at an intracellular site. Thus, there must be an unknown mechanism for transmitting the
effect of phosphorylation from the inside of the membrane to the outside.

FIG. 1. NMDA currents are up-regulated by protein tyrosine kinase and down-regulated by protein tyrosine phosphatase. NMDA
receptor-mediated currents recorded by using the whole-cell patch configuration were evoked by pressure application of L-aspartate
(200 M) at an interval of 1 min from a pressure pipette whose tip was positioned within 100 m from the recorded neuron. A shows
individual whole-cell current traces taken before (Control) or 20 min after the start of recording with PTK (src; 30 units/ml) or PTP
(truncated T cell PTP; 100 M) in the intracellular solution. B Normalized peak NMDA currents recorded with standard intracellular
solution (Control; n = 11 neurons), or intracellular solution supplemented with PTK (n = 8) or PTP (n = 7). For each neuron peak,
NMDA current is normalized by dividing the amplitude of current recorded at 20 min after the start of the recording (I20) by that of
the initial current (I1). (Bar = 2 sec and 200 pA.)
These observations foreshadow a new mechanism that has potential relevance to the general issue of the regulation of ion channels.
However, whether this mechanism applies to native NMDA receptors is doubtful, because it has been found that NMDA channel function
is up-regulated by tyrosine kinase activity even when Zn2+ is chelated ( 63 ), and NMDA channels with low single-channel conductance,
characteristic of NMDA channels that are insensitive to Zn2+ (64), are also up-regulated by PTKs ( 54 ). Therefore, it appears that removal
of Zn2+ inhibition is not the means by which the function of native NMDA receptors is up-regulated by tyrosine phosphorylation. Thus, for
native NMDA channels, the question of whether phosphorylation at the sites implicated by mutagenesis is the means for up-regulating
NMDA channel function remains open. It is alternatively possible that this up-regulation is through phosphorylation of other tyrosine
residues in the NMDA receptor subunit proteins or by phosphorylation of an associated protein, such as one of many proteins already
known to bind to the receptors ( 65 68 ).
Src Is an Endogenous PTK That Up-Regulates NMDA Receptor Function. Once it had been determined that NMDA receptor
function is regulated by PTKs and PTPs, a central question to be addressed was that of identifying the endogenous enzymes involved.
Notionally, this is not a trivial task, because the mammalian genome is expected to encode more than a thousand PTKs ( 69 ) and nearly as
many PTPs ( 70 ). Many of these enzymes are known to be expressed in the spinal cord and elsewhere in the CNS ( 20 , 71 ), providing
numerous potential candidates for the endogenous enzymes. Nevertheless, an endogenous PTK regulating NMDA receptors has been
identified, as described below, and there is preliminary evidence for a possible PTP ( 72 ).
It is well known that PTKs fall into two main categories: receptor and nonreceptor kinases ( 73 , 74 ). Within each of these categories,
there are numerous families with common features, in terms of primary sequence and domain structure. These common features have
permitted the development of pharmacological tools, including peptides and antibodies with activity against particular families of
enzymes, that have allowed the screening of broad groups of kinases. We took advantage of such reagents during our hunt to discover the
PTK regulating NMDA receptors ( 17 , 54 ). As a first step, we used a reagent that activates PTKs in the Src family: the phosphopeptide,
EPQ(pY)EEIPIA ( 75 ), which was found to enhance NMDA channel function. Conversely, channel function is depressed by an antibody,
anti-cst1, which inhibits Src-family PTKs ( 76 ). These results indicated that the endogenous PTK was a member of the Src family.
The family of Src kinases comprises a total of nine members, five of whichSrc, Fyn, Lyn, Lck, and Yesare known to be
expressed in the CNS. All members of the Src family contain highly homologous regionsthe C-terminal, catalytic, Src homology 2, and
the authoritative version for attribution.

Src homology 3 domains ( 77 ). The various members do, however, have substantial differences in a region of low sequence conservation
near the N terminus known as the unique domain. Therefore, reagents directed against this domain may distinguish between the various
Src family members.
Src was identified as the specific member of the family that regulates NMDA channel function by means of testing one such reagent,
the antibody, anti-src1, which selectively blocks the function of Src but not other members of the Src family ( 78 ). It was found that anti-
src1 caused a decrease in NMDA channel activity when this antibody was applied to the cytoplasmic face of membrane patches containing
NMDA channels (e.g., Fig. 2 A). In contrast, a control IgG had no effect. Moreover, anti-src1 prevented the enhancement of NMDA
channel activity by EPQ(pY)EEIPIA, indicating that Src is
 SRC, A MOLECULAR SWITCH GOVERNING GAIN CONTROL OF SYNAPTIC TRANSMISSION MEDIATED BY N-METHYL-D- 7699
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASPARTATE RECEPTORS

necessary for the effect of the activating peptide. As would be anticipated if Src indeed up-regulates NMDA receptors, we found that
applying recombinant pp60c-src increased NMDA channel activity ( Fig. 2 B), an effect not produced by heat-inactivating the kinase just
before use.

FIG. 2. Regulation of NMDA receptor single-channel activity in inside-out patches by Src and overlapping distribution of Src and
NMDA receptor subunit proteins. (A) A continuous record of NMDA channel-open probability (Po). Anti-src1 was applied to the
cytoplasmic face of the patch during the period indicated. Po was calculated in bins of 10 sec. (B) A record of NMDA channel Po
from a different inside-out patch with Src applied to the cytoplasmic side as indicated. (C) Confocal images show immunofluorescent
labeling of a dorsal horn primary culture by antibodies recognizing NR2A/B subunit proteins (green; courtesy of R. Wenthold,
National Institutes of Health, Bethesda, MD) or Src (red). Bottom shows the merged images; areas showing overlapping fluorescence
are yellow. We found similar colocalization when anti-NR1 and anti-Src antibodies were used. Also, experiments without primary
antibodies or with primary antibodies incubated with the respective immunogen peptides showed no labeling. (Bar = 10 m.)
Anti-src1 does not bind to the catalytic domain of Src, and we were therefore curious as to its mechanism of action. Along this line, it
was determined that applying a peptide, Src(4058), comprising the region in Src which the antibody recognizes, i.e., amino acids 4058,
to the cytoplasmic side of membrane patches reduced NMDA channel activity. A control peptide with the same amino acid composition
but in random order, scrambled Src(4058 ), had no effect on channel function. Because it was found that Src(4058) did not block in
vitro phosphorylation of a small substrate peptide by recombinant Src, we concluded that amino acids in the region 4058 may interact
with a component of the receptor complex and that this interaction is necessary for the effect of Src on NMDA channels. The region 4058
is within the unique domain of Src, and these results implicated this domain as being functionally important in this well known enzyme.
Through kinetic analysis of the NMDA channel activity in the patches, it was determined that the effect of Src is due to an increase in
channel gating during single activations of the receptor. This is relevant because synaptic responses mediated by NMDA receptors are
caused by single receptor activations ( 79 ). Thus, if NMDA receptors that are synaptically stimulated are affected similarly to the
receptors in the patches, which are by necessity extrasynaptic, it was predicted that Src should increase synaptic NMDA responses. This
was confirmed in studies of spontaneously occurring miniature excitatory postsynaptic currents (mEPSCs) ( 17 ). Consistent with these
electrophysiological results, it was found by using immunocytochemistry that the distribution of Src within neurons overlaps with that of
NMDA receptors ( Fig. 2 C) and that Src is localized at sites where NMDA receptors are highly enriched, presumably at synapses.
Whether Src is physically associated with NMDA channels could not be determined from any of the previous experiments:
membrane patches are large in comparison with the size of the proteins that comprise ion channels, and confocal microscopy does not
have sufficient spatial resolution. While it was therefore possible that Src was separate from the NMDA receptor complex, we found that
Src and NMDA receptor subunit proteins coprecipitate, demonstrating that Src is associated with the NMDA channel complex. The
coprecipitation might be via a direct interaction between Src and an NMDA receptor subunit protein or, alternatively, it is possible that Src
associates with NMDA channels by means of an intervening adaptor protein. Taking all of the information together, it was concluded that
Src is physically associated with and up-regulates the function of NMDA receptors.
Src is expressed at high levels within the CNS with a number of neuron-specific isoforms being generated by alternative splicing of
one or more cassettes ( 19 , 80 , 81 ) after amino acid 114, which is in the SH3 domain. In the nervous system, Src has been found to be
localized both pre- and postsynaptically ( 20 , 82 ). The postsynaptic localization is especially relevant to the modulation of NMDA
receptor function because Src has been found in the postsynaptic density ( 82 , 83 ), which is the main structural component of excitatory
synapses and is where glutamate receptors are concentrated.
Src Up-Regulation of NMDA Receptors in Hippocampal LTP. The studies described above implicating Src in the up-regulation of
NMDA receptor function were focused primarily on neurons from the spinal cord dorsal horn. Because NMDA receptors and Src are
widely expressed in the nervous system, there is the possibility that by regulating the activity of
the authoritative version for attribution.
 SRC, A MOLECULAR SWITCH GOVERNING GAIN CONTROL OF SYNAPTIC TRANSMISSION MEDIATED BY N-METHYL-D- 7700
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASPARTATE RECEPTORS

postsynaptic NMDA receptors, tyrosine phosphorylation/ dephosphorylation might modulate the efficacy of synaptic transmission in many
regions of the CNS. One region where Src is highly expressed is the CA1 region of the hippocampus ( 20 ). In this region, a lasting
enhancement of the efficacy of synaptic transmission, LTP, is induced by tetanic stimulation of the Schaffer collateral inputs to CA1
neurons ( 84 , 85 ). It has been established that LTP in the CA1 region is induced by a sequence of biochemical steps occurring in the
pyramidal neurons ( 86 , 87 ). Both PTK function ( 88 ) and NMDA receptor activation ( 89 ) have been found to be necessary for LTP
induction in these neurons. Therefore, we used the reagents characterized previously to determine whether Src participates in LTP in CA1.
It was found that administering anti-src1 or Src(4058) into CA1 neurons prevented the induction of LTP in an acute hippocampal
slice preparation ( 21 ). On the other hand, administering recombinant Src or activating Src by means of the EPQ(pY)EEIPIA peptide
induced a long-lasting enhancement of synaptic responses. This enhancement occluded the induction of LTP and vice versa, implying that
Src activation is sufficient to cause LTP. In addition, by measuring Src activity biochemically in vitro by using an immune-complex kinase
assay, it was found that tetanic stimulation, which produced LTP, caused an increase in the activity of Src within 1 min of the stimulation.
Thus, Src is up-regulated very rapidly as a consequence of tetanic stimulation. Because inhibiting Src did not affect basal synaptic
responses, it was concluded that rapid up-regulation of Src activity is necessary, as well as sufficient, for the induction of LTP.
The principal means by which LTP is expressed in CA1 neurons is enhancement of the -amino-3-hydroxy-5-methyl-4-
isoxazolepropionic acid (AMPA) component of synaptic responses and therefore, our conclusion seemed at odds with other findings ( 17 )
that the AMPA receptor-mediated synaptic response is not potentiated when Src is activated in dorsal horn neurons. However, in our
studies on dorsal horn neurons, intracellular Ca2+ was highly buffered, whereas in the experiments in hippocampal slices low intracellular
Ca2+ buffering was used, because this is required to induce LTP. When Ca2+ buffering was increased in the hippocampal neurons, Src no
longer potentiated the synaptic AMPA responses, but synaptic NMDA responses were still potentiated by Src, as was the case in the dorsal
horn neurons. Thus, enhancement of AMPA responses produced by activating Src depends on a rise in intracellular Ca2+. Because Src is
not a Ca2+-dependent enzyme ( 90 ), these results indicated that Src does not up-regulate AMPA receptors directly but rather does so
indirectly through one or more Ca2+-dependent steps.
In other experiments, it was determined that blocking NMDA receptors prevents but does not reverse Src-induced potentiation of
AMPA responses. Thus, like LTP induced by tetanus, NMDA receptors are necessary to produce, but not to maintain, the potentiation of
AMPA responses induced by activating Src directly. Together, these findings required the development of a new model where activation
of Src appears to be a biochemical mechanism gating the induction of LTP in CA1 neurons ( 54 ). It is hypothesized that during induction
of LTP, Src is rapidly activated, leading to enhanced NMDA receptor function, which boosts the entry of Ca2+ sufficiently to trigger the
downstream signaling cascade.
A key question opened up by this work is, what is the mechanism causing Src activation upon tetanic stimulation? Src has a number
of regulatory sites ( 18 ), and there are numerous biochemical pathways that converge to activate ( 91 , 92 ) or to inhibit ( 93 , 94 ) this
kinase. Determining whether the increase in Src activity is produced by stimulating an activating pathway or by blocking an inhibiting one
and identifying the specific biochemical steps are central goals of future work in this area.
Because the role of Src in LTP induction appears to be to enhance NMDA receptor function, one potential mechanism is
phosphorylation of one of the NMDA receptor subunit proteins, as discussed above. Indeed, it has been found that the level of tyrosine
phosphorylation of the NR2B NMDA receptor subunit is increased after induction of LTP in the dentate gyrus in the hippocampus ( 95 ,
96 ). Like CA1, the dentate is a region where LTP induction is NMDA receptor-dependent and is blocked by inhibitors of PTKs ( 97 ).
Another region where NMDA receptor-dependent synaptic plasticity has been associated with tyrosine phosphorylation of NMDA
receptors is the insular cortex, where it has been found that taste aversion conditioning causes an increase in phosphorylation of NR2B (
98 ). However, whether phosphorylation of NR2B mediates the plasticity in the dentate gyrus or in the insular cortex, and if so by what
mechanism, remains to be determined ( 54 ).
Tyrosine kinases were first implicated in the induction of LTP from experiments showing that LTP is blocked by bath-applied
tyrosine kinase inhibitors ( 88 ). It was found subsequently that mutant mice with targeted deletion of the src gene showed LTP in CA1,
which is a genetic argument against the absolute requirement for Src in the induction of LTP. Also, it was reported that in mice lacking the
Src-family kinase fyn, LTP is blunted but not abolished ( 99 ). The impairment in LTP is age-dependent in fyn / mice, with young fyn
/ animals showing LTP comparable to that in wild-type animals ( 100 ). The developmental time at which LTP becomes impaired in fyn
/ mice correlates with a large decline in the level of Src expression. Src and Fyn are known to substitute for each other in various
processes ( 101 ). Thus, from our evidence together with that from experiments using genetic manipulation, it appears that in wild-type
animals, Src is a required mediator of LTP induction, whereas in animals that develop without src, another member of the src family,
likely fyn, may substitute. It is possible that, for example, by being upstream of Src activation, Fyn might also be necessary for LTP
induction in the wild-type animal, and this possibility needs to be examined in acute experiments by using Fyn-specific manipulations.
Intracellular Sodium Regulates NMDA Receptors. During high levels of neuronal discharge activity, such as occur during the
induction of lasting changes in synaptic efficacy, there is a large influx of Na+, leading to substantial increases in the intracellular
concentration of Na+ ([Na+]i) ( 102 ). It has been found that during such activity, the level of [Na+]i may increase by 1520 mM in the
neuronal soma ( 103 ) and likely even more in the dendrites ( 102 ). While there can be no doubt that Na+ is the major carrier of electrical
charge responsible for producing action potentials and excitatory postsynaptic potentials, the possibility that raising [Na+]i may act as a
signaling factor in neurons had been virtually ignored. Thus, we recently set out to determine whether Na+ might regulate synaptic
function in postsynaptic neurons ( 22 ).
We first examined the effects of raising [Na+]i on whole-cell NMDA currents evoked by exogenously applied NMDA. Perfusing the
neurons with an elevated [Na+]i solution produced an increase in the amplitude of the NMDA currents of nearly 40% ( 22 ). In this case,
the intracellular perfusion contained [Na+]i of 50 mM, but we only achieved an increase in [Na+]i of about 30 mM, as determined by using
the fluorescent Na+-sensitive dye sodium-binding benzofuran isophthalate (SBFI), presumably because of the very active Na+ pumping in
the neurons. The increase in NMDA currents was not reproduced by perfusing Cs+ at 50 mM, indicating that not all monovalent cations
were able to cause the potentiation and, equally importantly, that the potentiation was not because of lowering intracellular [K+], which
was required to maintain proper osmolarity.
To characterize the effects of intracellular Na+ on NMDA channel gating, single-channel currents were recorded by using cell-
the authoritative version for attribution.

attached patches, and [Na+]i was varied by application of

 SRC, A MOLECULAR SWITCH GOVERNING GAIN CONTROL OF SYNAPTIC TRANSMISSION MEDIATED BY N-METHYL-D- 7701
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASPARTATE RECEPTORS

the Na+-ionophore monensin ( 104 ). Ratiometric measurement of [Na+]i was done under similar conditions so that single-channel activity
could be correlated to the actual change in [Na+]i. It was found that NMDA channel activity followed the level of [Na+]iincreasing when
Na+ was raised and falling when [Na+]i was reduced ( Fig. 3 ). We found no change in the single-channel conductance of the NMDA
channels, indicating that the increase in whole-cell current when Na+ was raised could be accounted for by increased NMDA channel
gating.
Neurons express a diversity of Na+-permeable channels, e.g., ionotropic glutamate receptors and voltage-gated Na+ channels, and
therefore we wondered whether Na+ influx through these various channels could affect NMDA channel function. This was investigated by
examining single-channel activity by enclosing single channels within a patch pipette attached to the cell and stimulating surrounding
channels in the cell by bath-applying activators. An important consideration in these experiments was to avoid the known voltage
dependence of NMDA channel gating ( 105 ), and this was done by adjusting the transmembrane potential of the patch so that it was at a
constant level with respect to the channel reversal potential. We thus found that bath-applying agonists to activate NMDA or non-NMDA
receptors outside the patch led to an increase in activity of the NMDA channels within the patch. This increase in activity was prevented
when Na+ was removed from the bath solution. Importantly, removal of Ca2+ from the bath did not alter the effect of stimulating the extra-
patch NMDA receptors. Also, depolarizing the cells by bath-applying a high-K+ solution to mimic the depolarization caused by applying
the agonists did not affect NMDA channel function. Thus, we concluded that NMDA receptor function may be up-regulated by Na+ influx
through neighboring glutamate receptors.

FIG. 3. Increases in [Na+]i by application of monensin potentiate single NMDA channel activity recorded in the cell-attached
configuration. A shows the recording configuration. (B) A continuous record of NMDA channel Po from neurons bathed with
extracellular solution containing 50 mM Na+. (C) shows representative single-channel currents before and during monensin
application. D Changes in Po and mean open time (to) versus [Na+]i during monensin application. For each Na+ concentration, six
patches were tested. *, P < 0.05; **, P < 0.01, MannWhitney U test when compared with the channel activity at 0 mM Na+.
Another main route for Na+ entry into neurons is via voltage-gated channels permeable to Na+, such as those responsible for the
generation of action potentials. To produce a consistent activity of voltage-gated Na+ channels and thereby attain a stable membrane
potential of the cell, as required to accurately record channel function in the cell-attached patches, we bath-applied the alkaloid veratridine
( 106 ). Veratridine caused a significant increase in NMDA channel activity, and the effects of veratridine were prevented when it was
applied in the presence of tetrodotoxin. Thus, by using veratridine as a surrogate for evoking action potentials, we concluded that influx of
Na+ through tetrodotoxin-sensitive voltage-gated Na+ channels is sufficient to increase NMDA channel activity.
The NMDA receptors studied by using the cell-attached recordings are of necessity localized extrasynaptically. To determine whether
synaptic NMDA receptors are affected by changing [Na+]i, we studied spontaneous mEPSCs. We found that applying Na+ into neurons via
the patch pipette significantly increased the NMDA receptor-mediated component of the mEPSCs. By contrast, the non-NMDA receptor
component of the mEPSCs was not altered by raising [Na+]i. Applying Cs+ into the cell did not affect either the NMDA or the non-NMDA
receptor-mediated components of the mEPSCs. Thus, increasing [Na+]i selectively amplifies synaptic responses mediated by NMDA but
not non-NMDA receptors. Taking into account the effects of varying Na+ levels on NMDA channel activity in cell-attached patches
described above, it appears that the efficacy of synaptic transmission through NMDA receptors tracks the level of Na+ in the postsynaptic
neuron.
A Src Kinase Controls the Enhancement of NMDA Channel Function by [Na+]i. To lay the foundation for understanding the
mechanisms and molecules by which intracellular Na+ alters NMDA channel function, we examined the effects of Na+ applied directly to
the authoritative version for attribution.

the cytoplasmic side of the membrane in inside-out patches. In contrast to the effects found in cell-attached recordings from intact neurons,
applying 50 mM Na+ to the cytoplasmic face of inside-out patches did not change NMDA channel activity. We reasoned that the action of
Na+ on NMDA receptors may depend on a molecule(s) lost from the excised patches or a biochemical process that had been disrupted.
Because regulation of NMDA receptors by protein phosphorylation is well established, we considered the possibility that phosphorylation
may be involved in the effects of intracellular Na+. To examine this, we used a broad-spectrum inhibitor of protein kinases, staurosporine,
which was found to abolish the increase in NMDA channel activity caused by raising Na+ through bath-applying monensin during cell-
attached patch experiments. Importantly, staurosporine did not prevent the rise in [Na+]i produced by the application of monensin. Thus, it
was concluded that protein kinase activity may be required for the up-regulation of NMDA receptor activity by raising intracellular Na+.
From these results, two mechanistic possibilities emerged: that the effect of Na+ is mediated by activating a kinase (or inhibiting a
phosphatase) that is present in the cell but is lost from the patches, or alternatively, but not mutually exclusively, that the activity of a
kinase is required for the effect of Na+ but is not directly a mediator. Because it had been established that endogenous Src could be up-
regulated in the excised patches, we wondered whether this up-regulation would permit an effect of Na+. Indeed, this was found to be the
case, because
 SRC, A MOLECULAR SWITCH GOVERNING GAIN CONTROL OF SYNAPTIC TRANSMISSION MEDIATED BY N-METHYL-D- 7702
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASPARTATE RECEPTORS

applying Na+ to the cytoplasmic face increased NMDA channel activity in patches treated with the Src family-activating phosphopeptide
EPQ(pY)EEIPIA ( Fig. 4 ). In contrast, the nonphosphorylated form of this peptide, which does not activate Src kinases, did not permit the
effect of Na+. We also found that at concentrations well above those expected under physiological conditions, Na+ increased NMDA
channel activity in the untreated inside-out patches ( 22 ). Thus, the Src-activating peptide shifted the concentration-response curve for Na+
to the left.

FIG. 4. NMDA channel activity is increased by raising the Na+ concentration on the cytoplasmic side of the membrane in inside-out
patches during activation of Src-family kinases. A shows the inside-out recording configuration. (B) A continuous record of NMDA
channel Po. The peptide EPQ(pY)EEIPIA was applied 35 min before the recording and throughout the recording period. Na+ (50
mM) was applied as indicated. (C) Left, changes in Po versus [Na+] on the cytoplasmic side in the presence of the peptide EPQ(pY)
EEIPIA; Right, effects of 50 mM Na+ on NMDA channels in control experiments and in the presence of the nonphosphorylated
peptide EPQYEEIPIA. *, P < 0.05, Wilcoxon test or paired t test.

FIG. 5. Diagram illustrating a working model for the regulation of NMDA channel gating by Src and Na +. See text for details.
Therefore, from the convergence of evidence described above, we have developed a working model that the sensitivity of NMDA
channels to intracellular Na+ is controlled by a channel-associated Src kinase. This model is represented diagrammatically in Fig. 5 . For
simplicity of presentation, we represent the entire complex of NMDA channel subunits and associated proteins together as one pair of
ovals forming the ion channel. This is not to imply that either the target of phosphorylation or the site of action of Na+ is necessarily one of
the NMDA channel subunit proteins. Rather, as alluded to above, the target of phosphorylation remains to be determined, as does the site
of action of Na+. Also, because the Src-activating peptide stimulates all Src-family kinases, it needs to be established whether the
enhanced sensitivity to Na+ is caused by Src itself or another member of the family.
Changing Src activity from the lowest to the highest level produces an 3- to 4-fold increase in NMDA channel function. Raising Na+
from 0 to 40 mM produces an 1.5- to 2-fold change in NMDA channel activity. The change produced by Na+ is over and above that
produced by Src and therefore, we estimate that the total range over which NMDA channel activity is regulated by Src and Na+ is about 5-
to 8-fold. This degree of change would represent a dramatic alteration in synaptic efficacy and would be expected to have sizable effects
the authoritative version for attribution.

on synaptic integration of individual neurons and on the behavior of neural networks. Moreover, because the influx of Ca2+ through
NMDA receptors follows the amplitude of the currents (e.g., ref. 12 ), then a 5-fold increase in current would result in a 5-fold increase in
Ca2+ influx. This boost in Ca2+ influx may then allow the resultant rise in the level of intracellular Ca2+ to be large enough to engage Ca2+-
activated signaling pathways in the cell.
We expect that the boost in NMDA channel function by coincidence of Src activation and raising [Na+]i may be relevant to the
induction of lasting enhancement of synaptic efficacy in phenomena such as LTP. The role of Src in this process has already been
established, but what about a role of Na+? High levels of action potential discharge, similar to those that have been shown to cause rises in
[Na+]i, occur as a result of tetanic stimulation. Action potentials initiated in the soma are known to propagate into the dendrites ( 107
109 ), and this action potential backpropagation may participate in the induction of LTP ( 110 , 111 ). It is thought that the
backpropagation of action potentials works by promoting Ca2+ entry
 SRC, A MOLECULAR SWITCH GOVERNING GAIN CONTROL OF SYNAPTIC TRANSMISSION MEDIATED BY N-METHYL-D- 7703
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASPARTATE RECEPTORS

through voltage-gated Ca2+ channels and by producing depolarization that relieves the Mg2+ block of NMDA channels. Our recent results,
however, open up a new possibility, i.e., that it is the rise in intracellular Na+ that is the important event. This rise in Na+, when coincident
with postsynaptic activation of Src, may lead to a large amplification of NMDA receptor function sufficiently large to set off the rest of the
intracellular signaling cascade.
What about implications for pain? Excitatory synaptic transmission in nociceptive pathways in the spinal cord is facilitated by
stimulation of C fiber nociceptors ( 112 114 ), apparently as a result of increased responsiveness of NMDA receptors ( 6 , 11 ). It is not
yet known whether Src is activated under such circumstances, but it seems likely given that Src can be activated through a number of
signaling pathways, such as stimulating G protein-coupled receptors and receptor tyrosine kinases ( 91 , 92 ), and these types of pathways
have been implicated in enhancement of nociceptive transmission ( 115 , 116 ). Moreover, the discharge rate of central nociceptive
neurons responding to noxious inputs is sufficiently high as to be expected to produce dramatic rises in [Na+]i. Therefore, it is possible that
the up-regulation of NMDA receptors by Src and sodium is involved in the central up-regulation of transmission in nociceptive pathways.
This may be relevant to human pain states because this central up-regulation appears to be sufficient to produce hyperalgesia and allodynia
(e.g., ref. 117 ).
CONCLUSIONS
The gene encoding Src was found over 20 years ago as the first protooncogene, with particular Src mutations causing cancer ( 118 ).
Since then, this enzyme has been found to have many roles in cell signaling in a variety of cell types. In neurons, Src up-regulates the
function of NMDA receptors and thereby gates the induction of a lasting enhancement of synaptic transmission, LTP in the CA1 region of
the hippocampus. Src also sensitizes NMDA channels to up-regulation by intracellular Na+. Thus, the coincidence of Src activation and a
rise in [Na+]i may be important for boosting synaptic NMDA receptor function and initiating the intracellular signaling cascades that
produce persistent alterations in synaptic function. Because NMDA receptors are implicated in a variety of pathophysiological conditions
in the CNS, the regulation by Src and Na+ represent potential targets for developing new types of therapeutic intervention in a variety of
CNS disorders.
We thank Yueqiao Huang and Jeff Gingrich for helpful comments on the manuscript. Work of the authors is supported by the
Medical Research Council of Canada and the Nicole Fealdman Memorial Fund. M.W.S. in an MRC Scientist and X.M.Y. is an MRC
Scholar.
1. Dubner, R. & Ruda, M. A. ( 1992 ) Trends Neurosci. 15 , 96103 .
2. Levine, J. D. ( 1998 ) Neuron 20 , 649654 .
3. Woolf, C. J. & Doubell, T. P. ( 1994 ) Curr. Opin. Neurobiol. 4 , 525534 .
4. Chiang, C. Y. , Park, S. J. , Kwan, C. L. , Hu, J. W. & Sessle, B. J. ( 1998 ) J. Neurophysiol. 80 , 26212631 .
5. Yu, X. M. , Sessle, B. J. , Haas, D. A. , Izzo, A. , Vernon, H. & Hu, J. W. ( 1996 ) Pain 68 , 169178 .
6. Woolf, C. J. & Thompson, S. W. N. ( 1991 ) Pain 44 , 293299 .
7. Neugebauer, V. , Lucke, T. & Schaible, H. G. ( 1993 ) J. Neurophysiol. 70 , 13651377 .
8. Dickenson, A. H. , Chapman, V. & Green, G. M. ( 1997 ) Gen. Pharmacol. 28 , 633638 .
9. Dickenson, A. H. & Sullivan, A. F. ( 1987 ) Neuropharmacology 26 , 12351238 .
10. Ma, Q. P. & Woolf, C. J. ( 1995 ) Pain 61 , 383390 .
11. Liu, X. G. & Sandkuhler, J. ( 1995 ) Neurosci. Lett. 191 , 4346 .
12. Wang, Y. T. & Salter, M. W. ( 1994 ) Nature (London) 369 , 233235 .
13. Fadool, D. A. , Holmes, T. C. , Berman, K. , Dagan, D. & Levitan, I. B. ( 1997 ) J. Neurophysiol. 78 , 15631573 .
14. Cataldi, M. , Taglialatela, M. , Guerriero, S. , Amoroso, S. , Lombardi, G. , Di Renzo, G. & Annunziato, L. ( 1996 ) J. Biol. Chem 271 , 94419446 .
15. Moss, S. J. , Gorrie, G. H. , Amato, A. & Smart, T. G. ( 1995 ) Nature (London) 377 , 344348 .
16. Lev, S. , Moreno, H. , Martinez, R. , Canoll, P. , Peles, E. , Musacchio, J. M. , Plowman, G. D. , Rudy, B. & Schlessinger, J. ( 1995 ) Nature
(London) 376 , 737745 .
17. Yu, X. M. , Askalan, R. , Keil, G. J. & Salter, M. W. ( 1997 ) Science 275 , 674678 .
18. Brown, M. T. & Cooper, J. A. ( 1996 ) Biochim. Biophys. Acta 1287 , 121149 .
19. Brugge, J. S. , Cotton, P. C. , Queral, A. E. , Barrett, J. N. , Nonner, D. & Keane, R. W. ( 1985 ) Nature (London) 316 , 554557 .
20. Sugrue, M. M. , Brugge, J. S. , Marshak, D. R. , Greengard, P. & Gustafson, E. L. ( 1990 ) J. Neurosci. 10 , 25132527 .
21. Lu, Y. M. , Roder, J. C. , Davidow, J. & Salter, M. W. ( 1998 ) Science 279 , 13631368 .
22. Yu, X.-M. & Salter, M. W. ( 1998 ) Nature (London) 396 , 469474 .
23. McBain, C. J. & Mayer, M. L. ( 1994 ) Physiol. Rev. 74 , 723760 .
24. Gasic, G. P. & Hollmann, M. ( 1992 ) Annu. Rev. Physiol. 54 , 507536 .
25. Seeburg, P. H. ( 1993 ) Trends Neurosci. 16 , 359365 .
26. Raymond, L. A. , Blackstone, C. D. & Huganir, R. L. ( 1993 ) Trends Neurosci. 16 , 147153 .
27. Benveniste, M. & Mayer, M. L. ( 1991 ) Biophys. J. 59 , 560573 .
28. Johnson, J. W. & Ascher, P. ( 1987 ) Nature (London) 325 , 529531 .
29. Hirai, H. , Kirsch, J. , Laube, B. , Betz, H. & Kuhse, J. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 60316036 .
30. Ivanovic, A. , Reilander, H. , Laube, B. & Kuhse, J. ( 1998 ) J. Biol. Chem. 273 , 1993319937 .
31. Laube, B. , Hirai, H. , Sturgess, M. , Betz, H. & Kuhse, J. ( 1997 ) Neuron 18 , 493503 .
32. Sucher, N. J. , Akbarian, S. , Chi, C. L. , Leclerc, C. L. , Awobuluyi, M. , Deitcher, D. L. , Wu, M. K. , Yuan, J. P. , Jones, E. G. & Lipton, S. A. (
1995 ) J. Neurosci. 15 , 65096520 .
33. Ciabarra, A. M. , Sullivan, J. M. , Gahn, L. G. , Pecht, G. , Heinemann, S. & Sevarino, K. A. ( 1995 ) J. Neurosci. 15 , 64986508 .
34. Mayer, M. L. & Westbrook, G. L. ( 1987 ) J. Physiol. (London) 394 , 501527 .
35. Ascher, P. & Nowak, L. ( 1988 ) J. Physiol. (London) 399 , 247266 .
36. Traynelis, S. F. & Cull-Candy, S. G. ( 1990 ) Nature (London) 345 , 347350 .
37. Christine, C. W. & Choi, D. W. ( 1990 ) J. Neurosci. 10 , 108116 .
38. Hollmann, M. , Boulter, J. , Maron, C. , Beasley, L. , Sullivan, J. , Pecht, G. & Heinemann, S. ( 1993 ) Neuron 10 , 943954 .
39. McGurk, J. F. , Bennett, M. V. L. & Zukin, R. S. ( 1990 ) Proc. Natl. Acad. Sci. USA 87 , 99719974 .
40. Aizenman, E. , Lipton, S. A. & Loring, R. H. ( 1989 ) Neuron 2 , 12571263 .
41. Rosenmund, C. & Westbrook, G. L. ( 1993 ) Neuron 10 , 805814 .
the authoritative version for attribution.

42. MacDonald, J. F. , Mody, I. & Salter, M. W. ( 1989 ) J. Physiol. (London) 414 , 1734 .
43. Chen, L. & Huang, L.-Y. M. ( 1991 ) Neuron 7 , 319326 .
44. Chen, L. & Huang, L. Y. M. ( 1992 ) Nature (London) 356 , 521523 .
45. Kelso, S. R. , Nelson, T. E. & Leonard, J. P. ( 1992 ) J. Physiol. (London) 449 , 705718 .
46. Durand, G. M. , Gregor, P. , Zheng, X. , Bennett, M. V. L. , Uhl, G. R. & Zukin, R. S. ( 1992 ) Proc. Natl. Acad. Sci. USA 89 , 93599363 .
47. Tingley, W. G. , Roche, K. W. , Thompson, A. K. & Huganir, R. L. ( 1993 ) Nature (London) 364 , 7073 .
48. Sigel, E. , Baur, R. & Malherbe, P. ( 1994 ) J. Biol. Chem. 269 , 82048208 .
49. Wang, L.-Y. , Orser, B. A. , Brautigan, D. L. & MacDonald, J. F. ( 1994 ) Nature (London) 369 , 230232 .
50. Raman, I. M. , Tong, G. & Jahr, C. E. ( 1996 ) Neuron 16 , 415421 .
51. Lieberman, D. N. & Mody, I. ( 1994 ) Nature (London) 369 , 235239 .
 SRC, A MOLECULAR SWITCH GOVERNING GAIN CONTROL OF SYNAPTIC TRANSMISSION MEDIATED BY N-METHYL-D- 7704
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

ASPARTATE RECEPTORS

52. Tong, G. , Shepherd, D. & Jahr, C. E. ( 1995 ) Science 267 , 15101512 .

53. Wang, Y. T. , Yu, X. M. & Salter, M. W. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 17211725 .
54. Salter, M. W. ( 1998 ) Biochem. Pharmacol. 56 , 789798 .
55. Lau, L. F. & Huganir, R. L. ( 1995 ) J. Biol. Chem. 270 , 2003620041 .
56. Moon, I. S. , Apperson, M. L. & Kennedy, M. B. ( 1994 ) Proc. Natl. Acad. Sci. USA 91 , 39543958 .
57. Khr, G. & Seeburg, P. H. ( 1996 ) J. Physiol. (London) 492 , 445452 .
58. Chen, S.-J. & Leonard, J. P. ( 1996 ) J. Neurochem. 67 , 194200 .
59. Zhou, S. Y. & Cantley, L. C. ( 1995 ) Trends Biochem. Sci. 20 , 470475 .
60. Zheng, F. , Gingrich, M. B. , Traynelis, S. F. & Conn, P. J. ( 1998 ) Nat. Neurosci. 1 , 195191 .
61. Paoletti, P. , Ascher, P. & Neyton, J. ( 1997 ) J. Neurosci. 17 , 57115725 .
62. Ascher, P. ( 1998 ) Nat. Neurosci. 1 , 173175 .
63. Pelkey, K. A. , Xiong, Z. G. , Lu, W. Y. , Salter, M. W. & MacDonald, J. F. ( 1999 ) Soc. Neurosci. Abstr. 25 , in press.
64. Farrant, M. , Feldmeyer, D. , Takahashi, T. & Cull-Candy, S. G. ( 1994 ) Nature (London) 368 , 335339 .
65. Kornau, H. C. , Schenker, L. T. , Kennedy, M. B. & Seeburg, P. H. ( 1995 ) Science 269 , 17371740 .
66. Muller, B. M. , Kistner, U. , Kindler, S. , Chung, W. J. , Kuhlendahl, S. , Fenster, S. D. , Lau, L. F. , Veh, R. W. , Huganir, R. L. , Gundelfinger, E.
D. & Garner, C. C. ( 1996 ) Neuron 17 , 255265 .
67. Wyszynski, M. , Lin, J. , Rao, A. , Nigh, E. , Beggs, A. H. , Craig, A. M. & Sheng, M. ( 1997 ) Nature (London) 385 , 439442 .
68. Ehlers, M. D. , Zhang, S. , Bernhardt, J. P. & Huganir, R. L. ( 1996 ) Cell 84 , 745755 .
69. Hunter, T. ( 1994 ) Semin. Cell Biol. 5 , 367376 .
70. Charbonneau, H. & Tonks, N. K. ( 1992 ) Annu. Rev. Cell Biol. 8 , 463493 .
71. Ross, C. A. , Wright, G. E. , Resh, M. D. , Pearson, R. C. & Snyder, S. H. ( 1988 ) Proc. Natl. Acad. Sci. USA 85 , 98319835 .
72. Pelkey, K. A. , Askalan, R. , Nguyen, T. H. , Salter, M. W. & Lombroso, P. J. ( 1998 ) Soc. Neurosci. Abstr. 24 , 1825.
73. Hunter, T. & Cooper, J. A. ( 1985 ) Annu. Rev. Biochem. 54 , 897930 .
74. Ullrich, A. & Schlessinger, J. ( 1990 ) Cell 61 , 203212 .
75. Liu, X. , Brodeur, S. R. , Gish, G. , Songyang, Z. , Cantley, L. C. , Laudano, A. P. & Pawson, T. ( 1993 ) Oncogene 8 , 11191126 .
76. Roche, S. , Koegl, M. , Barone, M. V. , Roussel, M. F. & Courtneidge, S. A. ( 1995 ) Mol. Cell Biol. 15 , 11021109 .
77. Superti-Furga, G. & Courtneidge, S. A. ( 1995 ) BioEssays 17 , 321330 .
78. Roche, S. , Fumagalli, S. & Courtneidge, S. A. ( 1995 ) Science 269 , 15671569 .
79. Edmonds, B. , Gibbs, R. A. & Colquhoun, D. ( 1995 ) Annu. Rev. Physiol. 57 , 495519 .
80. Martinez, R. , Mathey-Prevot, B. , Bernards, A. & Baltimore , D. ( 1987 ) Science 237 , 411415 .
81. Pyper, J. M. & Bolen, J. B. ( 1990 ) Mol. Cell. Biol. 10 , 20352040 .
82. Atsumi, S. , Wakabayashi, K. , Titani, K. , Fujii, Y. & Kawate, T. ( 1993 ) J. Neurocytol. 22 , 244258 .
83. Cudmore, S. B. & Gurd, J. W. ( 1991 ) J. Neurochem. 57 , 12401248 .
84. Bliss, T. V. & Collingridge, G. L. ( 1993 ) Nature (London) 361 , 3139 .
85. Malenka, R. C. & Nicoll, R. A. ( 1993 ) Trends Neurosci. 16 , 521527 .
86. Malenka, R. C. , Kauer, J. A. , Perkel, D. J. , Mauk, M. D. , Kelly, P. T. , Nicoll, R. A. & Waxham, M. N. ( 1989 ) Nature (London) 340 , 554557 .
87. Nicoll, R. A. & Malenka, R. C. ( 1995 ) Nature (London) 377 , 115118 .
88. ODell, T. J. , Kandel, E. R. & Grant, S. G. ( 1991 ) Nature (London) 353 , 558560 .
89. Collingridge, G. L. & Bliss, T. V. P. ( 1987 ) Trends Neurosci. 10 , 288293 .
90. Erpel, T. & Courtneidge, S. A. ( 1995 ) Curr. Opin. Cell Biol. 7 , 176182 .
91. Della, R. G. , Van Biesen, T. , Daaka, Y. , Luttrell, D. K. , Luttrell, L. M. & Lefkowitz, R. J. ( 1997 ) J. Biol. Chem. 272 , 1912519132 .
92. Dikic, I. , Tokiwa, G. , Lev, S. , Courtneidge, S. A. & Schlessinger, J. ( 1996 ) Nature (London) 383 , 547550 .
93. Okada, M. , Nada, S. , Yamanashi, Y. , Yamamoto, T. & Nakagawa, H. ( 1991 ) J. Biol. Chem. 266 , 2424924252 .
94. Klages, S. , Adam, D. , Class, K. , Fargnoli, J. , Bolen, J. B. & Penhallow, R. C. ( 1994 ) Proc. Natl. Acad. Sci. USA 91 , 2597 2601 .
95. Rosenblum, K. , Dudai, Y. & Richter-Levin, G. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 1045710460 .
96. Rostas, J. A. , Brent, V. A. , Voss, K. , Errington, M. L. , Bliss, T. V. & Gurd, J. W. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 1045210456 .
97. Abe, K. & Saito, H. ( 1993 ) Brain Res. 621 , 167170 .
98. Rosenblum, K. , Berman, D. E. , Hazvi, S. , Lamprecht, R. & Dudai, Y. ( 1997 ) J. Neurosci. 17 , 51295135 .
99. Grant, S. G. N. , ODell, T. J. , Karl, K. A. , Stein, P. L. , Soriano, P. & Kandel, E. R. ( 1992 ) Science 258 , 19031910 .
100. Kojima, N. , Wang, J. , Mansuy, I. M. , Grant, S. G. , Mayford, M. & Kandel, E. R. ( 1997 ) Proc. Natl. Acad. Sci. USA 94 , 4761 4765 .
101. Thomas, S. M. , Soriano, P. & Imamoto, A. ( 1995 ) Nature (London) 376 , 267271 .
102. Jaffe, D. B. , Johnston, D. , Lasser-Ross, N. , Lisman, J. E. , Miyakawa, H. & Ross, W. N. ( 1992 ) Nature (London) 357 , 244246 .
103. Rose, C. R. & Ransom, B. R. ( 1997 ) J. Physiol. (London) 499 , 573587 .
104. Sandeaux, R. , Sandeaux, J. , Gavach, C. & Brun, B. ( 1982 ) Biochim. Biophys. Acta 684 , 127132 .
105. Nowak, L. M. & Wright, J. M. ( 1992 ) Neuron 8 , 181187 .
106. Strichartz, G. , Rando, T. & Wang, G. K. ( 1987 ) Annu. Rev. Neurosci. 10 , 237267 .
107. Regehr, W. G. , Konnerth, A. & Armstrong, C. M. ( 1992 ) Proc. Natl. Acad. Sci. USA 89 , 54925496 .
108. Regehr, W. , Kehoe, J. S. , Ascher, P. & Armstrong, C. ( 1993 ) Neuron 11 , 145151 .
109. Spruston, N. , Schiller, Y. , Stuart, G. & Sakmann, B. ( 1995 ) Science 268 , 297300 .
110. Stuart, G. , Spruston, N. , Sakmann, B. & Hausser, M. ( 1997 ) Trends. Neurosci. 20 , 125131 .
111. Markram, H. , Lubke, J. , Frotscher, M. & Sakmann, B. ( 1997 ) Science 275 , 213215 .
112. Woolf, C. J. ( 1983 ) Nature (London) 306 , 686688 .
113. Randic, M. , Jiang, M. C. & Cerne, R. ( 1993 ) J. Neurosci. 13 , 52285241 .
114. Liu, X. G. , Morton, C. R. , Azkue, J. J. , Zimmermann, M. & Sandkuhler, J. ( 1998 ) Eur. J. Neurosci. 10 , 30693075 .
115. Xu, X.-J. , Dalsgaard, C.-J. & Wiesenfeld-Hallin, Z. ( 1992 ) Neuroscience 51 , 641648 .
116. Ma, Q. P. & Woolf, C. J. ( 1995 ) J. Physiol. (London) 486 , 769777 .
117. Torebjork, H. E. , Lundberg, L. E. & LaMotte, R. H. ( 1992 ) J. Physiol. (London) 448 , 765780 .
118. Levinson, A. D. , Oppermann, H. , Levintow, L. , Varmus, H. E. & Bishop, J. M. ( 1978 ) Cell 15 , 561572 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

PAIN PERCEPTION: IS THERE A ROLE FOR PRIMARY SOMATOSENSORY CORTEX? 7705

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Pain perception: Is there a role for primary somatosensory cortex?

