Biological Concepts and Techniques in Toxicology

An Integrated Approach 1st Edition E. Riviere

Jim pdf download

https://ebookname.com/product/biological-concepts-and-techniques-
in-toxicology-an-integrated-approach-1st-edition-e-riviere-jim/

Get the full ebook with Bonus Features for a Better Reading Experience on ebookname.com
Instant digital products (PDF, ePub, MOBI) available
Download now and explore formats that suit you...

Dermal Absorption Models in Toxicology and Pharmacology

1st Edition Jim E. Riviere

https://ebookname.com/product/dermal-absorption-models-in-
toxicology-and-pharmacology-1st-edition-jim-e-riviere/

Mechanical Design An Integrated Approach 1st Edition

Ansel Ugural

https://ebookname.com/product/mechanical-design-an-integrated-
approach-1st-edition-ansel-ugural/

Chemical Structure and Reactivity An Integrated

Approach 1st Edition James Keeler

https://ebookname.com/product/chemical-structure-and-reactivity-
an-integrated-approach-1st-edition-james-keeler/

Drug Metabolism Chemical and Enzymatic Aspects 1st

Edition Jack P. Uetrecht

https://ebookname.com/product/drug-metabolism-chemical-and-
enzymatic-aspects-1st-edition-jack-p-uetrecht/
Political Women in Japan The Search for a Place in
Political Life Susan J. Pharr

https://ebookname.com/product/political-women-in-japan-the-
search-for-a-place-in-political-life-susan-j-pharr/

Devils in Daylight Tanizaki

https://ebookname.com/product/devils-in-daylight-tanizaki/

Iberian Books Libros ibéricos IB Alexander S. Wilkinson

https://ebookname.com/product/iberian-books-libros-ibericos-ib-
alexander-s-wilkinson/

Structural Analysis Third Edition with CD ROM Aslam

Kassimali

https://ebookname.com/product/structural-analysis-third-edition-
with-cd-rom-aslam-kassimali/

The Dynamics of Persuasion Communication and Attitudes

in the 21st Century 2nd Edition Richard M. Perloff

https://ebookname.com/product/the-dynamics-of-persuasion-
communication-and-attitudes-in-the-21st-century-2nd-edition-
richard-m-perloff/
The Light and the Shadow How Breakthrough Innovation is
Shaping European Business 1st Edition Otto Kalthoff

https://ebookname.com/product/the-light-and-the-shadow-how-
breakthrough-innovation-is-shaping-european-business-1st-edition-
otto-kalthoff/
 Biological
Concepts and
Techniques in
Toxicology

DK4995_FM.indd 1 1/5/06 2:01:50 PM

Process CyanProcess
CyanProcess MagentaProcess
MagentaProcess Yellow
YellowProcess
Process Black
 Biological
Concepts and
Techniques in
Toxicology
An Integrated Approach

edited by
Jim E. Riviere
North Carolina State University
Raleigh, North Carolina, U.S.A.

DK4995_FM.indd 2 1/5/06 2:01:50 PM

Process CyanProcess
CyanProcess MagentaProcess
MagentaProcess Yellow
YellowProcess
Process Black
DK4995_Discl.fm Page 1 Wednesday, November 30, 2005 10:12 AM

Published in 2006 by
Taylor & Francis Group
270 Madison Avenue
New York, NY 10016
© 2006 by Taylor & Francis Group, LLC

No claim to original U.S. Government works

Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4 3 2 1

International Standard Book Number-10: 0-8247-2979-X (Hardcover)

International Standard Book Number-13: 978-0-8247-2979-0 (Hardcover)
Library of Congress Card Number 2005056861

This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with
permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish
reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials
or for the consequences of their use.

No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or
other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information
storage or retrieval system, without written permission from the publishers.

For permission to photocopy or use material electronically from this work, please access www.copyright.com
(http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive, Danvers, MA
01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For
organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

Biological concepts and techniques in toxicology : an integrated approach / edited by Jim E. Riviere.
p. ; cm.
Includes bibliographical references and index.
ISBN-13: 978-0-8247-2979-0 (hardcover : alk. paper)
ISBN-10: 0-8247-2979-X (hardcover: alk. paper)
1. Toxicology. 2. Molecular toxicology. 3. Drugs--Toxicology. I. Riviere, J. Edmond (Jim Edmond)
[DNLM: 1. Toxicology--methods. 2. Drug Toxicity. QV 600 B6146 2006]

RA1191.B52 2006
615.9--dc22 2005056861

Visit the Taylor & Francis Web site at

Taylor & Francis Group
http://www.taylorandfrancis.com
is the Academic Division of Informa plc.
Preface

Toxicology is a rapidly developing discipline whose advances are often based on uti-
lizing concepts and techniques developed from basic biomedical research. Earlier
revolutions in molecular biology and mathematical modeling have led to routine
use of in vitro models and toxicokinetics developed from these disciplines. In order
for toxicology to remain current and continue making significant advances, it is
important that the latest developments be presented in a context focused on how
they can be applied to different fields of investigation. How do concepts and new
paradigms migrate into routine practice? Several recent toxicology issues have
engendered attention and are deserving of a review that illustrates how new techni-
ques have been applied to actual problems of toxicological interest. How did these
migrate from the basic science laboratory to routine toxicology test protocols?
How will today’s revolution in the ‘‘-omics’’ affect the practice of toxicology in
the next decade? This book is meant to provide a timely introduction to these tech-
niques focused on how the latest science can be applied to existing problems in tox-
icology. It is also meant to overview some recent areas where progress has been made
and a story can begin to be told.
The initial chapters review new approaches or concepts that will have a major
impact on the practice of toxicology in the coming decade. These include the ‘‘omics’’
which together comprise systems biology as well as mathematical approaches
designed to link chemical structure to activity. This section ends with a cogent pre-
sentation of hormesis, a concept that has begun to alter the way toxicologists and
others interpret dose–response relationships, the cornerstone to the practice of clas-
sical toxicology. The four chapters comprising the next section of the book deal with
how biological data and test systems are integrated into the actual practice of
toxicology. What is the validation process that allows a novel idea to become a stan-
dard test? The final section of chapters reviews applications of toxicology to specific
and more focused topics that represent current issues facing society today, including
discussions of genetically modified food, nanomaterials, and pharmaceutics. Specific
disciplines, including inhalational and forensic toxicology are discussed, as are
current military issues that continue to draw attention.

iii
iv Preface

The goal of this book is not to reproduce comprehensive toxicology texts cov-
ering basic mechanisms of toxicity or target organ toxicology; nor is it a methods
text overviewing the nuts and bolts of new techniques. The goal is to highlight
new methods and concepts that may have a major impact on toxicology and present
how such concepts and techniques migrate into the mainstream of toxicology.

Jim E. Riviere
Raleigh, North Carolina
Contents

Preface . . . . iii
Contributors . . . . xi

1. Introduction and Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Jim E. Riviere

PART I. NEW CONCEPTS AND APPROACHES

2. Toxicogenomics: Gene Expression Analysis and

Computational Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Rupesh P. Amin, Hisham K. Hamadeh, J. Todd Auman, Lee Bennett,
Cynthia A. Afshari, Pierre R. Bushel, Richard S. Paules, and
J. Christopher Corton
Introduction . . . . 5
Microarray-Based Gene Expression Profiling . . . . 6
Selection of Differentially Expressed Genes . . . . 9
Finding Expression Patterns . . . . 11
Data Management . . . . 13
Biological Databases Useful for Toxicogenomics . . . . 14
Examples from an In Vivo Toxicogenomics Study . . . . 14
Future Prospects for Toxicogenomics . . . . 19
References . . . . 21

3. Proteomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Frank A. Witzmann
Introduction . . . . 25
Differential Expression Proteomic Technologies . . . . 27
A 2-D Gel-Based Toxicoproteomic Analysis of
Jet Fuel Exposure . . . . 39
Conclusion . . . . 41
References . . . . 42

v
vi Contents

4. The Potential of Metabonomics in Toxicology . . . . . . . . . . . . . . . 47

Julian L. Griffin
Introduction . . . . 47
Metabonomics and Genomics . . . . 49
Metabonomics and Disease . . . . 51
Metabonomics and Transcriptomics . . . . 51
MS and Metabonomics . . . . 55
Metabonomics and Systems Biology . . . . 56
Future Developments . . . . 56
Conclusions . . . . 58
References . . . . 59