M. C. BUSHNELL * , G. H. DUNCAN, R. K. HOFBAUER, B. HA, J.-I. CHEN, AND B. CARRIER

McGill University and Universit de Montral, Montreal, Quebec, Canada H3A 1A1
ABSTRACT Anatomical, physiological, and lesion data implicate multiple cortical regions in the complex experience of pain.
These regions include primary and secondary so matosensory cortices, anterior cingulate cortex, insular cortex, and regions of the
frontal cortex. Nevertheless, the role of different cortical areas in pain processing is controversial, particularly that of primary
somatosensory cortex (S1). Human brain-imaging studies do not consistently reveal pain-related activation of S1, and older studies
of cortical lesions and cortical stimulation in humans did not uncover a clear role of S1 in the pain experience. Whereas studies
from a number of laboratories show that S1 is activated during the presentation of noxious stimuli as well as in association with
some pathological pain states, others do not report such activation. Several factors may contribute to the different results among
studies. First, we have evidence demonstrating that S1 activation is highly modulated by cognitive factors that alter pain
perception, including attention and previous experience. Second, the precise somatotopic organization of S1 may lead to small
focal activations, which are degraded by sulcal anatomical variability when averaging data across subjects. Third, the probable
mixed excitatory and inhibitory effects of nociceptive input to S1 could be disparately represented in different experimental
paradigms. Finally, statistical considerations are important in interpreting negative findings in S1. We conclude that, when these
factors are taken into account, the bulk of the evidence now strongly supports a prominent and highly modulated role for S1
cortex in the sensory aspects of pain, including localization and discrimination of pain intensity.
The role of primary somatosensory cortex (S1) in pain perception has long been in dispute. In the early 20th century, Head and
Holmes ( 1 ) observed that patients with longstanding cortical lesions did not show deficits in pain perception. Similarly, Penfield and
Boldrey ( 2 ), based on studies of electrical stimulation of patients exposed cerebral cortices during epilepsy surgery, concluded that pain
probably has little or no cortical representation. In more recent studies of S1 cortex in monkeys, single-cell recordings in both anesthetized
( 3 ) and awake ( 4 ) monkeys revealed so few nociceptive neurons that their functional significance was uncertain.
Other evidence suggests that S1 cortex may indeed play an important role in pain perception. Despite the lack of profound deficits in
pain perception after widespread cortical lesions, patients do show at least transient deficits following cortical lesions ( 1 , 5 ).
Furthermore, Young and Blume ( 6 ) have reported that some patients with epileptic foci involving S1 cortex experience painful seizures.
Anatomical evidence from primates demonstrates that regions of thalamus containing nociceptive neurons project to S1 cortex ( 7 9 ),
thus providing the possible framework necessary to support the processing of nociceptive information within S1. Additionally, despite the
small numbers of nociceptive neurons observed in monkey S1 cortex, their responses parallel pain perception in humans ( 10 , 11 ). For
example, by using optical imaging and neuronal recording techniques, Tommerdahl et al. ( 11 ) showed that nociceptive activity in area 3a
of S1 cortex exhibited slow temporal summation and poststimulus response persistence after repeated cutaneous heat stimulation, which
parallel perceptual consequences of the stimulation in humans. Finally, bilateral ablation of S1 cortex in monkeys disrupts their ability to
discriminate intensities of noxious heat (D. R. Kenshalo, Jr., D. A. Thomas, and R. Dubner, unpublished observations).
Findings from human brain imaging studies have produced inconsistent results pertaining to the role of S1 cortex in pain perception.
The first three modern brain imaging studies of pain, published in the early 1990s, produced vastly different results in terms of S1 cortex.
By using positron emission tomography (PET) and repeated 5-sec heat stimuli presented to six spots on the arm, Talbot et al. ( 12 ) found
a significant activation focus in S1 cortex contralateral to the stimulated arm. By using similar heat stimuli, but repetitively presented to a
single spot on the dorsal hand, Jones et al. ( 13 ) failed to observe significant activation in S1 cortex. Finally, by using single photon-
emission computed tomography, Apkarian et al. ( 14 ) found that submerging the fingers in hot water for 3 minutes led to a decrease in S1
activity.
Jones and colleagues ( 15 , 16 ) postulated that the experimental procedures used by Talbot et al., particularly moving the stimulus
among six spots during the scans, differentially direct more attention to the pain stimulus than to the control stimulus, and thus produce an
attention-related modulation of S1 cortical activity. They further postulated that the presence or absence of pain, itself, is probably not a
main determinant of S1 activation. More recent studies support the idea that attention can significantly modulate pain-evoked S1 activity,
but little evidence supports the premise that pain is not a major determinant of S1 activity during painful stimulation.
Table 1 shows the methods and results of a number of human brain imaging studies of pain, by using PET, single photon-emission
computed tomography, functional MRI, and magnetoencephalographic imaging. In the various studies, pain stimuli include phasic and
tonic heat, cold, chemical irritants, electric shock, ischemia, visceral distension, headache, and neuropathic pain. As can be seen in Table
1 , there is little consistency among the studies as to whether S1 is activated by pain. Some studies involving thermal, chemical, or
electrical stimulation reveal S1 activation, whereas others using similar stimuli do not. Several factors that may contribute to these
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: S1, primary somatosensory cortex; PET, positron emission tomography; rCBF, regional cerebral blood flow.
* To whom reprint requests should be addressed. e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

PAIN PERCEPTION: IS THERE A ROLE FOR PRIMARY SOMATOSENSORY CORTEX? 7706

differential results include (i) influences of cognitive modulation in S1 activity; (ii) averaging-related degradation of the signal because of
variability of sulcal anatomy; (iii) a possible combination of excitatory and inhibitory effects of nociceptive input to S1; and (iv)
differences in statistical analyses and power.
Table 1. Methods and results of brain imaging studies
Study Ref. Modality Subject n Stimulation device Stimulation S1
Talbot et al. 12 PET H215O Healthy 8 Thermode 1 cm2 42, 4748C Yes
Jones et al. 13 PET 15CO2 Healthy 6 Thermode 2.5 cm 36.1, 41.3, and 46.4C No
5.0 cm
Apkarian et al. 14 SPECT Healthy Water bath Moderate heat pain Inhibition
Crawford et al. 24 PET 131Xe Healthy 11 Tourniquet Ischemia Yes
Coghill et al. 25 PET H215O Healthy 9 Thermode, 1 cm2 34, 4748C Yes
Di Piero et al. 26 SPET 133Xn Healthy 7 Water bath Cold pressor test Yes
Derbyshire et al. 27 PET H215O Facial pain 6 Thermode Ramp, 2543C No
Healthy 7
Rosen et al. 28 PET H215O Angina pectoris 12 Dobutamine infusion Angina No
Hsieh et al. 29 PET H215O Neuropathic 8 None Spontaneous pain No
Hsieh et al. 30 PET H215O Healthy 4 Intracutaneous Ethanol Yes
injection
Davis et al. 31 fMRI Healthy 9 Electrical nerve Median nerve, 50 Hz Yes
stimulator
Weiller et al. 32 PET H215O Migraine patients 9 None Spontaneous migraine No
Howland et al. 33 MEG Healthy 5 Electrical nerve Electric finger shock Yes
stimulator
Kitamura et al. 34 MEG Healthy 5 Electrical nerve Electric finger shock Yes
stimulator
Craig et al. 35 PET H215O Healthy 11 Thermal grill Bars 20 and 40C Yes
Casey et al. 36 PET H215O Healthy 9 Thermode 40 and 50C Yes
Casey et al. 36 PET H215O Healthy 9 Thermode 6, 20C Yes
Hsieh et al. 37 PET H215O Cluster headache 7 Sublinguial Headache No
nitroglycerin
Andersson et al. 21 PET H215O Healthy 6 Capsaicin Intracutaneous Yes
Antognini et al. 38 fMRI Healthy 5 Electrical stimulator Electric hand shock Yes
Aziz et al. 39 PET H215O Healthy 8 Balloon Esophageal distension Yes
Cluster headache 7
DiPiero et al. 40 SPECT Healthy 12 Water bath Cold pressor Yes
Rainville et al. 17 PET H215O Healthy 11 Water bath 35, 47C Yes
Silverman et al. 41 PET H215O Healthy 6 Balloon Rectal distension No
Svensson et al. 42 PET H215O Healthy 11 CO2 laser Cutaneous Yes, ns
Svensson et al. 43 PET H215O Healthy 11 Electrical stimulator Intramuscular Yes, ns
Xu et al. 44 PET H215O Healthy 6 CO2 laser Cutaneous Yes
Binkofski et al. 45 fMRI Healthy 5 Balloon Esophageal distension Yes
Derbyshire et al. 46 PET H215O Healthy 6 Thermode Ramp, 2543C No
Iadarola et al. 47 PET H215O Healthy Capsaicin Subcutaneous arm No
May et al. 48 PET H215O Healthy 7 Capsaicin Subcutaneous No
forehead
Oshiro et al. 49 fMRI Healthy 6 Electrical stimulator Finger, 8 Hz Yes
Paulson et al. 50 PET H215O Healthy 10 Thermode, 254 mm2 40 & 50C No

ns, not significant; SPECT, single photon-emission computed tomography; fMRI, functional MRI; MEG, magnetoencephalographic
imaging.

Cognitive Modulation of S1 Activity. As proposed by Jones and colleagues ( 15 , 16 ), S1 pain-related activation is highly
modulated by cognitive factors that alter pain perception, including attention and previous experience. In our laboratory, we have shown
that when the subjects attention is directed away from a painful stimulus, the activity of S1 cortex is dramatically reduced (B.C., P.
Rainville, T. Paus, G.H.D., and M.C.B., unpublished observations). Fig. 1 shows the results of this study, in which we used PET H215O
bolus methods to measure regional cerebral blood flow (rCBF) in nine subjects while they discriminated changes in thermal intensity or
auditory frequency. During all scans, concurrent sequences of tones and contact heat stimuli (pain, 46.548.5C or warm, 3238C on the
left arm) were presented. After each scan, subjects used a 100-mm visual analogue scale to rate the perceived level of pain associated with
the thermal stimuli. Statistical brain maps of pain-related activity, i.e., rCBF during the pain condition minus rCBF during the nonpainful
warm condition, were obtained. Pain intensity was rated higher in the thermal than in the auditory task (50.4 vs. 41.4, P = 0.01), indicating
that pain perception was modulated by the attentional demands of the discrimination tasks. Likewise, whereas in the thermal task there
was a significant pain-related rCBF increase within S1 (t = 4.42, P < 0.01), there was no significant change in pain-related rCBF within S1
during the auditory task ( Fig. 1 ). A direct comparison of pain-related S1 activity during the pain and auditory tasks showed that pain-
evoked rCBF was significantly larger in the thermal than in the auditory task (t = 3.92; P < 0.01). In this experiment, the behavioral task
the authoritative version for attribution.

used to direct attention toward the thermal stimuli involved the detection of a small change in the intensity of the heat stimulus. This task
probably served to specifically direct the subjects attention to sensory aspects of the pain, rather than to the unpleasantness or suffering.
Other data from our laboratory also support the idea that attention to sensory aspects of the pain experience can alter S1 activity. By
using hypnosis, we found that suggestions specifically directed toward increasing or decreasing the perceived intensity of the burning pain
sensation produced by submerging a subjects hand in painfully hot water modulated pain-related activity in S1 (R.K.H., P. Rainville,
G.H.D., and M.C.B., unpublished observations). In contrast, suggestions
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

PAIN PERCEPTION: IS THERE A ROLE FOR PRIMARY SOMATOSENSORY CORTEX? 7707

directed toward changing the unpleasantness of the pain had no effect on pain-related activity in S1, but produced instead a robust
modulation of activity in anterior cingulate cortex directly correlated with the subjects perception of unpleasantness (ANCOVA, P =
0.005) ( 17 ).

FIG. 1. Pain-related activity when attention is directed to the painful heat stimulus (Left) or to an auditory stimulus (Center) is
revealed by subtracting PET data recorded when a warm stimulus (3238C) was presented from those recorded when a painfully hot
stimulus (46.548.5C) was presented during each attentional state. Differences in pain-related activity during the two attentional
conditions are revealed (Right) by subtracting PET data recorded during the auditory task from that recorded during the heat-
discrimination task (using only painful stimulus trials46.548.5C). PET data, averaged across nine subjects, are illustrated against
an MRI from one subject. Horizontal and coronal slices through S1 are centered at the activation peaks. Red circles surround the
region of S1. Whereas there was a significant activation of S1 when subjects attended to the painful stimulus (Left), there was no
significant activation when subjects attended to the auditory stimulus (Center). However, there was a subsignificant activation in S1
during the auditory task, as shown in the Inset. The direct comparison of pain in the two attentional conditions (Right) shows a
significant difference in pain-related S1 activity during the two attentional states.
In these hypnosis experiments, we also found evidence that experience with the hypnotic suggestions may have produced long-term
changes in the subjects neural processing of pain. At least a week before participating in a PET scanning session, all subjects received the
same hypnotic induction, suggestions, and painful stimuli that were to be used during the scanning experiment. In subsequent PET
sessions, two scans using the painful heat and two using the nonpainful warm control stimulus were performed before the subjects
underwent hypnotic induction and suggestions. During these four scans, the subjects were simply instructed to relax and attend to the
thermal stimulusa control situation for identifying regions that could be examined for modulation related to hypnotic suggestions given
in subsequent scans. Although subjects in the two hypnosis experiments produced similar ratings of pain intensity and unpleasantness
during these control scans, those previously trained to attend to the intensity of the painful stimuli showed substantially greater pain-
related activity in S1 than did those who had been trained to attend to the unpleasantness of those stimuli [ Fig. 2 Upper (t = 5.05); Lower
(17) (t = 3.01)].
Attentional modulation within S1 cortex is not restricted to pain-related activity. Other investigators have found that rCBF in S1,
evoked by tactile stimuli, is reduced when subjects attend to another stimulus modality ( 18 ). Similarly, neuronal recordings in S1 cortex
of trained monkeys reveal low-threshold neurons whose activity is enhanced by attention to the tactile stimulus ( 19 , 20 ). Despite the
extensive nature of attentional modulation of S1 activity, there is little evidence that attention activates S1 neurons without the concurrent
presence of sensory-evoked activation. Anticipation of a painful stimulus has been shown to produce decreases in S1 rCBF ( 51 ) rather
than increases in rCBF that would reflect excitatory neuronal activity.
Variability of S1 Sulcal Anatomy. Both monkey and human data indicate that nociceptive activity in S1 cortex is somatotopically
organized. Neuronal recording studies in monkeys show a somatotopic organization for nociceptive neurons similar to that observed for
low-threshold cells ( 3 ). Similarly, by using PET to measure rCBF and capsaicin as a specific nociceptive stimulus, Andersson et al. ( 21 )
found distinct activation sites related to foot and hand pain, consistent with the known topographic organization of cutaneous receptive
fields within S1. The somatotopic organization of S1 cortex probably results in a small focal activation evoked by a localized pain
stimulus. Such focal activation is more easily observed with single-subject functional MRI studies than in PET studies involving a number
of subjects and data smoothing. Fig. 3 shows the focal activation produced in S1 cortex of three individual subjects when the leg was
stimulated with noxious heat (B.H., J.-I.C, B. Pike, G.H.D., and M.C.B., unpublished observations). The images show regions activated
during stimulation with painful heat, as compared with that observed during nonpainful warm. Fig. 3 also shows that the pain-related S1
activation sites, although on the posterior bank of the central sulcus in all subjects, varied in terms of their stereotaxic coordinates,
suggesting small intersubject differences in the localization of pain-related activity. This anatomical variability, although not of a large
magnitude, degrades a focal signal when data are averaged across subjects. Thus, because of the somatotopically organized focal
activation observed in S1, the rCBF signal arising from this area may be particularly susceptible to degradation when averaging data.
Inhibitory Effects of Noxious Stimuli in S1 activity. Tommerdahl et al. ( 11 ) found in monkey S1 cortex that the presence of
the authoritative version for attribution.

noxious heat reduced the intrinsic optical-imaging signal evoked by low-threshold mechanical stimulation of the skin.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

PAIN PERCEPTION: IS THERE A ROLE FOR PRIMARY SOMATOSENSORY CORTEX? 7708

These data are consistent with the findings of Apkarian et al. ( 14 ), which showed a decrease in blood flow to S1 cortex in human subjects
during the presentation of a tonic heat stimulus. Consonant with the idea that noxious stimulation produces inhibition of tactile sensitivity
in S1 cortex are psychophysical data showing that the presence of pain reduces tactile perception ( 22 ).

FIG. 2. Changes in pain-related activity associated with previous hypnotic training by using suggestions for modulating pain
sensation (Upper) or pain unpleasantness (Lower). Both images represent data from control scans, in which no hypnotic suggestions
were given. Each image represents the subtraction of PET data recorded when the hand was submerged in thermally neutral water
(35C) from data recorded when the hand was submerged in painfully hot water (47C). PET data were averaged across 10
experimental sessions in the sensory study (Upper) and, in a different group of subjects, 11 experimental sessions in the affective
study (Lower). The PET data are illustrated against the average MRI for that subject group. Coronal slices through S1 are centered at
the activation peaks, and red circles surround the region of S1.
Other evidence suggests that the inhibition of S1 tactile activity by noxious stimuli may take place at lower levels of the neuraxis
rather than through a direct inhibitory influence at the level of S1. In awake monkeys, the spontaneous activity of low-threshold neurons in
the ventroposterior thalamus is inhibited by topical application of capsaicin, which specifically excites C fibers (C.-C. Chen and M.C.B.,
unpublished observations). Similarly, capsaicin sometimes reduces the responses of spinothalamic tract neurons to noxious heat
stimulation ( 23 ). Thus, when a painful stimulus is presented in a human brain imaging study, the net effect of exciting some neurons and
inhibiting the spontaneous activity of others could have different effects on rCBF (as measured by PET) or on venous blood oxygenation
(as measured by functional MRI), depending on such variables as timing, duration, location, and intensity of the painful stimulus.
Procedural and Analytical Differences in Studies. In human brain imaging studies, many procedural variables can influence the
resultant data. Analytical techniques are not standardized across laboratories or across imaging methods. For example, different
approaches are used to compare stimulation conditions, including subtraction and regression comparisons across scans. Instructions to the
subjects, which can influence the cognitive state, vary among studies, as does the timing of stimulus variables. The statistical analyses,
including methods for calculating variance and assumptions about the nature of the data, also differ among laboratories. Although all
analyses rely on some type of statistical determination of significance, the method of accounting for multiple comparisons varies, and thus
the criteria for identifying an activation as significant are not uniform across studies. Finally, the power of any statistical test is influenced
by the number of subjects studied, which is another factor that varies greatly among studies.
the authoritative version for attribution.

FIG. 3. Functional MRI data from three subjects, using a 1.5-T scanner and standard head coil. Each horizontal and coronal image
represents the anatomical and functional data from a single subject during one session, which included a high-resolution anatomical
scan and five to eight runs of 120 whole-brain functional MRI scans ( 10 13 ) 7-mm slices acquired at 3-sec intervals. Thermal
stimuli were applied to the left calf on separate runs. Thermal runs consisted of 9 s alternating cycles of rest, painful (4546C), rest,
and neutral (35 36C) stimulation by using a 9 cm2 thermode. Activation maps were generated by using Spearmans rank order
correlation, comparing painful to neutral heat. The coronal and horizontal slices through S1 are centered at the activation peaks, and
red circles surround the region of S1.
As with any statistical test, the interpretation of negative results must be performed with caution. Thus, it would appear more fruitful
to identify regions that show activation across a
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

PAIN PERCEPTION: IS THERE A ROLE FOR PRIMARY SOMATOSENSORY CORTEX? 7709

number of pain studies than to rely on the data of any one study in isolation. Despite the wide methodological and analytical variation in
human brain imaging studies of pain, there is surprising consistency in the activation of a number of brain regions, including anterior
cingulate and insular cortices. Although the observation of pain-related activation in S1 is somewhat less consistent, the fact that at least
half of the human brain imaging studies have identified significant activation of this region when subjects perceive pain suggests that S1
has a significant role in nociceptive processing.
What Is the Role of S1 in Pain Processing? Anatomical, neurophysiological, and imaging data confirm a role of S1 cortex in pain
processing. Overall, the findings support the traditional view that S1 is primarily involved in discriminative aspects of somatic sensation
and extends this view to include discriminative aspects of somatic stimulation that is potentially tissue-damaging, e.g., painful. Single
neurons in monkey S1 code stimulus intensity, location, and duration, and their activity correlates with human perception. Human imaging
studies show activation of S1 by a range of noxious stimuli, including capsaicin, which selectively activates C fibers. These studies also
confirm the somatotopic organization of S1 pain responses, thus supporting the role of S1 in pain localization. Other imaging data that
implicate S1 in the sensory aspect of pain perception are findings that S1 activation is modulated by cognitive manipulations that alter
perceived pain intensity but not by manipulations that alter unpleasantness, independent of pain intensity. Nevertheless, despite the
probable role of S1 in the encoding of the various sensory features of pain, a considerable amount of evidence suggests that nociceptive
input to S1 may also serve to modulate tactile perception, described by Apkarian et al. as a touch gate ( 22 ). Thus, S1 cortex may be
involved in both perception and modulation of both painful and nonpainful somatosensory sensations.
We wish to express our appreciation to Ms. Francine Blanger for her help in preparing this manuscript and to Dr. Pierre Rainville for
his intellectual and technical contributions to these studies. Imaging studies were performed at the Montreal Neurological Institute with the
help and expertise of the its staff. This research is supported by operating grants from the Canadian Medical Research Council awarded to
M.C.B. and to G.H.D.
1. Head, H. & Holmes, G. ( 1911 ) Brain 34 , 102254 .
2. Penfield, W. & Boldrey, E. ( 1937 ) Brain 60 , 389443 .
3. Kenshalo, D. R., Jr. , & Isensee, O. ( 1983 ) J. Neurophysiol. 50 , 14791496 .
4. Kenshalo, D. R., Jr. , Chudler, E. H. , Anton, F. & Dubner, R. ( 1988 ) Brain Res. 454 , 378382 .
5. White, J. C. & Sweet, W. H. ( 1969 ) Pain and the Neurosurgeon: A Forty-Year Experience ( Thomas , Springfield, IL ).
6. Young, G. B. & Blume, W. T. ( 1983 ) Brain 106 , 537554 .
7. Gingold, S. I. , Greenspan, J. D. & Apkarian, A. V. ( 1991 ) J. Comp. Neurol. 308 , 467490 .
8. Kenshalo, D. R., Jr. , Giesler, G. J. , Leonard, R. B. & Willis, W. D. ( 1980 ) J. Neurophysiol. 43 , 15941614 .
9. Rausell, E. & Jones, E. G. ( 1991 ) J. Neurosci. 11 , 210225 .
10. Chudler, E. H. , Anton, F. , Dubner, R. & Kenshalo, D. R. , Jr. ( 1990 ) J. Neurophysiol. 63 , 559569 .
11. Tommerdahl, M. , Delemos, K. A , Vierck, C. J., Jr. , Favorov, O. V. & Whitsel, B. L. ( 1996 ) J. Neurophysiol. 75 , 26622670 .
12. Talbot, J. D. , Marrett, S. , Evans, A. C. , Meyer, E. , Bushnell, M. C. & Duncan, G. H. ( 1991 ) Science 251 , 13551358 .
13. Jones, A K. P. , Brown, W. D. , Friston, K. J. , Qi, L. Y. & Frackowiak, R. S. J. ( 1991 ) Proc. R. Soc. London Ser. B 244 , 3944 .
14. Apkarian A. V. , Stea R. A , Manglos S. H. , Szeverenyi N. M. , King R. B. & Thomas F. D. ( 1992 ) Neurosci. Lett. 140 , 141147 .
15. Jones, A. K. P. , Friston, K. & Frackowiak, R. S. J. ( 1992 ) Science 255 , 215215 .
16. Jones, A. K. P. & Derbyshire, S. W. G. ( 1996 ) Ann. Rheum. Dis. 55 , 411420 .
17. Rainville, P. , Duncan, G. H. , Price, D. D. , Carrier, B. , and Bushnell, M. C. ( 1997 ) Science 227 , 968971 .
18. Meyer, E. , Ferguson, S. G. , Zatorre, R. J. , Alivisatos, B. , Marrett, S. , Evans, A. C. & Hakim, A. M. ( 1991 ) Ann. Neurol. 29 , 440443 .
19. Hyvrinen, J. , Poranen, A. & Jokinen, Y. ( 1980 ) J. Neurophysiol. 43 , 870882 .
20. Poranen, A. & Hyvrinen, J. ( 1982 ) Electroencephalogr. Clin. Neurophysiol. 53 , 535537 .
21. Andersson, J. L. R. , Lilja, A , Hartvig, P. , Lngstrm, B. , Gordh, T. , Handwerker, H. & Torebjrk, E. ( 1997 ) Exp. Brain Res. 117 , 192199 .
22. Apkarian, A. V. , Stea, R. A. & Bolanowski, S. J. ( 1994 ) Somatosens. Mot. Res. 11 , 259267 .
23. Dougherty, P. M. , Schwartz, A. & Lenz, F. A. ( 1998 ) Neuroscience , 90 , 13771392 .
24. Crawford, H. J. , Gur, R. C. , Skolnick, B. , Gur, R. E. & Benson, D. M. ( 1993 ) Int. J. Psychophysiol. 15 , 181195 .
25. Coghill, R. C. , Talbot, J. D. , Evans, A. C. , Meyer, E. , Gjedde, A , Bushnell, M. C. & Duncan, G. H. ( 1994 ) J. Neurosci. 14 , 40954108 .
26. Di Piero, V. , Ferracuti, S. , Sabatini, U. , Pantano, P. , Cruccu, G. & Lenzi, G. L. ( 1994 ) Pain 56 , 167173 .
27. Derbyshire, S. W. G. , Jones, A. K. P. , Devani, P. , Friston, K. J. , Feinmann, C. , Harris, M. , Pearce, S. , Watson, J. D. G. & Frackowiak, R. S. J. (
1994 ) J. Neurol. Neurosurg. Psychiatry 57 , 11661172 .
28. Rosen, S. D. , Paulesu, E. , Frith, C. D. , Frackowiak, R. S. J. , Davies, G. J. , Jones, T. & Camici, P. G. ( 1994 ) Lancet 344 , 147150 .
29. Hsieh, J. C. , Belfrage, M. , Stone-Elander, S. , Hansson, P. & Ingvar, M. ( 1995 ) Pain 63 , 225236 .
30. Hsieh, J. C. , Sthle-Bckdahl, M. , Hgermark , Stone-Elander, S. , Rosenquist, G. & Ingvar, M. ( 1995 ) Pain 64 , 303314 .
31. Davis, K. D. , Wood, M. L. , Crawley, A. P. & Mikulis, D. J. ( 1995 ) NeuroReport 7 , 321325 .
32. Weiller, C. , May, A , Limmroth, V. , Juptner, M. , Kaube, H. , Schayck, R. v. , Coenen, H. H. & Diener, H. C. ( 1995 ) Nat. Med. 1 , 658660 .
33. Howland, E. W. , Wakai, R. T. , Mjaanes, B. A. , Balog, J. P. & Cleeland, C. S. ( 1995 ) Cognit. Brain Res. 2 , 165172 .
34. Kitamura, Y. , Kakigi, R. , Hoshiyama, M. , Koyama, S. , Shimojo, M. & Watanabe, S. ( 1995 ) Electroencephalogr. Clin. Neurophysiol. 95 , 463
474 .
35. Craig, A. D. , Reiman, E. M. , Evans, A. C. & Bushnell, M. C. ( 1996 ) Nature (London) 384 , 258260 .
36. Casey K. L. , Minoshima, S. , Morrow T. J. & Koeppe, R. A. ( 1996 ) Neurophysiology 76 , 571581 .
37. Hsieh, J. C. , Hannerz, J. & Ingvar, M. ( 1996 ) Pain 67 , 5968 .
38. Antognini, J. F. , Buonocore, M. H. , Disbrow, E. A. & Carstens, E. ( 1997 ) Life Sci. 61 , 349354 .
39. Aziz, Q. , Andersson, J. L. R. , Valind, S. , Sundin, A. , Hamdy, S. , Jones, A. K. P. , Foster, E. R. , Langstrom, B. & Thompson, D. G. ( 1997 )
Gastroenterology 113 , 5059 .
40. Di Piero, V. , Fiacco, F. , Tombari, D. & Pantano, P. ( 1997 ) Pain 70 , 185191 .
41. Silverman, D. H. S. , Munakata, J. A , Ennes, H. , Mandelkern, M. A. , Hoh, C. K. & Mayer, E. A ( 1997 ) Gastroenterology 112 , 6472 .
42. Svensson, P. , Rosenberg, B. , Beydoun, A. , Morrow, T. J. , and Casey, K. L. ( 1997 ) Somatosens. Mot. Res. 14 , 113118 .
43. Svensson, P. , Beydoun, A. , Morrow, T. J. & Casey, K. L. ( 1997 ) Exp. Brain Res. 114 , 390392 .
44. Xu, X. P. , Fukuyama, H. , Yazawa, S. , Mima, T. , Hanakawa, T. , Magata, Y. , Kanda, M. , Fujiwara, N. , Shindo, K. , Nagamine, T. & Shibasaki,
H. ( 1997 ) NeuroReport 8 , 555559 .
45. Binkofski, F. , Schnitzler, A , Enck, P. , Frieling, T. , Posse, S. , Seitz, R. J. & Freund, H.-J. ( 1998 ) Ann. Neurol. 44 , 811815 .
46. Derbyshire, S. W. G. , Vogt, B. A. & Jones, A. K. P. ( 1998 ) Exp. Brain Res. 118 , 5260 .
the authoritative version for attribution.

47. Iadarola, M. J. , Berman, K. F. , Zeffiro, T. A , Byas-Smith, M. G. , Gracely, R. H. , Max, M. B. & Bennett, G. J. ( 1998 ) Brain 121 , 931947 .
48. May, A , Kaube, H. , Bchel, C. , Eichten, C. , Rijntjes, M. , Jptner, M. , Weiller, C. & Diener, H. C. ( 1998 ) Pain 74 , 6166 .
49. Oshiro, Y. , Fuijita, N. , Tanaka, H. , Hirabuki, N. , Nakamura, H. & Yoshiya, I. ( 1998 ) NeuroReport 9 , 22852289 .
50. Paulson P. E. , Monoshima, S. , Morrow, T. J. & Casey, K. L. ( 1998 ) Pain 76 , 223229 .
51. Hsieh, J.-C . ( 1995 ) Ph.D. thesis ( Karolinska Institute , Stockholm ).
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

IMMUNE-TO-BRAIN COMMUNICATION IN SICKNESS 7710

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Implications of immune-to-brain communication for sickness and

pain

LINDA R. WATKINS* AND STEVEN F. MAIER

Department of Psychology, University of Colorado, Boulder, CO 80309
ABSTRACT This review presents a view of hyperalgesia and allodynia not typical of the field as a whole. That is, exaggerated
pain is presented as one of many natural consequences of peripheral infection and injury. The constellation of changes that results
from such immune challenges is called the sickness response. This sickness response results from immune-to-brain communication
initiated by proinflammatory cytokines released by activated immune cells. In response to signals it receives from the immune
system, the brain orchestrates the broad array of physiological, behavioral, and hormonal changes that comprise the sickness
response. The neurocircuitry and neurochemistry of sickness-induced hyperalgesia are described. One focus of this discussion is
on the evidence that spinal cord microglia and astrocytes are key mediators of sickness-induced hyperalgesia. Last, evidence is
presented that hyperalgesia and allodynia also result from direct immune activation, rather than neural activation, of these same
spinal cord glia. Such glial activation is induced by viruses such as HIV-1 that are known to invade the central nervous system.
Implications of exaggerated pain states created by peripheral and central immune activation are discussed.
Hyperalgesia and allodynia generally are viewed as purely neural phenomena that reflect changes in spinal cord dorsal horn neuronal
excitability brought about by changes in afferent inputs. The pharmacology of exaggerated pain states also typically is viewed in purely
neural terms, involving substances either released from sensory and/or centrifugal afferents of dorsal horn neurons or, like nitric oxide,
from the dorsal horn neurons themselves. This paper will present a different view. The work to be reviewed illustrates that non-neuronal
cells also can drive hyperalgesic and allodynic states. These non-neuronal cells are immune cells in the periphery and glia within the brain
and spinal cord. Substances released by these immune and immune-like cells can dramatically alter pain processing.
Until recently, the central nervous system and immune system were thought to operate independently of each other. However, they do
not. The first ideas about the dynamic inter-relationships of these two system arose from studies examining the cascade of events initiated
by exposure to stressors ( 1 , 2 ). Stress activates neural circuits in the brain. These stress-induced alterations in brain activity lead to
activation of brain-controlled outflow pathways to the periphery, such as the hypothalamo-pituitary-adrenal axis and sympathetic nervous
system. The hormones and transmitters released by these outflow pathways turned out to bind receptors expressed by immune cells and
immune organs, thereby dramatically altering immune function ( 1 , 2 ). Thus, the central nervous system proved to regulate immune
function.
Within just the past few years, it has been recognized that the inter-relationship between the central nervous system and immune
system is, in fact, bidirectional ( 1 ). That is, products of activated immune cells feed back to the brain to alter neural activity. The sections
that follow focus first on the broad view of how and why the immune system communicates to the brain. The manner in which immune-to-
brain communication impacts the pain response then will be explored. Finally, the role of immune-like glia in the spinal cord in
exaggerated pain responses will be described and implications discussed.
IMMUNE-TO-BRAIN COMMUNICATION IN SICKNESS
The immune system responds to infection in two related, but differing, ways. One is slow and selective; the other is rapid and
generalized ( 3 ). The slow response involves recognition of foreign invaders such as bacteria and viruses through binding to specific
receptors expressed on specialized types of immune cells, resulting in the slow and prolonged production of antibodies directed
specifically against that particular foreign entity. The other, very rapid and generalized, response is referred to as the sickness response or,
alternatively, as the acute phase response ( 4 ). This sickness response is triggered by the recognition of anything foreign to the host. It
serves as a rapid early defense mechanism until the much slower antibody response can be developed. The sickness response is an
organized constellation of responses initiated by the immune system but orchestrated and partially created by the brain ( 1 , 4 , 5 ). The
sickness response includes physiological responses (fever, alterations in plasma ions to suppress minerals required by bacteria/viruses to
replicate, increases in white blood cell replication, increased sleep, etc.), behavioral responses (decreased social interaction and
exploration, decreased sexual activity, decreased food and water intake, etc.), and hormonal responses (increased release of classic
hypothalamo-pituitary-adrenal and sympathetic hormones). It has been argued that much of this constellation of changes is in the service
of fever. Fever is a phylogenetically very old response that raises the core body temperature to the point where bacteria/viruses do not
multiple rapidly, bacteria cannot form protective outer coats, where white blood cells do multiple very rapidly, destructive enzymes key
for survival function most effectively, and so on. Every degree of fever requires a 1015% increase in energy, and most of the components
of the sickness response can be viewed as supporting this energetic requirement by either creating energy (hormones released by sickness
free energy from bodily stores) or saving energy (increased sleep, decreased exploration and sex, decreased foraging/stalking, etc.) ( 1 , 4 ,
5 ). This sickness response requires that immune
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

* To whom reprint requests should be addressed at: Department of Psychology, Campus Box 345, University of Colorado, Boulder, CO
80309-0345. e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

PAIN AS A NATURAL OUTCOME OF IMMUNE-TO-BRAIN COMMUNICATION 7711

to-brain communication must occur because only the brain can orchestrate such a pervasive array of changes.
The triggers for initiation of the sickness response are substances released by immune cells activated by the foreign entity. As a
group, these proteins are referred to as proinflammatory cytokines ( 3 ). This name reflects the fact that these proteins orchestrate and
augment inflammatory responses. Proinflammatory cytokines include IL-1, IL-6, and tumor necrosis factor. These proteins may be
necessary and sufficient for sickness, because preventing their actions (using receptor antagonists, etc.) block sickness responses, whereas
exogenous administration of these proteins can create sickness responses ( 1 , 5 ).
PAIN AS A NATURAL OUTCOME OF IMMUNE-TO-BRAIN COMMUNICATION
The constellation of responses reviewed above constitute the classic view of sickness. Although changes in pain responsivity have not
been considered in this classic view, it would be reasonable to suspect that hyperalgesia might be a natural part of this response profile
because recuperative behaviors supportive of healing would be produced by enhanced pain ( 6 ). Furthermore, recuperative behaviors
would be expected to decrease activity and so decrease energy expenditure, again in keeping with the view that sickness responses serve to
save energy for use in the production of fever ( 5 ).
If such an argument has merit, then one would expect that hyperalgesia should occur after administration of agents known to induce
sickness. In fact, hyperalgesia is produced both by i.p. administration of the cell walls of Gram-negative bacteria (endotoxin; also called
lipopolysaccharide) ( 7 ) and by i.p. live bacteria ( 8 ), both of which are known to elicit the release of proinflammatory cytokines from a
variety of immune cells. Hyperalgesia also can be elicited simply by administering either IL-1 or tumor necrosis factor alone ( 9 , 10 ). The
key importance of proinflammatory cytokines in sickness hyperalgesia is clear from the fact that it can be blocked by either an IL-1
receptor antagonist or tumor necrosis factor binding protein ( 9 , 11 ). Thus, these cytokines can be both necessary and sufficient for
sickness-induced hyperalgesia to occur.
The mechanism(s) by which proinflammatory cytokines activate the central nervous system is a matter of lively controversy. Both
blood-borne and peripheral nerve-generated signals have been proposed ( 12 ). For localized tissue infection/inflammation and localized
proinflammatory cytokine administration, at least, activation of peripheral nerves appears to be the most likely route ( 12 14 ). By using
i.p. administration of sickness agents as the example, signals to the brain appear to be carried via the subdiaphragmatic vagus because: (a)
sickness agents activate sensory vagal neurons, as supported by increases in cFos expression in these cells ( 15 ), and (b) subdiaphragmatic
vagotomy blocks hyperalgesia induced by endotoxin, IL-1, and tumor necrosis factor ( 9 , 10 ). Activation of the sensory vagus may be
secondary to activation of specialized sensory structures called paraganglia ( 16 ). These sensory cells are thought to be chemoreceptors,
based on their anatomy. They express IL-1 binding sites and are known to form afferent synapses onto sensory vagal fibers. Intriguingly,
these IL-1 binding paraganglia are in close physical proximity to dense populations of immune cells that can create and release IL-1 ( 17 ).
These accumulations of macrophages, mast cells, and dendritic cells embedded in connective tissue near paraganglia have been referred to
as nerve-associated lymphoid cells (NALC) ( 17 ). The location of these immune cells and their ability to express IL-1 suggest that this
NALC may rapidly recognize infection and signal the brain via IL-1 release onto neighboring paraganglia. In support of this notion, we
have found that levels of IL-1 in the NALC associated with the paraganglia rapidly and dramatically increase after i.p. illness agents ( 17 ).
Thus, this arrangement of peripheral cells may underlie immune-to-brain communication arising from the abdomen.
The pathway by which abdominal sickness signals elicit hyperalgesia has been at least partially mapped in the central nervous
system, by using a combination of discrete lesions and expression of cFos, an immediate-early gene product used as a neuronal activation
marker ( 18 ). From this work, the neural circuitry underlying sickness hyperalgesia was found to involve a nucleus tractus solitarius
nucleus raphe magnusspinal cord dorsolateral funiculus circuit. The involvement of the nucleus tractus solitarius is notable, in that this
appears to be a common hub within the brain for creating sickness responses ( 19 ).
Whether hyperalgesia induced by peripheral inflammation is mediated by the same general pathway is not yet clear. Although s.c.
inflammation (formalin) hyperalgesia is not affected by subdiaphragmatic vagotomy ( 10 ), it does require a brain-to-spinal cord circuit
that, like sickness hyperalgesia, involves the nucleus raphe magnus ( 20 ). Whether the nucleus tractus solitarius is involved is uncertain.
Although unilateral lesions of the nucleus tractus solitarius failed to affect s.c. formalin hyperalgesia ( 21 ), this may well be because
dorsal horn laminae responsive to s.c. formalin send bilateral projections to this medullary structure ( 22 ). Bilateral lesions could not be
tested given the survival problems such animals face. In addition to at least partial overlap in the neurocircuitry of inflammation- and
sickness-induced hyperalgesias, there are similarities in their neurochemistry as well. Both depend on nitric oxide, excitatory amino acids,
and substance P at the level of the spinal cord ( 20 ).
Role of Spinal Cord Glia in Exaggerated Pain
From the discussion above, it is clear that peripheral infection/ inflammation leads to activation of a brain-to-spinal cord pathway,
culminating in the creation of hyperalgesia. However, an intriguing aspect of this spinal circuitry is that it critically depends on activation
of spinal cord microglia and astrocytes. Indeed, hyperalgesia produced either by s.c. inflammation ( 23 ) or i.p. bacterial infection ( 8 ) can
be blocked by spinal administration of drugs that disrupt glial function. Further, anatomical examination of astrocytes and microglia show
them to be clearly activated by peripheral infection/inflammation, as evidenced immunohistochemically by increased expression of glia-
specific activation markers ( 8 ). Lastly, i.p. endotoxin at the same dose that elicits hyperalgesia rapidly increases dorsal spinal cord levels
of IL-1, a product of glial activation ( 24 ).
So how are these glia activated and what role do they play in exaggerated pain states? Regarding activation, the glia may be activated
by neurotransmitter(s) released by spinal projections of the nucleus raphe magnus. Candidate neurotransmitters to serve this role are
substance P and glutamate, because: (a) as noted above, both sickness-induced hyperalgesia and s.c. formalin hyperalgesia are mediated,
in part, by spinal cord substance P and excitatory amino acids, (b) the nucleus raphe magnus-to-spinal cord pathway mediating both
sickness-induced hyperalgesia and s.c. formalin hyperalgesia contains substance P and glutamate as neurotransmitters ( 25 ), (c) microglia
and astrocytes express substance P and glutamate receptors ( 26 ), and (d) glia are activated by substance P and glutamate in vitro ( 27 ).
Once activated, astrocytes and microglia form a positive feedback circuit whereby substances released from microglia activate astrocytes
to release substances that further stimulate microglia, and so forth ( 28 ). Many of the substances that can be released from microglia and
astrocytes are known to be key mediators of hyperalgesia, including nitric oxide, excitatory amino acids [both N-methyl-D-aspartate
the authoritative version for attribution.

(NMDA) and non-NMDA agonists], IL-1, pros

lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CONCLUSIONS AND IMPLICATIONS 7712

taglandins, and nerve growth factor ( 28 ). Thus, once spinal cord microglia and astrocytes are activated in a perseverative positive
feedback manner, the neuroexcitatory substances they release could drive exaggerated pain states.
However, astrocytes and microglia have not generally been viewed as cells whose major function is activation in response to
centrifugal hyperalgesia circuitry. Rather, astrocytes and microglia are immunocompetent cells and thus can respond like immune cells
within the central nervous system. Astrocytes and microglia express specific receptors for various bacteria and viruses and are activated on
binding to these infectious agents. An example of a neurotropic virus (that is, a virus that can home to the brain and spinal cord) is
HIV-1, which causes AIDS. HIV-1 invades the brain and spinal cord early in, and continuing throughout, disease progression ( 29 ) and
this invasion leads to the activation of microglia and astrocytes ( 30 ). One reason for the prolonged microglial and astrocyte activation by
HIV-1 is that currently available drugs used to treat AIDS do not readily penetrate the blood-brain barrier, so HIV-1 within the brain and
spinal cord are not disrupted by such treatments. Once within nervous tissue, HIV-1 binds to receptors on microglia and astrocytes that
recognize one specific portion of the virus, namely a glycoprotein (called gp120) expressed on the outer surface of the viral coat ( 31 ).
The arguments developed above predict that intrathecal delivery of gp120 should be sufficient to produce hyperalgesia, if immune
activation of glia initiates the same sort of positive feedback cascade as occurs for sickness-induced hyperalgesia. In our initial series of
studies of this issue, we found that gp120 delivered over lumbosacral cord caused dose-dependent hyperalgesia as measured by the tail-
flick test. Because the receptors identified on microglia and astrocytes require the complex three-dimensional conformational structure of
native gp120 for receptor activation to occur ( 32 ), it follows that irreversible heat denaturation of gp120s natural conformation should
disrupt the ability of this protein to create hyperalgesia, if microglia and astrocytes function as we hypothesize. In fact, such disruption of
the normal three-dimensional structure of gp120 dose abolish the effects of this viral protein on pain.
Results using the tail-flick test can, at times, be confounded by drug-induced alterations in tail skin temperature ( 33 ). We have
examined this issue by using two independent approaches. First, we monitored superficial tail temperature throughout tail-flick testing and
found that there was no correlation between superficial tail temperature and tail-flick latency ( 24 ). Second, we examined the effect of
intrathecal gp120 on withdrawal latency of the plantar surface of the hind paws in response to a radiant heat stimulus. This experiment was
important because the paws, unlike the tail, are not used to regulate the organisms temperature. Therefore, paw withdrawal latencies are
not subject to the skin temperature confounds inherent in the tail-flick test. Here again, robust hyperalgesia was observed, supporting the
conclusion that intrathecal gp120 produces thermal hyperalgesia ( 34 ).
Intrathecal gp120 produces mechanical allodynia as well as thermal hyperalgesia. The first indication of this effect came from pilot
studies in which we observed marked increases in vocalization of gp120-injected rats to light touch, compared with vehicle controls. These
initial observations were followed by experiments that used two standardized tests. First, we used calibrated VonFrey monofilaments to
examine withdrawal responses elicited from the plantar surface of the hind paws in response to low-threshold mechanical stimuli. This
procedure clearly demonstrated that gp120 induces allodynia ( 34 ). Second, we used light touch stimuli to the fur of the hindquarters and
found that gp120 induced touch-evoked agitation as well ( 34 ).
Importantly, activation of spinal cord microglia and astrocytes appear to be critical for the hyperalgesic and allodynic effects of
gp120. Pilot studies using immunohistochemistry suggest activation of glia in spinal cord after gp120. Furthermore, pretreatment of the
rats with a drug that disrupts glial function prevented both gp120-induced thermal hyperalgesia and mechanical allodynia ( 24 , 34 ).
The mechanisms underlying these effects are at present unknown. We are actively investigating this issue and predict that the effects
will be similar to those previously defined for sickness-induced hyperalgesia. Recall that IL-1 was previously noted to rapidly increase in
dorsal spinal cord after i.p. endotoxin, and that spinal cord IL-1 is a key mediator of hyperalgesia induced by peripheral inflammation.
Thus, our initial pilot studies have focused on potential gp120-induced changes in spinal cord IL-1. To date, these data indicate that
intrathecal gp120 produces a rapid and dramatic increase in dorsal spinal cord IL-1 mRNA and IL-1 protein ( 34 ). This gp120-induced
increase in IL-1 protein does not simply reflect intracellular content, but rather is indicative of increased IL-1 release into extracellular
space, because levels of IL-1 in lumbosacral cerebrospinal fluid dramatically increase as well ( 34 ). Although we have not yet directly
tested the effect of IL-1 receptor antagonist on gp120-induced effects, these preliminary data are certainly consistent with the view that
gp120-induced alterations in glial function are likely to be important for hyperalgesia and/or allodynia.
CONCLUSIONS AND IMPLICATIONS
The evidence reviewed above places exaggerated pain in a new framework. This view conceptualizes hyperalgesia not as an entity in
and of itself, but rather as a single element in a much larger constellation of physiological, behavioral, and hormonal changes, orchestrated
by the central nervous system in response to bodily infection and inflammation. This sickness response enhances survival of the organism
in the face of immune challenges. Hyperalgesia, like all of the other components of the sickness response, is triggered by proinflammatory
cytokines released by activated macrophages and other immune cells. For localized immune challenges, at least, this proinflammatory
cytokine release leads to activation of peripheral nerves that signal the brain. Activation of a centrifugal pathway then occurs, resulting in
the activation of microglia and astrocytes within the spinal cord dorsal horn. Neurotransmitters released by the centrifugal pathway
combined with neuroexcitatory substances released by astrocytes and microglia create exaggerated pain responses. Thus, for this form of
hyperalgesia at least, glia assume a new, pivotal role in the generation of exaggerated pain.
The importance of microglia and astrocytes in spinal cord hyperalgesia mechanisms has potentially important implications. Most
important of these is that glia may exaggerate pain when these immune-like cells respond to infection/ inflammation within the central
nervous system. Clearly, this is the case after intrathecal administration of the HIV-1 envelope protein, gp120. Given that a number of
viruses and bacteria can invade the central nervous system, such observations suggest that glia may have a far more important role in pain
modulation than previously recognized.
This work was supported by National Institutes of Health Grants MH55283, MH01558, MH00314, and MH45045 and the Council of
Research and Creative Works of the University of Colorado at Boulder.
the authoritative version for attribution.