5. Quantitative Structure–Toxicity Relationships Using

Chemodescriptors and Biodescriptors . . . . . . . . . . . . . . . . . . . . . . 61
Subhash C. Basak, Denise Mills, and Brian D. Gute
Introduction . . . . 61
The Molecular Structure Conundrum . . . . 62
The HiQSAR Approach . . . . 63
Statistical Methods . . . . 69
Applications in Toxicokinetics and Toxicodynamics . . . . 70
Cell Level Toxicity Estimation . . . . 71
Molecular Similarity and Predictive Toxicology . . . . 75
Discussion . . . . 76
References . . . . 79

6. Hormesis: A Key Concept in Toxicology . . . . . . . . . . . . . . . . . . . . 83

Edward J. Calabrese
What Does This Mean Today? . . . . 90
References . . . . 90

PART II. INTEGRATION INTO PRACTICE

7. Chemical Risk Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

Ronald E. Baynes and Jennifer L. Buur
Introduction . . . . 93
Hazard Identification . . . . 94
WOE . . . . 96
Dose–Response Assessment . . . . 98
Quantifying Risk for Noncarcinogenic
Effects: Hazard Quotient . . . . 103
Chemical Mixtures . . . . 104
Dermal RA . . . . 107
PBPK Modeling . . . . 108
Conclusion . . . . 114
References . . . . 114
Contents vii

8. Toxicokinetics: Fundamentals and Applications in

Drug Development and Safety Assessment . . . . . . . . . . . . . . . . . 117
Rakesh Dixit
Introduction . . . . 117
Pharmacokinetics and Toxicokinetics . . . . 117
Conclusions . . . . 156
References . . . . 157

9. Validation of In Vitro Methods for Toxicology Studies . . . . . . . . 161

William S. Stokes
Introduction . . . . 161
The Concept of Test Method Validation . . . . 162
Evolution Process for Test Methods . . . . 162
Validation and Regulatory Acceptance Criteria . . . . 164
The Validation Process . . . . 165
ICCVAM Role in Validation and Regulatory
Acceptance . . . . 170
Summary . . . . 173
References . . . . 174

10. Chemical Mixtures Risk Assessment and

Technological Advances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
M. M. Mumtaz, B. A. Fowler, C. T. De Rosa, P. Ruiz, M. Whittaker
and J. Dennison
Disease Conditions Caused by Chemical Mixtures . . . . 177
Assessment Approaches . . . . 184
New Developments . . . . 192
Research Advancements in Joint Toxicity of Metals . . . . 195
Perspectives and Future Needs . . . . 197
References . . . . 199

PART III. EXAMPLES APPLIED TO PROBLEMS

11. Risk-Based Approach to Foods Derived from Genetically

Engineered Animals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Larisa Rudenko, Kevin J. Greenlees, and John C. Matheson III
Introduction: The Need to Move Beyond
Traditional Dose-Exaggeration Studies . . . . 205
Underlying Risk Basis and Paradigm Setting . . . . 208
Toward a Risk-Based Approach . . . . 209
Hazard Analysis for GE Animals Traditionally
Consumed for Food . . . . 210
Looking Forward . . . . 213
References . . . . 214
viii Contents

12. Toxicology of Nanomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . 217

Nancy A. Monteiro-Riviere and Jessica P. Ryman-Rasmussen
Introduction . . . . 217
What Are Nanomaterials? . . . . 217
Nanomaterial Nomenclature . . . . 218
Characterization . . . . 219
Nanomaterials and Toxicity . . . . 219
Exposure and Risk Assessment . . . . 228
Environmental Impact . . . . 229
Conclusion . . . . 230
References . . . . 230

13. The Biological Basis of Experimental Toxicity of

Jet Propulsion Fuel–8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
Simon S. Wong and Mark L. Witten
Introduction . . . . 235
Pulmonary Toxicity . . . . 236
Dermal Toxicity . . . . 241
Immune Toxicity . . . . 245
References . . . . 247

14. Drug Safety Toxicology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

Joseph P. Hanig and Robert E. Osterberg
Introduction . . . . 249
Background of Drug Safety Regulation . . . . 249
U.S. Food and Drug Administration Toxicity
Testing Documents . . . . 250
Modern Approaches to the Development of Methods
for Drug Testing and Safety Evaluation . . . . 251
History of Regulatory Assay Review and/or
Acceptance . . . . 257
The Impact of FDA’s ‘‘Critical Path’’ Concept on
Future Development of Drug Safety Methodology . . . . 265
Crucial Issues and Challenges . . . . 266
Conclusions . . . . 267
References . . . . 268