Sham, H., Kempf, D., Molla, A., Marsh, K., Betebenner, D., Chen, X., Rosenbrook, W., Wideburg, N., Chen, C, Kati, W., et al.,
Fourth Conference on Retroviruses and Opportunistic Infections, Jan. 22 26, 1997, Washington, DC, abstr. 396.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CONCLUSIONS AND IMPLICATIONS 7713

1. Maier, S. F. & Watkins, L. R. ( 1998 ) Psychol. Rev. 105 , 83107 .

2. Maier, S. F. , Watkins, L. R. & Fleshner, M. ( 1994 ) Am. Psychol. 49 , 10041018 .
3. Kuby, J. ( 1992 ) Immunology ( Freeman , New York ).
4. Hart, B. L. ( 1988 ) Neurosci. Biobehav. Rev. 12 , 123137 .
5. Kent, S. , Bluthe, R.-M. , Kelley, K. W. & Dantzer, R. ( 1992 ) Trends Pharmacol. Sci. 13 , 2428 .
6. Watkins, L. R. , Maier, S. F. & Goehler, L. E. ( 1995 ) Pain 63 , 289302 .
7. Mason, P. ( 1993 ) Neurosci. Lett. 154 , 134136 .
8. Watkins, L. R. , Deak, T. , Silbert, L. , Martinez, J. , Goehler, L. , Relton, J. , Martin, D. & Maier, S. F. ( 1995 ) Proc. Soc. Neurosci. 21 , 357.1
(Abstr.) .
9. Watkins, L. R. , Wiertelak, E. P. , Goehler, L. E. , Smith, K. P. , Martin, D. & Maier, S. F. ( 1994 ) Brain Res. 654 , 1526 .
10. Watkins, L. R. , Goehler, L. E. , Relton, J. , Brewer, M. T. & Maier, S. F. ( 1995 ) Brain Res. 692 , 244250 .
11. Maier, S. F. , Wiertelak, E. P. , Martin, D. & Watkins, L. R. ( 1993 ) Brain Res. 623 , 321324 .
12. Watkins, L. R. , Maier, S. F. & Goehler, L. E. ( 1995 ) Life Sci. 57 , 10111026 .
13. Sommer, C , Lindenlaub, L. , Petrausch, S. & George, A. ( 1998 ) Proc. Soc. Neurosci. 24 , 495.4 (Abstr.) .
14. Sorkin, L. S. , Xiao, W.-H. , Wagner, R. & Myers, R. R. ( 1997 ) Neuroscience 81 , 255262 .
15. Goehler, L. E. , Gaykema, R. P. A. , Hammack, S. E. , Maier, S. F. & Watkins, L. R. ( 1998 ) Brain Res. 804 , 306310 .
16. Goehler, L. E. , Relton, J. K. , Dripps, D. , Kiechle, R. , Tartaglia, N. , Maier, S. F. & Watkins, L. R. ( 1997 ) Brain Res. Bull. 43 , 357364 .
17. Goehler, L. E. , Gaykema, R. P. A. , Nguyen, K. T. , Lee, J. E. , Tilders, F. J. H. , Maier, S. F. & Watkins, L. R. ( 1999 ) J. Neurosci. , 19 , 2799
2806 .
18. Watkins, L. R. , Wiertelak, E. P. , Goehler, L. , Mooney-Heiberger, K. , Martinez, J. , Furness, L. , Smith, K. P. & Maier, S. F. ( 1994 ) Brain Res.
639 , 283299 .
19. Ritter, S. , Ritter, R. C. & Barnes, C. D. ( 1992 ) Neuroanatomy and Physiology of Abdominal Vagal Afferents ( CRC, Ann Arbor, MI ).
20. Wiertelak, E. P. , Furness, L. E. , Horan, R. , Martinez, J. , Maier, S. F. & Watkins, L. R. ( 1994 ) Brain Res. 649 , 1926 .
21. Wiertelak, E. P. , Roemer, B. , Maier, S. F. & Watkins, L. R. ( 1996 ) Brain Res. 748 , 143150 .
22. Menetrey, D. & Basbaum, A. I. ( 1987 ) J. Comp. Neurol. 255 , 439450 .
23. Watkins, L. R. , Martin, D. , Ulrich, P. , Tracey, K. J. & Maier, S. F. ( 1997 ) Pain 71 , 225235 .
24. Watkins, L. R. & Maier, S. F. , eds. ( 1999 ) Cytokines and Pain ( Birkhauser , Amsterdam ), pp. 3958 .
25. Menetrey, D. & Basbaum, A. I. ( 1987 ) Neuroscience 23 , 173187 .
26. Porter, J. T. & McCarthy, K. D. ( 1997 ) Prog. Neurobiol 51 , 439455 .
27. Martin, F. C. , Anton, P. A. , Gornbein, J. A. , Shanahan, F. & Merrill, J. E. ( 1993 ) J. Neuroimmunol. 42 , 5360 .
28. Zielasek, J. & Hartung, H.-P. ( 1996 ) Adv. Neuroimmunol. 6 , 191222 .
29. Diederich, N. , Acermann, R. , Jurgens, R. , Ortseifen, M. , Thun, F. , Schneider, M. & Vukadinovic, I. ( 1988 ) Eur. Neurol. 28 , 93103 .
30. Tyor, W. R. , Glas, J. D. , Griffin, J. , Becker, P. S. , McArtheru, J. C. , Bezman, L. & Griffin, D. E. ( 1992 ) Ann. Neurol. 31 , 349360 .
31. Ma, M. , Geiger, J. D. & Nath, A. ( 1994 ) J. Virol. 68 , 68246828 .
32. Sundar, S. K. , Cierpial, M. A. , Kamaraju, L. S. , Long, S. , Hsieh, S. , Lorenz, C. , Aaron, M. , Ritchie, J. C. & Weiss, J. M. ( 1991 ) Proc. Natl.
Acad. Sci. USA 88 , 1124611250 .
33. Tjolsen, A. & Hole, K. ( 1993 ) Am. Pain Soc. J. 2 , 107111 .
34. Milligan, E. D. , Nguyen, K. T. , Hansen, M. , Xu, D. , Lee, J. , Martin, D. , Hinde, J. L. , Mehmert, K. , Bohl, M. , Maier, S. F. & Watkins, L. R. (
1998 ) Proc. Soc. Neurosci. 24 , 155.1 (Abstr.) .
the authoritative version for attribution.
 BRAIN-DERIVED NEUROTROPHIC FACTOR IS AN ENDOGENOUS MODULATOR OF NOCICEPTIVE RESPONSES IN THE 7714
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SPINAL CORD

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Brain-derived neurotrophic factor is an endogenous modulator of

nociceptive responses in the spinal cord

S. W. N. THOMPSON, D. L. H. BENNETT, B. J. KERR, E. J. BRADBURY, AND S. B. MCMAHON *

Division of Physiology, Sherrington Building, St. Thomas Hospital Campus, Lambeth Palace Road, London SE1 7EH, United
Kingdom
ABSTRACT The primary sensory neurons that respond to noxious stimulation and project to the spinal cord are known to
fall into two distinct groups: one sensitive to nerve growth factor and the other sensitive to glial cell-line-derived neurotrophic
factor. There is currently considerable interest in the ways in which these factors may regulate nociceptor properties. Recently,
however, it has emerged that another trophic factorbrain-derived neurotrophic factor (BDNF) may play an important
neuromodulatory role in the dorsal horn of the spinal cord. BDNF meets many of the criteria necessary to establish it as a
neurotransmitter/neuromodulator in small-diameter nociceptive neurons. It is synthesized by these neurons and packaged in
dense core vesicles in nociceptor terminals in the superficial dorsal horn. It is markedly up-regulated in inflammatory conditions
in a nerve growth factor-dependent fashion. Postsynaptic cells in this region express receptors for BDNF. Spinal neurons show
increased excitability to nociceptive inputs after treatment with exogenous BDNF. There are both electrophysiological and
behavioral data showing that antagonism of BDNF at least partially prevents some aspects of central sensitization. Together, these
findings suggest that BDNF may be released from primary sensory nociceptors with activity, particularly in some persistent pain
states, and may then increase the excitability of rostrally projecting second-order systems. BDNF released from nociceptive
terminals may thus contribute to the sensory abnormalities associated with some pathophysiological states, notably inflammatory
conditions.
Studies on null-mutant mice have provided ample evidence for the neurotrophic hypothesisthat discrete groups of neurons receive
from their targets specific trophic support that is essential for the survival of those neurons. For small-diameter primary sensory neurons of
neural-crest origin that project to the spinal cord, a member of the neurotrophin family, nerve growth factor (NGF) seems to fulfil this role
in development. However, as animals mature, these small-diameter neurons, which are virtually all nociceptive in nature, show several
changes in their trophic requirements: first, the cells loose their dependence on NGF for survival, a process that emerges in the first 2
weeks of postnatal life in the rat; second, the same cells change their expression of trophic factor receptors. At birth, virtually all these
small cells express the high-affinity NGF receptor trkA. Again, over the first 2 postnatal weeks, about one-half of the cells down-regulate
trkA and simultaneously up-regulate receptor components for another factor, glial cell-line-derived neurotrophic factor (GDNF). Thus, in
the mature state, the large group of small-diameter dorsal root ganglion (DRG) neurons forms two distinct groups, one sensitive to NGF
and the other sensitive to GDNF ( 1 ). It is therefore not surprising that much effort has been directed at elucidating the biological effects
and roles of these two factors on nociceptive systems ( 2 ).
A third member of the neurotrophin family, brain-derived neurotrophic factor (BDNF), seems to be a target-derived trophic factor for
many placode-derived sensory neurons (e.g., those in the nodose ganglion). Given the modest sensory neuron loss in BDNF null mutant
mice, BDNF does not seem to act in this fashion for the neural-crest-derived sensory neurons. However, BDNF does interact with this
population of sensory neurons in maturity. BDNF is important in mature animals for regulating the mechanosensitivity of slowly adapting
mechanoreceptors, myelinated fibers required for fine tactile discrimination ( 3 ). In addition, there is now evidence accruing that BDNF
has a distinct role in nociceptive processing in the adult nervous system, a role distinct from that played by NGF or GDNF. The evidence
to date suggests that BDNF may function as a central neuromodulator: not only is BDNF synthesized by some nociceptors, but it also may
be released in the spinal cord in an activity-dependent fashion, thereby regulating the excitability of rostrally projecting second-order
systems. This article will consider the evidence relating to the hypothesis that BDNF may function as a central modulator of pain.
BDNF Is Constitutively Expressed Within a Subpopulation of Primary Sensory Neurons. BDNF shares one of the defining
characteristics of the neurotrophin family of molecules in that it is retrogradely transported by responsive sensory neurons to the DRG (
4 ). In addition, BDNF shows an unusual property for this family of molecules in that it is constitutively expressed by adult sensory
neurons ( 5 8 ). Fig. 1 illustrates that, in the lumbar DRG of the rat, the distribution of BDNF protein is not uniform but is localized to a
restricted number of primary sensory neurons. Recently, the type of neurons that synthesize BDNF has been elucidated ( 8 ). Although,
under normal circumstances, BDNF mRNA and protein seem to be present in several different subtypes, the overwhelming contribution is
from those sensory neurons that express the neuropeptide CGRP and the high-affinity NGF receptor trkA. Perhaps surprisingly, cells that
express trkB (the high-affinity receptor for BDNF) do not, for the most part, express either BDNF mRNA or protein ( 8 ). This pattern of
localization has important functional implications, because those sensory neurons that express CGRP and trkA are considered to be
predominantly nociceptive in function. This nociceptive function is supported further by observations on the spinal localization of BDNF.
BDNF is anterogradely transported to both to the periphery and to the spinal cord where it accumulates in the central terminals in the
superficial laminae (I and II) and colocalizes with CGRP-containing arborizations (refs. 8
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: NGF, nerve grown factor; GDNF, glial cell-line-derived neurotrophic factor; DRG, dorsal root ganglion; BDNF, brain-
derived neurotrophic factor; VRP, ventral root potential; NMDA, N-methyl-D-aspartic acid; CGRP, calcitonin-gene-related peptide.
* To whom reprint requests should be addressed. e-mail: [email protected] .
 BRAIN-DERIVED NEUROTROPHIC FACTOR IS AN ENDOGENOUS MODULATOR OF NOCICEPTIVE RESPONSES IN THE 7715
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SPINAL CORD

abundant in intrinsic spinal-cord neurons nor does it coexist and 9 ; see Fig. 1 ). Notably, BDNF immunoreactivity is not within the
primary afferent terminals or cell bodies marked by binding of the lectin IB4. This population of sensory neurons does not constitutively
express neuropeptides but expresses GDNF receptor components and is responsive to this factor ( 10 ). BDNF immunoreactivity is present
only sparingly within deeper laminae. Thus, most of the BDNF protein in spinal cord seems to be localized in the terminals of primary
sensory nociceptors, specifically those that express trkA and are known to be sensitive to NGF ( 11 ). In these sensory neurons, BDNF is
located in dense core vesicles ( 12 ), suggesting that it is likely to be released with activity in nociceptors. The superficial laminae of the
spinal cord are of course, important sites for integration of nociceptive information. Full-length trkB receptors are present in the cells in
this region of the spinal cord ( 13 ).

FIG. 1. BDNF expression within the DRG and spinal dorsal horn, (a and c) Under normal circumstances, modest BDNF
immunoreactivity is observed within lumbar DRG and superficial dorsal horn, (e, bold arrow) Constitutively expressed BDNF
colocalizes with the neuropeptide calcitonin-gene-related peptide (CGRP) and is present mainly within small-diameter sensory
neurons. These neurons mostly do not express trkB. BDNF expression is regulated by NGF. (b and d) After systemic NGF treatment,
BDNF immunoreactivity is enhanced within lumbar ganglia and dorsal horn, (e, light arrow) After NGF treatment, 8090% of trkA-
expressing sensory neurons contain BDNF immunoreactivity. These neurons coexpress CGRP.
The possible interaction of BDNF with nociceptive pathways is highlighted further by the fact that the levels of BDNF are regulated
by another neurotrophin, NGF ( Fig. 1 ). Several recent studies have shown that BDNF levels within sensory neurons may be increased by
exogenous administration of NGF ( 6 , 8 ): NGF increases the number of DRG cells showing BDNF immunoreactivity and mRNA.
Although there is some debate as to which cells express BDNF after NGF treatment, it is apparent that considerable numbers of these
neurons express trkA (and are thus directly responsive to NGF). In one study ( 8 ) for instance, 8090% of those sensory neurons
expressing BDNF immunoreactivity after intrathecal NGF belonged to the trkA group.
The peripheral synthesis of NGF is known to be increased in a variety of inflammatory conditions, particularly after injury or
experimentally induced inflammation. Under such pathophysiological states BDNF may be up-regulated in nociceptors. Cho et al. ( 7 )
showed that intraplantar Freunds adjuvant injection was associated with a significant increase in BDNF mRNA with a time course
consistent with the known increase of NGF in this model. Moreover, the increase in BDNF mRNA was prevented by coinjection of NGF
antibody, suggesting that BDNF expression depends on peripheral NGF production. Because there is good evidence that some of the
sensory abnormalities associated with inflammation depend on NGF production, this finding in consistent with the notion that BDNF itself
contributes to altered sensory processing.
The expression of BDNF is altered in an interesting manner in another pathophysiological state, that associated with peripheral nerve
injury. After sciatic axotomy, BDNF is up-regulated in large-diameter DRG cells, and anterograde transport to the dorsal horn of the spinal
cord increases ( 14 ). Thus, with peripheral nerve injury, BDNF expression decreases in nociceptors but increases in large-diameter
sensory neurons (these large-diameter neurons are generally considered to subserve transmission of innocuous sensations from the
periphery). The signal for the expression in large cells is not yet known, although reduced retrograde transport of neurotrophin-3 by large-
diameter sensory neurons is one obvious possibility. The BDNF in damaged large neurons is also transported to the central terminals of
these fibers. It is tempting to speculate that this BDNF also may have similar postsynaptic effects to those described below. Of course, if
so, such effects are likely to be found in deeper dorsal horn laminae, the known normal termination sites of large afferents.
BDNF Has Postsynaptic Actions Within the Spinal Dorsal Horn. Modulation of spinal reflex excitability. Although there is
evidence for the activity-dependent secretion of neurotrophins from hippocampal slices ( 15 ), there is, as yet, no such evidence that BDNF
is released within the spinal cord after afferent fiber activation. The localization of BDNF and its injury-associated regulation detailed
above, however, are strongly consistent with a possible neuromodulatory role. Recently, we tested this hypothesis further by examining
the effect of exogenous BDNF delivery on spinal reflex excitability.
Because neurotrophins do not penetrate the spinal cord readily after systemic delivery, we used an in vitro spinal-cord preparation to
study BDNFs actions. The in vitro hemisected spinal cord is used routinely in our and other laboratories to monitor alterations in synaptic
excitability. Reflex responses to C fiber inputs may be recorded from the ventral root, close to the motoneuron cell bodies, with close-
fitting glass suction electrodes, and these reflex responses appear as prolonged ventral root potentials (VRPs). These extracellularly
recorded potentials are a good measure of spinal excitability ( 16 ) and, under control conditions, will remain stable for several hours (
Fig. 2 a). These preparations were superfused the with BDNF at 2- 1,000 ng/ml for 30 min. The C fiber-evoked responses were recorded
the authoritative version for attribution.

before BDNF exposure and for several hours after cessation of BDNF superfusion. A significant and sustained increase in the C fiber-
evoked reflex response was observed repeatedly ( Fig. 2 b; refs. 17 and 18 ). This
 BRAIN-DERIVED NEUROTROPHIC FACTOR IS AN ENDOGENOUS MODULATOR OF NOCICEPTIVE RESPONSES IN THE 7716
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SPINAL CORD

enhancement of C fiber-evoked excitability was sustained, outlasting the duration of BDNF administration, and was dosedependent. The
threshold dose to elicit an increased spinal excitability was in the range 100200 ng/ml BDNF, which suggests a specific and receptor-
mediated event. Coadministration of the BDNF-sequestering molecule trkB-IgG to spinal-cord preparations eliminates the facilitatory
effect of BDNF and further indicates that the effect of BDNF is likely to be mediated by interaction at the trkB receptor. Supervision of
spinal-cord preparations with NGF (200 ng/ml) does not produce similar effects on C fiber-evoked reflex excitability.

FIG. 2. Exogenous BDNF enhances postsynaptic activity in the spinal cord and endogenous BDNF contributes to reflex excitability,
(a) Spinal reflex excitability was assessed from the VRP evoked after electrical stimulation of C fibers in the dorsal root of an
isolated hemisected rat spinal-cord preparation. Under normal circumstances, the VRP evoked by C fiber activation is characterized
by a prolonged, slowly decaying potential, and this potential remains stable from trial to trial, (b) After a 30-min superfusion with
BDNF (200 ng/ml), the amplitude and duration of the C fiber-evoked VRP increased significantly relative to pretreatment activity
(arrow, control trace; arrowhead, trace after BDNF superfusion). (c) The contribution of endogenous BDNF to reflex excitability was
assessed after a 30-min superfusion of spinal-cord preparations with trkB-IgG (500 ng/ml). Under these circumstances C fiber-
evoked activity was reduced significantly compared with pretreatment control responses (arrow, control response; arrowhead,
response after trkB-IgG superfusion). (b and c; bars = 0.5 mV and 5.0 s.)
The mechanism of action of BDNF within the spinal cord is, at present, unknown. There is evidence from primary cultures of
hippocampal neurons that BDNF enhances spontaneous release of glutamate and acetylcholine, suggesting that BDNF functions there as a
retrograde modulator of presynaptic transmitter release ( 19 ). It is unlikely that such a mechanism will operate within the spinal cord,
because those sensory neurons that express BDNF do not in general display the trkB receptor and are therefore likely to be insensitive to
the synaptically released neurotrophin. There is also good evidence, again from the hippocampus, that postsynaptic injection of protein
kinase inhibitors will block the effect of BDNF on synaptic enhancement, suggesting that phosphorylation of postsynaptic receptors is a
critical step in the BDNF effect. Recently, it has been shown that BDNF rapidly and specifically enhances phosphorylation of the
postsynaptic N-methyl-D-aspartic acid (NMDA) receptor ( 20 ). Again in the hippocampus, BDNF will potentiate NMDA responses via a
3-fold increase in NMDA receptor open time ( 20 ). In cultured rat neurons in the central nervous system, BDNF will also induce changes
in NR2A and NR2B NMDA receptor subunit expression ( 21 ). NR2 NMDA receptor subunits define the pharmacological and biophysical
properties of this receptor ( 22 ). In the spinal cord, NMDA receptor activation plays a central role in the induction and maintenance of
central sensitization. In this well studied phenomenon, the recruitment of the NMDA receptor seems to

FIG. 3. BDNF induces alterations in immediate early gene expression within the spinal cord. (a) A large increase in c-fos
immunoreactivity is observed within the superficial dorsal horn 2 h after intrathecal administration of BDNF (5 g). (b) This increase
in immunoreactivity is not observed after intrathecal administration of NGF or saline.
the authoritative version for attribution.
 BRAIN-DERIVED NEUROTROPHIC FACTOR IS AN ENDOGENOUS MODULATOR OF NOCICEPTIVE RESPONSES IN THE 7717
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SPINAL CORD

be the pivotal event in increasing the sensitivity of nociceptive spinal circuits to sensory inputs ( 23 ).

FIG. 4. Endogenous BDNF release contributes to nociceptive behavioral responses in vivo. Animals were pretreated with NGF to
maximize nociceptor BDNF levels. Intraplantar formalin [50 l; 2% (vol/vol)] induces a characteristic two-phase nocifensor paw-
withdrawal behavior. Intrathecal administration of trkB-IgG (but not saline) 30 min before formalin injection significantly reduces
both phases of the formalin response.
Central sensitization is believed to underlie some aspects of altered sensibility found in chronic pain states. Thus, again, the available
data are consistent with the idea that BDNF may be a key mediator of central sensitization within the spinal cord via an interaction at the
NMDA receptor site. Several other molecules have been implicated in this process of central sensitization, most notably the tachykinin
substance P ( 23 ). It is possible that BDNF is simply one of a series of modulators capable of triggering postsynaptic changeperhaps via
a common mechanism of NMDA receptor phosphorylation. It is interesting that substance P is itself regulated in a very similar fashion to
that described above for BDNF. That is, it is up-regulated in trkA-expressing cells by exogenous NGF or by inflammatory stimuli in an
NGF-dependent fashion. Moreover, its expression falls in these cells after peripheral axotomy, but the same stimulus apparently can
induce expression in some large neurons ( 24 ). The clear effects of trkB-IgG coupled with the rather modest phenotype of substance P and
NK1 null-mutant animals ( 25 ) may indicate that BDNF is a more important mediator of these central changes, and antagonism of this
molecule may provide a novel therapeutic target.
Modulation of gene expression. In addition to the effect of BDNF on short-term modulation of the NMDA receptor, there is a strong
possibility that trk receptor activation and phosphorylation may underlie more prolonged changes in synaptic excitability. Kang and
Schuman ( 26 ) showed that some effects of BDNF on synaptic plasticity in the hippocampus may occur in the presence of the NMDA
receptor antagonist AP-5, suggesting that alternative mechanisms may exist. There is now evidence that BDNF will interact with signal
transduction pathways that underlie longer-lasting forms of synaptic plasticity ( 27 ). We have examined the ability of BDNF to induce
alterations in immediate early genes expressed in dorsal horn neurons in the spinal cord ( 18 ). Intrathecal administration of BDNF to adult
rats induces a large increase in c-fos immunoreactivity in the superficial dorsal horn ( Fig. 3 ), whereas NGF does not. In the spinal cord, it
is likely therefore that BDNF interacts with signal transduction pathways to induce long-term changes in synaptic excitability that depend
on gene transcription and protein synthesis. It is well recognized that activity in primary sensory neurons, and specifically in nociceptive
sensory neurons, is capable of inducing c-fos in spinal
the authoritative version for attribution.

FIG. 5. Potential mechanisms by which BDNF may influence spinal nociceptive processing in normal and inflammatory conditions.
See text for discussion. mem. pot., membrane potential; AMPA, - amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid.
 BRAIN-DERIVED NEUROTROPHIC FACTOR IS AN ENDOGENOUS MODULATOR OF NOCICEPTIVE RESPONSES IN THE 7718
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

SPINAL CORD

neurons. However, it is not yet clear whether the c-fos induced by BDNF represents some nociceptive action or it relates to some other
function.
Endogenous BDNF Contributes to Nociceptive Responses. A key question for the hypothesis put forward here is whether
endogenous BDNF contributes to behavioral responses after peripheral injury. To address this important question, we attempted to
neutralize synaptically released BDNF with the sequestering antibody trkB-IgG. This molecule is a recombinantly produced fusion protein
based on the extracellular domain of the trkB receptor. It is soluble and has the ability to bind with high affinity and thereby sequester
endogenous BDNF (as well as the other trkB ligand, neurotrophin-4/5). It effectively serves as an antagonist of BDNF.
We have examined the effects of trkB-IgG on nociceptive afferent-evoked reflex responses in the spinal cord. We have again used the
in vitro hemisected spinal-cord preparation and maximized levels of BDNF within the spinal cord by pretreating animals with NGF. Under
such circumstances, the afferentevoked response is reduced significantly after superfusion with trkB-IgG ( Fig. 2 c), strongly suggesting
that BDNF is released from afferent fibers under these conditions.
To date, most of the work that has examined the effects of BDNF on synaptic transmission has been undertaken on isolated
preparations in vitro. There is evidence, howeveragain from work on the hippocampusthat BDNF will induce long-lasting
enhancement of synaptic transmission in vivo ( 28 ). We have recently tested this idea more stringently by assessing the effect of trkB-IgG
on nociceptive behavioral responses in adult rats. Behavioral nociceptive responses evoked after subcutaneous injection of dilute formalin
into one hind paw were significantly reduced by prior intrathecal administration of trkB-IgG ( Fig. 4 ). Both phases of the formalin
response were affected, although the second phase, known to depend in part on the induction of central sensitization, was reduced to a
greater degree.
Thus, these findings show that under some conditions, specifically those that model persistent pain states (e.g., inflammation), BDNF
is released with activity from nociceptive afferent terminals and contributes to the postsynaptic responses in spinal-cord neurons.
CONCLUSIONS
As we review here, BDNF meets many of the criteria necessary to establish it as a neurotransmitter/neuromodulator in small-diameter
nociceptive neurons. Thus, it is synthesized by these neurons and packaged in dense core vesicles. The BDNF-expressing nociceptive
afferents terminate mostly in the superficial dorsal horn, and the postsynaptic cells in this region express fall-length trkB receptors.
However, the synaptic apposition of BDNF containing afferent profiles on neurons with trkB has not yet been established. Spinal neurons
are responsive to exogenous BDNF, as evidenced both histochemically (by the induction of c-fos) and electrophysiologically (by an
increased excitability to nociceptive inputs). BDNF is thus sufficient to elicit changes in postsynaptic excitability. Although the release of
BDNF from nociceptive fibers with activity has not been shown, there are both electrophysiological and behavioral data indicating that
antagonism of BDNF at least partially prevents some aspects of central sensitization. Thus, BDNF is also necessary for the full expression
of this phenomenon.
Currently, the mechanism of action of BDNF is not well understood. In Fig. 5 , we illustrate several events that may contribute to the
central modulatory role of BDNF. BDNF is constitutively expressed in some nociceptors. In normal animals, some BDNF is likely to
released with noxious stimulation; however, such release may have limited postsynaptic effects, because much of the trkB receptor protein
may not be present at the postsynaptic cell surface. Glutamate, also released with activity in nociceptors, may also have only restricted
postsynaptic actions because of the limited responsiveness of NMDA receptors in these normal states. With peripheral inflammation,
however, BDNF may play a much greater role. The expression is markedly up-regulated in trkA-expressing nociceptors in an NGF-
dependent fashion. More BDNF is likely to be released with activity. The increased activity of postsynaptic neurons in these states may
also lead to translocation of trkB receptors to the cell surface. The increased activation of trkB receptors in turn may lead to
phosphorylation and increased responsiveness of NMDA receptors, perhaps independently of induction of immediate early genes.
Together, these actions potentiate the efficacy of synaptic inputs from nociceptors and thus may contribute to the abnormal pain sensitivity
found in these inflammatory conditions.
We would like to thank Matt Ramer for useful comments regarding the manuscript. Some of the work described in this article was
supported by grants from the Medical Research Council of the United Kingdom, the Special Trustees of St. Thomas Hospital, and the
Physiological Society (London).
1. Snider, W. D. & McMahon, S. B. ( 1998 ) Neuron 20 , 629632 .
2. McMahon, S. B. , Bennett, D. L. H. , Michael, G. J. & Priestley J. V. ( 1997 ) in Neurotrophic Factors and Pain , eds. Jensen, T. S. , Turner, J. A. &
Wiesenfeld-Hallin, Z. ( IASP , Seattle ), pp. 353379 .
3. Carroll, P. , Lewin, G. R. , Koltzenburg, M. , Toyka, K. V. & Thoenen, H. ( 1998 ) Nat. Neurosci. 1 , 4246 .
4. DiStefano, P. S. , Friedman, B. , Radziejewski, C. , Alexander, C. , Boland, P. , Schick, C. M. , Lindsay, R. M. & Wiegand, S. J. ( 1992 ) Neuron 8 ,
983993 .
5. Ernfors, P. , Wetmore, C. , Olson, L. & Persson, H. ( 1990 ) Neuron 5 , 511526 .
6. Apfel, S. C. , Wright, D. E. , Wiideman, A. M. , Dormia, C. , Snider, W. D. & Kessler, J. A. ( 1996 ) Mol Cell Neurosci. 7 , 134142 .
7. Cho, H. J. , Kim, J. K. , Zhou, X. F. & Rush, R. A. ( 1997 ) Brain Res. 764 , 269272 .
8. Michael, G. J. , Averill, S. , Nitkunan, A. , Rattray, M. , Bennett, D. L. , Yan, Q. & Priestley, J. V. ( 1997 ) J. Neurosci. 17 , 84768490 .
9. Zhou, X. F. & Rush, R. A. ( 1996 ) Neuroscience 74 , 945951 .
10. Bennett, D. L. H. , Michael, G. J. , Ramachandran, N. , Munson, J. B. , Averill, S. , Yan, Q. , McMahon, S. B. & Priestley, J. V. ( 1998 ) J. Neurosci.
18 , 30593072 .
11. McMahon, S. B. & Bennett, D. L. H. ( 1999 ) in Text Book of Pain , eds. Wall, P. D. & Dubner, R. ( Churchill Livingston , Edinburgh ), 4th Ed. , in
press.
12. Michael, G. J. , Averill, S. , Nitkunan, A. , Rattray, M. , Bennett, D. L. H. , Yan, Q. & Priestley, J. V. ( 1997 ) J. Neurosci. 17 , 84768490 .
13. Bradbury, E. J. , King, V. , Simmons, L. J. , Priestley, J. V. & McMahon, S. B. ( 1998 ) Eur. J. Neurosci. 10 , 30583068 .
14. Tonra, J. R. , Curtis, R. , Wong, V. , Cliffer, K. D. , Park, J. S. , Timmes, A. , Nguyen, T. , Lindsay, R. M. , Acheson, A. & DiStefano, P. S. ( 1998 )
J. Neurosci. 18 , 43744383 .
15. Blochl, A. & Thoenen, H. ( 1995 ) Eur. J. Neurosci. 7 , 12201228 .
16. Thompson, S. W. N. , King, A. E. & Woolf, C. J. ( 1990 ) Eur. J. Neurosci. 2 , 638649 .
17. Thompson, S. W. N. ( 1998 ) Eur. J. Neurosci. 10 , Suppl. 10, 48.03 (abstr.) .
the authoritative version for attribution.

18. Kerr, B. J. , Bradbury, E. J. , Bennett, D. L. H. , Trivedi, P. M. , Dassan, P. , French, J. , Shelton, D. B. , McMahon, S. B. & Thompson, S. W. N. (
1999 ) J. Neurosci. 15 , in press.
19. Lessmann, V. , Gottmann, K. & Heumann, R. ( 1994 ) Neuroreport 6 , 2125 .
20. Levine, E. S. , Crozier, R. A. , Black, I. B. & Plummer, M. R. ( 1998 ) Proc. Natl. Acad. Sci. USA 95 , 1023510238 .
21. Small, D. L. , Murray, C. L. , Mealing, G. E. A. , Poulter, M. O. , Buchan, A. M. & Morley, P. ( 1998 ) Neurosci. Lett. 252 , 211214 .
22. Sucher, N. J. , Awobuluyi, M. , Choi, Y. B. & Lipton, S. A. ( 1996 ) Trends Pharmacol Sci. 17 , 348355 .
23. Urban, L. , Thompson, S. W. N. & Dray, A , ( 1994 ) Trends Neurosci. 17 , 432438 .
24. Noguchi, K. , Dubner, R. , De Leon, M. , Senba, E. & Ruda, M. A. ( 1994 ) J. Neurosci. Res. 37 , 596603 .
25. Woolf, C. J. , Mannion, R. J. & Neumann, S. ( 1998 ) Neuron 20 , 10631066 .
26. Kang, H. & Schuman, E. M. ( 1995 ) Science 267 , 16581662 .
27. Kang, H. , Welcher, A. A. , Shelton, D. & Schuman, E. M. ( 1997 ) Neuron 19 , 653664 .
28. Messaoudi, E. , Bardsen, K. , Srebo, B. & Bramham, C. R. ( 1998 ) J. Neurophysiol. 79 , 496499 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE POSTNATAL DEVELOPMENT OF SPINAL SENSORY PROCESSING 7719

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

The postnatal development of spinal sensory processing

MARIA FITZGERALD * AND ERNEST JENNINGS

Department of Anatomy and Developmental Biology, University College London, Gower Street, London WC1E 6BT, United Kingdom
ABSTRACT The mechanisms by which infants and children process pain should be viewed within the context of a developing
sensory nervous system. The study of the neurophysiological properties and connectivity of sensory neurons in the developing
spinal cord dorsal horn of the intact postnatal rat has shed light on the way in which the newborn central nervous system analyzes
cutaneous innocuous and noxious stimuli. The receptive field properties and evoked activity of newborn dorsal horn cells to single
repetitive and persistent innocuous and noxious inputs are developmentally regulated and reflect the maturation of excitatory
transmission within the spinal cord. These changes will have an important influence on pain processing in the postnatal period.
Increasing recognition of the importance of pain in infancy and childhood has focused attention on the basic neurobiology of
developing pain pathways. After birth, many regions of the somatosensory nervous system undergo changes in connectivity, leading to
transient functional stages before the adult pattern is achieved. Such changes are likely to determine pain and sensory processing at each
developmental stage.
The aim here is to discuss some of the changing features of sensory connections underlying pain processing in the dorsal horn of the
spinal cord over the course of postnatal development. The responses of newborn dorsal horn cells to single repetitive and persistent
innocuous and noxious inputs are shown to change over the postnatal period. These are discussed in terms of the maturation of excitatory
transmission within the spinal cord and how it may influence pain processing in the newborn.
Cutaneous Reflex Function in the Newborn. Although cutaneous reflexes are not evidence of pain perception as such, they can
provide information about the sensitivity and selectivity of the neonatal nervous system to nociceptive stimuli. A feature of the cutaneous
flexion reflex in the newborn rat, kitten, and human is that it is exaggerated compared with the adult ( 1 4 ). Thresholds to mechanical
skin stimulation are lower and the reflex muscle contractions more synchronized and long lasting ( 5 , 6 ). This feature is particularly
marked in the first postnatal week in rats, and then gradually changes until the third postnatal week when a rapid maturation takes place.
Repeated low-intensity skin stimulation results in hyperexcitability or sensitization of the reflex with lowered thresholds and generalized
movements of all limbs that becomes much less pronounced after 29- to 35-wk gestational age in the human and postnatal day 8 (P8) in
the rat ( 4 , 5 ). In addition, flexor reflex cutaneous receptive fields are larger ( 4 ) and less organized ( 6 ) than in the adult. Thresholds for
withdrawal from heat stimuli are also lower in younger animals ( 7 11 ), and the response to formalin has a 10-fold higher sensitivity in
neonatal rats compared with weanlings ( 12 , 13 ). The specific C-fiber irritant, mustard oil, is less effective at producing a nociceptive
response in the newborn; however, it gradually increases with postnatal age ( 14 , 15 ).
Because the thresholds of cutaneous mechanosensitive primary afferents are generally the same in the adult and the neonatal rat (
16 ), these postnatal changes in reflex sensitivity are likely to be caused by changes in central processing. The newborn spinal cord is
clearly in a generally more excitable state than in the adult, and one possibility is that both low- and high-intensity stimuli can activate
spinal pathways that are purely nociceptive in the adult, so neonatal A-fibers can evoke excitatory synaptic processes normally restricted
to C-fiber input in adults. Studies of the changing sensory connectivity in developing dorsal horn neurons have provided some insight into
this possibility.
Growth of Primary Afferent Terminals in the Newborn Spinal Cord. Although large-diameter dorsal root afferent collaterals
begin to grow into the dorsal grey at E15 (embryonic day 15) in the rat (refs. 17 , 18 ; A. Jackman and M.F., unpublished observations), C-
fibers grow into the spinal cord considerably later, at E19 onwards ( 19 ), and many chemical markers associated with C-fibers are not
apparent in the spinal cord until the perinatal period. C-type afferent terminals within synaptic glomeruli are not observed at electron
microscopy level until P5 ( 20 ).
The growth of both A- and C-fibers into the cord is somatotopically precise ( 19 , 21 , 22 ), but this is not true of the laminar
organization. Although in the adult, A afferents are restricted to laminae III and IV, in the neonate their terminals extend dorsally right up
into laminae I and II to reach the surface of the grey matter ( 23 , 24 ). This pattern is followed by a gradual withdrawal from the
superficial laminae over the first three postnatal weeks ( 23 ). C-fibers, on the other hand, grow specifically to laminae I and II, and for a
considerable postnatal period, these laminae are occupied by both A- and C-fiber terminals ( 23 ). During their occupation of superficial
laminae, A-fiber terminals can be seen to form synaptic connections at electron microscopy level ( 25 ). Furthermore, during this period c-
fos expression can be evoked in superficial dorsal horn cells in response to an innocuous or A-strength skin stimulus ( 26 ), whereas in
the adult, c-fos expression is normally induced only by noxious skin or A- and C-fiber nerve stimulation. Fos induction is triggered only
by innocuous stimulation in the adult under pathological conditions ( 27 ).
Postsynaptic Responses to Primary Afferent Stimulation in the Newborn Spinal Cord. The importance of A-fiber afferent input
in the newborn dorsal horn can be seen in an analysis of the extracellularly recorded spike activity evoked in individual cells in
anesthetized rat pups ( 28 ). Low-intensity electrical skin stimulation (100 A3.5 mA, 50200 s) sufficient to recruit A-fibers evokes
spike activity in both superficial and deep laminae at latencies that progressively decrease with age. At
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: P, postnatal day; E, embryonic day; NMDA, N-methyl-D-aspartate.
* To whom reprint requests should be addressed, e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE POSTNATAL DEVELOPMENT OF SPINAL SENSORY PROCESSING 7720

P3, the mean latency of the A-fiber-evoked response is 33.1 2.78 ms (n = 22), compared with 19.1 1.32 ms (n = 65) at P6, 13.5 0.8
ms (n = 53) at P10, and 7.3 0.3 ms (n = 35) at P21. Furthermore, the variation in the A-fiber latencies within the population of recorded
cells decreases with age ( 28 ).
In contrast to responses to A-fiber input, no long-latency C-fiber-evoked (15 mA, 500 s) spike responses are evoked in dorsal horn
cells in vivo in the first postnatal week ( 28 , 29 ). At P10, only 7 of 53 cells (13%) have a C-fiber input with a mean latency of 97.65
4.44 ms (n = 7), and at P21 the value is 32% with a mean latency of 107.0 10.12 ms (n = 10). These results do not, of course, provide
information about subthreshold C-fiber-evoked responses at this time.
Convergence of Afferent Inputs in the Postnatal Dorsal Horn Cells. Low-threshold inputs can also be seen to dominate the
newborn dorsal horn when the responses to natural stimulation are examined. Background activity is generally absent when neonatal cells
are initially isolated for recording, but strong responses can be evoked by mechanical stimulation of the skin of the receptive field. Some
cells respond to both innocuous brushing and noxious pinching of the skin, but the convergence of input to dorsal horn cells changes over
the postnatal period ( Table 1 ). The responses recorded from cells in the younger animals are elicited mainly by low-threshold
mechanoreceptors, and there are few cells with convergent input in the first week of life. This population gradually increases so that by
P21 the percentage of neurones with convergent primary afferent input is similar to that seen in the adult.
Receptive Fields of Postnatal Dorsal Horn Cells. The size of dorsal horn cell peripheral cutaneous receptive field decreases with
age ( 30 ). We have studied this in more detail using those cells with mechanoreceptive fields on the plantar surface of the hindpaw.
Receptive field areas were scanned into a computer, converted into pixels by using Adobe PHOTOSHOP software (Adobe Systems,
Mountain View, CA) and expressed as a percentage of the entire plantar surface of the hindpaw. At P3, the mean (SE) peripheral
receptive field occupies 50 5.6% of the plantar hindpaw. This value drops to 36 2.9% at P6, 20 1.9% at P10, and 15% 1.6 at P21.
The biggest change, therefore, occurs in the first postnatal week.
In the neonate, therefore, receptive fields are not only dominated by low-threshold inputs but are also larger and will therefore
overlap more than in the adult, increasing the chance of activation by peripheral skin stimulation.
Activity-Dependent Changes in Postnatal Dorsal Horn.
(i) Repetitive stimulation of receptive fields.
C-fiber-evoked activity is not observed in dorsal horn cells in the first postnatal week, and repetitive peripheral stimulation at C-fiber
strength also has no observable effects on dorsal horn cell spike responses. From P10, repetitive C-fiber stimulation at three times the C-
fiber threshold produces a classical wind-up, as reported in the adult dorsal horn ( 31 37 ) in 18% of cells. This percentage has
increased to 40% of cells by P21.
In contrast to the lack of C-fiber influence, stimulation of cells at twice the A-fiber threshold at a frequency of 0.5 Hz through pin
electrodes placed in the center of the peripheral receptive field on the hindlimb can produce considerable sensitization of the dorsal horn
cells in the neonate ( 28 ). This sensitization takes the form of a buildup of background activity in the cells during repetitive stimulation
that outlasts the stimulation period, thereby producing a prolonged after-discharge of up to 138 s. It is particularly apparent in younger
animals, and at P6, 19 of 57 cells (33%) display background firing during, and a prolonged after-discharge of, 70.6 18 s after repetitive
A-fiber stimulation. At P10, 3 of 48 cells showed this type of sensitization (6%) with an after-discharge of 63 s, whereas at P21, it was not
seen in any cells (n = 31) ( 28 ).
Table 1. Convergence of afferent input to dorsal horn cells at different postnatal ages
Cells with different input
Age Brush Pinch Brush and Pinch
P3 (n = 22) 20 (91%) 1 (4.5%) 1 (4.5%)
P6 (n = 65) 54 (83%) 7 (11%) 4 (16%)
P10 (n = 53) 22 (42%) 12 (22%) 19 (36%)
P21 (n = 35) 10 (29%) 5 (14%) 20 (57%)
A-fiber-induced sensitization is not accompanied by an increase in the direct A-fiber-evoked spike discharge, but during the
stimulation period, the sensitized units show a significant increase in activity outside of this short-latency evoked burst ( 28 ). The mean
activity during the stimulation period, measured in the 40- to 2,000-ms period between stimuli, is 2.6 0.16 spikes in P6 sensitizing cells,
significantly greater (p < 0.0001) than the 0.4 0.04 spikes in nonsensitizing cells. At P10, there is a similar pattern; the mean background
activity for sensitizing cells was 15.7 0.84 spikes, whereas that of nonsensitized cells was 1.3 0.13 spikes, another significant
difference (p < 0.0001).
(ii) Experimental inflammation in rat pups.
Carrageenan is reported to be a reliable agent in modeling inflammation in adults ( 38 44 ). After subcutaneous injection, edema
develops rapidly, followed by hyperalgesia, which peaks at 34 hr and decreases to baseline by 2472 hr ( 40 42 ). In some cases, the
period of hyperalgesia can last 1014 days ( 40 ). In view of the differences in sensory processing in the newborn compared with the adult
dorsal horn, we examined the responses of newborn dorsal horn cells to a carrageenan inflammatory stimulus.
The allodynia or drop in mechanical threshold that follows carrageenan injection ( 11 , 40 ) and hyperalgesia after mustard oil
application ( 15 ) is clear, but smaller in amplitude, in neonates. Carrageenan-induced inflammation produces a parallel fall in von Frey
thresholds at P3 and P10, whereas P21 animals show a significantly greater effect. At all ages, the effect increases with time, reaching a
maximum at 4 hr postinjection, but is still maintained at 5 hr postinjection ( 11 ).
This finding agrees with the responses of dorsal horn cells at this time. After carrageenan injection, dorsal horn cell receptive fields
recorded in anesthetized rat pups in vivo were measured and expressed as a percentage of the plantar foot area ( Fig. 1 ). There was a
significant increase in the size of peripheral receptive fields in animals in both the P10 and P21 age groups (P < 0.0001 in both cases). At
P10, the size of the peripheral receptive fields increased 2.5-fold, and at P21 the increase was 3.4-fold. Mean size SEM (as a percentage
of the plantar hindpaw area) of the peripheral receptive field in the inflamed group at P10 was 47.2 6.4%, and that of the control was
19.1 2.0%. At P21, the mean size of the receptive field in the inflamed group was 51.8 12.2% and that of the control group was 14.9
the authoritative version for attribution.

1.6%.
The receptive fields of adult dorsal horn neurons observed between 4 and 8.5 hr after injection of complete Freunds adjuvant
expanded to 2.4 times their original size ( 45 ). This increase of the receptive field may be responsible for hyperalgesia. Because the
receptive fields are larger, there is a greater degree of overlap, and so a single stimulus would activate many more neurons than in the
control state, a summative effect ( 46 ).
Effects of Primary Afferent Stimulation. The magnitude of the evoked response (number of spikes) after electrical stimulation at
twice A-fiber threshold directly to the nerve significantly increased after inflammation in P10 cells. Mean SEM evoked response for the
control animals was 3.2 0.25 spikes, and for the inflamed animals it was 7.6 0.21 spikes. The Students t test gives a p < 0.0001 when
comparing these
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE POSTNATAL DEVELOPMENT OF SPINAL SENSORY PROCESSING 7721

two groups, suggesting that the difference is extremely significant.

FIG. 1. Peripheral receptive fields: change with age and after inflammation. Representations of the peripheral receptive fields. Those
in black represent cells receiving afferents from inflamed skin, whilst those in grey are the control. The mean receptive field size
SEM (as a percentage of the total plantar foot area) is 50 5.6% at P3, 36 2.9% at P6, 20 1.97% at P10, and 15 1.62% at P21.
For those with inflamed feet, the receptive field sizes are 47 6.4% at P10 and 52 12.2% at P21.
In addition, the number of cells with C-fiber-evoked activity increased from 0/6 in the control group to 4/7 in the inflamed group.
Even such small numbers suggest an important effect of the carrageenan-induced central excitation on unmasking C-fiber-evoked spike
activity.
At P21, the mean magnitude of the evoked response for cells receiving an A afferent input was 6.8 0.32 spikes in the control group
and 6.6 0.21 spikes in the inflamed group. The difference between these two groups is not significant (P = 0.59). For those cells
responding to C afferent input, the magnitude of response was 4.9 0.5 spikes for the control group and 10.1 0.67 spikes for the
inflamed group. These two results were significantly different, with a P < 0.0001.
A Possible Role for N-methyl-D-aspartate (NMDA) Receptors. The results demonstrate important differences in the synaptic
connectivity underlying sensory processing in the newborn spinal cord. The slow maturation of C-fiber afferent input appears to result in a
predominance of A-fiber-evoked activity, such that processes that are exclusively activated by small-diameter nociceptive inputs in adults
can be triggered by low-threshold large-diameter inputs in the first postnatal weeks.
There are likely to be several mechanisms underlying this transient state of A-fiber-induced excitation, but we would like to propose
that one important one could be in the developmental regulation of NMDA receptors.
The neonatal spinal cord has a higher concentration of NMDA receptors in the grey matter than that observed in older animals ( 47 ).
All laminae in the dorsal horn are uniformly labeled with NMDA-sensitive [3H]glutamate until day 1012, when higher densities
gradually appear in substantia gelatinosa so that by P30, binding is similar to that in the adult. Furthermore, the affinity of the receptors for
NMDA and the NMDA-evoked calcium efflux in rat substantia gelatinosa is high in the first postnatal week and then declines ( 48 ). This
maturation is delayed by neonatal capsaicin treatment, suggesting that C-fiber afferent activity regulates the postnatal maturation of
NMDA receptors ( 48 ). There is also considerable rearrangement of the subunit composition of the NMDA channel complex during spinal
cord development ( 49 ).
This study was supported by a European Community Biomedicine and Health grant (BMH4-CT95-0172). E.A.J. is a graduate student
supported by the Medical Research Council of Great Britain and the Department of Anatomy, University College London.
1. Ekholm, J. ( 1967 ) Acta Physiol Scand. Suppl. 297 , 1130 .
2. Issler, H. & Stephens, J. A. ( 1983 ) J. Physiol. (London) 335 , 643654 .
3. Fitzgerald, M. & Gibson, S. ( 1984 ) Neuroscience 13 , 933944 .
4. Andrews, K. & Fitzgerald, M. ( 1994 ) Pain 56 , 95101 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE POSTNATAL DEVELOPMENT OF SPINAL SENSORY PROCESSING 7722

5. Fitzgerald, M. , Shaw, A. & MacIntosh, N. ( 1988 ) Dev. Med. Child Neurol. 30 , 520526 .
6. Marsh, D. , Dickenson, A. H. , Hatch, D. & Fitzgerald, M. ( 1999 ) Pain , in press.
7. Holmberg, H. & Schouenborg, J. ( 1996 ) J. Physiol. (London) 493 , 239252 .
8. Lewin, G. R. , Ritter, A. M. & Mendell, L. M. ( 1993 ) J. Neurosci. 13 , 21362148 .
9. Falcon, M. , Guendellman, D. , Stolberg, A. , Frenk, H. & Urca, G. ( 1996 ) Pain 67 , 203208 .
10. Hu, D. , Hu, R. & Berde, C. B. ( 1997 ) Anesthesiology 86 , 957965 .
11. Marsh, D. , Dickenson, A. H. , Hatch, D. & Fitzgerald, M. ( 1999 ) Pain , in press.
12. Guy, E. R. & Abbott, F. V. ( 1992 ) Pain 51 , 8190 .
13. Teng, C. J. & Abbott, F. V. ( 1998 ) Pain 76 , 337347 .
14. Fitzgerald, M. & Gibson, S. ( 1984 ) Neuroscience 13 , 933944 .
15. Jiang, M. C. & Gebhart, G. F. ( 1998 ) Pain 77 , 305313 .
16. Fitzgerald, M. ( 1987 ) J. Physiol. (London) 383 , 7992 .
17. Smith, C. L. ( 1983 ) J. Comp. Neurol. 220 , 2943 .
18. Fitzgerald, M. , Reynolds, M. L. & Benowitz, L. I. ( 1991 ) Neuroscience 41 , 187199 .
19. Fitzgerald, M. ( 1987 ) J. Comp. Neurol. 261 , 98104 .
20. Pignatelli, D. , Ribeiro da Silva, A. & Coimbra, A. ( 1989 ) Brain Res. 491 , 3344 .
21. Fitzgerald, M. & Swett, J. ( 1983 ) Neurosci. Lett. 43 , 149154 .
22. Ozaki, S. & Snider, W. D. ( 1997 ) J. Comp. Neurol. 380 , 215229 .
23. Fitzgerald, M. , Butcher, T. & Shortland, P. ( 1994 ) J. Comp. Neurol. 348 , 225233 .
24. Mirnics, K. & Koerber, H. R. ( 1995 ) J. Comp. Neurol. 355 , 601614 .
25. Coggeshall, R. E. , Jennings, E. A. & Fitzgerald, M. ( 1996 ) Brain Res. Dev. Brain Res. 92 , 8190 .
26. Jennings, E. & Fitzgerald, M. ( 1996 ) Pain 68 , 301306 .
27. Ma, Q. P. & Woolf, C. J. ( 1996 ) Pain 67 , 307316 .
28. Jennings, E. & Fitzgerald, M. ( 1998 ) J. Physiol. (London) 509 , 859868 .
29. Fitzgerald, M. ( 1988 ) Neurosci. Lett. 86 , 161166 .
30. Fitzgerald, M. ( 1985 ) J. Physiol. (London) 364 , 118 .
31. Mendell, L. M. & Wall, P. D. ( 1965 ) Nature (London) 206 , 9799 .
32. Mendell, L. M. ( 1966 ) Exp. Neurol. 16 , 316332 .
33. Woolf, C. J. , Thompson, S. W. & King, A. E. ( 1988 ) J. Physiol. (Paris) 83 , 255266 .
34. Davies, S. N. , Martin, D. , Millar, J. D. , Aram, J. A. , Church, J. & Lodge, D. ( 1988 ) Eur. J. Pharmacol. 145 , 141151 .
35. Dickenson, A. H. & Sullivan, A. F. ( 1990 ) Brain Res. 506 , 3139 .
36. Thompson, S. W. , Woolf, C. J. & Sivilotti, L. G. ( 1993 ) J. Neurophysiol. 69 , 21162128 .
37. Woolf, C. J. & Thompson, S. W. ( 1991 ) Pain 44 , 293299 .
38. Xu, X. J. , Elfvin, A. & Wiesenfeld Hallin, Z. ( 1995 ) Neurosci. Lett. 196 , 116118 .
39. Winter, C. A. , Risley, E. A. & Nuss, G. W. ( 1962 ) Proc. Soc. Exp. Biol. Med. 111 , 544547 .
40. Kayser, V. & Guilbaud, G. ( 1987 ) Pain 28 , 99107 .
41. Hargreaves, K. , Dubner, R. , Brown, F. , Flores, C. & Joris, J. ( 1988 ) Pain 32 , 7788 .
42. Hylden, J. L. , Thomas, D. A. , Iadarola, M. J. , Nahin, R. L. & Dubner, R. ( 1991 ) Eur. J. Pharmacol. 194 , 135143 .
43. Stanfa, L. C. , Sullivan, A. F. & Dickenson, A. H. ( 1992 ) Pain 50 , 345354 .
44. Stanfa, L. C. , Misra, C. & Dickenson, A. H. ( 1996 ) Brain Res. 737 , 9298 .
45. Hylden, J. L. , Nahin, R. L. , Traub, R. J. & Dubner, R. ( 1989 ) Pain 37 , 229243 .
46. Dubner, R. & Ruda, M. A. ( 1992 ) Trends Neurosci. 15 , 96103 .
47. Gonzalez, D. L. , Fuchs, J. L. & Droge, M. H. ( 1993 ) Neurosci. Lett. 151 , 134137 .
48. Hori, Y. & Kanda, K. ( 1994 ) Dev. Brain Res. 80 , 141148 .
49. Watanabe, M. , Mishina, M. & Inoue, Y. ( 1994 ) J. Comp. Neurol. 345 , 314319 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TRANSCRIPTIONAL AND POSTTRANSLATIONAL PLASTICITY AND THE GENERATION OF INFLAMMATORY PAIN 7723

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Transcriptional and posttranslational plasticity and the generation

of inflammatory pain

CLIFFORD J. WOOLF * AND MICHAEL COSTIGAN

Neural Plasticity Research Group, Department of Anesthesia and Critical Care, Massachusetts General Hospital and Harvard Medical
School, Massachusetts General HospitalEast, Charlestown, MA 02129
ABSTRACT Inflammatory pain manifests as spontaneous pain and pain hypersensitivity. Spontaneous pain reflects direct
activation of specific receptors on nociceptor terminals by inflammatory mediators. Pain hypersensitivity is the consequence of
early posttranslational changes, both in the peripheral terminals of the nociceptor and in dorsal horn neurons, as well as later
transcription-dependent changes in effector genes, again in primary sensory and dorsal horn neurons. This inflammatory
neuroplasticity is the consequence of a combination of activity-dependent changes in the neurons and specific signal molecules
initiating particular signal-transduction pathways. These pathways phosphorylate membrane proteins, changing their function,
and activate transcription factors, altering gene expression. Two distinct aspects of sensory neuron function are changed as a
result of these processes, basal sensitivity, or the capacity of peripheral stimuli to evoke pain, and stimulus-evoked
hypersensitivity, the capacity of certain inputs to generate prolonged alterations in the sensitivity of the system. Posttranslational
changes largely alter basal sensitivity. Transcriptional changes both potentiate the system and alter neuronal phenotype.
Potentiation occurs as a result of the up-regulation in the dorsal root ganglion of centrally acting neuromodulators and
simultaneously in the dorsal horn of their receptors. This means that the response to subsequent inputs is augmented, particularly
those that induce stimulus-induced hypersensitivity. Alterations in phenotype includes the acquisition by A fibers of
neurochemical features typical of C fibers, enabling these fibers to induce stimulus-evoked hypersensitivity, something only C
fiber inputs normally can do. Elucidation of the molecular mechanisms responsible provides new opportunities for therapeutic
approaches to managing inflammatory pain.
Pain is a state-dependent sensory experience. Normally, it is generated only by activation of a specific subset of high-threshold
peripheral sensory neurons, the nociceptors, hence nociceptionthe detection of noxious or tissue-damaging stimuli. Nociception is
important for being aware of and reacting to potentially or actually damaging stimuli in the environment. Absence of this capacity, as in
individuals with congenital analgesia, results in ongoing tissue damage. Pain is also, of course, a major clinical problem. After
inflammation or nerve injury, dramatic alterations in the somatosensory system occur, amplifying responses and increasing sensitivity to
peripheral stimuli so that pain can now be activated by normally innocuous or low-intensity stimuli. Clinical pain is an expression, then, of
plasticity in the somatosensory system, operating at multiple sites and due to diverse mechanisms. The purpose of this paper is to highlight
key features of the plasticity of primary sensory neurons and of the synaptic contacts they make with dorsal horn neurons, the mechanisms
that operate to produce this plasticity, and how this relates to the pathogenesis of clinical pain hypersensitivity. The emphasis will be on
posttranslational and transcriptional changes in primary sensory and dorsal horn neurons initiated by electrical activity and peripheral
inflammation. We will show that sensory inputs not only produce sensations, painful and non-painful (basal sensitivity), but also alter the
somatosensory system by inducing neural plasticity, and that this changes the responsiveness of the system to subsequent stimuli (stimulus-
induced hypersensitivity).
STATE-DEPENDENT SENSORY PROCESSING IN THE SOMATOSENSORY SYSTEM
Fig. 1 defines four distinct states of the somatosensory system. The first state is present under normal or physiological circumstances
and is responsible for the normal processing of low- and high-intensity peripheral stimuli into innocuous and painful sensations,
respectively. The second state follows a noxious C fiber activating input of sufficient intensity or duration to initiate immediate
posttranslational changes in membrane-bound receptors of dorsal horn neurons, modifying their excitability and thereby altering
sensitivity to subsequent low- and high-intensity peripheral stimuli, the phenomenon of central sensitization. The third state is established
several hours after a substantial noxious stimulus, which, by virtue of the activity-dependent changes in the expression of functional genes
in sensory and dorsal horn neurons it produces, results in a potentiated system. This transcription-dependent potentiated system now
displays an enhanced responsiveness to subsequent central sensitization-inducing inputs. Finally, the last state follows peripheral
inflammation, where a combination of activity-dependent and signal molecule-mediated posttranslational and transcriptional changes in
the sensory and dorsal horn neurons results in a fundamentally altered system, one in which the transduction sensitivity of the peripheral
terminals is reduced (peripheral sensitization) ( 1 ), the excitability of dorsal horn neurons is increased (central sensitization) ( 2 ), and the
phenotype of sensory neurons is altered ( 3 ) such that both low-intensity A fiber and C fiber inputs can initiate central sensitization. Here
we will discuss each of these states in turn, discussing the molecular mechanisms within the
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: DRG, dorsal root ganglion; NGF, nerve growth factor; TTX, tetrodotoxin; TTXr, TTX-resistant; TTXs, TTX-sensitive;
BDNF, brain-derived neurotrophic factor; AMPA, -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid; CREB, cAMP responsive
element-binding protein; MAP, mitogen-activated protein; NMDA, N-methyl-D-aspartate.
* To whom reprint requests should be addressed at: Department of Anesthesia & Critical Care, Massachusetts General Hospital and

Harvard Medical School, 149 13th Street, Room 4310, Charlestown, MA 02129. e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TRANSCRIPTIONAL AND POSTTRANSLATIONAL PLASTICITY AND THE GENERATION OF INFLAMMATORY PAIN 7724

dorsal root ganglion (DRG) and dorsal horn that elicit these changes, from the mediators to the receptors through the intracellular signal-
transduction cascades and finally culminating on the posttranslational and transcriptional changes that result.

FIG. 1. State-dependent sensory processing in the somatosensory system. (A) The basal sensitivity of the system is normally such that
only a high-intensity C or A stimulus produces pain. (B) C fiber inputs induce immediate posttranslational changes in the dorsal
horn, resulting in central sensitization and altering basal sensitivity such that low intensity stimuli result in pain. (C) Conditioning C
fiber strength inputs induce activity-dependent transcriptional changes in the DRG and dorsal horn, resulting in a potentiated system
augmenting responsiveness to subsequent C fiber inputs. (D) Inflammation results in both a potentiated system and phenotypic
switches such that both C fiber and low-intensity A fiber inputs can evoke central sensitization.
State One: Nociception (Normal Transmission). The activation of high-threshold nociceptive primary afferent neurons, mainly
unmyelinated C fibers but A6 fibers as well, by intense but nondamaging stimuli, results within seconds in a transient, well localized pain
that does not require long-term nontranscriptional or transcriptional changes. Primary sensory neurons have three clear functions with
respect to their role in nociception: transduction of noxious but not low-intensity peripheral stimuli; conduction of action potentials to the
central nervous system; and transmission to central neurons. Distinct transducers, receptors, ion channels, and transmitters mediate these
functions. Tables 1 and 2summarize the key ion channels and receptors expressed in dorsal root ganglion cells, the functions they mediate,
and their posttranslational modulation or transcriptional regulation.
Nociceptive transduction. Peripheral nociceptive transduction involves the detection of hot and cold (but not warm or cool) thermal
stimuli and intense (but not innocuous) mechanical stimuli, as well as a sensitivity to chemical irritants such as the pungent ingredient in
chili peppers, capsaicin. This sensitivity is mediated by multiple specialized receptors, including heat-sensitive ion channels like the
vanniloid receptor VR1, which is gated by protons and activated by capsaicin ( 4 , 5 ), channels sensitive to protons and possibly
mechanical stimuli, the Na+/degenerin family including ASIC (acid-sensing ionic channel) and DRASIC (dorsal root acid-sensing ionic
channel) ( 6 ), and receptors sensitive to chemical stimuli alone (the histamine, bradykinin, purine, and serotonin receptors) ( 7 10 ).
Chemical stimuli are a very prominent component of inflammatory pain, where they are generated as part of the inflammatory response
(see below). A chemical component of nociception occurs only on exposure to nondamaging external chemical irritants, such as plant or
insect stings. The relative expression of these transduction elements in specific subsets of nociceptor sensory neurons is beginning to
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TRANSCRIPTIONAL AND POSTTRANSLATIONAL PLASTICITY AND THE GENERATION OF INFLAMMATORY PAIN 7725

reveal specialization of modality sensitivity. The purine receptor P2X3, for example, is expressed only on glial cell linederived
neurotrophic factor-responsive c-Ret-expressing C fibers ( 11 ), whereas VR1 is expressed both in these cells and in the TrkA-expressing
nerve growth factor (NGF)-responsive neurons ( 5 ).

Nociceptive conduction. Voltage-gated sodium channels, which are responsible for the rising phase of the action potential and play a
key role, with potassium channels, in determining the excitability of the sensory neurons, mediate conduction in nociceptors by
transferring input from the peripheral nerve terminals to the spinal cord. Neuronal voltage-gated sodium channels can be classified into
two types: those sensitive to nanomolar concentrations of the puffer fish toxin (tetrodotoxin-sensitive, TTXs) and those resistant to all but
micromolar concentrations of tetrodotoxin (tetrodotoxinresistant, TTXr) ( 12 ). DRG neurons express several distinct kinetic types of
sodium current. Small-diameter, highthreshold nociceptor neurons coexpress a rapidly inactivating, fast TTXs current and a slowly
activating and inactivating TTXr sodium current ( 13 , 14 ). Large diameter cells only express a TTXs sodium current ( 13 ).
Within the DRG, several voltage-gated sodium channels are known to be responsible for the TTXs current (PN1, rSCP6/ PN4, rBI,
rBII, and rBIII), and these show a wide distribution across large and small neuronal cells ( 15 17 ). Two sensory neuron-specific TTXr
sodium channels have been cloned [SNS/PN3 ( 18 , 19 ) and SNS2/NaN ( 20 , 21 )]. SNS2/NaN is found only in small sensory neurons,
where it is colocalized with SNS/PN3 ( 20 ), with which it may form functional heteromultimers, generating the native TTXr current. SNS/
PN3 is also found without SNS2 in a subpopulation of larger neurons ( 20 ). The expression of TTXr sodium channels in nociceptive
neurons points to a specific role for these channels in contrib
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TRANSCRIPTIONAL AND POSTTRANSLATIONAL PLASTICITY AND THE GENERATION OF INFLAMMATORY PAIN 7726

uting to the excitability of nociceptors, although the TTXs current may be sufficient for conduction, and the TTX current plays more of a
modulatory role in the peripheral terminals and in generating ectopic inputs.
Nociceptive transmission. Transfer of synaptic input from nociceptors to specific laminae in the dorsal horn is highly topographically
organized and activates particular subsets of second-order projection neurons ultimately leading, following activation of specific brain
centers, to the sensation of acute pain as well emotional, cognitive, and autonomic responses. Primary afferent neurons, by virtue of their
peripheral transduction specialization, central termination site, or temporal characteristics, encode stimulus modality, intensity, location,
and duration ( 22 ). This is relayed with great fidelity to second-order neurons. Synaptic transmission for C fibers is mediated by glutamate
acting on -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors on dorsal horn neurons, generating fast excitatory
postsynaptic potentials with a slow trailing edge contributed to by the N-methyl-D-aspartate (NMDA) receptor ( 23 , 24 ). If the stimulus
intensity is high enough, release of neuropeptides from dense core vesiclesparticularly substance Pwill occur, which will, via an
activation of the NK1 receptor, generate a greater postsynaptic response, reflecting the greater input and representing, therefore, a form of
intensity coding ( 25 ). Antagonism or deletion of the NMDA, NK1, or mGluR receptor does not normally interfere, though, with the
nociceptive response to normal non-tissue-damaging noxious stimuli ( 26 ); this is largely mediated by glutamate acting on its AMPA
receptor and reflects the necessity for a high-safety, fast early warning protective pain system. The NMDA, NK1, or mGluR receptors play
a key role in the initiation or maintenance of changes in synaptic transmission that constitute central sensitization ( 27 32 ), where the
specific link between a noxious stimulus and pain is lost and pain can be produced by normally innocuous inputs ( 2 , 33 35 ).
Neural plasticity and pain. Sensory neurons can undergo functional, chemical, and structural changes in response to changes in their
environment that modify their transduction, conduction, and transmission and move them from their specific role in mediating normal
nociceptive transmission to a new modified condition contributing to an altered state of sensibility. The key issues, then, are what produces
these changes, how and where they manifest, and what the sensory consequences are. We will analyze this in terms of two forms of
plasticity, activity- and signal molecule-dependent, and two kinds of change, that induced immediately by posttranslational modifications
in the neurons and that caused by an alteration in expression of key effector molecules, which takes some time to manifest. The functional
significance of this plasticity, both for basal sensibility and stimulus-evoked hypersensitivity, will be discussed.
State Two: Immediate Changes (Posttranslational Modulation of the Dorsal Horn). The first example of stimulusevoked
hypersensitivity is that of C fiber-mediated central sensitization ( 2 ). Nociceptive pain is characterized by a clear distinction in the sensory
responses to a noxious or low-intensity stimulus, pain, and an innocuous sensation, respectively. Activation of C fibers, can, however,
result in a situation where basal sensitivity changes such that normal low-threshold A inputs begin to evoke pain and tactile allodynia and
noxious inputs provoke a greater pain response, hyperalgesia (refs. 2 and 33 35 ; Fig. 1 ). Central sensitization, which in normal
noninflamed situations, can be produced only by C fiber input ( 36 ), is the direct manifestation of a complex set of activity-dependent
posttranslational changes in dorsal horn neurons. Relatively brief C fiber inputs (lasting only tens of seconds) can initiate very rapid
changes in membrane excitability, which manifest both as a progressive increase in excitability during the course of the stimulus, windup,
and poststimulus changes that can last for several hours (central sensitization) ( 2 , 36 ). Windup is a manifestation of the removal by
successive synaptic depolarization of the voltage-dependent Mg2+ block of the NMDA receptor amplifying the response to each
subsequent input. Central sensitization is quite different. Fig. 2 summarizes some of the salient features. Presynaptic transmitter/
neuromodulator release [glutamate, substance P, and brain-derived neurotrophic factor (BDNF)] results in changes in signal-transduction
pathways in dorsal horn neurons as a result of the activation of ligand-gated ion channels (NMDA-Rglutamate), metabotropic receptors
(mGluR glutamate, NK1substance P) and tyrosine kinase receptors (TrkBBDNF). Not shown in Fig. 2 but likely to have major effects
on posttranslational changes in the dorsal horn are the prostanoids. These act through prostaglandin E and prostacyclin and IP receptors
and may be released pre- and postsynaptically (reviewed in ref. 37 ).
Activation of these multiple receptors results in an increase in intracellular calcium, both via calcium inflow and release from
intracellular stores ( 38 ) and a consequent activation of calcium-dependent enzymes (protein kinase C and calcium calmodulin kinase),
protein kinase A (via G protein coupled-receptors), and tyrosine kinases (via the TrkB receptor) (see Fig. 2 ). These pathways converge
and diverge in a complex fashion and are linked such that TrkB, which itself is a tyrosine kinase, also activates other tyrosine kinases src
and protein kinase C ( 39 , 40 ). The targets for these different kinases are membrane-bound receptors/ion channels, of which the NMDA
and AMPA are best characterized, although others are certainly involved, including activation of neuronal nitric-oxide synthase and
prostaglandin production, with generation of retrograde signals to the presynaptic terminal. Phosphorylation of the NMDA receptor at
either serine/threonine residues on its NR1 subunit ( 41 ) or on tyrosine residues on its NR2 subunit ( 42 , 43 ) is a major factor. This
posttranslational modification results in dramatic changes in NMDA-receptor channel kinetics and a reduction in its voltage-dependent
Mg2+ block ( 42 , 44 , 45 ). Both of these changes augment subsequent responsiveness to synaptically released glutamate, increasing
synaptic strength and enabling previously subthreshold inputs to drive the output of the cell ( 46 ). This
the authoritative version for attribution.

FIG. 2. Posttranslational changes within dorsal horn neurons after release of transmitters from C fiber central terminals. These
transmitters/neuromodulators act on receptors and ion channels in the dorsal horn to activate protein kinases that phosphorylate
membrane-bound NMDA and AMPA receptors and alter their functional properties, increasing membrane excitability and thereby
eliciting central sensitization.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TRANSCRIPTIONAL AND POSTTRANSLATIONAL PLASTICITY AND THE GENERATION OF INFLAMMATORY PAIN 7727

effective increase in gain alters receptive field properties ( 46 48 ) and pain sensitivity, causing tactile allodynia and pin prick
hyperalgesia, far beyond the site of the C fiber-activating stimulus that initiated the central sensitization ( 49 ).
State Three: Early Changes in Transcription. C fiber inputs, in addition to the generation of central sensitization, which occurs
within seconds of the appropriate activation of dorsal horn neurons, also generate an activity-dependent change in transcription in dorsal
root ganglion and dorsal horn neurons, which has functional consequences that take hours to manifest. BDNF mRNA levels, for example,
increase in the DRG 2 hours after C fiber stimuli, whereas such inputs also increase TrkB message in the dorsal horn (unpublished
observations). These changes are very likely to result from an increase in calcium influx through voltage-gated calcium channels in the
DRG, secondary to increased electrical activity. Such changes in intracellular free calcium will cause changes in the phosphorylation and
therefore activation of a whole host of transcription factors, a well characterized example of which in neuronal systems is the cAMP
responsive element-binding protein (CREB) ( Fig. 3 ). Seconds to minutes after calcium influx, neuronal CREB is phosphorylated at
Ser-133 ( 50 ). Activity-dependent CREB Ser-133 phosphorylation has also been demonstrated in mouse embryonic DRG cells in culture
as has activity-dependent regulation of mitogen-activated protein (MAP) kinase ( 51 ). Multiple signal-transduction cascades converge on
CREB Ser-133 phosphorylation (reviewed in ref. 52 ). In hippocampal neurons, phosphorylation of CREB is mediated by the enzyme
CaMKIV (calciumcalmodulin-dependent protein kinase IV). Activation of this enzyme in turn is mediated by CaMK kinase (CamKK).
Interestingly, Ji et al. ( 53 ) have shown that CamKIV immunostaining is present mostly in small neurons of the DRG. Calcium-dependent
BDNF transcription in cortical neurons is mediated through CREB ( 54 , 55 ). It is likely, then, that at least some of the activity-dependent
up-regulation of BDNF within the DRG is also CREB-dependent. Calcium influx into PC12 cells, mediated by P2X2 receptors as opposed
to voltage-gated calcium channels, can activate the MAP kinase cascade via Pyk2 ( 56 ). Pyk2 provides a link between free cellular
calcium and the Map kinase cascade ( 57 ).

FIG. 3. Putative DRG activity- and signal molecule-dependent second-messenger cascades leading to transcription.
Activitydependent calcium flux through voltage-gated calcium channels may increase BDNF transcription via CREB
phosphorylation on Ser-133. NGF acts via TrkA to activate the MAP kinase cascade through ras. Myc phosphorylation through this
pathway can result in E box transactivation via its association with P Max. The preprotachykinin promoter contains multiple E box
units and thus maybe regulated by these transcription factors. RSK2 links the MAP kinase cascade to CREB phosphorylation. ATP
can stimulate the MAP kinase cascade via Pyk2.
BDNF produced in TrkA-expressing small neurons in the DRG is transported to their central terminals, where BDNF is located in
dense core vesicles ( 58 ) and contributes not to basal sensitivity, which appears to be the result primarily of glutamate acting on AMPA
receptors, but to C fiber-induced central sensitization (unpublished observations) ( Fig. 2 ). An increase in the amount of this presynaptic
neuromodulator in C fibers as a result of prior activity in these fibers, combined with an increase in its high-affinity receptor in the dorsal
horn, will result in what we have termed a potentiated system. A potentiated system is one where the same stimulus applied a second time,
some considerable time after the first, will produce a greater response, because the first stimulus has changed the system after a slow onset,
but for a relatively long period ( Fig. 1 ). There are features of this change into a potentiated sensory system that are reminiscent of the
activity-dependent transcriptional changes that contribute to the persistence of long-term potentiation in cortical cells and imply that this
state constitutes a kind of pain memory. The potentiation of the sensory system needs, though, to be differentiated from windup, where a
progressive increase in responsiveness occurs rapidly (within 1 or 2 seconds), during the course of a train of inputs. Potentiation refers to
an effect that only manifests several hours after the input and is due to a change in transcription. Although the system is potentiated, this is
still only to C fiber inputs, which evoke an augmented hypersensitivity because of changes intrinsic only to these fibers and their receptive
dorsal horn neurons. This situation differs from inflammation where A fibers acquire the capacity to induce central hypersensitivity (see
below).
State Four: Inflammatory Changes in DRG and Dorsal Horn Neurons. Inflammation is associated with tissue damage, which
results in the leak of intracellular contents into the extracellular fluid, the recruitment of inflammatory cells, and the production and release
of a broad range of neuroactive agents by inflammatory and noninflammatory cells, including ions (K+ and H+) ( 6 ), amines (5-
hydroxytryptamine, histamine) ( 9 ), kinins (bradykinin) ( 8 ), prostanoids (PGE2) ( 37 ), purines (ATP) ( 7 ), NO, cytokines (IL-1, TNF,
Il-6) ( 9 , 10 ) and growth factors (leukemia inhibitory factor, NGF) ( 10 ). Some of these inflammatory agents may be sufficient by
the authoritative version for attribution.

themselves to activate the peripheral nerve endings of those nociceptors that express the appropriate receptors (Tables 1 and 2 ), generating
inward currents and sensory inflow. Most of these agents, however, act by changing the sensory neuron rather than directly activating it.
These changes include early posttranslational changes both in the peripheral terminals of nociceptors, altering transduction sensitivity
(peripheral sensitization), and in dorsal horn neurons secondary to activity in C fibers (central sensitization). Both peripheral sensitization
and central sensitization alter basal sensitivity to noxious and normally innocuous stimuli. There are, in addition, later and longer-lasting
transcription-dependent changes in the DRG and in the dorsal horn that are due to a complex combination of activity and retrograde
transport of specific signal molecules produced as a result of the inflammation. These changes result both in a potentiated nociceptive
system, as described above, and one in which phenotypic switches alter the central responses elicited by low-threshold A fiber inputs, the
phenomenon of progressive tactile hypersensitivity. Both the potentiation of the system and the phenotypic changes manifest as a change
in stimulus-evoked rather than basal hypersensitivity.
Posttranslational changes. A defining feature of nociceptors is their normal high threshold for activation. After inflamma
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TRANSCRIPTIONAL AND POSTTRANSLATIONAL PLASTICITY AND THE GENERATION OF INFLAMMATORY PAIN 7728

tion, or even after repeated noxious stimuli, however, the peripheral terminal threshold drops such that lower intensity stimuli can now
initiate activity in the nociceptors. This peripheral sensitization, which can be detected within a very short period, is the result of changes
either in the transducing receptor molecules themselves or in sodium channels in the terminal ( Fig. 4 ). A change in the transducer
molecule is best exemplified by VR1, where repeated heat stimuli by themselves result in a progressively augmenting inward current
through the ion channel, and exposure to protons does the same thing, by mechanisms currently not known ( 4 , 5 ). The other contributor
to peripheral sensitization is phosphorylation of membrane-bound receptors/ion channels. Many inflammatory mediators activate either
protein kinase A or C (see Fig. 4 ), both of which can phosphorylate the TTXr sodium channel SNS/PN3 and contribute to a greater
sodium current in the terminal ( 59 61 ). The recently discovered TTXr channel SNS2 has different protein kinase C and protein kinase
A consensus regions on its intracellular loops ( 20 ), and its contribution to the change in sodium current is undetermined. These
sensitizing changes occur locally in the peripheral terminal, independent of transcription in the DRG. However, transcription of the very
elements that show posttranslational changes in these terminals is highly regulated in the cell soma. Inflammation up-regulates both VR1
and SNS/SNS2 ( 5 , 20 , 62 ). Inflammation is associated with an increase in peripheral NGF levels ( 63 ), and this may be the key signal
molecule for many of the transcriptional changes (see below). Although peripheral sensitization does not itself require transcription, an
increase in the substrate for such sensitization is very likely to amplify the phenomenon. Because of the delay inherent in the initiation of
changes in expression and transport of proteins,

FIG. 4. Posttranslational changes in transduction mechanisms/ion channels at nociceptor peripheral terminals induced by
inflammatory mediators increases sensitivity and reduces threshold (peripheral sensitization), which occurs as a result of changes in
the transducer proteins themselves (e.g., VR1) and as a result of a PKC- and PKA-mediated phosphorylation of TTXr sodium
channels. Both activity and retrograde transport of signal molecules can also induce transcriptional changes in the DRG, which
increase transducer molecules (VR1), ion channels (SNS/SNS2), and synaptic neuromodulators (BDNF/substance P) both altering
phenotype and potentiating the system.
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TRANSCRIPTIONAL AND POSTTRANSLATIONAL PLASTICITY AND THE GENERATION OF INFLAMMATORY PAIN 7729

such transcription-dependent augmentation will only manifest many hours after the onset of inflammation ( Fig. 4 ).
Apart from the increased sensitivity of the peripheral terminals of nociceptors on exposure to inflammatory mediators, which will
result in an area of increased sensitivity to thermal and mechanical stimuli localized to the site of inflammation, any C fiber input activated
at the time of or during inflammation will also initiate central sensitization. This C fiber input will result in an NMDA receptor-sensitive
increase in responsiveness to low- and high-intensity stimuli, both when applied to the site of the inflammation and in the contiguous
noninflamed area. Tactile allodynia and pin prick hyperalgesia in the zone of secondary hyperalgesia ( 64 ), for example, are characteristic
NMDA-receptor-mediated ( 65 ) features of central sensitization.
Transcriptional changes. In addition to these posttranslational changes, an alteration in the expression of effector molecules in the
DRG ( 66 ) and dorsal horn ( 67 69 ) is a prominent feature of inflammation and can be initiated in two ways. The first is as a result of an
activity-dependent activation of the CREB transcription factor both in DRG and dorsal horn neurons (ref. 70 ; Fig. 3 ). As discussed
above, this will result, after a delay of several hours for transcription and translation in the DRG and protein transport to central terminals,
in a potentiated system in which the C fibers are primed to exert a greater effect on dorsal horn neurons as a result of an increased
expression of neuromodulators like BDNF. In addition, the dorsal horn is simultaneously made hyperresponsive to such neuromodulators
as a result of an activity-dependent increased expression of the TrkB receptor. The second way that transcriptional changes occur after
inflammation is via the production in the inflamed tissue of specific signal molecules that bind to receptors on nociceptor sensory
terminals. The ligandreceptor complex is then internalized and retrogradely transported to the cell body, where it activates specific signal-
transduction cascades ( Fig. 3 ). NGF is the prototypical example of such a signal molecule whose level increases substantially in inflamed
tissue and neutralization of which massively reduces inflammatory hypersensitivity (reviewed in ref. 71 ).
Signals from extracellular growth factors such as NGF are transduced into intracellular responses through the small G protein Ras and
MAP kinase (reviewed in refs. 72 74 ). The Ras cascade (see Fig. 3 ) involves the sequential activation of Ras, Raf, MAP kinase kinase,
and MAP kinase itself [p44/p42 MAP kinase is also known as ERK 1/2 (extracellular signal regulated kinase)]. Subsequent to
phosphorylation, a fragment of activated MAP kinase translocates to the nucleus, where it then regulates gene expression through the
phosphorylation, and therefore activation, of various transcription factors including c-Myc, Elk-1, c-Fos, and c-Jun ( 72 , 73 ).
Phosphorylation of these transcription factors results in their association into complexes. These complexes activate transcription of many
downstream genes via their association with response elements present in the response gene promoter regions ( Fig.3 ). The MAP kinase
cascade can also phosphorylate and thus activate CREB within neurons via the activation of Rsk2 by MAP kinase itself ( 75 , 76 ).
One consequence of transcriptional changes in DRG neurons after inflammation is that some low-threshold A neurons acquire the
chemical phenotype typical of C fibers. For example, the neuropeptide substance P is normally found almost exclusively in a subset of the
TrkA-expressing C fibers, with only a very small number in TrkA expressing small myelinated A fibers ( 77 , 78 ). After inflammation,
there is a NGF-dependent increase in substance P expression in C fibers ( 63 ), but also a novel expression of this neuropeptide in some
large A fibers as well ( 3 ). This new expression, together with the inflammation-induced increase in NK1 receptors in the dorsal horn (
79 ), will result not only in a potentiated system, but one in which the specific type of stimulus that can evoke central sensitization has
changed. Stimulus-induced hypersensitivity can thus be mediated by low-intensity A inputs as well as high-intensity C fiber inputs (
Fig. 1 ). This heightened sensibility manifests as progressive tactile hypersensitivity, where low-intensity mechanical stimulation of
inflamed skin (light touch) produces a progressively incrementing increase in the excitability of spinal neurons, something they never can
elicit in the normal situation, and this can be mimicked by A fiber stimulation ( 3 , 80 83 ).
CONCLUSION
A clear distinction needs to be made between nociception, the detection only of intense noxious stimuli, and inflammatory pain,
which is evoked by normally innocuous stimuli and incorporates an exaggerated response to noxious stimuli. The transition from one state
to the other involves multiple changes, some early and others late, some mediated by posttranslational changes and others reliant on
altered gene expression, some evoked by activity and others in response to specific inflammatory mediators/signal molecules. Enormous
progress has been made in unraveling what changes occur, when they occur, and the molecular mechanisms involved. Although much still
remains to be determined, this new insight is translating into novel and more targeted approaches to treating inflammatory pain.
The reviewed work was supported by the Medical Research Council G9431792, the Human Frontier Science Program RG73/96, the
Wellcome Trust 039614, the European Union BMH4-CT 95 0172, and the National Institute of Neurological Disorders and Stroke R01
NS3825301.
1. Bessou, P. & Perl, E. R. ( 1969 ) J. Neurophysiol. 32 , 10251043 .
2. Woolf, C. J. ( 1983 ) Nature (London) 306 , 686688 .
3. Neumann, S. , Doubell, T. P. , Leslie, T. A. & Woolf, C. J. ( 1996 ) Nature (London) 384 , 360364 .
4. Caterina, M. J. , Schumacher, M. A. , Tominaga, M. , Rosen, T. A. , Levine, J. D. & Julius, D. ( 1997 ) Nature (London) 389 , 816824 .
5. Tominaga, M. , Caterina, M. J. , Malmberg, A. B. , Rosen, T. A. , Gilbert, H. , Skinner, K. , Raumann, B. E. , Basbaum, A. I. & Julius, D. ( 1998 )
Neuron 21 , 531543 .
6. Waldmann, R. & Lazdunski, M. ( 1998 ) Curr. Opin. Neurobiol. 8 , 418424 .
7. Burnstock, G. & Wood, J. N. ( 1996 ) Curr. Opin. Neurobiol. 6 , 526532 .
8. Walker, K. , Perkins, M. & Dray, A. ( 1995 ) Neurochem. Int. 26 , 117 .
9. Dray, A. ( 1995 ) Br. J. Anaesth. 75 , 125131 .
10. Levine, J. D. & Taiwo, Y. O. ( 1994 ) in Textbook of Pain , eds. Wall, P. D. & Melzack, R. ( Churchill Livingstone , New York ), pp. 4546 .
11. Bradbury, E. J. , Burnstock, G. & McMahon, S. B. ( 1998 ) Mol. Brain Res. 12 , 256268 .
12. Catterall, W. A. ( 1992 ) Physiol. Rev. 72 , S1548 .
13. Caffrey, J. M. , Eng, D. L. , Black, J. A. , Waxman, S. G. & Kocsis, J. D. ( 1992 ) Brain Res. 592 , 283297 .
14. Roy, M. L. & Narahashi, T. ( 1992 ) J. Neurosci. 12 , 21042111 .
15. Waxman, S. G. , Kocsis, J. D. & Black, J. A. ( 1994 ) J. Neurophysiol. 72 , 466470 .
16. Black, J. A. & Waxman, S. G. ( 1996 ) Dev. Neurosci. 18 , 139152 .
17. Black, J. A. , Dib-Hajj, S. , McNabola, K. , Jeste, S. , Rizzo, M. A. , Kocsis, J. D. & Waxman, S. G. ( 1996 ) Brain Res. Mol. Brain Res. 43 , 117
131 .
18. Akopian, A. N. , Sivilotti, L. & Wood, J. N. ( 1996 ) Nature (London) 379 , 257262 .
the authoritative version for attribution.