15. Statins and Toxicity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273

John Farmer
Introduction . . . . 273
Pathogenesis of Atherosclerosis and the Potential of
Statin Toxicity . . . . 274
Statin Toxicity . . . . 277
References . . . . 292
Contents ix

16. Inhalation Toxicology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297

Roger O. McClellan, Michele A. Medinsky, and
M. Burton Snipes
Introduction . . . . 297
Planning Inhalation Toxicology Research . . . . 299
Characteristics of Toxicants and Targets . . . . 302
Fate of Inhaled Materials . . . . 309
Exposure Systems, Generation of Atmosphere, and
Characterization of Atmosphere . . . . 321
Study Design . . . . 327
Respiratory Tract Responses . . . . 330
Assessing Human Risks of Inhaled Particles . . . . 334
Quality Control and Quality Assurance . . . . 355
Additional References . . . . 355
Summary . . . . 356
References . . . . 357

17. Modern Gas Chromatography–Mass Spectrometry in Human

Forensic Toxicology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Beat Aebi and Werner Bernhard
Introduction . . . . 365
Types of GC–MS and Experimental Setups . . . . 366
Gas Chromatography . . . . 366
Mass Spectrometry with Quadrupole Analyzers . . . . 367
Sector Field Instruments, Tandem and Hybrid Mass
Spectrometers . . . . 368
TOF-MS . . . . 369
Screening Analyses with GC–MS . . . . 370
Alternative Matrices and Samples . . . . 372
Quality of Analytical Data . . . . 372
Limitations of GC–MS, Traps and Pitfalls . . . . 373
Summary and Outlook, Expanding LC–MS,
New Technologies . . . . 377
References . . . . 378

Index . . . . 387
Contributors

Beat Aebi Institute of Legal Medicine, University of Berne, Berne, Switzerland

Cynthia A. Afshari National Institute of Environmental Health Sciences,

Research Triangle Park, North Carolina, U.S.A.

Rupesh P. Amin National Institute of Environmental Health Sciences,

Research Triangle Park, North Carolina, U.S.A.

J. Todd Auman National Institute of Environmental Health Sciences,

Research Triangle Park, North Carolina, U.S.A.

Subhash C. Basak Natural Resources Research Institute, University of Minnesota

Duluth, Duluth, Minnesota, U.S.A.

Ronald E. Baynes Center for Chemical Toxicology Research and

Pharmacokinetics, College of Veterinary Medicine, North Carolina State University,
Raleigh, North Carolina, U.S.A.

Lee Bennett National Institute of Environmental Health Sciences,

Research Triangle Park, North Carolina, U.S.A.

Werner Bernhard Institute of Legal Medicine, University of Berne,

Berne, Switzerland

Pierre R. Bushel National Institute of Environmental Health Sciences,

Research Triangle Park, North Carolina, U.S.A.

Jennifer L. Buur Center for Chemical Toxicology Research and Pharmacokinetics,

College of Veterinary Medicine, North Carolina State University, Raleigh,
North Carolina, U.S.A.

Edward J. Calabrese Environmental Health Sciences, University of Massachusetts,

Amherst, Massachusetts, U.S.A.

J. Christopher Corton U.S. Environmental Protection Agency,

Research Triangle Park, North Carolina, U.S.A.

J. Dennison Colorado State University, Fort Collins, Colorado, U.S.A.

xi
xii Contributors

C. T. De Rosa Division of Toxicology and Environmental Medicine, Agency for

Toxic Substances and Disease Registry, Atlanta, Georgia, U.S.A.

Rakesh Dixit Toxicology Department, Johnson and Johnson Pharmaceutical

Research and Development, L.L.C., San Diego, California, U.S.A.

John Farmer Section of Cardiology, Baylor College of Medicine,

Houston, Texas, U.S.A.

B. A. Fowler Division of Toxicology and Environmental Medicine, Agency for

Toxic Substances and Disease Registry, Atlanta, Georgia, U.S.A.

Kevin J. Greenlees USFDA Center for Veterinary Medicine,

Rockville, Maryland, U.S.A.

Julian L. Griffin Department of Biochemistry, University of Cambridge,

Cambridge, U.K.

Brian D. Gute Natural Resources Research Institute, University of Minnesota

Duluth, Duluth, Minnesota, U.S.A.