19. Sangameswaran, L. , Delgado, S. G. , Fish, L. M. , Koch, B. D. , Jakeman, L. B. , Stewart, G. R. , Sze, P. , Hunter, J. C. , Eglen, R. M. & Herman, R.
C. ( 1996 ) J. Biol. Chem. 271 , 59535956 .
20. Tate, S. , Benn, S. , Hick, C. , Trezise, D. , John, V. , Mannion, R. J. , Costigan, M. , Plumpton, C. , Grose, D. , Gladwell, Z. , Kendall, G. , Dale, K. ,
Bountra, C. & Woolf, C. J. ( 1998 ) Nat. Neurosci. 1 , 653655 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

TRANSCRIPTIONAL AND POSTTRANSLATIONAL PLASTICITY AND THE GENERATION OF INFLAMMATORY PAIN 7730

21. Dib-Hajj, S. , Tyrrell, L. , Black, J. A. & Waxman, S. G. ( 1998 ) Proc. Natl. Acad. Sci. USA 95 , 89638968 .
22. Willis, W. D. & Coggeshall, R. E. ( 1991 ) Sensory Mechanisms of the Spinal Cord ( Plenum , New York ).
23. Yoshimura, M. & Jessell, T. M. ( 1990 ) J. Physiol. (London) 430 , 315335 .
24. King, A. E. , Thompson, S. W. N. , Urban, L. & Woolf, C. J. ( 1988 ) Neurosci. Lett. 89 , 286292 .
25. Cao, Y. Q. , Mantyh, P. W. , Carlson, E. J. , Gillespie, A. M. , Epstein, C. J. & Bausbaum, A. I. ( 1998 ) Nature (London) 392 , 390394 .
26. Woolf, C. J , Mannion, R. J. & Neumann, S. ( 1998 ) Neuron 20 , 10631066 .
27. Woolf, C. J. & Thompson, S. W. N. ( 1991 ) Pain 44 , 293299 .
28. Ma, Q.-P. & Woolf, C. J. ( 1995 ) J. Physiol. (London) 486.3 , 769777 .
29. Dubner, R. ( 1991 ) in Neuronal Plasticity and Pain Following Peripheral Tissue Inflammation or Nerve Injury , eds. Bond, M. R. , Charlton, J. E. &
Woolf, C. J. ( Elsevier , Amsterdam ), pp. 263276 .
30. McMahon, S. B. , Lewin, G. R. & Wall, P. D. ( 1993 ) Curr. Opin. Neurobiol. 3 , 602610 .
31. Coderre, T. J. & Melzack, R. ( 1992 ) J. Neurosci. 12 , 36653670 .
32. Ren, K. & Dubner, R. ( 1993 ) Neurosci. Lett. 163 , 2226 .
33. Woolf, C. J. , Shortland, P. & Sivilotti, L. G. ( 1994 ) Pain 58 , 141155 .
34. Koltzenburg, M. , Torebjork, H. E. & Wahren, L. K. ( 1994 ) Brain 117 , 579591 .
35. Kilo, S. , Schmelz, M. , Koltzenburg, M. & Handwerker, H. O. ( 1994 ) Brain 117 , 385396 .
36. Woolf, C. J. & Wall, P. D. ( 1986 ) J. Neurosci. 6 , 14331443 .
37. Bley, K. R. , Hunter, J. C. , Eglen, R. M. & Smith, J. A. ( 1998 ) Trends Pharmacol. Sci. 19 , 141147 .
38. Heath, M. J. , Womack, M. D. & MacDermott, A. B. ( 1994 ) J. Neurophysiol. 72 , 11921198 .
39. Zirrgiebel, U. , Ohga, Y. , Carter, B. , Berninger, B. , Inagaki, N. , Thoenen, H. & Lindholm, D. ( 1995 ) J. Neurochem. 65 , 2241 2250 .
40. Iwasaki, Y. , Gay, B. , Wada, K. & Koizumi, S. ( 1998 ) J. Neurochem. 71 , 106111 .
41. Suen, P.-C , Wu, K. , Xu, J.-L. , Lin, S. Y. , Levine, E. S. & Black, I. B. ( 1998 ) Brain Res. Mol. Brain Res. 59 , 215228 .
42. Yu, X. M. , Askalan, R. , Keil, G. J.2. & Salter, M. W. ( 1997 ) Science 275 , 674678 .
43. Lin, S. Y. , Wu, K. , Levine, E. S. , Mount, H. T. , Suen, P.-C. & Black, I. B. ( 1998 ) Brain Res. Mol. Brain Res. 55 , 2027 .
44. Wang, T. & Salter, M. W. ( 1994 ) Nature (London) 369 , 233235 .
45. Chen, L. & Huang, L.-Y. M. ( 1992 ) Nature (London) 356 , 521523 .
46. Woolf, C. J. & King, A. E. ( 1990 ) J. Neurosci. 10 , 27172726 .
47. Cook, A. J. , Woolf, C. J. , Wall, P. D. & McMahon, S. B. ( 1987 ) Nature (London) 325 , 151153 .
48. Simone, D. A. , Baumann, T. K. , Collins, J. G. & LaMotte, R. H. ( 1989 ) Brain Res. 486 , 185189 .
49. Torebjork, H. E. , Lundberg, L. E. R. & LaMotte, R. H. ( 1992 ) J. Physiol. (London) 448 , 765780 .
50. Ginty, D. D. , Kornhauser, J. M. , Thompson, M. A. , Bading, H. , Mayo, K. E. , Takahashi, J. S. & Greenberg, M. E. ( 1993 ) Science 260 , 238241 .
51. Fields, R. D. , Eshete, F. , Stevens, B. & Itoh, K. ( 1997 ) J. Neurosci. 17 , 72527266 .
52. Bito, H. , Deisseroth, K. & Tsien, R. W. ( 1997 ) Curr. Opin. Neurobiol. 7 , 419429 .
53. Ji, R. R. , Shi, T.-J. , Xu, Z.-Q. , Zhang, Q. , Sakagami, H. , Tsubochi, H. , Kondo, H. & Hokfelt, T. ( 1996 ) Brain Res. 721 , 167173 .
54. Tao, X. , Finkbeiner, S. , Arnold, D. B. , Shaywitz, A. J. & Greenberg, M. E. ( 1998 ) Neuron 20 , 709726 .
55. Shieh, P. B. , Hu, S. C. , Bobb, K. , Timmusk, T. & Ghosh, A. ( 1998 ) Neuron 20 , 727740 .
56. Swanson, K. D. , Reigh, C. & Landreth, G. E. ( 1998 ) J. Biol. Chem. 273 , 1996519971 .
57. Lev, S. , Moreno, H. , Martinez, R. , Canoll, P. , Peles, E. , Musacchio, J. M. , Plowman, G. D. , Rudy, B. & Schlessinger, J. ( 1995 ) Nature
(London) 376 , 737745 .
58. Michael, G. J. , Averill, S. , Nitkunan, A. , Rattray, M. , Bennett, D. L. H. , Yan, Q. & Priestley, J. V. ( 1997 ) J. Neurosci. 17 , 84768490 .
59. Gold, M. S. , Levine, J. D. & Correa, A. M. ( 1998 ) J. Neurosci. 18 , 1034510355 .
60. England, S. , Bevan, S. & Docherty, R. J. ( 1996 ) J. Physiol. (London) 495 , 429440 .
61. Gold, M. S. , Reichling, D. B. , Schuster, M. J. & Levine, J. D. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 11081112 .
62. Okuse, K. , Chaplan, S. R. , McMahon, S. B. , Luo, Z. D. , Calcutt, N. A. , Scott, B. P. , Akopian, A. N. & Wood, J. N. ( 1997 ) Mol. Cell. Biol. 10 ,
196207 .
63. Woolf, C. J. , Safieh-Garabedian, B. , Ma, Q.-P. , Crilly, P. & Winter, J. ( 1994 ) Neuroscience 62 , 327331 .
64. Koltzenburg, M. , Lundberg, L. E. R. & Torebjork, H. E. ( 1992 ) Pain 51 , 207220 .
65. Stubhaug, A. , Breivik, H. , Eide, P. K. , Kreunen, M. & Foss, A. ( 1997 ) Acta Anaesthesiol. Scand. 41 , 11241132 .
66. Leslie, T. A. , Emson, P. C. , Dowd, P. M. & Woolf, C. J. ( 1995 ) Neuroscience 67 , 753761 .
67. Noguchi, K. , Dubner, R. & Ruda, M. A. ( 1992 ) Neuroscience 46 , 561570 .
68. Noguchi, K. , Kowalski, K. , Traub, R. , Solodkin, A. , Iadarola, M. J. & Ruda, M. A. ( 1991 ) Mol. Brain Res. 10 , 227233 .
69. Ruda, M. A. , Iadarola, M. J. , Cohen, L. V. & Young, W. S., III. ( 1988 ) Proc. Natl. Acad. Sci. USA 85 , 622626 .
70. Ji, R. R. & Rupp, F. ( 1997 ) J. Neurosci. 17 , 17761785 .
71. Woolf, C. J. ( 1996 ) Philos. Trans. R. Soc. London B 351 , 441448 .
72. Davis, R. J. ( 1993 ) J. Biol. Chem. 268 , 1455314556 .
73. Seger, R. & Krebs, E. G. ( 1995 ) FASEB J. 9 , 726735 .
74. Fukunaga, K. & Miyamoto, E. ( 1998 ) Mol. Neurobiol. 16 , 7995 .
75. Xing, J. , Ginty, D. D. & Greenberg, M. E. ( 1996 ) Science 273 , 959963 .
76. Impey, S. , Obrietan, K. , Wong, S. T. , Poser, S. , Yano, S. , Wayman, G. , Deloulme, J. C. , Chan, G. & Storm, D. R. ( 1998 ) Neuron 21 , 869883 .
77. Averill, S. & McMahon, S. B. ( 1998 ) Eur. J. Neurosci. 7 , 1484 1494 .
78. McMahon, S. B. , Armanini, M. P. , Ling, L. H. & Phillips, H. S. ( 1994 ) Neuron 12 , 11611171 .
79. Krause, J. E. , DiMaggio, D. A. & McCarson, K. E. ( 1995 ) Can. J. Physiol. Pharmacol. 73 , 854859 .
80. Ma, Q.-P. & Woolf, C. J. ( 1996 ) Pain 67 , 97106 .
81. Ma, Q.-P. & Woolf, C. J. ( 1997 ) NeuroReport 8 , 807810 .
82. Ma, Q.-P. , Allchorne, A. J. & Woolf, C. J. ( 1998 ) Pain 77 , 4957 .
83. Ma, Q.-P. & Woolf, C. J. ( 1997 ) Eur. J. Pharmacol. 322 , 165171 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CELLULAR MECHANISMS OF NEUROPATHIC PAIN, MORPHINE TOLERANCE, AND THEIR INTERACTIONS 7731

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Cellular mechanisms of neuropathic pain, morphine tolerance, and

their interactions

(hyperalgesia/protein kinase C/excitatory amino acid/analgesia/nociception)

DAVID J. MAYER , JIANREN MAO , JASON HOLT, AND DONALD D. PRICE
Department of Anesthesiology, Medical College of Virginia, Virginia Commonwealth University, Richmond, VA 23298;
Department of Cell Biology, Duke University School of Medicine, Durham, NC 27710; and Departments of Oral Surgery and

Neuroscience, University of Florida, Gainesville, FL 32610

ABSTRACT Compelling evidence has accumulated over the last several years from our laboratory, as well as others,
indicating that central hyperactive states resulting from neuronal plastic changes within the spinal cord play a critical role in
hyperalgesia associated with nerve injury and inflammation. In our laboratory, chronic constriction injury of the common sciatic
nerve, a rat model of neuropathic pain, has been shown to result in activation of central nervous system excitatory amino acid
receptors and subsequent intracellular cascades including protein kinase C translocation and activation, nitric oxide production,
and nitric oxide-activated poly(ADP ribose) synthetase activation. Similar cellular mechanisms also have been implicated in the
development of tolerance to the analgesic effects of morphine. A recently observed phenomenon, the development of dark
neurons, is associated with both chronic constriction injury and morphine tolerance. A site of action involved in both
hyperalgesia and morphine tolerance is in the superficial laminae of the spinal cord dorsal horn. These observations suggest that
hyperalgesia and morphine tolerance may be interrelated at the level of the superficial laminae of the dorsal horn by common
neural substrates that interact at the level of excitatory amino acid receptor activation and subsequent intracellular events. The
demonstration of interrelationships between neural mechanisms underlying hyperalgesia and morphine tolerance may lead to a
better understanding of the neurobiology of these two phenomena in particular and pain in general. This knowledge may also
provide a scientific basis for improved pain management with opiate analgesics.
A number of studies, both from our laboratory as well as others, indicate that central hyperactive states resulting from neuronal
plastic changes within the spinal cord play a critical role in hyperalgesia associated with nerve injury and inflammation. We have recently
shown in a rat model of neuropathic pain that chronic constrictive injury (CCI) of the common sciatic nerve can result in activation of
central nervous system excitatory amino acid receptors and subsequent intracellular cascades including protein kinase C translocation and
activation, nitric oxide (NO) production, and NO-activated poly(ADP ribose) synthetase (PARS) activation. Similar cellular mechanisms
also have been implicated in the development of tolerance to the analgesic effects of morphine. Of particular interest is that morphological
changes in the spinal cord dorsal horn, the development of so called dark neurons, are associated with both CCI and morphine tolerance.
A site of action involved in both hyperalgesia and morphine tolerance has also been shown to be in the superficial laminae of the spinal
cord dorsal horn. We will first summarize recent evidence indicating central mechanisms of hyperalgesia and morphine tolerance. The
main focus of this article will be on the recent development in our understanding of the involvement of PARS activation in central
mechanisms of morphine tolerance. A working hypothesis with regard to interactions between hyperalgesia and morphine tolerance within
the spinal cord dorsal horn will be discussed. Finally, clinical implications of such interactions will be addressed.
Central Mechanisms Subserving Hyperalgesia. Recent insights into neural mechanisms of hyperalgesia are based on knowledge of
the involvement of the N-methyl-D-aspartate (NMDA) receptor and associated intracellular cascades. These insights result from studies of
central nervous system (CNS) neuronal plasticity in general and from studies of spinal cord mechanisms of neurogenic and inflammatory
hyperalgesia in particular ( 1 8 ). Hyperalgesia after tissue injury and inflammation may reflect central sensitization resulting from
prolonged and excessive activation of spinal cord excitatory amino acid receptors and subsequent intracellular cascades. Tonic activation
of NMDA receptors activates second-messenger systems, an ultimate result of which is phosphorylation and hence sensitization of ion
channel complexes, including that of the NMDA receptor ( 9 ). These central changes are initiated by abnormal and often tonic input to the
spinal cord. This tonic input, in turn, may result from peripheral nerve injury or tissue inflammation. Potential peripheral generators of this
input include nerve injury-induced impulse discharges ( 3 ), generation of ectopic nerve action potentials ( 10 ), aberrant sympathetic
influences ( 11 ), and/or sensitization of peripheral nociceptors ( 2 , 11 ). Of the several central changes initiated by tonic nociceptive
afferent input, protein kinase C (PKC) translocation/activation and/or NO production are pivotal intracellular events within spinal cord
neurons. Translocation/activation of PKC enhances postsynaptic neuronal excitability by means of increasing the efficacy of receptorion
channel complexes ( 12 15 ). Similar changes may also occur at presynaptic sites via activation of presynaptic NMDA receptors
localized on primary afferent fibers ( 16 ) and/or via the effects of extracellular NO ( 17 ). Central hyperactive states reflect the combined
effects of these pre- and postsynaptic mechanisms. Direct support of the involvement of PKC in these mechanisms is provided by an
experiment in which a phorbol ester (a PKC activator) increased spontaneous and stimulusevoked activity of dorsal horn spinothalamic
tract neurons ( 18 ) and by an experiment in which a PKC activator produced a long-lasting increase in the amplitude and duration of
excitatory postsynaptic potentials evoked in dorsal horn neurons by orthodromic dorsal root stimulation ( 19 ). Both experiments show that
activation of PKC can indeed modulate activity of spinal cord
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: PARS, poly(ADP ribose) synthetase; CCI, chronic constrictive injury; NMDA, N-methyl-D-aspartate; CNS, central
nervous system; PKC, protein kinase C; i.t., intrathecal.
To whom reprint requests should be addressed, e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CELLULAR MECHANISMS OF NEUROPATHIC PAIN, MORPHINE TOLERANCE, AND THEIR INTERACTIONS 7732

nociceptive neurons. These experiments are complemented by studies that demonstrate elevated dorsal horn levels of translocated PKC or
increased biosynthesis of PKC in neuropathic rats ( 20 , 21 ).
Contributions of Central Excitation to Hyperalgesia. It has been proposed that central sensitization may be manifested as
increases in spontaneous and stimulus-evoked neuronal activity within the spinal cord, which, in turn, contribute to the development and
maintenance of neurogenic and inflammatory pain syndromes. There may be an increase in spontaneous neural activity of pain
transmission pathways and hence spontaneous pain. This increased spontaneous action potential activity includes that of spinal cord dorsal
horn neurons ( 22 , 23 ), and that of pain-related thalamic neurons ( 24 ) of CCI rats with neuropathic hyperalgesia ( 25 ). These extensive
elevations in neural activity, which have also been mapped by using the 2-deoxyglucose metabolic technique ( 26 , 27 ), occur within the
spinal cord and a variety of pain-related brain regions even in the absence of overt somatic stimulation. Furthermore, responses of central
pain-related neurons of CCI rats show exaggerated responses to innocuous and noxious peripheral stimulation. This is indicated by
increased responses of spinothalamic neurons to mechanical or thermal stimulation ( 23 ). An exaggerated response to innocuous
stimulation may contribute to allodynia, whereas the enhanced response to noxious stimulation is likely related to hyperalgesia. Peripheral
receptive fields also expand after tissue inflammation ( 28 ). Given the role of receptive field size in neuronal recruitment, expanded
receptive fields would mean that a nociceptive stimulus would activate more central neurons than would normally occur, leading to
exaggerated pain and exaggerated spatial radiation of the painful sensation. Taken together, elevated spontaneous discharges, increased
stimulus-evoked impulse frequencies, and expanded receptive fields are likely to operate in concert to cause persistent hyperalgesia,
spontaneous pain, allodynia, and radiation of pain.
Contributions of Central Disinhibition to Hyperalgesia. Excitatory amino acid-induced PKC translocation/activation and NO
production may also result in disinhibitory processes associated with excitotoxic consequences including neuronal death within the CNS.
The involvement of such excitotoxic processes in mechanisms of injury-induced pain syndromes has been suggested by several
experimental observations. There is histological evidence showing that peripheral nerve injury induces excitotoxic transsynaptic
morphological changes of superficial dorsal horn (laminae III) neurons (dark neurons), which have been proposed to be inhibitory
interneurons ( 29 ). Importantly, a PARS has been shown to be activated in CCI rats, which is likely to contribute to the development of
dark neurons ( 30 ). The activation of PARS occurs in the presence of DNA fragmentation, a process that may be initiated by the
production of NO and that may eventually lead to programmed cell death. These morphological changes may reflect a pathological
process in which injury-induced central responses result in a persistent imbalance of the excitatoryinhibitory circuitry within the spinal
cord dorsal horn. This possibility is supported by the demonstration of A-mediated mechanical allodynia in neuropathic pain patients (
31 ) and the demonstration that pharmacological blockade of the spinal cord -aminobutyric acid and glycine inhibitory system augments
thermal hyperalgesia in rats with sciatic nerve injury ( 32 ). Thus, in addition to direct central sensitzation described above, the possible
disinhibition resulting from the loss of function of spinal cord inhibitory interneurons may also contribute to central hyperexcitability after
peripheral nerve injury and inflammation.
Central Mechanisms Subserving Morphine Tolerance. Similar mechanisms, including the activation of NMDA receptors,
translocation and activation of PKC, and the production of NO have been implicated in the development and maintenance of morphine
tolerance. Detailed discussion of these mechanisms of morphine tolerance is presented elsewhere ( 9 ). The following sections will
emphasize new evidence indicating the involvement of PARS in mechanisms of morphine tolerance.
As discussed above for the development of hyperalgesia, a consequence of excessive NMDA receptor activation is the initiation of a
cascade of intracellular events leading to PARS activation ( 33 ). In view of the evidence implicating NMDA receptor activation and the
subsequent intracellular events resulting from this, we have conducted a series of experiments to examine the possible role of PARS in the
development of morphine tolerance. In addition, because a consequence of PARS activation may include alterations in cell morphology,
we examined the development of dark neurons as a marker of altered cellular morphology.
For these experiments, adult male SpragueDawley rats were prepared for intrathecal (i.t.) injection by implanting a polyethylene
(PE-10) tube into the lumbar spinal cord. The vehicle or drugs were delivered slowly (13 minutes) through the i.t. catheter followed by 10
l of saline, which flushed the catheter. To examine the antinociceptive effects of morphine, the tail-flick test was used. The light bulb
intensity of the tail-flick device was set to produce baseline tail-flick latencies between 3.5 and 4.5 sec. To ensure that no tissue damage
occurred, the light bulb was automatically turned off after 8 sec, even if no tail-flick occurred. An average of three tail-flick trials each
separated by a 1-min intertrial interval constituted the mean baseline latency (BL). The antinociceptive effects of morphine and the
development of tolerance were then determined by measuring test (tail-flick) latencies (TL) after drug administration. Thus, data were
expressed as percent maximal possible antinociceptive effect (%MPAE) using the equation %MPAE = [(TL BL)/(8 BL)] 100.
The procedure for the histological examination of dark neurons was similar to that described previously ( 29 , 30 ). It has been
reported by Sugimoto et al. ( 29 ) that there are three principal characteristics of dark neurons: (i) irregular cellular outlines; (ii) increased
amounts of chromatin throughout the nucleoplasm and cytoplasm, which is why they are called dark neurons; and (iii) intensely and
homogeneously stained nucleoplasm with almost indiscernible heterochromatin. Dorsal horn neurons that are normal sometimes show
increased cytoplasmic staining also, but they do not exhibit enhanced nucleoplasmic staining. Some glial cells, especially
oligodendrocytes, also exhibit chromophilia that is as dark as that seen in dark neurons. Although this is true, oligodendrocytes are
distinguishable from dark neurons, because glial cells have aggregates of heterochromatin that are seen under high-power magnification,
whereas dark neurons do not exhibit this characteristic. Neurons that exhibited all three of the above-mentioned characteristics were
counted as dark neurons, whereas other cells were excluded.
The spinal cord sections were divided into three zones for examination by microscope. They were divided, by using Rexeds laminar
system, into laminae III, IIIIV, and VVI. Sections were examined with the experimenter blind to the treatment regimen. Dark neurons
were counted in each zone mentioned above under medium-power magnification. If there was a dark neuron that was in question, it was
then viewed under high-power magnification to discern if it was indeed a dark neuron. At least three 1-m spinal cord sections were
viewed for each rat. This yielded an average number of dark neurons for each subdivision.
The tail-flick data were analyzed by using two-way ANOVA to discern differences among treatment groups. When main effects were
seen, a WallerDuncan D ratio t test was performed to determine the source of variations between the groups. Dark neurons were counted
for the left and right sides of the dorsal horn in laminae IVI. The numbers were then averaged and analyzed by using a two-way
the authoritative version for attribution.

ANOVA. The total number of dark neurons from a sampled region was analyzed to determine (i) differences in the number of dark
neurons between the left and right side of the dorsal horn; (ii) differences in numbers of dark neurons among sampled dorsal horn regions
(i.e., laminae III);
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CELLULAR MECHANISMS OF NEUROPATHIC PAIN, MORPHINE TOLERANCE, AND THEIR INTERACTIONS 7733

and (iii) differences in the number of dark neurons among treatment groups.
An initial experiment was performed to determine the effect of the chronic administration of morphine on the development of dark
neurons. Three groups of rats (n = 59 per group) were used: (i) rats treated with saline i.t. once a day; (ii) rats treated with 10 g of
morphine i.t. once daily; and (iii) rats treated with 20 g of morphine i.t. once daily. The doses of morphine given have previously been
shown to induce the development of tolerance to the antinociceptive effects of morphine ( 34 ). All of the agents were given once daily for
8 days.
On day 8, those rats receiving morphine and saline showed the development of tolerance to the analgesic effects of morphine. Rats
that were made tolerant to morphine exhibited a reliable increase in the number of dark neurons in the dorsal horn of the lumbar spinal
cord (P < 0.01). Several features characterized this increase in dark neurons. Dark neurons were primarily located in laminae III and to a
much lesser degree to laminae IIIIV. Also, there was no statistical difference in the number of dark neurons observed on the left and right
sides of the spinal cord (P > 0.05). Because chronic administration of morphine induced tolerance and the development of dark neurons,
we examined the effect of the selective PARS inhibitor benzamide, which has been shown to prevent the development of hyperalgesia in
the CCI model ( 30 ), on the development of morphine tolerance and dark neurons resulting from chronic morphine administration. For
this experiment, seven groups of rats were used. They included rats receiving 100, 200, or 400 nmol benzamide and 20 g of morphine on
days 18, rats receiving 400 nmol benzamide and saline on days 18, rats receiving 20 g morphine and saline on days 18, rats receiving
only saline on days 18, and rats receiving saline on days 17 and 20 g of morphine on day 8.
As shown in Fig. 1 , coadministration of 20 g of morphine with 200 or 400 nmol (not 100 nmol) benzamide for 7 days reliably
attenuated the development of tolerance (P < 0.01). Neither baseline tail-flick latency nor the response to a single injection of 20 g of
morphine changed after repeated saline treatment for 7 days. Coadministration of 20 g of morphine with benzamide (100400 nmol) for
7 days also reliably prevented the increase in dark neurons (P < 0.01; Fig. 2 ). Neither repeated benzamide (400 nmol) treatment alone nor
a single injection of 20 g of morphine on day 8 (the 20*/0 group) affected the occurrence of dark neurons as compared with the saline
group.

FIG. 1. Effect of benzamide on morphine tolerance. Tolerance to the antinociceptive effect of morphine developed in rats treated with
20 g of morphine for 7 days. Coadministration of 20 g of morphine with 200 or 400 nmol (not 100 nmol) benzamide for 7 days
reliably attenuated the development of tolerance. Neither baseline tail-flick latency nor the response to a single injection of 20 g of
morphine changed after repeated saline treatment for seven days. **, P < 0.01, as compared with that of day 1 in each corresponding
group. Neither repeated benzamide (400 nmol) treatment alone nor a single injection of 20 of g morphine on day 8 (the 20*/0 group)
affected the degree of tolerance as compared with the saline group. MPAE%, percent of maximal possible antinociceptive effect.
The data from the previous experiment showed that benzamide was effective in inhibiting the development of morphine tolerance and
dark neurons. The specificity of this effect to PARS inhibition was examined by utilizing other PARS inhibitors. For this experiment, four
groups of rats were used; they included rats receiving 400 nmol benzamide and 20 g of morphine on days 18, rats receiving 200 nmol 3-
aminobenzamide and 20 g of morphine on days 18, rats receiving 1 mol niacinamide (nicotinamide) and 20 g of morphine on days 1
8, and rats receiving 20 g of morphine and saline on days 18. Coadministration of 20 g of morphine with either 200 nmol 3-
aminobenzamide or 1 mol niacinamide (nicotinamide) for 7 days reliably (P < 0.01) attenuated the development of tolerance as
compared with that of day 1 in the same group ( Fig. 3 ). As shown in Fig. 4 , coadministration of 20 g of morphine with either 200 nmol
3-aminobenzamide or 1 mol niacinamide for 7 days also reliably (P < 0.05 and P < 0.01 for the drugs respectively) prevented the
increase in dark neurons as compared with the morphine + saline group.
To confirm that the development of tolerance and dark neurons in morphine-treated rats is associated with the activation of opioid
receptors, we examined the effect of the opioid receptor antagonist, naltrexone, on the ability of morphine to produce tolerance and dark
neurons. For this experiment, two groups of rats were used; they included rats receiving 10 mg/kg naltrexone intraperitoneally 5 min
before 20 g of morphine i.t. on days 18 and rats receiving 20 g of morphine and saline on days 18. Coadministration of 20 g of
morphine with 10 mg/kg naltrexone for 7 days reliably prevented both the development of the antinociceptive tolerance (P < 0.01) and the
increase in dark neurons (P < 0.01) as compared with the morphine + saline group.
The major findings of this series of studies are (i) the incidence of dark neurons increased significantly within the spinal cord dorsal
horn, particularly the superficial laminae III, of rats injected daily for 8 days with i.t. morphine; (ii) benzamide and other PARS inhibitors
reduced or prevented the development of
the authoritative version for attribution.

FIG. 2. Effect of benzamide on incidence of dark neurons. Coadministration of 20 g of morphine with benzamide (100400 nmol)
for 7 days reliably prevented the increase in dark neurons. Neither repeated benzamide (400 nmol) treatment alone nor a single
injection of 20 g of morphine on day 8 (the 20*/0 group) affected the occurrence of dark neuron as compared with the saline group.
**, P < 0.01, as compared with each of the rest groups.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CELLULAR MECHANISMS OF NEUROPATHIC PAIN, MORPHINE TOLERANCE, AND THEIR INTERACTIONS 7734

analgesic tolerance and dark neurons; and (iii) the development of dark neurons after repeated morphine injection is opioid receptor-
mediated, because concurrent administration of naltrexone with morphine prevents the development of dark neurons. These studies
demonstrate that in vivo administration of morphine for 8 days produces dark neurons similar in morphology and location to transsynaptic
alterations resulting from chronic constrictive injury of the sciatic nerve. The morphological characteristics of these cells are consistent
with those of cells undergoing programmed cell death.

FIG. 3. Effect of nicotinamide and 3-aminobenzamide on morphine tolerance. Coadministration of 20 g of morphine with either 200
nmol 3-aminobenzamide or 1 mol niacinamide (nicotinamide) for 7 days reliably attenuated the development of tolerance. **, P <
0.01, as compared with that of day 1 in the same group.
These results, along with others, indicate that the cascade of events leading to the formation of dark neurons in the spinal cord after
morphine administration involves activation of the NMDA receptor, PKC, NO, and PARS. Thus, morphine tolerance may be reduced by
intervening at any one of these steps. These findings suggest that neurotoxicity contributes to neural mechanisms underlying opioid
tolerance and that PARS inhibitors are protective from such neurotoxicity. The results from these studies, combined with our results in
similar studies using the CCI model ( 30 ), suggest that opioids may exacerbate the excitotoxicity underlying at least certain types of
neuropathic pain. A corollary of this is that the excitotoxicity from neuropathic pain may, under some circumstances, reduce the response
to opioids.

FIG. 4. Effect of nicotinamide and 3-aminobenzamide on incidence of dark neurons. Coadministration of 20 g of morphine with
either 200 nmol 3-aminobenzamide or 1 mol niacinamide (nicotinamide) for 7 days also reliably prevented the increase in dark
neurons. a, P < 0.05; b, P < 0.01, as compared with the MSO4 + saline group.
Several perplexing observations about neuropathic pain, opioid tolerance, and the interactions between them led us to propose an
early model of the events involved ( 34 ) and now lead us to propose a revised version of this model here ( Fig. 5 ). These observations
include the following. (i) Neuropathic pain syndromes often present with symptoms indicative of both hyperexitability (e.g., hyperalgesia)
and disinhibition (e.g., spontaneous pain, allodynia). Some of these symptoms, such as hyperalgesia, can be at least partially reversed in
animal models by NMDA receptor antagonists ( 35 ), whereas others, such as allodynia, cannot ( 36 ). (ii) NMDA receptor antagonists
prevent, but do not acutely reverse, tolerance to opioids ( 37 ). (iii) Our model assumes, for numerous reasons ( 34 ), that postsynaptic
opioid and NMDA receptors are on the same neurons, at least in the spinal cord. In fact, we have reported direct immunohistochemical
evidence of this ( 38 ). Opioids are known to produce hyperpolarization via an inwardly rectifying K+ channel. Despite the hyperpolarized
state in these cells, voltage/ligand-gated NMDA receptors must be activated during the process of tolerance development, because NMDA
receptor antagonists block the development of tolerance. (iv) At the spinal cord level, it is likely, except in the presence of nociceptive
input, that presynaptic elements release only small amounts of glutamate onto postsynaptic elements in the superficial laminae of the
dorsal horn that contain NMDA receptors related to opioid tolerance, yet tolerance to opioids occurs in the absence of nociceptive input.
(v) PKC translocation to the cell membrane is greatly increased by chronic as compared with acute morphine treatment ( 34 ). (vi) Chronic
opioid administration produces hyperalgesia, which can be reversed by NMDA receptor antagonists ( 27 ). (vii) CCI, which produces
hyperalgesia causes a rightward shift in the dose response curve to morphine ( 39 ). (viii) NO is involved in opioid tolerance ( 40 ). (ix)
CCI-induced hyperalgesia ( 30 ) and opioid tolerance ( 39 ) are associated with the development of dark neurons ( 30 ). As reviewed in this
article, the development of hyperalgesia, morphine tolerance, and dark neurons can be prevented by inhibiting the nuclear repair enzyme
PARS.
A model of our current working hypothesis concerning the development of neuropathic pain, opioid tolerance, and their interactions
that is consistent with these observations is presented in Fig. 5 . In this model, excessive release of glutamate resulting from peripheral
events, such as those occurring in the CCI model, initiates a series of intracellular events, which, via different messenger systems, leads to
(i) at least partially NMDA antagonist reversible hyperexcitability that results from a PKC-mediated alteration of NMDA receptors
(pathway 1, Fig. 5 ); (ii) events such as allodynia, which are not reversible with NMDA antagonists ( 36 ) but are prevented by inhibition
of the NO/PARS pathway ( 30 ) and thus may be mediated by cellular dysfunction resulting from depletion of cellular energy stores
the authoritative version for attribution.

(pathway 2, Fig.5 ). This cellular dysfunction may be morphologically manifested by the appearance of dark neurons, which can also be
prevented by inhibition of the NO/PARS pathway ( 30 ); and (iii) the rightward shift of the morphine doseresponse curve resulting from
CCI ( 39 ), which has been hypothesized to result from relatively long duration, PKC-mediated alterations in gene expression (pathway 3,
Fig. 5 ), which may result in PKC-mediated opioid receptor/K+ channel uncoupling ( 41 ).
With regard to opioid tolerance, in this model, activation of the -opioid receptor may initiate PKC translocation to the membrane (
42 ). This PKC translocation allows the NMDA receptor to function as a ligand-gated channel by removal of the voltagedependent Mg2+
blockade ( 15 ). The removal of the Mg2+ blockade from the NMDA receptor allows for an increased influx of Ca2+ despite membrane
hyperpolarization by -opioids and low levels of presynaptic glutamate release. This influx of Ca2+ has two effects. It activates either a
separate pool of PKC (PKC2) or much greater amounts of the original pool of PKC1. The other
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CELLULAR MECHANISMS OF NEUROPATHIC PAIN, MORPHINE TOLERANCE, AND THEIR INTERACTIONS 7735

pool of PKC may be translocated directly to the membrane, modifying various excitatory amino acid and/or other receptors (pathway 1),
and/or it may modify nuclear transcription (pathway 3), the products of which result in delayed and persistent changes in cellular function
such as opioid receptor/K+ channel uncoupling ( 41 ). This consequence cannot be reversed by acute administration of NMDA antagonists.
A second effect of the influx of Ca2+ is that it activates NO synthase, which increases the production of NO and superoxide ( 43 ) (pathway
2). The simultaneous generation of these two molecules favors the production of peroxynitrite (ONOO) ( 44 ). ONOO is a very potent
initiator of DNA strand breakage ( 45 ). Thus, ONOO initiates the production of the nuclear repair enzyme, PARS. Pronounced activation
of PARS can result in cell dysfunction and eventually cell death because of inhibition of mitochondrial respiration and depletion of
cellular energy stores ( 33 ). This then leads to the formation of dark neurons perhaps by way of programmed cell death. Such an
excitotoxic cascade may underlie at least some aspects of opioid tolerance. A third series of events initiated by influx of Ca2+ is the
activation of either a separate pool of PKC (PKC2, Fig. 5 ) or much greater amounts of the original pool of PKC1. This other pool of PKC
may function as a transcription factor resulting in the production of more PKC (PKC3, Fig. 5 ), which can modulate -opioid receptor
responsiveness ( 46 ), resulting in desensitization of the -opioid receptor coupling through its associated G protein to an inwardly
rectifying K+ channel, a mechanism know to exist in vitro ( 40 ).

FIG. 5. A proposed model for the excitotoxic formation of dark neurons in the dorsal horn of the spinal cord from peripheral nerve
injury or repeated morphine administration. Excessive excitation of the NMDA receptor and subsequent influx of Ca2+ occurs either
directly by glutamate release from primary afferent input (CCI model) or indirectly by activation of -opioid receptors (repeated
opiate administration). Activation of the -opioid receptor results in indirect NMDA receptor activation by initiating a second-
messenger PKC translocation to the membrane( 16 ) This PKC translocation activates the NMDA receptor by removal of the Mg2+
blockade( 17 ). The removal of the Mg2+ blockade from the NMDA receptor allows for an increased influx of Ca2+. The influx of Ca2
+, via direct or indirect activation of the NMDA receptor, has several effects ( 5 ). It activates either a separate pool of PKC (PKC ) or
2
much greater amounts of the original pool of PKC1. This second pool of PKC may be translocated directly to the membrane,
modifying various excitatory amino acid or other receptors. It also may function as a transcription factor, resulting in the production
of more PKC (PKC3), which can result in uncoupling of the -opioid receptor from its associated G protein. Another effect of the
influx of Ca2+ is that it activates NO synthase, which increases the production of NO. In addition, the influx of Ca2+ results in the
production of superoxide from mitochondria. The simultaneous generation of these two molecules favors the production of
peroxynitrite (ONOO), a very potent initiator of DNA strand breakage, which, in turn, initiates the production of the nuclear repair
enzyme, PARS. Pronounced activation of PARS can result in cell dysfunction and eventually cell death because of inhibition of
mitochondrial respiration and depletion of cellular energy stores, which in turn may lead to the formation of dark neurons, perhaps by
way of programmed cell death. PKCx, various pools of protein kinase C; GPro, heterotrimeric guanine nucleotide binding protein;
NOS, nitric oxide synthase.
Although we have focused here on NMDA receptor activation as a primary initiator of neuropathic pain, opioid tolerance, and their
interactions, these events, as we have pointed out previously (34), are likely to involve additional factors. It is likely that Ca2+/calmodulin,
cAMP, and other second and third messengers, non-NMDA receptors, and cholecystokinin ( 47 ) and other nonglutamate receptors also
participate. Nevertheless, it is clear from the data reviewed here that the complex events resulting from NMDA receptor activation are of
critical importance in neuropathic pain, opioid tolerance, and their interactions.
Clinical Implications. Recent progress in investigating neural mechanisms subserving neuropathic and inflammatory pain as well as
opioid tolerance has significantly advanced our knowledge about pain and pain modulation. These studies represent two important
frontiers in pain research. First, these studies have led to the concept that pathological pain may reflect a disease process with both
the authoritative version for attribution.

dynamic and progressive changes during its course ( 1 , 2 , 9 ). A key feature of this process is that neuronal plastic changes occur within
the CNS in association with the progress of pathological pain states. This concept provides, at least in part, a basis for explaining
pathological pain that often persists long after the initial insults. Second, mechanisms of opioid tolerance may involve neuroplastic
changes within the CNS as well ( 9 , 39 ). Neuroplastic changes in relation to opioid tolerance have much in common with those of
pathological pain, both of which begin with the activation of NMDA receptors ( 9 ). Evidence also exists indicating that interactions do
indeed occur between cellular and intracellular mechanisms of pathological pain and opioid tolerance, and such interactions are likely to
be a contributing factor to a generally weak analgesic effect of opioids in pathological pain states ( 39 ).
Thus far, little information exists with regard to interactions between hyperalgesia and analgesic tolerance in man. It is conceivable
that reduced morphine analgesia after repeated administration to patients with chronic pain could result from the development of both
pharmacological tolerance to morphine analgesia and tolerance-associated hyperalgesia. Because of the coincidental development of
morphine tolerance and tolerance-associated hyperalgesia, progressively higher morphine doses may be needed to overcome both
conditions. In turn, a vicious cycle may be initiated involving higher opiate doses, more tolerance, and greater hyperalgesia. Thus, the need
for higher opiate doses in a clinical setting of opiate treatment could be partly due to the
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

CELLULAR MECHANISMS OF NEUROPATHIC PAIN, MORPHINE TOLERANCE, AND THEIR INTERACTIONS 7736

development of hyperalgesia that may result from repeated opiate administration. Conceivably, the development of tolerance-associated
hyperalgesic states may also contribute to withdrawal signs of opioid dependence. This possibility also suggests that an inappropriate
opiate treatment schedule in pain management may precipitate unexpected hyperalgesic responses to preexisting pain conditions and could
be a source of the clinical complexity with respect to the responsiveness to opiate treatment.
On the other hand, evidence indicating hyperalgesia-associated reduction of morphine antinociception ( 27 ) has bearing on the
controversy concerning opiate effects or lack of effects on neuropathic pain states in man. It is conceivable that the diversity of clinical
response patterns to opiate treatment in neuropathic pain patients may result from varying degrees of CNS neuronal plastic changes
initiated by nerve injury or injury to other tissues. Such neuronal plastic changes can underlie the development of neuropathic pain
syndromes and result in reduced morphine analgesia even before opiate treatment starts ( 9 ). To complicate matters further, neuropathic
pain syndromes as well as other chronic pain states (such as cancer pain) often present a dynamic and progressive course that demands
increased opiate doses for adequate pain relief.
The complexity of opioid tolerance, hyperalgesia, and their interactions calls for a new look into some clinical issues of pain
management. Because the development of pathological pain states often involves the activation of NMDA receptors and because there
exists an intimate relationship between the NMDA and opioid receptor systems that may lead to changes in responsiveness to opioid
analgesics, early recognition of clinical conditions that may lead to the development of pathological pain states would be of utmost
importance. Such clinical conditions may include, but may not be limited to, nerve injury, tissue inflammation, and prolonged and ongoing
peripheral nociceptive input (such as those seen in persistent postoperative pain). Some of these conditions (e.g., postoperative pain) may
be treated preemptively to produce a favorable outcome. However, nerve injury and inflammatory tissue diseases often occur in an
unpredictable manner, thereby obviating the possibility of preemptive treatment. Thus, effective early interruption of ongoing nociceptive
input from the injured site (e.g., using nerve block or field block), thereby reducing CNS activation of NMDA receptors, could be the key
to preventing or minimizing changes in NMDA and opioid systems that may eventually lead to persistent pain and reduced opioid
effectiveness for treating such pain states.
As discussed above, interactions between NMDA and opioid receptors could occur in both directions ( 9 ). Thus, any condition that
results in activation of NMDA receptors within the CNS could modulate opioid receptors, causing reduced efficacy of opioid analgesia;
conversely, repeated treatment with opioids could set up a condition mimicking ongoing nociceptive input through interactions between
opioid and NMDA receptors ( 9 , 39 ). This concept is the basis for recommending a combined use of opioids and clinically available
NMDA receptor antagonists ( 9 ). Importantly, such a strategy should be integrated into the treatment regimen not only for chronic pain
management (treating an existing pain condition) but also for preventing an evolving pain condition, such as that after nerve injury. By the
same token, effective nerve block or field block should be part of an integrated therapeutic regimen for treating clinical conditions that
may later lead to the development of intractable chronic pain states. It should be noted that, although early treatment of pain after tissue
injury and inflammation with opioids often provides satisfactory clinical pain relief, opioids alone offer little help for stopping the process
of an evolving pathological pain state, because evidence presented in this article and elsewhere suggests that opioids alone could actually
contribute to the development of neuronal plastic changes via interactions with NMDA receptors. It can be anticipated that our further
understanding of neural mechanisms subserving hyperalgesia, opioid tolerance, and their interactions would advance and improve clinical
management of debilitating and intractable pain syndromes.
Portions of this work were supported by Public Health Service Grants DA08835 and NS24009.
1. Woolf, C. J. & Thompson, S. W. N. ( 1991 ) Pain 44 , 293299 .
2. Dubner, R. ( 1991 ) in Proceedings of 5th World Congress on Pain , ed. Bond, M. , Charlton, E. & Woolf, C. J. ( Elsevier , Amsterdam ), Vol. 5 , pp.
263276 .
3. Seltzer, Z. , Cohn, S. , Ginzgurg, R. & Berlin, B. ( 1991 ) Pain 45 , 6975 .
4. Davar, G. , Hama, A. , Deykin, A. , Vos, B. & Maciewicz, R. ( 1991 ) Brain Res. 553 , 327330 .
5. Mao, J. , Mayer, D. J. , Hayes, R. L. , Lu, J. & Price, D. D. ( 1992 ) Brain Res. 598 , 271278 .
6. Mao, J. , Price, D. D. , Mayer, D. J. , Lu, J. & Hayes, R. L. ( 1992 ) Brain Res. 576 , 254262 .
7. Mao, J. , Price, D. D. , Hayes, R. L. , Lu, J. , Mayer, D. J. & Frenk, H. ( 1993 ) Brain Res. 605 , 164168 .
8. Tal, M. & Bennett, G. J. ( 1993 ) Neurosci. Lett. 151 , 107110 .
9. Mao, J. , Price, D. D. & Mayer, D. J. ( 1995 ) Pain 62 , 259274 .
10. Kajander, K. C. , Wakisaka, S. & Bennett, G. J. ( 1992 ) Neurosci. Lett. 138 , 225228 .
11. Sato, J. & Perl, E. R. ( 1991 ) Science 251 , 16081610 .
12. Kaczmarek, L. K. ( 1987 ) Trends Neurosci. 10 , 3034 .
13. Numann, R. , Catterall, W. A. & Scheuer, T. ( 1991 ) Science 254 , 115118 .
14. West, J. W. , Numann, R. , Murphy, B. J. , Scheuer, T. & Catterall, W. A. ( 1991 ) Science 254 , 866868 .
15. Chen, L. & Huang, L. Y. M. ( 1992 ) Nature (London) 356 , 521523 .
16. Liu, H. , Wang, H. , Sheng, L. Y. , Jan, Y. N. & Basbaum, A. I. ( 1994 ) Soc. Neurosci. Abstr. 20 , 482 .
17. Kitto, K. F , Haley, J. E. & Wilcox, G. L. ( 1992 ) Neurosci. Lett. 148 , 15 .
18. Palecek, J. , Paleckova, V. , Dougherty, P. M. & Willis, W. D. ( 1994 ) J. Neurophysiol. 71 , 529537 .
19. Gerber, G. , Kangraga, I. , Ryu, P. D. , Larew, J. S. A. & Randic, M. ( 1989 ) J. Neurosci. 9 , 36063617 .
20. Mao, J. , Mayer, D. J. , Hayes, R. L. & Price, D. D. ( 1993 ) J. Neurophysiol. 70 , 470481 .
21. Mao, J. , Price, D. D. , Phillips, L. L. , Lu, J. & Mayer, D. J. ( 1995 ) Brain Res. 677 , 257267 .
22. Laird, J. M. A. & Bennett, G. J. ( 1993 ) J. Neurophysiol. 69 , 20712085 .
23. Palecek, J. , Paleckova, V. , Dougherty, P. M. , Carlton, S. M. & Willis, W. D. ( 1992 ) J. Neurophysiol. 67 , 15621573 .
24. Guilbaud, G. , Benoist, J. M. , Jazat, F. & Gautron, M. ( 1990 ) J. Neurophysiol. 64 , 15371554 .
25. Bennett, G. J. & Xie, Y. K. ( 1988 ) Pain 33 , 87107 .
26. Mao. J. , Coghill, R. C. , Mayer, D. J. & Hayes, R. L. ( 1992 ) Pain 50 , 89100 .
27. Mao, J. , Mayer, D. J. & Price, D. D. ( 1993 ) J. Neurosci. 13 , 26892702 .
28. Hylden, J. L. K. , Nahin, R. L. , Traub, R. J. & Dubner, R. ( 1989 ) Pain 37 , 329243 .
29. Sugimoto, T. , Bennett, G. J. & Kajander, K. C. ( 1990 ) Pain 42 , 205213 .
30. Mao, J. , Price, D. D. , Zhu, J. , Lu, J. & Mayer, D. J. ( 1997 ) Pain 72 , 355366 .
31. Price, D. D. , Long, S. & Huitt, C. ( 1992 ) Pain 49 , 163174 .
the authoritative version for attribution.

32. Yamamoto, T. & Yaksh, T. L. ( 1992 ) Pain 49 , 121128 .

33. Zhang, J. , Dawson, V. L. , Dawson, T. M. & Snyder, S. H. ( 1994 ) Science 263 , 687689 .
34. Mayer, D. J. , Mao, J. & Price, D. D. ( 1995 ) Pain 61 , 365374 .
35. Mao, J. , Price, D. D. , Mayer, D. J. , Lu, J. & Hayes, R. L. ( 1992 ) Brain Res. 576 , 254262 .
36. Tal, M. & Bennett, G. J. ( 1994 ) NeuroReport 5 , 14381440 .
37. Trujillo, K. A. & Akil, H. ( 1991 ) Science 251 , 8587 .
38. Keniston, L. , Mao, J. , Price, D. D. , Lu, J. & Mayer, D. J. ( 1998 ) Soc. Neurosci. Abstr. 24 , 390 .
39. Mao, J. , Price, D. D. & Mayer, D. J. ( 1995 ) Pain 61 , 353364 .
40. Kolesnikov, Y. A. , Pick, C. G. & Pasternak, G. W. ( 1992 ) Eur. J. Pharmacol. 221 , 399400 .
41. Chen, Y. & Yu, L. ( 1994 ) J. Biol. Chem. 269 , 78397842 .
42. Chen, L. & Huang, L. Y. ( 1991 ) Neuron 7 , 319326 .
43. Pou, S. , Pou, W. S. , Bredt, D. S. , Snyder, S. H. & Rosen, G. M. ( 1992 ) J. Biol Chem. 267 , 2417324176 .
44. Pryor, W. A. & Squadrito, G. L. ( 1995 ) Am. J. Physiol. 268 , L699 L722 .
45. Szabo, C. ( 1996 ) Free Radical Biol. Med. 21 , 855869 .
46. Harada, H. , Ueda, H. , Katada, T. , Ui, M. & Satoh, M. ( 1990 ) Neurosci. Lett. 113 , 4749 .
47. Watkins, L. R. , Kinscheck, I. B. & Mayer, D. J. ( 1984 ) Science 224 , 395396 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

DOES A NEUROIMMUNE INTERACTION CONTRIBUTE TO THE GENESIS OF PAINFUL PERIPHERAL NEUROPATHIES? 7737

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Does a neuroimmune interaction contribute to the genesis of painful

peripheral neuropathies?

GARY J. BENNETT *
Department of Neurology, MCP Hahnemann University, Philadelphia, PA 19102
ABSTRACT Painful peripheral neuropathies are precipitated by nerve injury from disease or trauma. All such injuries will
be accompanied by an inflammatory reaction, a neuritis, that will mobilize the immune system. The role of the inflammation itself
is difficult to determine in the presence of structural damage to the nerve. A method has been devised to produce a focal neuritis in
the rat sciatic nerve that involves no more than trivial structural damage to the nerve. This experimental focal neuritis produces
neuropathic pain sensations (heat- and mechano-hyperalgesia, and cold- and mechano-allodynia) in the ipsilateral hind paw. The
abnormal pain sensations begin in 12 days and last for 46 days, with a subsequent return to normal. These results suggest that
there is a neuroimmune interaction that occurs at the outset of nerve injury (and perhaps episodically over time in slow developing
conditions like diabetic neuropathy) that produces neuropathic pain. The short duration of the phenomena suggest that they may
prime the system for more slowly developing mechanisms of abnormal pain (e.g., ectopic discharge in axotomized primary
afferent neurons) that underlie the chronic phase of painful neuropathy.
Painful peripheral neuropathies begin with nerve injury caused by disease or trauma. This injury will result in an inflammatory
reaction, a neuritis, that will mobilize the immune system. It is important to note that this will occur not only in cases of injury caused by
infection and autoimmune disorder (e.g., herpes zoster and Guillain-Barr syndrome) but also in cases of sterile injury because cellular
debris is an inflammatory and immune stimulus. It is difficult to study the role of the inflammation and the immune system response when
it presents together with structural damage to axons because the structural damage itself gives rise to pathogenic mechanisms that lead to
pain: for example, ectopic discharge in injured nociceptors. We have devised a method for producing a focal neuritis in the rat that is
accompanied by little or no structural injury to the nerve ( 1 ). We find that this neuritis produces neuropathic pain sensations in the
ipsilateral hind paw, even though the inflammation is at mid-thigh level.
We have used adult male SpragueDawley rats. The neuritis is produced by loosely wrapping the nerve at mid-thigh level with
hemostatic oxidized cellulose (Oxycell, Parke-Davis) that then is saturated with an inflammatory stimulus. The Oxycell does not constrict
the nerve; it serves merely as a sponge for the inflammatory stimulus. As the stimulus, we have used both carrageenan and complete
Freunds adjuvant with about equal effects; the results described below were obtained with complete Freunds adjuvant. As a control, we
(i) have treated the opposite nerve with Oxycell saturated with saline (ii) and have examined animals with unilateral Oxycell/saline
treatment. These control procedures do not evoke abnormal pain sensations in the hind paw. As a control for the general effects of a
painful thigh, and for the possibility of a systemic response to the inflammatory stimulus, we created an experimental unilateral myositis
by implanting Oxycell/complete Freunds adjuvant in a pocket made in biceps femoris at the same level as the nerve treatment. Animals
with the myositis did not have abnormal pain responses in the hind paw.
Rats with the focal neuritis have heat- and mechano-hyperalgesia and cold- and mechano-allodynia on the ipsilateral hind paw.
Responses from the contralateral hind paw are normal regardless of whether the contralateral side is untreated or treated with Oxycell/
saline.
Heat-hyperalgesia was measured with the paw-flick method of Hargreaves et al. ( 2 ). Abnormal sensitivity was noted within 12
days of treatment and reached peak severity 34 days after treatment. Responsiveness returned to normal within 56 days. The maximum
severity of the heat-hyperalgesia was slightly less than what we have seen with an experimental traumatic nerve injury [the chronic
constriction injury (CCI) model of Bennett and Xie ( 3 )]. All animals with the neuritis developed obvious heat-hyperalgesia.
Mechano-hyperalgesia was measured with the pin-prick method, and mechano-allodynia was measured with the von Frey hair
method as described by Tal and Bennett ( 4 ). Mechano-hyperalgesia and mechano-allodynia were noted within 12 days of treatment,
reached peak severity after 34 days, and resolved to normal within 56 days. The maximum severity of both was comparable to that seen
in CCI rats ( 3 ). All animals with the neuritis developed obvious mechano-hyperalgesia and mechano-allodynia.
Cold-allodynia was assayed with a slight modification of the method described by Choi et al. ( 5 ): 0.15 ml of acetone is sprayed onto
the plantar hind paw while the animal stands on a floor made of screening. On our own forearm skin, this stimulus produces a strong but
non-painful cooling sensation (as the acetone evaporates). Normal rats either ignore the stimulus, or it produces a very brief and small
withdrawal reflex. Neuropathic animals react with a large and prolonged withdrawal response (painful peripheral neuropathy patients with
cold-allodynia complain that this stimulus produces a severe burning pain sensation). Approximately one-half of the neuritis rats displayed
neuropathic responses to cold. This is in contrast to the CCI model, in which nearly every rat has an abnormal response. When present,
cold-allodynia was detected within 23 days of treatment, peaked within 34 days, and resolved within 45 days. Light- and electron-
microscopic analyses of the treated region of the nerve harvested at the time of peak symptom severity (34 days after treatment)
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviation: CCI, chronic constriction injury.
* To whom reprint requests should be addressed at: Department of Neurology, MCP Hahnemann University, Broad Street and Vine

Street (Mail Stop 423), Philadelphia, PA 19102-1192. e-mail: [email protected] .

lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

DOES A NEUROIMMUNE INTERACTION CONTRIBUTE TO THE GENESIS OF PAINFUL PERIPHERAL NEUROPATHIES? 7738

showed that some cases had no detectable axonal injury. In one case only (of three that were examined), we detected 20 degenerating
axons confined to a small patch just below the epineurium. The presence of a few tens of degenerating axons is trivial and is very highly
unlikely to produce the marked signs of neuropathic pain that were present in every animal. In all cases, there were clear signs of an
endoneurial inflammatory reaction (even though the inflammatory stimulus was applied to the outside of the nerve). The spaces between
axons was greater than normal, indicating an edematous reaction, and immune cells (macrophages, polymorphonuclear leukocytes, and
monocytes) were present throughout the endoneurial compartment. Immunocytochemical staining identified CD4 and CD8 T-lymphocytes
amongst this infiltrate. The T-cells surrounded the nerve on the outside surface of the epineurium (to be expected because this was where
the inflammatory stimulus was applied), but they were also within the nerve. The endoneurial cells were most abundant toward the center.
This suggests that the cells within the nerve arrived via the endoneurial vasculature because, if they had migrated from the outside, they
would have been concentrated just beneath the epineurium.
The neuropathic pain produced by the neuritis lasted for only a few days. It therefore cannot be the sole mechanism for chronic
painful peripheral neuropathies. It is possible, however, that it is of clinical importance. First, there may be conditions under which nerves
are chronically inflamed or suffer from repeated episodes of inflammation. Of importance, the inflammation may be in structures near
nerves that expose the nerve as an innocent bystander to an inflammatory milieu. For example, in diabetes, nerves may experience
repeated episodes of inflammation as the underlying microvascular and metabolic disease processes wax and wane. Nerves that are near
tumors may be bathed in an inflammatory soup of tumor-products and inflammatory cytokines directed against the malignancy. It is
noteworthy that the nucleus pulposus is a very potent inflammatory stimulus ( 6 ), so that a dorsal root lying near a leaky intervertebral
disc may be exposed to an inflammatory environment. Second, in those conditions under which the neuritis is likely to be acute (e.g., post-
herpetic neuralgia), the neuritis-evoked neuropathic pain may prime the nervous system such that pathogenic mechanisms that develop
later (e.g., spontaneous ectopic discharge in injured nociceptors, sprouting sympathetic efferent axons in the dorsal root ganglia, A- low-
threshold mechanoreceptors invading laminae I and II) are able to produce chronic neuropathic pain.
Current work with this model is attempting to determine the relative roles of the inflammatory mediators derived from the
cycloxygenase and lipoxygenase cascades and the proinflammatory cytokines. We have not found any effect in the neuritis model with
indomethacin, suggesting that the arachidenic acid pathways are not involved (J.-E. Baos, S. Shiiba, and G.J.B., unpublished results). It
has already been shown that tumor necrosis factor is found in CCI nerves and that tumor necrosis factor injected into the nerve
produces neuropathic pain symptoms ( 7 ). Sommers et al. ( 8 ) have shown that inhibition of tumor necrosis factor release with
thalidomide reduces neuropathic pain in the CCI model. We have replicated the results of Sommers et al. in CCI rats but have not found
any effect of thalidomide in the neuritis model (J.-E. Baos, S. Shiiba, and G.J.B., unpublished results). The difference may be attributable
to differing immune system responsesprimarily to cellular debris in the case of the CCI model but to bacterial epitopes in the neuritis
model. We have found that other immune suppressants are effective in the neuritis model; for example, cyclosporin A works well (S.
Shiiba, J.-E. Baos, and G.J.B., unpublished results).
1. Eliav, E. , Herzberg, U. , Ruda, M. A. & Bennett, G. J. ( 1999 ) Pain , in press.
2. Hargreaves, K. , Dubner, R. , Brown, F. , Flores, C. & Joris, J. ( 1988 ) Pain 32 , 7788 .
3. Bennett, G. J. & Xie, Y.-K. ( 1988 ) Pain 33 , 87107 .
4. Tal, M. & Bennett, G. J. ( 1994 ) Pain 57 , 375382 .
5. Choi, Y. , Yoon, Y. W. , Na, H. S. , Kim, S. H. & Chung, J. M. ( 1994 ) Pain 59 , 369376 .
6. Olmarker, K. , Blomquist, J. , Strmberg, J. , Nannmark, U. , Thomsen, P. & Rydevik, B. ( 1995 ) Spine 20 , 665669 .
7. Wagner, R. & Myers, R. R. ( 1996 ) NeuroReport 7 , 28972901 .
8. Sommer, C. , Marziniak, M. & Myers, R. R. ( 1998 ) Pain 74 , 8391 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

DISTINCT NEUROCHEMICAL FEATURES OF ACUTE AND PERSISTENT PAIN 7739

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

Distinct neurochemical features of acute and persistent pain

ALLAN I. BASBAUM *
Departments of Anatomy and Physiology and W. M. Keck Foundation Center for Integrative Neuroscience, University of California,
San Francisco, CA 94143
ABSTRACT To address the neurochemistry of the mechanisms that underlie the development of acute and persistent pain,
our laboratory has been studying mice with deletions of gene products that have been implicated in nociceptive processing. We
have recently raised mice with a deletion of the preprotachykinin-A gene, which encodes the peptides substance P (SP) and
neurokinin A (NKA). These studies have identified a specific behavioral phenotype in which the animals do not detect a window of
pain intensities; this window cuts across thermal, mechanical, and chemical modalities. The lowered thermal and mechanical
withdrawal thresholds that are produced by tissue or nerve injury, however, were still present in the mutant mice. Thus, the
behavioral manifestations of threshold changes in nociceptive processing in the setting of injury do not appear to require SP or
NKA. To identify relevant neurochemical factors downstream of the primary afferent, we are also studying the dorsal horn second
messenger systems that underlie the development of tissue and nerve injury-induced persistent pain states. We have recently
implicated the isoform of protein kinase C (PKC) in the development of nerve injury-induced neuropathic pain. Acute pain
processing, by contrast, is intact in the PKC-null mice. Taken together, these studies emphasize that there is a distinct
neurochemistry of acute and persistent pain. Persistent pain should be considered a disease state of the nervous system, not merely
a prolonged acute pain symptom of some other disease conditions.
Recent studies have focused attention on the organization of two major small-diameter primary afferent systems that target dorsal
horn neurons. Broadly speaking, they can be divided into a peptide-containing class of C fibers and one that is generally not associated
with peptides ( 1 ). The peptide-containing neurons can be demonstrated by immunostaining dorsal root ganglia with antisera directed
against the preprotachykinin-A product substance P (SP). These neurons also coexpress calcitonin gene-related peptide (CGRP) and the
neurotrophin receptor TrkA. The second group of afferents can be distinguished because they express a cell surface -galactosyl epitope,
which can be stained with the lectin IB4. Most of the latter neurons also express a fluoride-resistant acid phosphatase (FRAP) ( 2 ) and the
P2X3 subtype of purinergic receptor ( 3 ). It has been suggested, however, that the failure of the peptide-containing group to immunostain
for the P2X3 receptor reflects the inability of antisera to recognize the epitope in that class of neurons, because of coassembly with another
subtype of purinergic receptor (E. W. McCleskey, personal communication). Perhaps of greater relevance to the functional differences of
these two broad classes of afferent is that their patterns of axon termination in the spinal cord dorsal horn differ greatly. The terminals of
the peptide-containing afferents are concentrated in laminae I and outer II of the superficial dorsal horn. The IB4-labeled afferents target
the inner part of lamina II. The same region can be stained for the P2X3 receptor and for FRAP.
Because these afferents are of small diameter, it has generally been assumed that they are nociceptors, even though some
nonnociceptive, low-threshold mechanoreceptive, small-diameter afferents have been identified ( 4 ). Interestingly, the latter group did not
express SP. Given that many neurons in the inner part of lamina II respond to nonnoxious stimuli ( 5 , 6 ), it has been suggested that this
region does not receive input from nociceptors. Despite these observations, we recently provided strong evidence in support of the
hypothesis that the majority of the small-diameter afferents are nociceptors. Specifically, we showed that neurons in both populations
express the vanilloid-1 receptor (VR1) to which capsaicin, the pungent ingredient in hot pepper, binds with high affinity ( 7 ). The fact that
these neurons express VR1 suggests that, at the minimum, these afferents are responsive to a noxious chemical stimulus. Furthermore,
because the capsaicin receptor can be activated by noxious thermal stimuli ( 8 ), it is also likely that the predominantly nonpeptide
population of C fiber includes heat nociceptors. Importantly, not all of the neurons were double-labeled, and we identified a population of
dorsal root ganglion (DRG) cells that was VR1 positive, but stained for neither SP nor IB4. Thus additional classes of primary afferent
nociceptor remain to be characterized. The complexity of the DRG population is highlighted even further by the recent observation that
there are at least two classes of heat nociceptors. Because only one of these responds to capsaicin ( 9 ) it will be of interest to determine the
neurochemical phenotype of the non-VR1 heat nociceptor.
These results establish that the two major classes of small-diameter primary afferent include nociceptors, but that is just the first step
in establishing their contribution to the generation of pain. For example, it is not at all clear to what extent they differ in the types of pain
provoked by their activation. We also do not know whether coincident activity in the different classes of afferent affects the type/quality of
pain that is provoked. Given that the afferents target very different populations of neurons in the superficial dorsal horn, it is also of
interest to address the contribution of these downstream neuronal populations. Importantly, many of the neurons targeted by the peptide
population are projection neurons, which transmit the nociceptive message to brainstem and/or thalamus. These are clustered in lamina I of
the dorsal horn. Indeed, the large majority of neurons that express the neurokinin-1 (NK-1) receptor, which is targeted by SP-containing
afferents, are projection neurons ( 10 ). By contrast, the nonpeptide group of small-diameter afferents targets a region of the superficial
dorsal horn that exclusively contains interneurons, namely the inner part of lamina II.
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: SP, substance P; VR1, vanilloid-1 receptor; NK-1, neurokinin-1; NKA, neurokinin A; PPT-A, preprotachykinin-A;
PKC, isoform of protein kinase C.
* To whom reprint requests should be addressed. e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

DISTINCT NEUROCHEMICAL FEATURES OF ACUTE AND PERSISTENT PAIN 7740

The problem is not a simple one to address, as it is presently not possible to selectively remove one or the other class of afferent or
postsynaptic neuron. The only exception to this are the very recent studies of Mantyh and colleagues ( 11 ), who used a SP-saporin
conjugate to selectively destroy superficial dorsal horn neurons that express the NK-1 receptor. Our approach to this problem has been to
study the phenotype of mice in which the genes for critical neurotransmitters, neurotransmitter receptors, or second messenger systems
have been deleted. In the following discussion, I will highlight some of our progress in attributing different properties of the pain response
to distinct gene products in afferents and second-order neurons.
PREPROTACHYKININ-A PRODUCTS AND ACUTE PAIN
Our first studies evaluated the phenotype of mice in which we deleted the preprotachykinin-A (PPT-A) gene, which encodes the two
tachykinin peptides, SP and neurokinin A (NKA) ( 12 ). These animals appeared normal in simple motor tests, including running on a
rotating rod and when examined in an open field. As expected, we also found that plasma extravasation (PE) in response to injection of
capsaicin was almost abolished in the mutant mice. That result was expected, as peripheral release of SP/NKA is generally required to
evoke PE ( 13 ), a hallmark of neurogenic inflammation. In tests of acute pain, using mild to moderate noxious stimuli, the mutant mice
also responded as the wild-type mice did. By contrast, when the stimulus intensity was increased we observed increased pain responses
only in the wild type. For example, when the heat stimulus was increased from 52.5C to 55C, only the wild-type mice showed a
decreased latency to respond. Decreased responsiveness in the setting of very intense stimuli was true for noxious chemical, thermal, and
mechanical stimuli. These results indicate that SP/NKA contribute to the intensity coding of stimuli across modalities, but that the coding
range is limited to one in which the stimulus is very intense. The fact that very intense stimuli are required to reveal the phenotype is
consistent with our previous studies, which found that internalization of the NK-1 receptor (which provides a measure of tachykinin
release) occurs only when the stimulus is very intense ( 14 ). This result is also consistent with neurochemical studies that found that
higher intensities and higher frequencies of stimulation are required to evoked primary afferent release of SP, compared with glutamate (
15 , 16 ). The latter is a major neurotransmitter of small-diameter primary afferents, and is found in the same terminals, albeit in a different
population of synaptic vesicles ( 17 ). We presume that as the threshold for activating C-fiber nociceptors is crossed, glutamate is released
at central synapses. Via an action at an -amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor, the glutamate
depolarizes postsynaptic projection neurons and interneurons of the dorsal horn, resulting in an acute pain message.
What about the contribution of SP/NKA to persistent pain conditions? Allodynia refers to the production of pain (including reflex
withdrawal) by nonnoxious stimuli. Hyperalgesia refers to an exaggerated pain response produced by noxious stimuli. Although most
studies have reported that the development of allodynia and hyperalgesia results from injury-induced long-term changes/central
sensitization of dorsal horn neurons secondary to activity at the N-methyl-D-aspartate (NMDA) receptor ( 18 , 19 ), other studies have
provided evidence for a contribution of tachykinins, notably SP ( 20 22 ). It was thus surprising that we found no change in the
development of mechanical or thermal allodynia after injury. Importantly, the tests of allodynia involve stimuli that are at or near
threshold. For example, within 24 hr of injecting complete Freunds adjuvant into the hindpaw, which produces a profound hindlimb
inflammation, there is a dramatic drop in the thermal and mechanical withdrawal threshold (allodynia) of the affected limb. We found that
comparable allodynia was present in both wild-type and PPT-A-null mice. The thermal and mechanical allodynia produced after partial
sciatic nerve section ( 23 ) was also intact, despite the loss of SP/NKA. Importantly, we did not specifically test for hyperalgesia, as this
would have required using a noxious stimulus in the setting of injury.
To determine whether the loss of the PPT-A gene products differentially affects the development of allodynia and hyperalgesia, we
are using an electrophysiological approach to study the response of dorsal horn neurons under conditions of inflammation, and in response
to noxious stimuli. In this setting the mice are anesthetized and thus it is possible to administer noxious stimuli, even to an inflamed
hindpaw. Consistent with the behavioral analysis, our electrophysiological results to date suggest that SP/NKA are predominantly
involved under acute conditions, when the stimulus is very intense. We do not have an explanation for the differences between our results
and those of others that implicate SP and NKA in the development of allodynic states. A particularly interesting discrepancy concerns the
observations of Neumann et al. ( 24 ), who reported that in the setting of persistent inflammation, SP is synthesized de novo in large-
diameter afferents and that it contributes to the A (i.e., large fiber-mediated) mechanical allodynia.
Our results with the PPT-A-null mice also bear on the receptors that are targeted by SP and NKA. It is generally assumed that SP
exerts its effects via the NK-1 receptor ( 25 ). It is, thus, of great interest that the phenotypes of the PPT-A mice and of mice with a
deletion of the NK-1 receptor ( 26 ) differ. Although we found that high-intensity stimuli are not processed normally in the PPT-A mutant
mice, the response to acute noxious stimuli appears to be intact in the NK-1 receptor mutant animals. Furthermore, although the NK-1
receptor mutants showed reduced second-phase formalin test responses, the PPT-A mutant mice had only reduced first-phase behavior.
Although there is disagreement concerning the extent to which central sensitization contributes to second-phase pain in the formalin test (
27 , 28 ), the possibility that the time course and magnitude of central sensitization differ in the wild-type and mutant mice needs to be
tested.
We hypothesize that the differences between the PPT-A mutant and NK-1 receptor mutant mice are indicative of an action of SP and/
or NKA at an additional tachykinin receptor. Although SP predominantly targets the NK-1 receptor, NKA has a higher affinity for the
NK-2 receptor and has reasonable affinity for the NK-3 site ( 25 ). Importantly, there is evidence for an independent contribution of the
NK-2 receptor (and thus of NKA) to nociceptive processing ( 29 31 ). Of particular interest, iontophoresis of NK-2 antagonists reduces
the hyperexcitability of dorsal horn neurons in the setting of inflammation and blocks the central effects of NKA, but not of SP ( 32 ).
Fleetwood-Walker and colleagues ( 33 ) found that NK-2 antagonists were far more effective than NK-1 antagonists at reducing
inflammation-induced up-regulation of dynorphin message in superficial dorsal horn. In preliminary studies, we found significant
induction of Fos expression in laminae I and II after intrathecal injection of NKA in mice with a deletion of the NK-1 receptor. Finally,
Duggan et al. ( 34 ) provided evidence that the central effects of SP and NKA differ, in part because of their differential susceptibility to
the endopeptidases that degrade them. Thus, if NKA targets a receptor to which SP has much lower affinity, its loss in the PPT-A-null
mouse could have profound consequences independent of the loss of SP.
Paradoxically, despite the pharmacological evidence for an independent NK-2 receptor contribution, neither in situ hybridization nor
immunocytochemistry, with one recent exception ( 35 ), has provided convincing evidence that the NK-2
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

DISTINCT NEUROCHEMICAL FEATURES OF ACUTE AND PERSISTENT PAIN 7741

receptor is present in the dorsal horn. One possibility is that there is another receptor to which NKA binds, different from the cloned NK-2
receptor, but sensitive to NK-2 antagonists. A somewhat comparable situation exists for opioid receptors ( 36 ), which are best
distinguished pharmacologically. If we are to understand the mechanisms through which SP/NKA influence short- and long-term
nociceptive processing in the dorsal horn, it is essential that this paradox be addressed. To this end we are now generating mice that
encode either SP or NKA, but not both. We will study these mice in both behavioral and electrophysiological tests of nociception and
compare their phenotypes with those of the PPT-A-null and NK-1 receptor-null mice. Hopefully, this approach will address the differential
contribution of SP and NKA to nociceptive processing.
PROTEIN KINASE C AND PERSISTENT PAIN
The results described above provide important information on the contribution of the peptide-containing class of primary afferents to
nociceptive processing. Much more difficult is the analysis of the contribution of the nonpeptide class of small-diameter afferents.
Although these neurons express the P2X3 receptor, it is difficult to selectively activate this receptor in vivo so as to address its contribution
to the generation of pain. Because of the difficulty in directly evaluating the afferents, we turned to a likely central target of the afferents,
namely neurons in the inner part of the lamina II.
These studies involved mice with a deletion of the gene that encodes the isoform of protein kinase C (PKC) ( 37 ). These mice
were of interest for several reasons. Most importantly, there is considerable evidence that PKC contributes to the development of the long-
term changes that underlie injury-associated allodynia and hyperalgesia ( 38 40 ). Unfortunately, there are no selective antagonists
available for the many isoforms of PKC; thus it is impossible to determine their individual contribution. Furthermore, because the different
isoforms are differentially distributed in the dorsal horn (see below), an understanding of the mechanisms through which they operate at
the dorsal horn circuit level is very difficult to achieve. Thus, the opportunity to study mice with a deletion of one isoform is particularly
useful. These mice were also of also interest because, unlike most PKC isoforms, PKC is not expressed until after birth; thus the
likelihood that the deletion is associated with major developmental abnormalities, or for compensatory responses to its loss during
development, is significantly reduced. Any compensatory responses to the deletion would have to have occurred postnatally. The PKC-
null mice are also of interest because studies in hippocampus revealed that the mice show a reduction of long-term potentiation (LTP) (
41 ). Although LTP and long-term changes in spinal cord nociceptive processing secondary to injury are not identical, it was reasonable to
hypothesize that pain-related behaviors that have been associated with alterations in the excitability of dorsal horn nociresponsive neurons
would be altered in these mice.

FIG. 1. (Left) Micrograph from the lumbar spinal cord of the rat. It illustrates the restricted distribution of PKC immunoreactivity to
interneurons in the inner part of lamina II of the superficial dorsal horn. (Right) Distribution of Fos-immunoreactive neurons in the
L4 cord dorsal horn evoked by an injection of formalin into the ipsilateral hindpaw. The rat was killed 1 hr after the formalin
injection. Note that there are many labeled neurons in the most superficial laminae (I and outer II), but that there is a distinct band
(corresponding to the inner part of lamina II; arrows) in which there is very little Fos expression. Thus although the small-diameter
afferents that innervate lamina II are chemonociceptors, their activation appears not to induce Fos in the neurons that they target.
(60.)
Finally, and perhaps, most importantly, we found that the dorsal horn distribution of PKC differs greatly from that of the other
isoforms. First, to our knowledge, PKC is the only isoform that is not found in dorsal root ganglia. Furthermore, PKC immunostaining in
the spinal cord is restricted to a subpopulation of interneurons in the inner part of lamina II ( Fig. 1 ). This distribution greatly differs from
that of other isoforms (e.g., PKC, 1, and 2), which are found throughout the superficial dorsal horn as well as more ventrally. The
presence of PKC in the inner part of lamina II raises the possibility that the phenotype of the deletion mutant is, at least in part, related to
loss of the consequences of activation of the nonpeptide population of primary afferents that target this region. Importantly, the restriction
of PKC to a single interneuron population significantly limits the circuits through which it can influence the development of persistent
pain conditions. It also offers the hope that further neuroanatomical studies at the light and electron microscopic level will be able to
identify the major afferent inputs to and the connections made by these neurons so that the circuits through which these neurons exert their
effects can be understood.
Our studies proved incredibly informative. In distinct contrast to the mice with a deletion of the PPT-A gene, we found that in the
PKC-null mice there was no defect in the processing of acute pain messages. For example, the PKC-null and wild-type mice behaved
identically in tests of thermal pain. By contrast, when we tested the mice in models that involved persistent injury, for example, after
partial sciatic nerve injury, we found a significant decrease in the magnitude of the mechanical and thermal allodynia that developed. This
the authoritative version for attribution.

difference persisted for the duration of the experiment. We concluded that in the absence of PKC, partial nerve injury
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

DISTINCT NEUROCHEMICAL FEATURES OF ACUTE AND PERSISTENT PAIN 7742

does not induce the hyperexcitability of dorsal horn neurons that is at the basis of the subsequent allodynia.
Given that PKC is concentrated in interneurons of the inner part of lamina II, we presume that changes in the circuits that involve
these neurons are critical to the failure of allodynia to develop after nerve injury. On the other hand, PKC is present throughout the
neuroaxis. Thus, the defect in nerve injury-induced processing of nociceptive messages need not reflect the loss of enzyme function in the
spinal cord. In preliminary studies using electrophysiological methods in the mouse spinal cord ( 42 ), however, we found that injury in the
PKC-null mice does not lead to an enhancement of the response of nociceptive neurons to nonnoxious stimuli, as it does in the wild-type
mice. This finding more directly points to the cord as the locus of the defect.
Finally, we found that the spinal cord neuroanatomical consequences of nerve injury are also significantly reduced in the PKC-null
mice. For example, numerous studies have demonstrated that there is a decrease in immunoreactivity for SP ( 43 ) and an increase in
staining for the NK-1 receptor after sciatic nerve section ( 44 ). Although comparable changes were readily demonstrated in the wild-type
mice, the changes were much reduced, and in some cases absent, in the mutant mice. Similarly, although we found that partial sciatic
nerve section induces an up-regulation of neuropeptide Y immunoreactivity in the superficial dorsal horn, this was also significantly
reduced in the PKC-null mice. Taken together, these results suggest either that the signal that is sent from the site of nerve injury to the
dorsal root ganglia and spinal cord is not transmitted properly in the PKC-null mice, or that the response to the signal by dorsal horn
neurons is altered. We favor the latter hypothesis and suggest that it is a signal in the nonpeptide population of small-diameter nociceptors
that is not processed normally by the PKC interneurons of the inner part of lamina II. How this comes about, and the extent to which such
changes can be prevented or reversed, remains to be determined.
The fact that the PKC-null mice behaved normally in tests of acute pain suggests that the contribution of these neurons is manifest
only when noxious inputs persist, i.e., in the setting of injury. In fact, it is remarkably difficult to demonstrate that neurons in the inner part
of lamina II, are activated by acute noxious stimuli. As noted above, electrophysiological studies have emphasized the responsiveness of
many of the neurons in inner II to nonnoxious mechanical stimuli ( 5 , 6 ). Furthermore, even a very intense noxious stimulus rarely
induces Fos expression in the interneurons of inner lamina II ( 45 ). This is true even when formalin ( Fig. 1 ) or capsaicin is used as the
stimulus. Because the bulk of the small-diameter afferents that target inner lamina II express the capsaicin receptor, it is likely that this
stimulus activates the interneurons of this region. We have tried repeatedly to find a stimulus that will induce Fos in the PKC population
of neurons, but have not succeeded. To our knowledge, the only consistent stimulus that induces Fos in neurons of inner II is kainate
injection in the raphe magnus of the medulla ( 46 ). Studies in our laboratory, however, found that even this stimulus rarely induces Fos
expression in the PKC population of interneurons. It is, of course, possible that the PKC neurons do not express Fos; that alone indicates
that they are unusual, at least when compared with other dorsal horn neurons that receive inputs from primary afferent nociceptors.
Given the nature of the afferents to the inner part of lamina II (i.e., that they express VR1) we find it very difficult to accept the
notion that this region exclusively, or even predominantly, subserves a nonnociceptive function. Rather, it appears that the conditions
under which these neurons are activated remains to be established. In part to address this question, we have established an in vitro spinal
cord slice preparation in which translocation of PKC in these interneurons (which occurs when the enzyme is activated) can be
monitored. We hope that this will permit a more extensive analysis of the neurochemistry of the inputs that activate these neurons. That
information will then be used to study the neurons in the intact animal.
In summary, our studies using mice with deletions of specific genes have demonstrated that particular features of the responses to
acute and persistent injury conditions are differentially influenced by SP/NKA and PKC. Our present studies are directed at identifying
the circuits through which these distinct phenotypes are generated. In one series of studies we are generating a very detailed analysis of the
circuits in which PKC-containing interneurons participate and a comprehensive description of their neurochemical phenotype ( 47 ). That
information will provide a better understanding of the mechanisms through which these neurons influence nociceptive processing in the
setting of injury and will hopefully be useful in the development of approaches to treating these conditions.
This work was supported by National Institutes of Health Grants NS 21445, DA 08377, DE 08973, and NS 14627.
1. Snider, W. D. & McMahon, S. B. ( 1998 ) Neuron 20 , 629632 .
2. Silverman, J. D. & Kruger, L. ( 1988 ) Somatosens. Res. 5 , 219246 .
3. Vulchanova, L. , Riedl, M. S. , Shuster, S. J. , Buell, G. , Surprenant, A. , North, R. A. & Elde, R. ( 1997 ) Neuropharmacology 36 , 12291242 .
4. Lawson, S. N. , Crepps, B. A. & Perl, E. R. ( 1997 ) J. Physiol (London) 505 , 177191 .
5. Light, A. R. , Trevino, D. L. & Perl, E. R. ( 1979 ) J. Comp. Neurol. 186 , 325330 .
6. Bennett, G. J. , Abdelmoumene, M. , Hayashi, H. & Dubner, R. ( 1980 ) J. Comp. Neurol. 194 , 809827 .
7. Tominaga, M. , Caterina, M. J. , Malmberg, A. B. , Rosen, T. A , Gilbert, H. , Skinner, K. , Raumann, B. E. , Basbaum, A. I. & Julius, D. ( 1998 )
Neuron 21 , 531543 .
8. Caterina, M. J. , Schumacher, M. A , Tominaga, M. , Rosen, T. A. , Levine, J. D. & Julius, D. ( 1997 ) Nature (London) 389 , 816824 .
9. Nagy, I. & Rang, H. ( 1998 ) Neuroscience 88 , 995998 .
10. Marshall, G. E. , Shehab, S. A. S. , Spike, R. C. & Todd, A. J. ( 1996 ) Neuroscience 72 , 255263 .
11. Mantyh, P. W. , Rogers, S. D. , Honor, P. , Allen, B. J. , Ghilardi, J. R. , Li, J. , Daughters, R. S. , Lappi, D. A , Wiley, R. G. & Simone, D. A. (
1997 ) Science 278 , 275279 .
12. Cao, Y. Q. , Mantyh, P. W. , Carlson, E. J. , Gillespie, A.-M. , Epstein, C. J. & Basbaum, A. I. ( 1998 ) Nature (London) 392 , 390394 .
13. Lembeck, F. , Donnerer, J. & Bartho, L. ( 1982 ) Eur. J. Pharm. 85 , 171176 .
14. Abbadie, C. , Trafton, J. , Liu, H. , Mantyh, P. W. & Basbaum, A. I. ( 1997 ) J. Neurosci. 17 , 80498060 .
15. Marvizon, J. C. G. , Martinez, V. , Grady, E. F. , Bunnett, N. W. & Mayer, E. A. ( 1997 ) J. Neurosci. 17 , 81298136 .
16. Duggan, A. W. , Riley, R. C. , Mark, M. A , MacMillan, S. J. A. & Schaible, H. G. ( 1995 ) Neuroscience 65 , 849858 .
17. De Biasi, S. & Rustioni, A. ( 1988 ) Proc. Natl. Acad. Sci. USA 85 , 78207824 .
18. Dickenson, A. H. , Chapman, V. & Green, G. M. ( 1997 ) Gen. Pharmacol. 28 , 633638 .
19. Woolf, C. J. & Thompson, S. ( 1991 ) Pain 44 , 293299 .
20. Dougherty, P. M. , Palecek, J. , Zorn, S. & Wiffis, W. D. ( 1993 ) Brain Res. Rev. 18 , 227246 .
21. Xu, X. J. , Dalsgaard, C. J. & Wiesenfeld, H. Z. ( 1992 ) Neuroscience 51 , 641648 .
22. Ma, Q. P. & Woolf, C. J. ( 1997 ) Eur. J. Pharmacol. 322 , 165171 .
23. Malmberg, A B. & Basbaum, A. I. ( 1998 ) Pain 76 , 215222 .
24. Neumann, S. , Doubell, T. P. , Leslie, T. & Woolf, C. J. ( 1996 ) Nature (London) 384 , 360364 .
25. Maggi, C. A. ( 1995 ) Gen. Pharmacol. 26 , 911944 .
the authoritative version for attribution.