Hisham K. Hamadeh National Institute of Environmental Health Sciences,

Research Triangle Park, North Carolina, U.S.A.

Joseph P. Hanig Food and Drug Administration, Center for Drug Evaluation and
Research, Silver Spring, Maryland, U.S.A.

John C. Matheson III USFDA Center for Veterinary Medicine, Rockville,

Maryland, U.S.A.

Roger O. McClellan Inhalation Toxicology and Human Health Risk Analysis,

Albuquerque, New Mexico, U.S.A.

Michele A. Medinsky Toxicon, Durham, North Carolina, U.S.A.

Denise Mills Natural Resources Research Institute, University of Minnesota

Duluth, Duluth, Minnesota, U.S.A.

Nancy A. Monteiro-Riviere Center for Chemical Toxicology Research

and Pharmacokinetics, College of Veterinary Medicine, North Carolina
State University, Raleigh, North Carolina, U.S.A.

M. M. Mumtaz Division of Toxicology and Environmental Medicine, Agency for

Toxic Substances and Disease Registry, Atlanta, Georgia, U.S.A.

Robert E. Osterberg Food and Drug Administration, Center for Drug Evaluation
and Research, Silver Spring, Maryland, U.S.A.

Richard S. Paules National Institute of Environmental Health Sciences, Research

Triangle Park, North Carolina, U.S.A.
Contributors xiii

Jim E. Riviere Center for Chemical Toxicology Research and Pharmacokinetics,

Biomathematics Program, North Carolina State University, Raleigh,
North Carolina, U.S.A.

Larisa Rudenko USFDA Center for Veterinary Medicine, Rockville,

Maryland, U.S.A.

P. Ruiz Division of Toxicology and Environmental Medicine, Agency for Toxic

Substances and Disease Registry, Atlanta, Georgia, and Oak Ridge Institute for
Science and Education, Oak Ridge, Tennessee, U.S.A.

Jessica P. Ryman-Rasmussen Center for Chemical Toxicology Research and

Pharmacokinetics, College of Veterinary Medicine, North Carolina State University,
Raleigh, North Carolina, U.S.A.

M. Burton Snipes Inhalation Toxicology Research Institute, Tijeras,

New Mexico, U.S.A.

William S. Stokes National Toxicology Program Interagency Center for the

Evaluation of Alternative Toxicological Methods, National Institute of
Environmental Health Sciences, National Institutes of Health, Department of
Health and Human Services, Research Triangle Park, North Carolina, U.S.A.

M. Whittaker ToxServices, Washington, D.C., U.S.A.

Mark L. Witten Department of Pediatrics, University of Arizona Health Sciences

Center, Tucson, Arizona, U.S.A.

Frank A. Witzmann Department of Cellular and Integrative Physiology,

Biotechnology Research and Training Center, Indiana University School of
Medicine, Indianapolis, Indiana, U.S.A.

Simon S. Wong Department of Pediatrics, University of Arizona Health Sciences

Center, Tucson, Arizona, U.S.A.
1
Introduction and Overview

Jim E. Riviere
Center for Chemical Toxicology Research and Pharmacokinetics,
Biomathematics Program, North Carolina State University, Raleigh,
North Carolina, U.S.A.

Toxicology is the science of poisons and dates back to the earliest annals of recorded
history. This history is steeped in both using and developing animal- and plant-
derived poisons, as well as treating the victims who succumbed to these practices.
Early practitioners were also integrated into the developing arts of medicine and
pharmacy where chemicals, mostly isolated from plant sources, began to be used
to treat human and animal diseases. In the early 15th century, the basic tenet of
the dose–response relationship was articulated by the alchemist–physician Paracelsus
when he stated ‘‘All substances are poisons; there is none which is not a poison. The
proper dose separates a poison from a remedy.’’ To the present, this rubric has
guided the practice of toxicology in most of its varied applications.
With the explosion of chemistry in the 19th and 20th centuries, toxicology
likewise blossomed. A similar phase of growth is now upon us. It started mid-
century with the advent of molecular biology, which was facilitated by the revolu-
tion in computer technology, and is now propelled by the growth of genomics that we
are presently experiencing. Toxicology has come of age. The public views toxicol-
ogy not from these expanding scientific roots, but rather from its application to
practical problems. In recent decades toxicology has become synonymous with
the science focused on assessing the safety of new drugs and determining the risk
to environmental or occupational chemical exposure. The working definition of
toxicology is generally defined as the study of adverse effects of chemicals on bio-
logical systems.
Toxicology has been viewed by many as an orphan discipline, because it relies
so heavily on fundamental advances in chemistry, biology, and mathematics. How-
ever, this view is oversimplified because modern toxicology has evolved into a
number of niches where elucidating the basic mechanisms of chemical interactions
with living systems is only conducted by scientists who can best be classified as
toxicologists. These include studies of the mechanism of cancer, chemical biotrans-
formation, inhalational and dermal exposure analysis, as well as defining the biolo-
gical response to specific chemical classes including metals and pesticides.
Subdisciplines focused on natural product toxicology, drug safety, forensics, as
well as occupational and environmental risk assessment have emerged. The use of