26. De Felipe, C. , Herrero, J. F. , OBrien, J. A , Palmer, J. A , Doyle, C. A , Smith, A. J. , Laird, J. M. , Belmonte, C. , Cervero, F. & Hunt, S. P. (
1998 ) Nature (London) 392 , 394397 .
27. Malmberg, A. B. & Yaksh, T. L. ( 1992 ) Science 257 , 12761279 .
28. Taylor, B. K. , Peterson, M. A. & Basbaum, A. I. ( 1997 ) J. Pharmacol. Exp. Ther. 280 , 876883 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

DISTINCT NEUROCHEMICAL FEATURES OF ACUTE AND PERSISTENT PAIN 7743

29. Liu, X. G. & Sandkhler, J. ( 1997 ) J. Neurophysiol 78 , 19731982 .

30. Sluka, K. A. , Milton, M. A. , Willis, W. D. & Westlund, K. N. ( 1997 ) Br. J. Pharmacol 120 , 12631273 .
31. Urban, L. , Dray, A. , Nagy, I. & Maggi, C. A. ( 1993 ) Regul. Pept. 46 , 413414 .
32. Neugebauer, V. , Rmenapp, P. & Schaible, H. G. ( 1996 ) Eur. J. Neurosci. 8 , 249260 .
33. Munro, F. E. , Fleetwood-Walker, S. M. , Parker, R. M. C. & Mitchell, R. ( 1993 ) Neuropeptides 25 , 299305 .
34. Duggan, A. W. , Hope, P. J. , Jarrott, B. , Schaible, H. G. & Fleetwood, W. S. ( 1990 ) Neuroscience 35 , 195202 .
35. Zerari, F. , Karpitskiy, V. , Krause, J. , Descarries, L. & Couture, R. ( 1998 ) Neuroscience 84 , 12331246 .
36. Lai, J. , Bilsky, E. J. , Rothman, R. B. & Porreca, F. ( 1994 ) NeuroReport 5 , 10491052 .
37. Malmberg, A. B. , Brandon, E. P. , Idzerda, R. L. , Liu, H. , McKnight, G. S. & Basbaum, A. I. ( 1997 ) J. Neurosci. 17 , 74627470 .
38. Meller, S. T. , Dykstra, C. & Gebhart, G. F. ( 1996 ) Neuroscience 71 , 327335 .
39. Mao, J. , Mayer, D. J. , Hayes, R. L. & Price, D. D. ( 1993 ) J. Neurophysiol. 70 , 470481 .
40. Coderre, T. J. ( 1992 ) Neurosci. Lett. 140 , 181184 .
41. Abeliovich, A. , Chen, C. , Goda, Y. , Silva, A. J. , Stevens, C. F. & Tonegawa, S. ( 1993 ) Cell 75 , 12531262 .
42. Martin, W. J. & Basbaum, A. I. ( 1998 ) Neurosci. Abstr. 24 , 880 .
43. Villar, M. J. , Corts, R. , Theodorsson, E. , Wiesenfeld, H. Z. , Schalling, M. , Fahrenkrug, J. , Emson, P. C. & Hkfelt, T. ( 1989 ) Neuroscience
33 , 587604 .
44. Abbadie, C. , Brown, J. L. , Mantyh, P. W. & Basbaum, A. I. ( 1996 ) Neuroscience 70 , 201209 .
45. Mentrey, D. , Gannon, A. , Levine, J. D. & Basbaum, A. I. ( 1989 ) J. Comp. Neurol. 285 , 177195 .
46. Bett, K. & Sandkhler, J. ( 1995 ) Neuroscience 67 , 497504 .
47. Martin, W. J. , Liu, H. , Wang, H. , Malmberg, A. B. & Basbaum, A. I. ( 1999 ) Neuroscience 88 , 12671274 .
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE GENETIC MEDIATION OF INDIVIDUAL DIFFERENCES IN SENSITIVITY TO PAIN AND ITS INHIBITION 7744

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

The genetic mediation of individual differences in sensitivity to pain

and its inhibition

(nociception/gene mapping/strain differences)

JEFFREY S. MOGIL *
Department of Psychology and Neuroscience Program, University of Illinois at Urbana-Champaign, Champaign, IL 61820
ABSTRACT The underlying bases of the considerable interindividual variability in pain-related traits are starting to be
revealed. Although the relative importance of genes versus experience in human pain perception remains unclear, rodent
populations display large and heritable differences in both nociceptive and analgesic sensitivity. The identification and
characterization of particularly divergent populations provides a powerful initial step in the genetic analysis of pain, because these
models can be exploited to identify genes contributing to the behavior-level variability. Ultimately, DNA sequence differences
representing the differential alleles at pain-relevant genes can be identified. Thus, by using a combination of top-down and
bottom-up strategies, we are now able to genetically dissect even complex biological traits like pain. The present review
summarizes the current progress toward these ends in both humans and rodents.
Pain is considered both a sensation and an emotion ( 1 ) and shows considerable complexity and subjectivity, especially when
compared with other sensory systems. Physical injury is neither necessary nor sufficient for the experience of pain, and in both clinical and
laboratory settings, the perception of pain bears a notoriously poor relationship to the intensity of the noxious stimulus. Mildly noxious
stimuli can be perceived as very painful whereas very noxious stimuli can produce no pain whatsoever ( 2 ). Substantial advances in
understanding why this might be have been made in the past few decades, including the discovery of pain-modulatory mechanisms
(producing both analgesia and antianalgesia) and nervous system plasticity following exposure to noxious stimuli.
Even as our understanding of pain physiology has been facilitated by such paradigmatic advancements, pain perception in humans
and animals nonetheless displays considerable, and unexplained, interindividual variability. The focus of this review is to examine the
scope of individual differences in pain and analgesia in both humans and laboratory animals and to consider what is known and what
remains to be determined regarding the genetic contributions to such variability. Null mutants (e.g., transgenic knockouts) will not be
considered presently, because they have been recently reviewed ( 3 ).
THE SCOPE OF INDIVIDUAL VARIABILITY TO PAIN AND ANALGESIA
Human Studies. Variability in pain-related traits (or phenotypes)experimental pain sensitivity, propensity to develop painful
pathologies, and sensitivity to analgesic manipulationshas long been appreciated, although empirical validation of such variability is
somewhat limited. In 1934, Libman ( 4 ) reported that pressure applied to the mastoid bone in the direction of the styloid process produces
marked pain in 6070% of his human patients, but no or little pain in 3040%. Subsequent investigations over the next few decades with
more quantitative methodology confirmed the presence of large individual differences in threshold sensitivity and tolerance to noxious
pressure ( 5 8 ), heat ( 7 , 9 , 10 ), and electrical current ( 7 , 11 ) applied to cutaneous tissues, and tolerance to visceral ( 6 ) and deep
muscle ( 9 , 10 ) pain. Impressive individual differences in sensitivity to opioid analgesics were also documented during this period,
typified by Lasagna and Beecher ( 12 ) observing a success rate of only 65% of the standard clinical dose of morphine, 10 mg (see also
ref. 13 ).
Modern studies have confirmed that interindividual variability in pain thresholds greatly exceeds intraindividual variability for
pressure pain threshold (e.g., refs. 14 and 15 ), cold pressor tolerance (e.g., ref. 16 ), and pain associated with manual palpation of tender
points in fibromyalgia sufferers ( 17 ). In the cold pressor test, Chen et al. ( 16 , 18 ) consistently observe dichotomous responses, with a
minority pain-sensitive group tolerating the test for a mean duration of 50 sec and a majority pain-tolerant group able to tolerate the
test for the full 300-sec duration.
Similar findings have been obtained for opiate analgesics, which display large clinical (e.g., refs. 19 and 20 ) and experimental (e.g.,
ref. 21 ) variability in their efficacies, side effects, and tolerance liability. The analgesic actions of morphine on dental extraction pain also
show evidence of bimodality, with morphine responder and nonresponder groups being easily identified at a range of doses ( 22 ).
Individual differences in the actions of an opiate antagonist, naloxone, have also been demonstrated. Buchsbaum et al. ( 23 ), after
dividing human subjects into pain-sensitive and pain-insensitive groups based on their sensitivity to electric shock to the forearm, noted
that intravenous injection of 2.0 mg of naloxone produced hyperalgesia in the latter group and (paradoxically) analgesia in the former.
Animal Studies. In the concluding comment of Libmans ( 4 ) seminal study of human pain variability, he posits that studies of
animals may prove of value, because in all the work hitherto performed it has been taken for granted that sensitiveness is the same in all
animals of a given species (p. 341). Indeed, this is the working assumption of much biological research: that findings obtained from some
presumed randomly bred sample of Rattus rattus will be representative of the universal rat, and subsequently generalizable to all rats,
and mice, and maybe to all humans. It remains an empirical question as to just how untenable this assumption is. Unfortunately for pain
researchers (but fortunately for pain genet
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviations: RI, recombinant inbred; SIA, stress-induced analgesia; CIP, congenital insensitivity to pain; cM, centimorgan; QTL,
quantitative trait locus; 5-HT, serotonin.
* To whom reprint requests should be addressed, e-mail: [email protected] .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE GENETIC MEDIATION OF INDIVIDUAL DIFFERENCES IN SENSITIVITY TO PAIN AND ITS INHIBITION 7745

icists!), the existing data regarding the nociceptive and analgesic sensitivity of laboratory rodent populations reveal great, and in some
cases qualitative, variability.
Rodent populations of use for genetic analysis can either be produced (e.g., inbred strains, recombinant inbred strains, artificially
selected lines, transgenics) or identified (e.g., spontaneous mutants). The characteristics of these genetic models have been reviewed
elsewhere (e.g., ref. 24 ). The most studied genetic rodent models of relevance to pain include the recombinant inbred (RI) CXBK mouse
strain ( 25 , 26 ), the High Analgesia/Low Analgesia (HA/LA) mouse lines selectively bred for swim stress-induced analgesia (SIA) ( 27 ,
28 ), the High Analgesic Response/Low Analgesic Response (HAR/LAR) mouse lines selectively bred for levorphanol analgesia ( 29 ),
the High Autotomy/Low Autotomy (HA/LA) rat lines selectively bred for autotomy ( 30 ), and the normotensive Wistar Kyoto versus
Spontaneously Hypertensive Rat (WKY/SHR). The former models have been thoroughly reviewed ( 24 ). The literature regarding the
relationship between nociception and genetic or experimentally induced hypertension has also been recently reviewed ( 31 ).
Strain differences. Although less well studied at present, the comparison of inbred strain responses is more relevant to the issue of the
scope of individual differences than the aforementioned models. Inbred strains are derived by repeated (>20 generation) full-sibling (i.e.,
brother sister) mating ( 32 ). Mating of individuals with common ancestors increases the probability of offspring inheriting two copies of
the same allele identical-by-descent. During inbreeding, therefore, genetic heterozygosity is progressively lost as alleles of initially
segregating genes are fixed into a homozygous state.
Table 1. Rat strain differences of relevance to pain
Trait Parameters Administration Stimulus Strain Difference * Ref.
Nociceptive sensitivity
Thermal TW WAG > F344 117
TF LE = LEW = WIS > F344 = SD 118
HP LEW > F344 119
Electrical FJ F344 > SD 120
Mechanical CD LEW > F344 121
VF F344 = LEW = WIS > SD 42
Chemical CFA LEW > SD 122
CFA LEW > AVN 123
Neuropathic NT SD > WKY 124
NT LE = SAB = SD = WKY > LEW 125
NT BUF = SD > BN > WIS > LEW 126
Analgesia
Morphine 015 mg/kg i.p. FJ SD > F344 120
010 mg/kg i.p. TF SD > WIS 127
010 mg/kg i.p. TW F344 > WAG 117
020 mg/kg i.p. FT F344 > LEW 41
010 mg/kg i.p. TF LE = SD LEW = WIS F344 118
010 mg/kg s.c. HP SD > WKY 128
Codeine 50400 mol/kg s.c. TF SD > DA 116
Clonidine 060 g/kg i.p. FT SD > WKY 129
10 g i.c.v. TW WAG > F344 130
TRH 1 mg/kg i.p. TW WAG > F344 131
Serotonin 0300 g/kg i.v. TF SD > WKY 132
0.5 g i.c.v. TW F344 > WAG 130
Footshock 1.5mA TW F344 > SD 133
Restraint 30 min in tubes TF LE = SD > F344 = LEW = WIS 118
Acupuncture 100 Hz, 13 mA TF P77PMC > WIS 134
Analgesic tolerance
Morphine 14 510 mg/kg i.p. TF SD > WIS 127
8 10 mg/kg s.c. HP WKY > SD 128
Strain Abbreviations: BN, Brown Norway; BUF, Buffalo; DA, Dark Agouti; F344, Fischer 344; LE, LongEvans (outbred); LEW,
Lewis; SAB, Sabra (outbred); SD, SpragueDawley (outbred); WAG, Wistar Albino Glaxo (WAG/GSto); WIS, Wistar (outbred); WKY,
Wistar Kyoto. Genealogical origins of all inbred rat strains can be found at http://www.informatics.jax.org/bin/strains/search .
Other Abbreviations: CD, colorectal distention; CFA, complete Freunds adjuvant; FJ, flinchjump test; FT, formalin test; HP, hot-
plate test; i.c.v., intracerebroventricular, i.p., intraperitoneal; NT, sciatic and saphenous nerve transection; TF, radiant heat tail-flick test;
TRH, thyrotropin-releasing hormone; TW, hot water tail-immersion/withdrawal test.
* Only studies with significant strain differences are reported. Excluded are studies involving selected lines [including the
spontaneously hypertensive rat (SHR)] and mutants.
Morphine analgesia was significantly attenuated by pretreatment with p-chlorophenylalanine in SD, but not F344 rats.
Clonidine analgesia was naloxone-reversible in SD rats but naloxone-insensitive in WKY rats.
F344 rats also developed increased conditioned analgesia to footshock relative to SD rats.

Tables 1 and 2 present some existing data regarding inbred (and outbred) strain differences of relevance to pain in rats and mice,
respectively. The only obvious generalizations that can be made from Table 1 are the nociceptive sensitivity of the Lewis (LEW) inbred
rat strain and the sensitivity to a wide variety of analgesic manipulations of the outbred Sprague Dawley (SD) strain. Multistrain
comparisons (strain surveys) are far more common in the mouse because of the ready availability of over 30 major inbred strains.
the authoritative version for attribution.

Obvious generalizations from mouse strain surveys are thus harder to make. One exception is the voluminous research demonstrating the
relative sensitivity of the DBA/2 (D2) strain to opioid analgesia compared with the C57BL/6 (B6) strain (not shown in Table 2 , but
reviewed in refs. 33 and 34 ). Although D2 mice display high, and B6 mice display low magnitudes of analgesia, the B6 strain is markedly
more sensitive than the D2 strain to other opioid-mediated phenomena, including locomotor activation, learning/memory, and muscular
rigidity (Straub tail).
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE GENETIC MEDIATION OF INDIVIDUAL DIFFERENCES IN SENSITIVITY TO PAIN AND ITS INHIBITION 7746

Strain Abbreviations (substrain identifiers are omitted; in most cases, inbred strains were obtained from The Jackson Laboratory: B6,
C57BL/6; BALB, BALB/c; C3H, C3H/He; CD-1, Hsd:ICR (outbred); CF-1, Hsd:NSA (outbred); CFW, HsdWin:CFW1 (outbred); D2,
DBA/2; ICR, Institute for Cancer Research stock (many suppliers; outbred); SW, Swiss Webster (outbred). Genealogical origins of all
inbred mouse strains can be found in Festing (37) or at http://www.informatics.jax.org/bin/strains/search .
Other Abbreviations: AC, abdominal constriction (writhing) test; CAR, carrageenan; FT, formalin test; i.c.v., intracerebroventricular;
i.p., intraperitoneal; i.v., intravenous; HP, hot-plate test; NalBzoH, naloxone benzoylhydrazone (a 3-opioid agonist); NT, sciatic and
saphenous nerve transection; PNI, peripheral nerve injury (Chung model); TF, radiant heat tail-flick test; TW, hot water tail-immersion/
withdrawal test; VF, von Frey fiber test.
* Only studies with significant strain differences are reported. Excluded are studies involving selected lines, mutants, recombinant
inbred (RI) strains, and non-extant populations. Also excluded are studies specifically comparing the B6 and D2 strains.
Strain differences observed varied with sex.
No analgesic tolerance whatsoever developed in the 129/SvEv substrain used.

Qualitative strain differences. In addition to the quantitative strain differences compiled in Tables 1 and 2 , some very intriguing
qualitative strain differences of relevance to pain have been noted. For instance, a number of investigations have suggested that certain
strains activate opioid analgesic systems after exposure to stress, whereas other strains produce approximately equivalent amounts of SIA,
but of a non-opioid (i.e., naloxone-insensitive) character ( 35 37 ). Vaccarino et al. ( 38 ) reported that naloxone injection produced
paradoxical analgesia on the formalin test in BALB/c mice, but not B6 or outbred CD-1 mice. Fujimoto and colleagues ( 39 ) have
demonstrated convincingly that heroin analgesia is mediated by -opioid receptor activation in outbred Institute for Cancer Research stock
the authoritative version for attribution.

(ICR) mice, but by -receptor activation in outbred Swiss Webster (SW) mice. The same workers have recently ( 40 ) identified -, -, and
-type heroin responders among inbred mouse strains. Vaccarino and Couret, Jr. ( 41 ) observed that the presence of formalin-induced pain
during tolerance induction wholly prevented tolerance development in the Fischer 344 (F344) strain but not in the LEW strain. Lee et al. (
42 ) observed a complete blockade of neuropathic mechanical allodynia after treatment with the -adrenergic receptor antagonist,
phentolamine, in LEW rats; this treatment was wholly ineffective in F344 rats. The authors concluded that this neuropathy was
sympathetically maintained in the former strain only. Finally, Proudfits laboratory has demonstrated ( 43 ) that electrical stimulation-
produced analgesia is reversed by 2-adrenergic antagonists in SD rats from
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE GENETIC MEDIATION OF INDIVIDUAL DIFFERENCES IN SENSITIVITY TO PAIN AND ITS INHIBITION 7747

the now defunct Sasco (Omaha, NE), but not in SD rats from HarlanSpragueDawley. This difference may be explained by the
differential projection routes and dorsal horn termination fields (laminae VIIX versus laminae IIV, respectively) of pontospinal
noradrenergic neurons in these substrains ( 44 ).
HERITABILITY OF PAIN-RELATED TRAITS
That individual differences in pain-related traits exist, of course, does not imply that these differences are necessarily attributable to
genetic factors. Familial aggregation of pain pathologies and extremes of pain sensitivity has been repeatedly noted in humans, but such
findings have almost uniformly been attributed to shared environmental variance and/or familial modeling (e.g., refs. 45 49 ). To
separate genetic and environmental factors, twin studies have been conducted (some even featuring adoption), comparing the concordance
rates of pain-related traits in monozygotic (identical) versus dizygotic (fraternal) twins (see ref. 50 ). Heritability (i.e., the proportion of
overall phenotypic variance accounted for by genetic factors) has been estimated as 3958% for migraine ( 51 53 ), 55% for menstrual
pain ( 54 ), 50% for back pain ( 55 ), and 21% for sciatica ( 56 ).
Only one report exists of a twin study of nonpathological, basal pain sensitivity. MacGregor et al. ( 57 ) assessed forehead pressure
pain thresholds in 269 monozygotic pairs and 340 dizygotic pairs, and observed only a slight excess correlation in monozygotic versus
dizygotic twins (r = 0.57 vs. 0.51, respectively). This excess corresponds to a heritability of this trait of only 10% and suggests that the
twin correlations in pain thresholds are largely due to shared environmental factors.
In contrast to this last study in humans, heritability estimates for nociceptive and analgesic sensitivity in mice are fairly high, ranging
from 28% to 76% ( 29 , 58 60 )certainly well within the range considered for further genetic analysis. Because none of the murine
nociceptive assays used are similar to the forehead pressure pain test used by MacGregor et al. ( 57 ), it is difficult at the present time to
evaluate whether humans and mice truly differ in the contribution of genes to pain sensitivity. Even if genetic factors are ultimately
demonstrated to play only a minor role in the determination of individual pain sensitivity in humans, genetic studies of pain may still prove
highly valuable. Such studies, for example, may illuminate those components of pain processing circuitry in mice and humans that are
especially amenable to alteration, knowledge likely to be useful for the development of novel analgesic strategies.
GENETIC CORRELATIONS AMONG PAIN PHENOTYPES
The tools of classical genetics can be used, even in advance of the identification of the relevant genes, to determine whether traits
share genetic mediation (see ref. 61 ). The fact that a given gene can influence more than one trait is known in genetic parlance as
pleiotropy. Pleiotropic actions of genes result in the genetic correlation of traits, because allelic variation in a gene will influence all traits
in which that gene participates. The determination of genetic correlation has proven to be very heuristic, leading to novel theories
regarding the underlying physiological mediation of traits. A number of genetic correlations of pain-related phenotypes have been noted
using selected lines and inbred strains (see ref. 24 ). I would like to focus on two intriguing findings, as follows.
Genetic Correlation of Nociception and Opiate Analgesia. It has been demonstrated by several groups that a negative genetic
correlation exists between initial nociceptive sensitivity and subsequent morphine analgesia ( 27 , 62 , 63 ). That is, mice that are initially
sensitive to noxious stimuli tend to exhibit modest analgesic responses to morphine, whereas mice that are relatively resistant to basal
nociception exhibit robust morphine analgesia. In two separate studies using multiple inbred strains, this correlation was estimated as r =
0.63 to 0.85 ( 62 ) and r = 0.61 ( 58 ). Thus, mice are doubly advantaged or doubly disadvantaged with respect to nociception and
its opiate inhibition. Interestingly, this negative correlation (albeit with n = 2 only) can also be observed when comparing the sexes.
Genetic Correlations Among Nociceptive Assays. We recently tested 11 inbred strains on 12 separate measures of nociception in
common use in the mouse ( 59 , 60 ). The assays used can be placed on a number of dimensions, including etiology (nociceptive,
inflammatory, neuropathic), modality (thermal, chemical, mechanical), duration (acute, tonic, chronic), and location (cutaneous,
subcutaneous, visceral). We reasoned that inbred strain variation could be exploited to identify clusters of genetically correlated
nociceptive assays. Similar genetic mediation implies similar physiological mediation of assays, suggesting that they measure the same
type of pain as defined mechanistically. Essentially, we were attempting to produce a natural rather than artificial taxonomy of
nociception in the mouse, similar to that called for by Woolf et al. ( 64 ).
The results of this effort were variously expected and surprising. By using multivariate analyses, we identified three obvious clusters
of pain tests, in which within-cluster genetic correlations greatly exceeded between-cluster correlations: thermal (Hargreaves test,
hotplate test, tail-immersion/ withdrawal test, and, surprisingly, autotomy), chemical (acetic acid abdominal constriction, magnesium
sulfate abdominal constriction, acute- and tonic-phase formalin test), and mechanical + hypersensitivity (von Frey test, carrageenan
thermal hypersensitivity, peripheral nerve injury thermal, and mechanical hypersensitivity) ( 59 , 60 ). Thus, the stimulus modality
dimension accounted for the obtained genetic correlations to a far greater degree than any other factor. The presence or absence of
neuropathy or inflammation was found to be essentially irrelevant as was the site or duration of the stimulus.
IDENTIFICATION OF PAIN-RELATED GENES
The holy grail of pain genetics, of course, is the actual identification of pain-related genes and the polymorphisms within or near such
genes that account for trait variability. Note that pain-related gene could be broadly defined as any gene encoding a protein of known
pain relevance or of a gene whose null mutant exhibits a pain-related phenotype. Defined in this way, a large number of pain-related genes
are known. However, if more properly defined as one in which allelic variation directly produces individual differences or pathology, only
a handful of pain-related genes have been identified.
Techniques. Essentially, there are two ways to identify genes associated with trait variability: (i) linkage analysis, including classical
model-based linkage techniques and allele-sharing methods in humans and test-crosses in animals, and (ii) association studies (reviewed in
ref. 65 ). Linkage analyses follow familial inheritance patterns, whereas association studies compare allele frequencies in defined
populations. Given the increasingly large number of genes already cloned and mapped in Homo sapiens and Mus musculus, linkage studies
may lead immediately to the identification of candidate genes, which can then be studied by using the latter approach once allelic variants
are found. Candidate genes, of course, can also be evaluated by using nongenetic means, by investigating the physiology of the proteins
they encode. Failing the identification of an already cloned candidate gene, positional cloning techniques (e.g., ref. 66 ) can be used to
the authoritative version for attribution.

narrow the 20-centimorgan (cM)-wide chromosomal region identified by linkage down to the <1-cM-wide region required to realistically
attempt DNA sequencing.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE GENETIC MEDIATION OF INDIVIDUAL DIFFERENCES IN SENSITIVITY TO PAIN AND ITS INHIBITION 7748

Human Studies. A handful of single gene pain pathologies have recently been, or are on the verge of being, explained on the DNA
sequence level.
Congenital insensitivity to pain. Over 40 cases of congenital insensitivity to pain (CIP) with preservation of all other sensory
modalities have been reported since the original description of a carnival performer known as The Human Pincushion (see ref. 67 ).
Recently, the genetic basis of this neuropathy (CIP with anhidrosis; hereditary sensory and autonomic neuropathy type IV) was elucidated.
Based in part on the striking similarities between CIP and the phenotype of null-mutant mice lacking the Ntrk1 gene encoding the high
affinity, nerve growth factor-specific tyrosine kinase receptor, Indo et al. ( 68 ) considered the homologous human gene, NTRK1
(previously known as TRKA), as a candidate for CIP. Direct sequencing of the coding region of TRKA in four unrelated CIP patients
revealed three separate, exonic mutations: a single base (C) deletion, an A C transversion, and a G C transversion ( 68 ).
Other sensory neuropathies. Nicholson et al. ( 69 ) performed a microsatellite marker genome screen on 102 members of four
Australian kindreds with multiple individuals with hereditary sensory neuropathy type I, a disease featuring loss of all sensory modalities
but especially pain and temperature. Linkage was established to a series of markers encompassing a 5-cM region of chromosome 9. An
obvious candidate gene, NTRK2, encoding the tyrosine kinase receptor type 2 that is bound by brain-derived neurotrophic factor and
neurotrophin-4, was excluded by informative recombination events. Positional cloning efforts facilitated by the development of a yeast
artificial chromosome-based transcript map are ongoing ( 70 ).
Another hereditary sensory neuropathy (type II), featuring loss of pain sensation and autoamputation, was recently subjected to
genetic analysis in a consanguineous family with two affected sisters ( 71 ). These investigators used exclusion mapping, a mixed linkage/
association strategy, to exclude a variety of known neurotrophin-related genes as candidates for this disorder.
Migraine. An important recent finding of relevance to a more prevalent painful condition, migraine, has been the attribution of
familial hemiplegic migraine and migraine-like episodic ataxia type 2 to mutations in the P/Q-type, calcium channel 1 subunit gene,
CACNL1A4 ( 72 ). Previous linkage studies mapped the gene for these disorders to chromosome 19p13 (e.g., ref. 73 ). By using a
technique called exon trapping, Ophoff et al. ( 72 ) cloned the 47-exon-long CACNL1A4 gene in this region, and identified four different
missense point mutations in affected individuals that segregated with the disease in five families. Although familial hemiplegic migraine
and episodic ataxia type 2 are rare, genetic factors are known to play a role in normal migraine as well, and common allelic variants of
this or other ion channel genes may (or may not, ref. 74 ) contribute to its etiology ( 75 ). Of interest as well is a report of a familial
migraine susceptibility locus on the X chromosome, which may explain the preponderance of this condition in females ( 76 ).
Pathologies linked to human lymphocyte antigens. Other preliminary genetic investigations of painful pathologies with familial
aggregationincluding reflex sympathetic dystrophy/ complex regional pain syndrome ( 77 ), rheumatoid arthritis (e.g., ref. 78 ), and
fibromyalgia (e.g., ref. 79 )have demonstrated at least provisional linkage to or association with various human lymphocyte antigen
regions or antigens. This system does not seem to be linked, however, to familial predisposition to discogenic low-back pain ( 80 ). It is
expected that full-scale genetic investigations of these and other complex pain traits are ongoing or imminent. What are less likely to occur
are investigations of nonpathological pain traits, owing to the added complexity introduced by a continuous phenotype. For genetic
investigation of normal pain sensitivity and sensitivity to analgesia, animal studies provide much-needed statistical power.
Animal Studies. At the present time, three published studies exist (although many more are ongoing in my laboratory and others)
demonstrating linkage of a pain-related trait to chromosomal locations in the mouse. A two-step quantitative trait locus (QTL) mapping
approach has been chosen in each case (see refs. 81 and 82 for a detailed description). First, RI strains of the 26-strain BxD set, developed
by Taylor ( 83 ) from an F2 intercross between B6 and D2 mice, were phenotyped to identify provisional linkages. These linkages were
then independently confirmed or disconfirmed by using new (B6 D2) F2 hybrids.
Morphine analgesia. Belknap and Crabbe ( 84 ) tested BxD RI strains for a number of systemic morphine responses, including
analgesia on the hot-plate test. Of eight broad chromosomal regions provisionally found to be linked with morphine analgesic magnitude,
two have subsequently been confirmed beyond the level of suggestive linkage (P < .0016) as proposed by Lander and Kruglyak ( 85 ):
the Mpmv5 region (020 cM) of mouse chromosome 10 ( 86 ) and the Myo5a region (3050 cM) of mouse chromosome 9 ( 87 ).
The results to date of this ongoing QTL mapping study nicely illustrate the utility of the approach. In the Mpmv5 region lies the Oprm
gene (8 cM) encoding the mouse -opioid receptor type. Oprm is an obvious candidate gene for morphine analgesic magnitude, implicated
via pharmacological (see ref. 88 ) and transgenic ( 89 , 90 ) studies. Allelic variation at this QTL accounts for 2833% of the observed
genetic variability, and F2 mice inheriting two copies of the D2 allele at this QTL exhibit 4-fold more analgesia from a 16 mg/kg dose of
morphine than do F2 mice inheriting two copies of the B6 allele ( 86 ). This QTL has been statistically associated with other opioid traits
as well, including morphine consumption ( 91 ) and whole-brain [3H]naloxone binding ( 86 ).
One may have expected a priori that the gene encoding the -opioid receptor would be associated with morphine analgesic
magnitude. The candidate gene on mouse chromosome 9 is perhaps more heuristic. Within this region lies the Htr1b gene (46 cM),
encoding the serotonin-1B (5-HT1B) receptor subtype (the mouse analog of the human 5-HT1D receptor). Based on the hypothesis that
Htr1b might represent the QTL for morphine analgesia on chromosome 9, we conducted a series of pharmacological experiments that
provided substantial support for the involvement of spinal 5-HT1B receptors ( 87 ). Although data exist indicating a relationship between 5-
HT1B receptors and opioid analgesia (e.g., ref. 92 ), there is still much confusion regarding the specific role of the 5-HT1A versus 5-HT1B
subtypes, and until very recently subtype-specific ligands were decidedly lacking ( 93 ). Thus, it is unlikely that the studies we conducted
would have been conceived of in the absence of the QTL mapping data.
Basal nociceptive sensitivity. By using similar methodology to that described above, we recently mapped basal thermal nociceptive
sensitivity by using the hot-plate test ( 94 ). The most promising of six putative linkages in BxD RI strainsthe D4Mit71 region (5070
cM) of mouse chromosome 4was largely confirmed by using F2 mice. This QTL displayed evidence of sex specificity, with a combined
BxD/F2 P value of 0.005 for males but only 0.085 for females. The identification of a candidate gene in this region, the Oprd1 gene (65
cM) encoding the mouse -opioid receptor type, inspired a simple pharmacological experiment in which B6 and D2 mice of both sexes
were administered ,-, -, and -specific antagonists prior to assessment of hot-plate sensitivity. As predicted by the hypothesis of Oprd1
as a male-specific QTL for this trait, we observed a sex- and strain-dependent pattern of responses, with the -specific antagonist,
naltrindole, lowering nociceptive latencies in the following order of efficacy: D2 male > B6 male > D2 female > B6 female ( 94 ).
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE GENETIC MEDIATION OF INDIVIDUAL DIFFERENCES IN SENSITIVITY TO PAIN AND ITS INHIBITION 7749

Nonopioid SIA. It has long been known that many forms of analgesia are resistant to antagonism by naloxone, representing the
recruitment of non-opioid mechanisms ( 95 ). To shed light on the mediation of these powerful but little understood systems, we conducted
a QTL mapping experiment of nonopioid SIA resulting from 3-min forced swims in 15C water ( 96 ). Six putative QTLs were identified
in the BxD RI phase, of which four were subsequently disconfirmed by using F2 hybrids. Of the two remaining QTLs, one on
chromosome 8 (5080 cM) was confirmed beyond Lander and Kruglyaks ( 85 ) threshold for significant linkage. This QTL (dubbed
Siafq1) exhibited compelling evidence of sex specificity, reaching a combined BxD/F2 P value of 0.00000012 for females but only 0.038
for males. Female F2 mice inheriting two copies of the D2 allele at this locus displayed 3-fold more SIA than those inheriting two copies
of the B6 allele ( 96 ). This finding of a female-specific QTL for SIA is of special interest because we ( 36 , 97 ) and others (e.g., ref. 98 )
had previously demonstrated the existence of qualitative sex differences in the neurochemical mediation of this trait.
SEX-SPECIFIC GENETIC MEDIATION OF PAIN AND ANALGESIA
The two findings of sex-specific QTLs described abovea male-specific QTL for baseline thermal nociceptive sensitivity and a
female-specific QTL for non-opioid SIAexemplify a phenomenon we and others have found repeatedly, i.e., sex/ genotype interactions
of relevance to nociception and its modulation. The discovery of sex-specific QTLs on autosomes, first reported by Melo et al. ( 99 ) for
alcohol preference, was surprising to many, but now more and more examples are being uncovered. It should be emphasized that the
existence of autosomal, sex-specific QTLs does not imply that the sexes possess or express different genes, but rather that different genes
are associated with trait variability in each sex. The existence of sex-specific QTLs does, however, imply that males and females possess
at least partially independent physiological mechanisms underlying the traits in question.
Sex differences in nociception and analgesia are controversial, but when differences are found, males of a number of species
consistently display higher thresholds, tolerance, and analgesic sensitivity (see refs. 100 and 101 for reviews). The inability of some to
observe these sex differences has been attributed to estrus cycle variability, test specificity, and experimental parameters (e.g., ref. 102 ).
Recent data from my laboratory suggest that an important factor contributing to variable results in this literature has been overlooked, i.e.,
genotype of the test subjects. For example, a recent survey of supraspinal morphine analgesia in 11 inbred strains revealed no significant
sex differences in morphine analgesic potency in seven of these strains ( 58 ). In three strains (AKR, B6, and SWR), males exhibited 3.5-
to 7-fold higher sensitivity to intracerebroventricularly administered morphine than their female counterparts. Finally, one strain (CBA)
was identified in which females were 5-fold more sensitive to morphine than males. In another, just completed study specifically
comparing outbred mouse strains, we found that a large male-vs.-female difference in baseline tail-flick latencies can be seen in SW mice
obtained from Simonsen Laboratories (Gilroy, CA), but the analogous sex difference in SW and CD-1 mice from HarlanSpragueDawley
is either absent or too small to detect statistically with n = 1632 (unpublished data). This vendor effect between SW mice from two
different suppliers is likely due to genetic factors, because both populations have been bred in my vivarium for several generations.
Ultimately, then, the failure of some to detect sex differences may simply be due to the fact that in the subject population chosen, there is
no sex difference to detect.
Another intriguing sex/genotype interaction is that demonstrated by Rady and Fujimoto ( 103 ). As described above, these
investigators have determined that heroin analgesia is mediated by -opioid receptors in ICR mice of both sexes but by -opioid receptors
in SW mice of both sexes ( 39 ). An analysis of reciprocal (ICR SW) F1 hybrid mice revealed that male offspring displayed an ICR-like
phenotype and female offspring displayed a SW-like phenotype. That is, in F1 males, heroin analgesia was blocked by -opioid- but not -
opioid-specific antagonists, whereas the reverse was true for F1 females ( 103 ). This is likely an example of sex-influenced autosomal
dominant inheritance (see ref. 104 ). Further analysis of this phenomenon, if it can be replicated in inbred strains, may help to illuminate
the basis of sex/genotype interactions in analgesia. Also potentially enlightening is our ongoing mapping study of supraspinal morphine
analgesia in (AKR CBA) F2 mice, focusing specifically on the identification of sex-specific QTLs. Male mice of these two strains
exhibit equipotent analgesic sensitivity to morphine, whereas the of female mice differ by a factor of 35 ( 58 ).
ALLELIC VARIANTS OF PAIN-RELATED GENES
Once the genes mediating a trait have been positively identified, a crucial task still remainsthe identification of alternate alleles of
that gene giving rise to the original phenotypic difference between individuals and/or populations. This effort is again rendered more
difficult when considering complex genetic traits, because the allelic variants are more likely to be single base pair changes (single-
nucleotide polymorphisms) than chromosomal rearrangements or large deletions. Also, whereas the mutations giving rise to disease
phenotypes are likely to occur in the coding region of a gene, allelic variation causing subtle changes in gene expression can occur outside
(even far outside) the coding region. Nonetheless, some success has been reported.
Opioid Receptor Genes. The coding and much of the regulatory and intronic regions of human opioid receptor genes have been
sequenced (e.g., ref. 105 ). By using direct sequencing of hundreds of individuals, three separate investigations identified two common
variants of the OPRM gene coding for the -opioid receptor ( 106 108 ), with allele frequencies estimated to be 6.611%. An A118G
variant (i.e., an A G substitution in nucleotide 118 of exon 1, resulting in a Asn Asp change in amino acid residue 40) was found to
be present in a lower proportion of opioid-dependent subjects than controls, whereas the C17T variant was more common in opioid-
dependent subjects ( 106 , 107 ). The A118G variant was found not to be associated with susceptibility to alcohol dependence ( 108 ). An
A118G -opioid receptor constructed by using site-directed mutagenesis and stably transfected into cell lines displayed higher binding
affinity for -endorphin than the more common wild-type receptor ( 106 ).
An allelic variant (T307C) of the OPRD gene encoding the -opioid receptor has also been found ( 109 ). Although the amino acid
sequence remains unchanged by this substitution, the investigators found that heroin addicts were significantly more likely than controls to
possess a CC genotype and less likely to possess a TT genotype. It was concluded that, although by unknown mechanisms, the C allele
predisposes to heroin abuse. The direct relevance of any such opioid receptor variants to pain or analgesic sensitivity is as yet unpublished,
although this work is no doubt underway in several laboratories.
Cytochrome P450. One genetic polymorphism of well documented relevance to pain is of the gene coding for the neuronal
cytochrome P450IID6 (CYP2D6; sparteine/ debrisoquine oxygenase) enzyme (see ref. 110 for review). This enzyme is in fact absent in
710% of Caucasians, who are thus unable to convert codeine to morphine by O-demethylation ( 111 ). Because much evidence indicates
the authoritative version for attribution.

that codeine produces analgesic effects by being biotransformed to morphine, these poor metabolizers will receive minimal therapeutic
benefit from administration of codeine but are generally subject nonetheless to its side effects ( 112 , 113 ). It has been shown as well that
poor metabolizers report increased pain compared with extensive metabolizers in the cold pressor test ( 114 ). An animal model of this
phenomenon exists, with the female Dark Agouti (DA) rat showing a poor metabolizer phenotype ( 115 , 116 ). This well known example
should serve to remind that much individual variability in drug response may be due to polymorphisms related to pharmacokinetics rather
than pharmacodynamics.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE GENETIC MEDIATION OF INDIVIDUAL DIFFERENCES IN SENSITIVITY TO PAIN AND ITS INHIBITION 7750

FUTURE DIRECTIONS
The construction of high-density genomic maps and the initial priority of the Mouse and Human Genome Projects will greatly
facilitate (and already has) the identification of the 50,000100,000 mammalian genes. Given the high degree of redundancy and
pleiotropy known to exist in biological systems, the determination of which genes participate in which physiological mechanisms will
remain a daunting task, occupying scientists for decades to come. New, high-throughput genomic technologies (e.g., gene chips) may
further accelerate the rate of discovery. It is likely that the focus in Homo sapiens will be on pathology, whereas animal models like the
mouse will continue to be used to investigate more subtle questions involving the normal range of behavior. Although the use of genetic
techniques naturally garners much excitement, it must be borne in mind that even variability in pain pathologies is largely determined by
environmental factors. Thus, investigations into the psychosocial determinants of pain tolerance and pain behaviors must continue
unabated. Nonetheless, knowledge of the genetic bases of pain-related traits may have important scientific and clinical implications,
facilitating both the development of novel analgesic strategies and improved, idiosyncratic treatment of pain using conventional therapies.
The author is supported by National Institutes of Health Grants DA11394 and DE12735.
1. IASP Subcommittee on Toxicology ( 1979 ) Pain 6 , 249 .
2. Beecher, H. K. ( 1959 ) Measurement of Subjective Responses ( Oxford Univ. Press , New York ).
3. Mogil, J. S. & Grisel, J. E. ( 1998 ) Pain 77 , 107128 .
4. Libman, E. ( 1934 ) J. Am. Med. Assoc. 102 , 335341 .
5. Sherman, E. D. ( 1943 ) Can. Med. Assoc. J. 45 , 437441 .
6. Chapman, W. P. & Jones, C. M. ( 1944 ) J. Clin. Invest. 23 , 8191 .
7. Clark, J. W. & Bindra, D. ( 1956 ) Can. J. Psychol. 10 , 6976 .
8. Woodrow, K. M. , Friedman, G. D. , Siegelaub, A. B. & Collen, M. F. ( 1972 ) Psychosom. Med. 34 , 548556 .
9. Wolff, B. B. & Jarvik, M. E. ( 1963 ) Can. J. Psychol. 17 , 3744 .
10. Wolff, B. B. & Jarvik, M. E. ( 1964 ) Am. J. Psychol. 77 , 589599 .
11. Lanier, L. H. ( 1943 ) Science 97 , 4950 .
12. Lasagna, L. & Beecher, H. K. ( 1954 ) J. Am. Med. Assoc. 156 , 230234 .
13. Wolff, B. B. , Kantor, T. G. , Jarvik, M. E. & Laska, E. ( 1965 ) Clin. Pharmacol. Ther. 7 , 224238 .
14. Brennum, J. , Kjeldsen, M. , Jensen, K. & Jensen, T. S. ( 1989 ) Pain 38 , 211217 .
15. Isselee, H. , De Laat, A. , Lesaffre, E. & Lysens, R. ( 1997 ) Eur. J. Oral Sci. 105 , 583587 .
16. Chen, A. C. N. , Dworkin, S. F. , Haug, J. & Gehrig, J. ( 1989 ) Pain 37 , 129141 .
17. Jacobs, J. W. G. , Geenan, R. , van der Heide, A. , Rasker, J. J. & Bijlsma, J. W. J. ( 1995 ) Scand. J. Rheumatol. 24 , 243247 .
18. Chen, A. C. N. , Dworkin, S. F. & Haug, J. ( 1989 ) Pain 37 , 143160 .
19. Galer, B. S. , Coyle, N. , Pasternak, G. W. & Portenoy, R. K. ( 1992 ) Pain 49 , 8791 .
20. Portenoy, K. M. , Foley, K. M. & Inturrisi, C. E. ( 1990 ) Pain 43 , 273286 .
21. Chapman, C. R. , Hill, H. F. , Saeger, L. & Gavrin, J. ( 1990 ) Pain 43 , 4755 .
22. Levine, J. D. , Gordon, N. C , Smith, R. & Fields, H. L. ( 1981 ) Pain 10 , 379389 .
23. Buchsbaum, M. S. , Davis, G. C. & Bunney, W. E., Jr. ( 1977 ) Nature (London) 270 , 620622 .
24. Mogil, J. S. , Sternberg, W. F. , Marek, P. , Sadowski, B. , Belknap, J. K. & Liebeskind, J. C. ( 1996 ) Proc. Natl. Acad. Sci. USA 93 , 30483055 .
25. Bailey, D. W. ( 1971 ) Transplantation 11 , 325327 .
26. Shuster, L. , Webster, G. W. , Yu, G. & Eleftheriou, B. E. ( 1975 ) Psychopharmacologia 42 , 249254 .
27. Panocka, I. , Marek, P. & Sadowski, B. ( 1986 ) Brain Res. 397 , 152155 .
28. Panocka, I. , Marek, P. & Sadowski, B. ( 1986 ) Brain Res. 397 , 156160 .
29. Belknap, J. K. , Haltli, N. R. , Goebel, D. M. & Lam, M. ( 1983 ) Behav. Genet. 13 , 383396 .
30. Devor, M. & Raber, P. ( 1990 ) Pain 42 , 5167 .
31. Ghione, S. ( 1996 ) Hypertension 28 , 494504 .
32. Festing, M. F. W. ( 1996 ) in Genetic Variants and Strains of the Laboratory Mouse , eds. Lyon, M. F. , Rastan, S. & Brown, S. D. M. ( Oxford
Univ. Press , Oxford ), Vol. 3 , pp. 15371576 .
33. Belknap, J. K. & OToole, L. A. ( 1991 ) in The Genetic Basis of Alcohol and Drug Actions , eds. Harris, R. A. & Crabbe, J. C. ( Plenum , New
York ), pp. 225252 .
34. Frischknecht, H.-R. , Siegfried, B. & Waser, P. G. ( 1988 ) Experientia 44 , 473481 .
35. Helmstetter, F. J. & Fanselow, M. S. ( 1987 ) Behav. Neurosci. 101 , 735737 .
36. Mogil, J. S. & Belknap, J. K. ( 1997 ) Pharmacol. Biochem. Behav. 56 , 6166 .
37. Urca, G. , Segev, S. & Sarne, Y. ( 1985 ) Brain Res. 329 , 109116 .
38. Vaccarino, A. L. , Tasker, R. A. R. & Melzack, R. ( 1988 ) Neurosci. Lett. 84 , 103107 .
39. Rady, J. J. , Roerig, S. C. & Fujimoto, J. M. ( 1991 ) J. Pharmacol. Exp. Ther. 256 , 448457 .
40. Fujimoto, J. M. , Rady, J. J. & Elmer, G. I. ( 1998 ) Soc. Neurosci. Abstr. 24 , 770 .
41. Vaccarino, A. L. & Couret, L. C., Jr. ( 1995 ) Pain 63 , 385389 .
42. Lee, D. H. , Chung, K. & Chung, J. M. ( 1997 ) Neuroreport 8 , 34533456 .
43. West, W. L. , Yeomans, D. C. & Proudfit, H. K. ( 1993 ) Brain Res. 626 , 127135 .
44. Clark, F. M. & Proudfit, H. K. ( 1992 ) Brain Res. 591 , 4453 .
45. Bachiocco, V. , Scesi, M. , Morselli, A. M. & Carli, G. ( 1993 ) Clin. J. Pain 9 , 266271 .
46. Edwards, P. W. , Zeichner, A. , Kuczmierczyk, A. R. & Boczkowski, J. ( 1985 ) Pain 21 , 379384 .
47. Violon, A. & Giurgea, D. ( 1984 ) Pain 18 , 199203 .
48. Pollard, C. A. ( 1985 ) Psychol. Rep. 57 , 813814 .
49. Lester, N. , Lefebvre, J. & Keefe, F. ( 1994 ) Clin. J. Pain 10 , 282289 .
50. Martin, N. , Boomsma, D. & Machin, G. ( 1997 ) Nat. Genet. 17 , 387392 .
51. Ziegler, D. K. , Hur, Y. M. , Bouchard, T. J., Jr. , Hassanein, R. S. & Barter, R. ( 1998 ) Headache 38 , 417422 .
52. Honkasalo, M.-L. , Kaprio, J. , Winter, T. , Heikkila, K. , Sillanpaa, M. & Koskenvuo, M. ( 1995 ) Headache 35 , 7078 .
53. Larsson, B. , Bille, B. & Pedersen, N. L. ( 1995 ) Headache 35 , 513519 .
54. Treloar, S. A. , Martin, N. G. & Heath, A. C. ( 1998 ) Behav. Genet. 28 , 107116 .
55. Bengtsson, B. & Thorson, J. ( 1991 ) Acta Genet. Med. Gemmellol. (Roma) 40 , 8390 .
56. Heikkila, J. K. , Koskenvuo, M. , Heliovaara, M. , Kurppa, K. , Riihimaki, H. , Heikkila, K. , Rita, H. & Videman, T. ( 1989 ) Ann. Med. 21 , 393
398 .
the authoritative version for attribution.

57. MacGregor, A. J. , Griffiths, G. O. , Baker, J. & Specter, T. D. ( 1997 ) Pain 73 , 253257 .

58. Kest, B. , Wilson, S. G. & Mogil, J. S. ( 1999 ) J. Pharmacol. Exp. Ther. 289 , 13701375 .
59. Mogil, J. S. , Wilson, S. G. , Bon, K. , Lee, S. E. , Chung, K. , Raber, P. , Pieper, J. O. , Hain, H. S. , Belknap, J. K. , Hubert, L. , et al. ( 1999 ) Pain
80 , 6782 .
60. Mogil, J. S. , Wilson, S. G. , Bon, K. , Lee, S. E. , Chung, K. , Raber, P. , Pieper, J. O. , Hain, H. S. , Belknap, J. K. , Hubert, L. , et al. ( 1999 ) Pain
80 , 8393 .
61. Crabbe, J. C. , Phillips, T. J. , Kosobud, A. & Belknap, J. K. ( 1990 ) Alcohol. Clin. Exp. Res. 14 , 141151 .
62. Elmer, G. I. , Pieper, J. O. , Negus, S. S. & Woods, J. H. ( 1997 ) Pain 75 , 129140 .
63. Mogil, J. S. , Kest, B. , Sadowski, B. & Belknap, J. K. ( 1996 ) J. Pharmacol. Exp. Ther. 276 , 532544 .
64. Woolf, C. J. , Bennett, G. J. , Doherty, M. , Dubner, R. , Kidd, B. , Koltzenburg, M. , Lipton, R. , Loeser, J. D. , Payne, R. & Torebjork, E. ( 1998 )
Pain 77 , 227229 .
65. Lander, E. S. & Schork, N. J. ( 1994 ) Science 265 , 20372048 .
66. Darvasi, A. ( 1998 ) Nat. Genet. 18 , 1924 .
67. Thrush, D. C. ( 1973 ) Brain 96 , 369386 .
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

THE GENETIC MEDIATION OF INDIVIDUAL DIFFERENCES IN SENSITIVITY TO PAIN AND ITS INHIBITION 7751

68. Indo, Y. , Tsurata, Y. , Karim, M. A. , Ohta, K. , Kawano, T. , Mitsubuchi, H. , Tonoki, H. , Awaya, Y. & Matsuda, I. ( 1996 ) Nat. Genet. 13 , 485488 .
69. Nicholson, G. A. , Dawkins, J. L. , Blair, I. P. , Kennerson, M. L. , Gordon, M. J. , Cherryson, A. K. , Nash, J. & Bananis, T. ( 1996 ) Nat. Genet. 13 , 101
104 .
70. Blair, I. P. , Hulme, D. , Dawkins, J. L. & Nicholson, G. A. ( 1998 ) Genomics 51 , 277281 .
71. Davar, G. , Shalish, C , Blumenfeld, A. & Breakefield, X. O. ( 1996 ) Pain 67 , 1351359 .
72. Ophoff, R. A. , Terwindt, G. M. , Vergouwe, M. N. , van Eijk, R. , Oefner, P. J. , Hoffman, S. M. G. , Lamerdin, J. E. , Mohrenweiser, H. W. , Bulman, D.
E. , Ferrari, M. , et al. ( 1996 ) Cell 87 , 543552 .
73. Joutel, A. , Bousser, M.-G. , Biousse, V. , Labauge, P. , Chabriat, H. , Nibbio, A. , Maciazek, J. , Meyer, B. , Bach, M. A. , Weissenbach, J. , et al. ( 1993 )
Nat. Genet. 5 , 4045 .
74. Baloh, R. W. , Yue, Q. , Jen, J. C. & Nelson, S. F. ( 1998 ) Soc. Neurosci. Abstr. 24 , 1266 .
75. Peroutka, S. J. ( 1998 ) Clin. Neurosci. 5 , 3437 .
76. Nyholt, D. R. , Dawkins, J. L. , Brimage, P. J. , Goadsby, P. J. , Nicholson, G. A. & Griffiths, L. R. ( 1998 ) Hum. Mol. Genet. 7 , 459463 .
77. Mailis, A. & Wade, J. ( 1994 ) Clin. J. Pain 10 , 210217 .
78. Khan, M. A. , Kushner, I. , Ballou, S. P. & Braun, W. E. ( 1979 ) Lancet 1 , 921922 .
79. Burda, C. D. , Cox, F. R. & Osborne, P. ( 1986 ) Clin. Exper. Rheumatol. 4 , 355357 .
80. Postacchini, F. , Lami, R. & Pugliese, O. ( 1988 ) Spine 13 , 14031406 .
81. Gora-Maslak, G. , McClearn, G. E. , Crabbe, J. C. , Phillips, T. J. , Belknap, J. K. & Plomin, R. ( 1991 ) Psychopharmacology 104 , 413424 .
82. Belknap, J. K. , Dubay, C. , Crabbe, J. C. & Buck, K. J. ( 1996 ) in Handbook of Psychiatric Genetics , eds. Blum, K. & Noble, E. P. ( CRC , New York ),
pp. 435453 .
83. Taylor, B. A. ( 1978 ) in Origin of Inbred Mice , ed. Morse, H. C., III ( Academic , New York ), pp. 423428 .
84. Belknap, J. K. & Crabbe, J. C. ( 1992 ) Ann. N.Y. Acad. Sci. 654 , 311323 .
85. Lander, E. S. & Kruglyak, L. ( 1995 ) Nat. Genet. 11 , 241247 .
86. Belknap, J. K. , Mogil, J. S. , Helms, M. L. , Richards, S. P. , OToole, L. A. , Bergeson, S. E. & Buck, K. J. ( 1995 ) Life Sci. 57 , PL117124 .
87. Hain, H. S. , Belknap, J. K. , Hen, R. & Mogil, J. S. ( 1998 ) J. Pharmacol. Exp. Ther. , in press .
88. Pasternak, G. W. ( 1993 ) Clin. Neuropharmacol. 16 , 118 .
89. Matthes, H. W. D. , Maldonado, R. , Simonin, F. , Valverde, O. , Slowe, S. , Kitchen, I. , Befort, K. , Dierich, A. , Le Meur, M. , Dolle, P. , et al. ( 1996 )
Nature (London) 383 , 819823 .
90. Sora, I. , Takahashi, N. , Funada, M. , Ujike, H. , Revay, R. S. , Donovan, D. M. , Miner, L. L. & Uhl, G. R. ( 1997 ) Proc. Natl. Acad. Sci. USA 94 , 15441549
.
91. Berrettini, W. H. , Ferraro, T. N. , Alexander, R. C. , Buchberg, A. M. & Vogel, W. H. ( 1994 ) Nat. Genet. 7 , 5458 .
92. Alhaider, A. A. & Wilcox, G. L. ( 1993 ) J. Pharmacol. Exp. Ther. 265 , 378385 .
93. Glennon, R. A. & Dukat, M. ( 1991 ) Pharmacol. Biochem. Behav. 40 , 10091017 .
94. Mogil, J. S. , Richards, S. P. , O Toole, L. A. , Helms, M. L. , Mitchell, S. R. & Belknap, J. K. ( 1997 ) Pain 70 , 267277 .
95. Lewis, J. W. , Cannon, J. T. & Liebeskind, J. C. ( 1980 ) Science 208 , 623625 .
96. Mogil, J. S. , Richards, S. P. , OToole, L. A. , Helms, M. L. , Mitchell, S. R. , Kest, B. & Belknap, J. K. ( 1997 ) J. Neurosci. 17 , 79958002 .
97. Mogil, J. S. , Sternberg, W. F. , Kest, B. , Marek, P. & Liebeskind, J. C. ( 1993 ) Pain 253 , 1725 .
98. Kavaliers, M. & Choleris, E. ( 1997 ) Brain Res. 768 , 3036 .
99. Melo, J. A. , Shendure, J. , Pociask, K. & Silver, L. M. ( 1996 ) Nat. Genet. 13 , 147153 .
100. Berkley, K. J. ( 1997 ) Behav. Brain Sci. 20 , 371380 .
101. Kest, B. , Sarton, E. , Mogil, J. S. & Dahan, A. ( 1998 ) in Physiology and Pharmacology of Cardio-respiratory Control , eds. Dahan, A. , Teppema, L. J. &
van Beek, E. J. ( Kluwer, Dordrecht , The Netherlands ), pp. 7582 .
102. Bartok, R. E. & Craft, R. M. ( 1997 ) J. Pharmacol. Exp. Ther. 282 , 769778 .
103. Rady, J. J. & Fujimoto, J. M. ( 1997 ) Pharmacogenetics 7 , 429433 .
104. McClearn, G. E. ( 1995 ) Recent Dev. Alcohol. 12 , 217221 .
105. Wendel, B. & Hoehe, M. R. ( 1998 ) J. Mol. Med. 76 , 525532 .
106. Bond, C. , LaForge, K. S. , Tian, M. , Melia, D. , Zhang, S. , Borg, L. , Gong, J. , Schluger, J. , Strong, J. A. , Leal, S. M. , et al. ( 1998 ) Proc. Natl. Acad.
Sci. USA 95 , 96089613 .
107. Berrettini, W. H. , Hoehe, M. R. , Ferrada, T. N. & Gottheil, E. ( 1997 ) Addict. Biol. 2 , 303308 .
108. Bergen, A. W. , Kokoszka, J. , Peterson, R. , Long, J. C. , Virkkunen, M. , Linnoila, M. & Goldman, D. ( 1997 ) Mol. Psychiatry 2 , 490494 .
109. Mayer, P. , Rochlitz, H. , Rauch, E. , Rommelspacher, H. , Hasse, H. E. , Schmidt, S. & Hollt, V. ( 1997 ) NeuroReport 8 , 25472550 .
110. Meyer, U. A. ( 1994 ) J. Pharm . Pharmacol. 46 , 409415 .
111. Alvan, G. , Bechtel, P. , Iselius, L. & Gundert-Remy, U. ( 1990 ) Eur. J. Clin. Pharmacol. 39 , 533537 .
112. Caraco, Y. , Sheller, J. & Wood, A. J. J. ( 1996 ) J. Pharmacol. Exp. Ther. 278 , 11651174 .
113. Poulsen, L. , Brosen, K. , Arendt-Nielsen, L. , Gram, L. F. , Elbaek, K. & Sindrup, S. H. ( 1996 ) Eur. J. Clin. Pharmacol. 51 , 289295 .
114. Sindrup, S. H. , Poulsen, L. , Brosen, K. , Arendt-Nielsen, L. & Gram, L. F. ( 1993 ) Pain 53 , 335349 .
115. Mikus, G. , Somogyi, A. A. , Bochner, F. & Eichelbaum, M. ( 1991 ) Biochem. Pharmacol. 41 , 757762 .
116. Cleary, J. , Mikus, G. , Somogyi, A. A. & Bochner, F. ( 1994 ) J. Pharmacol. Exp. Ther. 271 , 15281534 .
117. Sudakov, S. K. , Goldberg, S. R. , Borisova, E. V. , Surkova, L. A. , Turina, I. V. , Rusakov, D. J. & Elmer, G. I. ( 1993 ) Psychopharmacology 112 , 183188 .
118. Woolfolk, D. R. & Holtzman, S. G. ( 1995 ) Pharmacol. Biochem. Behav. 51 , 699703 .
119. Stohr, T. , Wermeling, D. S. , Szuran, T. , Pliska, V. , Domeney, A. , Welzl, H. , Weiner, I. & Feldon, J. ( 1998 ) Pharmacol. Biochem. Behav. 59 , 799805 .
120. Tilson, H. A. & Rech, R. H. ( 1974 ) Psychopharmacologia 35 , 4560 .
121. Gschossmann, J. M. , Cheng, S. H. , Martinez, V. , Tache, Y. & Mayer, E. A. ( 1996 ) Digest. Dis. Sci. J. , in press.
122. Rosenthale, M. E. ( 1970 ) Arch. Int. Pharmacodyn. Ther. 188 , 1422 .
123. Zidek, Z. & Perlik, F. ( 1971 ) J. Pharm. Pharmacol. 23 , 389340 .
124. Wiesenfeld, Z. & Hallin, R. G. ( 1981 ) Physiol. Behav. 27 , 735740 .
125. Seltzer, Z. & Shir, Y. ( 1991 ) J. Basic Clin . Physiol. Pharmacol. 2 , 1761 .
126. Carr, M. M. , Best, T. J. , Mackinnon, S. E. & Evans, P. L ( 1992 ) Ann. Plast. Surg. 28 , 538544 .
127. Kasson, B. G. & George, R. ( 1984 ) Life Sci. 34 , 16271634 .
128. Hoffmann, O. , Plesan, A. & Wiesenfeld-Hallin, Z. ( 1998 ) Brain Res. 806 , 232237 .
129. Kunos, G. , Mosqueda-Garcia, R. , Mastrianni, J. A. & Abbott, F. V. ( 1987 ) Can. J. Physiol. Pharmacol. 65 , 16241632 .
130. Tiurina, I. V. , Rusakov, D. I. & Sudakov, S. K. ( 1995 ) Eksp. Klin. Farmakol. 58 , 2224 .
131. Borisova, E. V. & Sudakov, S. K. ( 1995 ) Zh. Vyssh. Nerv. Deiat. Im. I. P. Pavlova 45 , 11741181 .
132. Meller, S. T. , Lewis, S. J. , Brody, M. J. & Gebhart, G. F. ( 1992 ) Brain Res. 587 , 8894 .
133. Rosecrans, J. A. , Robinson, S. E. , Johnson, J. H. , Mokler, D. J. & Hong, J.-S. ( 1986 ) Brain Res. 382 , 7180 .
134. Zhang, L.-X. , Li, X.-L. , Wang, L. & Han, J.-S. ( 1997 ) Brain Res. 745 , 158164 .
135. Eidelberg, E. , Erspamer, R. , Kreinick, C. J. & Harris, J. ( 1975 ) Eur. J. Pharmacol. 32 , 329336 .
136. Ramabadran, K. , Michaud, G. & Jacob, J. J. C. ( 1982 ) Ind. J. Exp. Biol. 20 , 7476 .
the authoritative version for attribution.

137. Brown, D. M. & Hughes, B. O. ( 1962 ) J. Pharm. Pharmacol. 14 , 399405 .

138. Mogil, J. S. , Lichtensteiger, C. A. & Wilson, S. G. ( 1998 ) Pain 76 , 115125 .
139. Navarro, X. , Buti, M. & Verdu, E. ( 1994 ) Restor. Neurol. Neurosci. 6 , 151157 .
140. Defrin, R. , Zeitoun, I. & Urca, G. ( 1996 ) Brain Res. 711 , 241244 .
141. Gebhart, G. F. & Mitchell, C. L. ( 1973 ) Arch. Int. Pharmacodyn. 201 , 128140 .
142. Brase, D. A. , Loh, H. H. & Way, E. L. ( 1977 ) J. Pharmacol. Exp. Ther. 201 , 368374 .
143. Ho, I. K. , Loh, H. H. & Way, E. L. ( 1977 ) J. Pharm. Pharmacol. 29 , 583584 .
144. Navarro, M. , Leza, J.-C , Lizasoain, I. & Lorenzo, P. ( 1991 ) Gen. Pharmacol. 22 , 713716 .
145. Pick, C. G. , Cheng, J. , Paul, D. & Pasternak, G. W. ( 1991 ) Brain Res. 566 , 295298 .
146. Ninomiya, Y. , Kawamura, H. , Nomura, T. , Uebayashi, H. , Sabashi, K. & Funakoshi, M. ( 1990 ) Comp. Biochem. Physiol. C Comp. Pharmacol .
97 , 341343 .
147. Quock, R. M. , Mueller, J. L. & Vaughn, L. K. ( 1993 ) Brain Res. 614 , 5256 .
148. Pick, C. G. ( 1996 ) Eur. Neuropsychopharmacol. 6 , 201205 .
149. Seale, T. W. , Nael, R. & Basmadjian, G. ( 1996 ) NeuroReport 8 , 191195 .
150. Flores, C. M. , Wilson, S. L. & Mogil, J. S. ( 1999 ) Pharmacogenetics , in press.
151. Kolesnikov, Y. , Jain, S. , Wilson, R. & Pasternak, G. W. ( 1998 ) J. Pharmacol. Exp. Ther. 284 , 455459 .
 THE OPIATE RECEPTOR AS A CANDIDATE GENE FOR PAIN: POLYMORPHISMS, VARIATIONS IN EXPRESSION, 7752
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NOCICEPTION, AND OPIATE RESPONSES

This paper was presented at the National Academy of Sciences colloquium The Neurobiology of Pain, held December 1113,
1998, at the Arnold and Mabel Beckman Center in Irvine, CA.

The opiate receptor as a candidate gene for pain: Polymorphisms,

variations in expression, nociception, and opiate responses

GEORGE R. UHL * , ICHIRO SORA *, AND ZAIJIE WANG *

* Molecular Neurobiology Branch, Intramural Research Program, National Institute on Drug Abuse, National Institutes of Health,
Baltimore, MD 21224; and Departments of Neurology and Neuroscience, Johns Hopkins University School of Medicine, Baltimore, MD
21224
ABSTRACT There are differences between human individuals and between mouse strains in levels of opiate receptor
(OR) expression, responses to painful stimuli, and responses to opiate drugs. One of the best candidates for contributing to these
differences is variation at the OR gene locus. Support for this idea comes from analyses of the human and murine OR genes.
Assessments of individual differences in human OR expression add further support. Studies with mice, including knockout-
transgenic, quantitative trait locus, and strain-comparison studies, also strongly support the possibility that OR gene alleles
would be strong candidates for contributing to individual differences in human nociception and opiate drug responses. This paper
reviews current analyses of the murine and human OR genes, their important variants, and correlations between these variants
and opiate influences on pain.
Opiates remain major weapons in pain therapy, but individual differences in the effectiveness of these drugs and in their side effects
can be a major limitation for effective pain treatment for many patients. A number of lines of evidence now indicate convincingly that the
morphine-preferring opiate receptor (OR) is the major site for the analgesic action of most clinically important opiate drugs.
The powerful analgesic effects of morphine and related drugs focus attention on morphine-preferring ORs and their endogenous and
exogenous agonists. A number of laboratories, including our own, have had success in cloning OR cDNAs and genomic sequences from
several species ( 1 4 ), thereby opening new avenues from which to approach this receptors neurobiology and its relationships with
nociceptive responses. This work has laid substantial groundwork for genetic analyses, although it remains incomplete (see below).
Data from animal models provide powerful motivation to search out and understand possible genetic bases for individual differences
in levels of human OR gene expression. Recent data from transgenic mice provide important information about the role of OR
expression levels in mouse models of human pain ( 4 , 5 ). The data indicate strongly that the OR gene product is the principal route for
opiate effects on nociception. Morphine is not analgesic without ORs. Prototypical and agonists can also function poorly without
ORs (refs. 6 and 7 ; I.S. and G.R.U., unpublished observations).
Several studies of the mice that lack OR provide evidence that ORs are important for baseline nociception (ref. 4 ; see also ref. 5 ).
Nociceptive thresholds vary in gene dose-dependent fashions in such mice. Mice with no ORs have lower nociceptive thresholds than
heterozygous knockouts that have 50% of wild-type receptor densities. These heterozygous mice, in turn, have lower nociceptive
thresholds than wild-type mice with intact ORs.
Mouse-strain comparisons and studies in recombinant inbred mouse lines also provide powerful models for possible sources and
consequences of genetic variation in humans. Strain-comparison studies have identified both reduced antinociceptive responses to
morphine and lower levels of OR expression in some mouse strains, although these are correlations that do not directly document causal
relationships between differences in OR expression levels and observed differences in morphine responses ( 8 10 ). Differences in
morphine responses between DBA and C57 mice as well as the BD recombinant inbred strains derived from these parental lines can be
mapped by using quantitative trait locus approaches ( 11 14 ). Berrettini and coworkers have mapped a significant portion of the genetic
variance in morphine preference to the vicinity of the OR locus by using quantitative-trait-locus approaches ( 13 , 14 ). Belknap et al. (
15 ) have also found that markers near this chromosomal locus correlate with mouse analgesic responses to 16 mg/kg morphine in hot-
plate test assays. This data set derives from a genomic marker somewhat distant from the mouse OR locus, a single analgesic measure,
and a single, relatively high morphine dose. However, the data do fit with those from morphine-preference studies. They also correlate
with maximum bound determinations for [3H]naloxone binding densities in the brains of the same species, performed under binding
conditions that should predominantly label ORs.
Recently, we have identified a murine OR gene 5
flanking-region polymorphism that lies much closer to candidate OR promoter/
enhancer regions (I.S. and G.R.U., unpublished observations). This simple sequence repeat has striking correlations with both levels of
OR expression and the extent of morphine antinociception in the BD recombinant inbred lines (see below). Preliminary analyses
suggest that the allelic status at this marker correlates with the baseline nociceptive thresholds for hot-plate assays in mice of eight tested
strains (see below). Replication of this finding and its extension to more strains and to opiate responsiveness in them could provide
striking evidence that a nearby region has sequence variants that have functional consequences for the level of OR expression and/or its
regulation.
Data from murine studies thus document (i) that ORs may well be key both for normal nociception and for normal opiate drug
responses, (ii) that changes in OR densities of 50%, or even less, can produce differences in both nociceptive re
the authoritative version for attribution.

PNAS is available online at www.pnas.org .

Abbreviation: OR, opiate receptor.
Data deposition: The sequence reported in this paper has been deposited in the GenBank database (accession no. AF153500).
To whom reprint requests should be addressed, e-mail: [email protected] .
 THE OPIATE RECEPTOR AS A CANDIDATE GENE FOR PAIN: POLYMORPHISMS, VARIATIONS IN EXPRESSION, 7753
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NOCICEPTION, AND OPIATE RESPONSES

sponses and in their modulation by opiates, and (iii) that allelic variants at the OR locus are strong candidates for contributing to these
differences in mice and attractive candidates for producing such effects in humans.
Humans differ in their individual responses to pain and to opiate drugs. Recent studies of twins document that individual differences
in several types of pain are likely to have substantial genetic determinants. Genetic components to susceptibility to migraine pain are
documented in studies of thousands of twin pairs, although family studies document substantial genetic heterogeneity in this disorder ( 16
18 ). Studies of concordance for self-reported menstrual pain also identify substantial genetic components ( 19 ). Interestingly, the
heritabilities documented in these human studies (0.5) fit nicely with those identified in murine strain-comparison and quantitative trait
locus studies ( 20 , 21 ).
Humans also differ from one another in OR densities. Binding studies to postmortem brain samples and in vivo positron-emission
tomography radioligand analyses both suggest 3050% or even larger ranges of individual human differences in OR densities. For
example, Pfeiffer et al. ( 22 ) reported that median OR binding in human frontal cortex was 2.3 pmol/g (SD = 0.52). Frost and coworkers
( 23 , 24 ) noted that a measure of thalamic OR binding with [11C]carfentanil was 3.8 pmol/g (SD = 1.4). If individuals in the upper third
of the population are characterized by these data, they should express >45% ( 22 ) or >74% ( 23 , 24 ) more ORs than the individuals in
the lower third of the population. Mouse studies document that genetic differences of this magnitude in OR expression can influence both
baseline nociception and morphine responses, as noted above. Elucidation of the genetic bases for these differences in receptor expression
would thus represent a substantial advance in our understanding of individual differences in nociceptive behaviors and drug responses.
Levels of expression of many, if not most, human genes differ from individual to individual. Many of these differences are thought to
be based on differences in the cis-acting DNA sequences that normally act to provide cell-type-specific, appropriately regulated gene
expression ( 25 ). Many of these DNA promoter and enhancer sequences are typically found in the 5
ends of genes and often serve as
recognition sites for regulatory DNA binding proteins.
Searches for the functional polymorphisms that contribute to these individual differences in gene expression can involve several
steps. Cloning appropriate genomic sequences and characterizing the site(s) for transcriptional initiation so that 5
flanking and other
potential regulatory regions can be determined with confidence represents an important initial step. Identifying polymorphisms in these
regions provides a second series of challenges ( 26 ). Seeking relationships between these polymorphisms and differences in levels of gene
expression is a third step. We can then ask whether the identified polymorphic sequences predict differences not only in levels of OR
expression but also in opiate responses.
The information currently available in GenBank describes 2 kilobases (kb) of murine and 0.2 kb of human OR genomic sequences
5
to the OR translational start site ( Fig. 1 ). Studies of rapid amplification of cDNA 5
ends have suggested to other workers that two
nearby regions provide the sites at which primer extension products terminate, which are thus potential transcriptional initiation sites
(793 and 268 bp from the translational start site; refs. 27 and 28 ). Sequences from each of these two regions can support some
expression of reporter genes in heterologous cell-expression systems. These sequences can even have enhanced expression in the SHY5Y
cells that normally express OR at modest levels.
None of the reported primer extension products have the modified bases characteristic of mRNA capping, however. None of these
sequences provide the 5
untranslated-region length characteristic of most long mRNAs with relatively short coding sequences (see
below). These 268-bp and 793-bp sequences might thus serve as true promoter/enhancer regions. If so, then human polymorphisms in
these regions should be sought out, as only a moderate number have been reported thus far.
Initial searches in our laboratory, as well as more extensive work by Goldman and coworkers ( 29 ) and by L. Yu (personal
communication), have failed to identify common human OR protein coding-sequence variants that dramatically change the receptors
function, although a modest alteration in affinity for the opioid peptide -endorphin has been noted by Yu and coworkers ( 30 ).
These data are in accord with studies that document no convincing individual differences in OR affinities among humans. The data
also fit with the substantial OR coding sequence conservation among species ( 2 , 4 , 22 ). Such information suggests that genetic
components may be unlikely to provide commonly encountered individual differences through functionally different OR protein
sequences. The information contrasts with the abundant data, noted above, documenting frequent individual differences in levels of OR
expression.
Studies by Ko et al. ( 27 ) and by Liang and Carr ( 28 ) indicate that searches for possible promoter-region sequences must include
the sequences located between 268 and 793 bp 5
to the translational initiation sites tentatively identified by these workers. However,
recently, we have also developed interesting results from comparisons of murine and human 5
flanking sequences. These data could also
suggest other sites at which to seek potential promoter-region polymorphisms in humans. Scatterplot comparisons of these species OR
sequences clearly show the area of high cross-species conservation at the 268/793-bp region identified by Ko et al. ( 27 ) and by Liang
and Carr ( 28 ). These analyses also find another highly conserved region that seems to extend from 2,500 to 4,500 bp 5
to the
translational start ( Fig. 2 ). Conceivably,

FIG. 1. Human opioid gene structure. Exons (Exn) are indicated by boxes. The codon (ATG) and stop codon (TAA for OR and
TAG for the less abundant variant XDR1A) are indicated, as are Alu repeats. (GT)n, dinucleotide repeat; (GTT)n, trinucleotide
repeat; SSR, short simple repeats; SNP, sample single nucleotide polymorphisms (54 G/T; 17 C/T; 118 A/G; 440 C/G; 12 C/G; 912
CG/GC); dashed line, sequence to be elucidated.
the authoritative version for attribution.
 THE OPIATE RECEPTOR AS A CANDIDATE GENE FOR PAIN: POLYMORPHISMS, VARIATIONS IN EXPRESSION, 7754
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NOCICEPTION, AND OPIATE RESPONSES

each of these regions could represent unusually highly conserved promoter/enhancer sequences. Polymorphisms from these two conserved
regions would thus be likely candidates for marking functional, level-of-expression allelic OR variants. Alternatively, these highly
conserved regions could also reflect additional exon sequences. Under this scenario, even more 5
sequences would suggest more
prominent candidates for contributing to OR regulation.

FIG. 2. Scatterplot of the nucleotide sequence comparisons between the mouse (x axis) and human (y axis) -OR 5
flanking
sequences, with the translational initiation site at the upper right. Oblique lines represent regions of sequence conservation as
described in the text.
We and others have also identified a number of interesting OR gene polymorphic markers. We identified a human polymorphism:
an MspI restriction fragment length polymorphism ( 2 ). We used PCR to amplify the DNA and have sequenced >1 kb of DNA containing
the 268/793-bp region from 12 unrelated human individuals (volunteers who gave informed consent for studies conducted for the
Intramural Research Program of the National Institute on Drug Abuse in Baltimore; Z.W. and G.R.U., unpublished observations).
Sequence comparisons identify repetitive sequences in these regions that were not polymorphic in initial screens. Two other types of
sequence variation have been identified. More than 20 single nucleotide polymorphisms have been identified in these sequences (see ref.
31 ). We are working to confirm these sequences and will study their frequencies in larger samples to establish their utility for correlations
with receptor-expression densities and opiate-drug responses. Altering 5
untranslated mRNA sequences could readily explain different
levels of OR mRNA stability or even translational efficacy and could contribute to the expression of differing levels of this protein in
different individuals or cell types.
We also have identified a polymorphic repetitive element in >8 kb of murine 5
flanking sequence (I.S. and G.R.U., unpublished
observations). This murine polymorphism lies 2 kb 5
to the translational start site, close to sequences recently identified as candidate OR
promoter/enhancer elements (see below).
Because data from comparisons of more 5
human and mouse OR genomic sequences also suggest the presence of additional exon
(s) and more 5
sites for transcriptional initiation or highly conserved regulatory regions, searches in more 5
genomic regions also make
sense. Workers are currently undertaking approaches consisting of cloning additional 5
genomic sequence, searching for simple sequence
repeat and single nucleotide polymorphisms, characterizing the individual differences in these polymorphic sites, and applying these
polymorphisms to seek correlations with OR expression levels and nociceptive responses. Most genes promoters have much of their
functional anatomy within several thousand base pairs of their transcriptional initiation sites. However, other genes have promoter regions
that extend for >10 or even >20 kb. Analyses of further OR genomic 5
flanking sequences could make great contributions to
understanding this gene and would be quite likely to identify many of its important regulatory elements.
Information about OR gene polymorphisms that can predict the likelihood of high or low levels of expression in an individual
could allow drug treatments to be individualized. These data could aid in selecting analgesic agents and in optimizing dose ranges. They
could thus improve pain management for individuals with acute or long-term pain problems. These data could suggest new therapeutic
specificities and efficacies to even this well established opiate drug class that remains a major weapon for amelioration of pain states.
the authoritative version for attribution.
 THE OPIATE RECEPTOR AS A CANDIDATE GENE FOR PAIN: POLYMORPHISMS, VARIATIONS IN EXPRESSION, 7755
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NOCICEPTION, AND OPIATE RESPONSES

Thus, the extensive work required to identify such markers should be worthwhile.
1. Wang, J. B. , Imai, Y. , Eppler, C. M , Gregor, P. , Spivak, C. E. & Uhl, G. R. ( 1993 ) Proc. Natl. Acad. Sci. USA 90 , 1023010234 .
2. Wang, J. B. , Johnson, P. S. , Persico, A. M. , Hawkins, A. L. , Griffin, C. A. & Uhl, G. R. ( 1994 ) FEBS Lett. 338 , 217222 .
3. Kaufman, D. L. , Keith, D. J. , Anton, B. , Tian, J. , Magendzo, K. , Newman, D. , Tran, T. H. , Lee, D. S. , Wen, C. , Xia, Y. R. , et al. ( 1995 ) J.
Biol. Chem. 270 , 1587715883 .
4. Sora, L , Takahashi, N. , Funada, M. , Ujike, H. , Revay, R. , Donovan, D. , Miner, L. & Uhl, G. R. ( 1997 ) Proc. Natl. Acad. Sci. USA 94 , 1544
1549 .
5. Matthes, H. W. , Maldonado, R. , Simonin, F. , Valverde, O. , Slowe, S. , Kitchen, I. , Befort, K. , Dierich, A. , Le, M. M. , Dolle, P. , et al. ( 1996 )
Nature (London) 383 , 819823 .
6. Sora, I. , Funada, M. & Uhl, G. R. ( 1997 ) Eur. J. Pharmacol. 324 , R1R2 .
7. Sora, I. , Li, X. F. , Funada, M. , Kinsey, S. & Uhl, G. R. ( 1999 ) Eur. J. Pharmacol 366 , R3R5 .
8. Baran, A. , Shuster, L. , Eleftheriou, B. E. & Bailey, D. W. ( 1975 ) Life Sci. 17 , 633640 .
9. Moskowitz, A. S. , Terman, G. W. , Carter, K. R. , Morgan, M. J. & Liebeskind, J. C. ( 1985 ) Brain Res. 361 , 4651 .
10. Vaught, J. L. , Mathiasen, J. R. & Raffa, R. B. ( 1988 ) J. Pharmacol. Exp. Ther. 245 , 1316 .
11. Brase, D. A. , Loh, H. H. & Way, E. L. ( 1977 ) J. Pharmacol. Exp. Ther. 201 , 368374 .
12. Lander, E. S. & Botstein, D. ( 1989 ) Genetics 121 , 185199 .
13. Berrettini, W. H. , Alexander, R. , Ferraro, T. N. & Vogel, W. H. ( 1994 ) Psychiatr. Genet. 4 , 8186 .
14. Berrettini, W. H. , Ferraro, T. N. , Alexander, R. C. , Buchberg, A. M. & Vogel, W. H. ( 1994 ) Nat. Genet. 7 , 5458 .
15. Belknap, J. K. , Mogil, J. S. , Helms, M. L. , Richards, S. P. , O Toole, L. A. , Bergeson, S. E. & Buck, K. J. ( 1995 ) Life Sci. 57 , PL117PL128 .
16. Nyholt, D. R. , Lea, R. A. , Goadsby, P. J. , Brimage, P. J. & Griffiths, L. R. ( 1998 ) Neurology 50 , 14281432 .
17. Peroutka, S. J. ( 1998 ) Clin. Neurosci. 5 , 3437 .
18. Ziegler, D. K. , Hur, Y. M. , Bouchard, T. J. , Hassanein, R. S. & Barter, R. ( 1998 ) Headache 38 , 417422 .
19. Treloar, S. A. , Martin, N. G. & Heath, A. C. ( 1998 ) Behav. Genet. 28 , 107116 .
20. Mogil, J. S. , Kest, B. , Sadowski, B. & Belknap, J. K. ( 1996 ) J. Pharmacol. Exp. Ther. 276 , 532544 .
21. Mogil, J. S. , Richards, S. P. , OToole, L. A. , Helms, M. L. , Mitchell, S. R. , Kest, B. & Belknap, J. K. ( 1997 ) J. Neurosci. 17 , 79958002 .
22. Pfeiffer, A. , Pasi, A. , Mehraein, P. & Herz, A. ( 1982 ) Brain Res. 248 , 8796 .
23. Frost, J. J. , Mayberg, H. S. , Fisher, R. S. , Douglass, K. H. , Dannals, R. F. , Links, J. M. , Wilson, A. A. , Ravert, H. T. , Rosenbaum, A. E. &
Snyder, S. H. ( 1988 ) Ann. Neurol. 23 , 231237 .
24. Frost, J. J. , Douglass, K. H. , Mayberg, H. S. , Dannals, R. F. , Links, J. M. , Wilson, A. A. , Ravert, H. T. , Crozier, W. C. & Wagner, H. J. ( 1989 )
J. Cereb. Blood Flow Metab. 9 , 398409 .
25. Uhl, G. R. , Gold, L. H. & Risch, N. ( 1997 ) Proc. Natl. Acad. Sci. USA 94 , 27852786 .
26. Collins, F. S. , Guyer, M. S. & Charkravarti, A. ( 1997 ) Science 278 , 15801581 .
27. Ko, J. L. , Minnerath, S. R. & Loh, H. H. ( 1997 ) Biochem. Biophys. Res. Commun. 234 , 351357 .
28. Liang, Y. & Carr, L. G. ( 1997 ) Brain Res. 769 , 372374 .
29. Bergen, A. W. , Kokoszka, J. , Peterson, R. , Long, J. C. , Virkkunen, M. , Linnoila, M. & Goldman, D. ( 1997 ) Mol. Psychiatry 2 , 490494 .
30. Bond, C , LaForge, K. S. , Tian, M. , Melia, D. , Zhang, S. , Borg, L. , Gong, J. , Schluger, J. , Strong, J. A. , Leal, S. M. , et al. ( 1998 ) Proc. Natl.
Acad. Sci. USA 95 , 96089613 .
31. Wendel, B. & Hoehe, M. R. ( 1998 ) J. Mol. Med. 76 , 525532 .
the authoritative version for attribution.
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
the authoritative version for attribution.
NOCICEPTION, AND OPIATE RESPONSES
THE OPIATE RECEPTOR AS A CANDIDATE GENE FOR PAIN: POLYMORPHISMS, VARIATIONS IN EXPRESSION,
7756
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NAS COLLOQUIA 7757

NAS COLLOQUIA

SCHEDULED
Virulence and Defense in Host-Pathogen Interactions: Common Features between Plants and Animals
December 9-11, 1999
Organizers: Jim Cook, Noel T. Keen,
Brian J. Staskawicz, John J. Meckalanos, and Frederick M. Ausubel
Incentives To Promote Human Capital
December 17-19, 1999; Irvine, California
Organizer: James Heckman
Variation and Evolution in Plants and Microorganisms; Fifty Years after Stebbins
January 27-29, 2000; Irvine, California
Organizers: Francisco Ayala and Walter Fitch
The Biotic Crisis and the Future of Evolution
March 16-19, 2000; Irvine, California
Organizers: Norman Myers, Andrew Knoll
Auditory Neuroscience: Development, Transduction, and Integration
May 19-21, 2000; Irvine, California
Organizers: James Hudspeth and Mark Konishi
Links between Recombination and Replication
November 10-12, 2000; Irvine, California
Organizer: Charles Radding
COMPLETED
Industrial Ecology
May 20-21, 1991; Washington, D.C.
Organizer: C. Kumar N. Patel
Proceedings: February 4, 1992
Images of Science: Science of Images
January 13-14, 1992; Washington, D.C.
Organizer: Albert Crewe
Proceedings: November 3, 1993
Physical Cosmology
March 27-29, 1992; Irvine, California
Organizer: David Schramm
Proceedings: June 3, 1993
Molecular Recognition
September 10-11, 1992; Washington, D.C.
Organizer: Ronald Breslow
Proceedings: February 16, 1993
Human-Machine Communication by Voice
February 8-9, 1993; Irvine, California
Organizer: Lawrence Rabiner
Proceedings: October 24, 1995
Voice Communication Between Humans and Machines:
August 1994
Changing Human Ecology and Behavior: Effects on Infectious Diseases
September 27-28, 1993; Washington, D.C.
Organizer: Bernard Roizman
Proceedings: March 29, 1994
Infectious Diseases in an Age of Change: January 1995
The Tempo and Mode of Evolution
January 27-29, 1994; Irvine, California
Organizers: Francisco Ayala, Walter Fitch
Proceedings: July 19, 1994
Tempo and Mode in Evolution: January 1995
Chemical Ecology: The Chemistry of Biotic Interaction
March 25-26, 1994; Washington, D.C.
Organizers: Thomas Eisner, Jerrold Meinwald
the authoritative version for attribution.

Proceedings: January 3, 1995

Physics: The Opening to Complexity
June 25-27, 1994; Irvine, California
Organizer: Philip Anderson
Proceedings: July 18, 1995
Self Defense by Plants: Induction and Signaling Pathways
September 15-17, 1994; Irvine, California
Organizers: Andr Jagendorf, Clarence Ryan
Proceedings: May 9, 1995
Earthquake Prediction
February 10-11, 1995; Irvine, California
Organizer: Leon Knopoff
Proceedings: April 30, 1996
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NAS COLLOQUIA 7758

Quasars and Active Galaxies: High Resolution

Radio Imaging
March 24-25, 1995; Irvine, California
Organizers: Marshall Cohen, Kenneth Kellerman
Proceedings: December 5, 1995
Vision: From Photon to Perception
May 21-22, 1995; Irvine, California
Organizers: John Dowling, Lubert Stryer, and Torsten Wiesel
Proceedings: January 23, 1996
Science, Technology, and the Economy
October 20-22, 1995; Irvine, California
Organizers: James Heckman, Ariel Pakes, and Kenneth Sokoloff
Proceedings: November 12, 1996
Developmental Biology of Transcription Control
October 25-28, 1995; Irvine, California
Organizers: Roy Britten, Eric Davidson, and Gary Felsenfeld
Proceedings: September 3, 1996
Carbon Dioxide and Climate Change
November 13-15, 1995; Irvine, California
Organizer: Charles Keeling
Proceedings: August 5, 1997
Memory: Recording Experience in Cells and Circuits
February 17-20, 1996; Irvine, California
Organizer: Patricia Goldman-Rakic
Proceedings: November 26, 1996
Elliptic Curves and Modular Forms
March 15-17, 1996; Washington, D.C.
Organizers: Barry Mazur, Karl Rubin
Proceedings: October 14, 1997
Symmetries Throughout the Sciences
May 10-12, 1996; Irvine, California
Organizer: Ernest Henley
Proceedings: December 15, 1996
Genetic Engineering of Viruses and Viral Vectors
June 9-11, 1996; Irvine, California
Organizers: Peter Palese, Bernard Roizman
Proceedings: October 15, 1996
Genetics and the Origin of Species
January 30-February 1, 1997; Irvine, California
Organizers: Francisco Ayala, Walter Fitch
Proceedings: July 22, 1997
The Age of the Universe: Dark Matter and Structure Formation
March 21-23, 1997; Irvine, California
Organizers: David Schramm, P. J. E. Peebles
Proceedings: January 6, 1998
Neuroimaging and Human Brain Function
May 29-31, 1997; Irvine, California
Organizers: Michael Posner, Marcus Raichle
Proceedings: February 3, 1998
Protecting Our Food Supply: The Value of Plant Genome Initiatives
June 2-4, 1997; Irvine, California
Organizers: Michael Freeling, Ronald Phillips, and John Axtell
Proceedings: March 5, 1998
Computational Biomolecular Science
September 11-14, 1997; Irvine, California
Organizers: Peter G. Wolynes, Russell Doolittle,
J. A. McCammon
Proceedings: May 26, 1998
A Library Approach to Chemistry
October 19-21, 1997; Irvine, California
Organizer: Peter Schultz, Jonathan Ellman
Geology, Mineralogy, and Human Welfare
November 7-9, 1998; Irvine, California
Organizers: Joseph Smith, Malcolm Ross, Peter Buseck
Proceedings: March 30, 1999
Plants and Population: is there time?
the authoritative version for attribution.

December 5-6, 1998; Irvine, California

Organizers: Nina Fedoroff, Joel Cohen
Proceedings: May 25, 1999
The Neurobiology of Pain
December 11-13, 1998; Irvine, California
Organizer: Ronald Dubner, Michael Gold
Proceedings: July 6, 1999
Non-linear Differential Equations and Computation
January 4-8, 1999; Irvine, California
Organizer: Haim Brezis, Felix Browder,
Louis Nirenberg, James Serrin
Proteolytic Processing and Physiological Regulation
February 20-21, 1999; Irvine, California
Organizer: Hans Neurath, Charles Craik
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

NAS COLLOQUIA 7759

NATIONAL ACADEMY OF SCIENCES COLLOQUIA

BOUND REPRINTS NOW AVAILABLE
In 1991, the National Academy of Sciences (NAS) inaugurated a series of scientific colloquia, several of which are
held each year under the auspices of the NAS Council Committee on Scientific Programs. These colloquia address
scientific topics of broad and topical interest that cut across two or more traditional disciplines. Typically two days long,
these colloquia are international in scope and bring together leading scientists in their field.
Papers presented at these colloquia are published in the Proceedings of the National Academy of Sciences
(PNAS) and are available on-line ( www.pnas.org ). Because they have generated much interest, these papers now will
be available in the form of collected bound reprints, which may be ordered through the National Academy Press.
Currently available are:
Carbon Dioxide and Climate Change ($11.00)
Held November 13-15, 1995 (Irvine, CA)
Computational Biomolecular Science ($16.00)
Held September 12-13, 1997 (Irvine, CA)
Earthquake Prediction ($16.00)
Held February 10-11, 1995 (Irvine, CA)
Elliptic Curves and Modular Forms ($7.00)
Held March 15-17, 1996 (Washington, DC)
Genetic Engineering of Viruses and Viral Vectors ($21.00)
Held June 9 -11, 1996 (Irvine, CA)
Genetics and the Origin of Species ($8.00)
Held January 31-February 1, 1997 (Irvine, CA)
New! Geology, Mineralogy and Human Welfare ($11.00)
Held November 8-9, 1998 (Irvine, CA)
New! Neurobiology of Pain ($8.00)
Held December 11-13, 1998 (Irvine, CA)
Neuroimaging of Human Brain Function ($17.00)
Held May 29-31, 1997 (Irvine, CA)
New! Plants and Population: Is There Time? ($8.00)
Held December 5-6, 1998 (Irvine, CA)
Protecting Our Food Supply: The Value of Plant Genome Initiatives ($13.00)
Held May 29-31, 1997 (Irvine, CA)
Science, Technology, and the Economy ($12.00)
Held November 20-22, 1995 (Irvine, CA)
The Age of the Universe, Dark Matter, and Structure Formation ($13.00)
Held March 21-23, 1997 (Irvine, CA)
Papers from future colloquia will be available for purchase after they appear in PNAS.
Shipping and Handling Charges
In the U.S. and Canada please add $4.50 for the first reprint ordered and $0.95 for each additional reprint.
Ordering Information
Telephone orders will be accepted only when charged to VISA, Mastercard, or American Express accounts.
TO ORDER CALL TOLL-FREE 1-800-624-6242
OR ORDER ON-LINE (
WWW.NAP.EDU
) AND RECEIVE A 20% DISCOUNT
the authoritative version for attribution.
lengths, word breaks, heading styles, and other typesetting-specific formatting, however, cannot be retained, and some typographic errors may have been accidentally inserted. Please use the print version of this publication as
About this PDF file: This new digital representation of the original work has been recomposed from XML files created from the original paper book, not from the original typesetting files. Page breaks are true to the original; line

1999 ORDER FORM 7760

PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA

1999 ORDER FORM

Visit PNAS Online at www.pnas.org .

For 1999 the print and the online editions are included in one price!
ANNUAL PRINT AND ONLINE SUBSCRIPTION RATES (Calendar Year 1999)
USA Online Only Print Plus Online
Institutional n/a $800
Individual * $100 $165
Student/Postdoctoral * $20 $110
ELSEWHERE Online Only Print Plus Online
Institutional n/a $995 (includes expedited delivery)
$290
Individual * $100 $235
Student/Postdoctoral * $20
Expedited Delivery Rates for print are included in the institutional foreign prices. The individual foreign expedited rate is $300.
The U.S. first class rate is $160.
* Individual, Student, and Postdoctoral subscriptions are for personal use and are not to be used in a library. To qualify for these
rates, you must pay by personal check, money order, or credit card, and state that your subscription is for personal use only. To qualify for
the Student or Postdoctoral rate, your order must also be accompanied by a letter from your advisor or supervisor, on institutional
letterhead, stating that you are enrolled in a program of study leading to a degree or that you are within four years of receiving a doctorate.
Tax information: Canadian subscribers add 7% GST (Registration Number R-133130880). DC and Florida subscribers add
appropriate sales tax: in DC add 5.75% sales tax, in Florida add 6% sales tax and local surtax, if you are in a taxing county. Tax exempt
institutions must include a copy of the tax exemption certificate with orders. Subscriptions paid at the individual rate are taxable.
Subscribers in Japan: Please send all orders to our exclusive agent, USACO Corporation, 13-12, Shimbashi, 1-Chome, Minato-ku,
Tokyo 105 Japan. Phone 81-3-3502-6471, Fax 81-3-3508-0770, e-mail [email protected]
Individual Issues: Available from Volume 88 (January 1991) and onward. Cost per issue: USA $35, elsewhere $45. No discount
allowed. Volume 92, Numbers 16, 21, 24 and 26, and issues published prior to Volume 88 are out of print.
Subscription Type: New Subscription Renewal
Amount Remitted (Required) Subscription Price $_______________
Expedited Delivery (optional) $_______________
Sales Tax (if applicable) $_______________
TOTAL $_______________
Check or money order enclosed payable to the Proceedings of the National Academy of Sciences. Foreign remittance must be by
draft on a U.S. Bank in U.S. dollars or by International Money Order.

Charge to: MasterCard VISA American Expiration

Express Date____________ _
Card Number___________________
Cardholder Authorized
Name___________________ Signature____________________
Shipping Information (please print or
type):
Name ______________________ Phone/Fax
_______________________
Address E-mail ____________________
_________________________________
_________________________________
_________________________________
Send All Orders and Correspondence to:
Proceedings of the National Academy of Sciences
PO Box 631008
Baltimore, MD 21263-1008 USA
If you have any questions, you may call 877-314-2253 or 615-377-3322, fax 615-377-0525, or E-mail [email protected] .
the authoritative version for attribution.