1
2 Riviere

physiologically based pharmacokinetic models, once applied to pharmaceutics, has

now been largely developed and put to practice in the field of toxicology.
Toxicology is thus a vibrant discipline in which novel techniques of other
sciences (e.g., analytical chemistry, molecular biology, computer-aided modeling,
and genomics/proteomics/metabonomics) are explored and then become main-
stream tools in its practice. The nature of this migration is the focus of this present
text. How does a basic technique or insight into mechanism of action move from a
science-discipline laboratory to the laboratory of a toxicologist exploring chemical
effects to an approved tool in the risk assessor’s arsenal? When do these migrations
work and when do they not? How are products of one discipline (e.g., genomics or
analytical chemistry) interfaced to another (e.g., drug-induced alteration of genetic
function, coupling of low-level chemical detection to a biological effect, etc.)?
Three different approaches to exploring this phenomenon are taken in this text.
The first is to closely examine new disciplines from within to see where potential
applications to toxicology may occur. These include detailed presentations in geno-
mics, proteomics, and metabonomics. Exciting developments are occurring in all of
these disciplines. What is now required is an understanding of how to interpret them,
define their true relevance, and incorporate mechanistic insights gleaned from these
studies into the theoretical foundation of the underlying science of toxicology. Two
other areas have had a major impact on this discipline. The first is the general area of
quantitative structure–activity relationships. This field bridges chemistry to toxicolo-
gical effects. The second, hormesis, a concept that many consider a challenge to both
Paracelsus and the dose–response paradigm itself, suggests that at very low doses,
some chemicals demonstrate a beneficial rather than the expected lack of response.
Incorporation of hormesis into the risk assessment paradigm that has evolved over
toxicology’s history is truly a challenging problem.
The next section examines these same issues from a slightly different perspec-
tive. In these cases, the fields of toxicology that actually accomplish this incorpora-
tion of science into practice are discussed. Insights into how new approaches to
assessing biological effects are used in chemical and drug risk assessments are
explored. In addition, how specific toxicological testing systems, developed from
basic toxicology research studies, are validated for regulatory use provide a different
perspective to the hurdles that face widespread integration of science into the prac-
tice of toxicology.
The third approach examines this issue by looking at specific technologies,
drugs, chemicals, issues, and disciplines to see how new scientific techniques have
been incorporated into these areas. The first involves two exciting areas of technol-
ogy that have potential for widespread exposure to humans. These are genetically
modified foods and nanomaterials. Although both come from very different
advances in science and technology, they are now being examined by toxicologists
using the tools that have evolved to assess chemical toxicity. Because they have been
developed using many of the modern techniques of biotechnology and material engi-
neering, they are also being examined by some of the newer toxicology techniques
that were not available to earlier workers.
The remainder of the text deals with more specific applications. Current issues
confronting the military, including jet fuel toxicology are reviewed and provide a
good overview of how the various tools of toxicology and exposure assessment have
been employed. Two chapters deal with the application of toxicology to drug safety.
Additional chapters are subdiscipline-oriented, where recent advances in inhala-
tional and forensic toxicology are presented.
Introduction and Overview 3

Reading this text gives one a true feeling of the intellectual diversity of modern
toxicology and how various seasoned practitioners have integrated basic chemistry,
biology, and mathematical models into specific applications. These developing fields
and issues define the scope of toxicology in the 21st century as it evolves to assure
that novel products of technology benefit and do no harm to either our species or
environment.
PART I. NEW CONCEPTS AND APPROACHES

2
Toxicogenomics: Gene Expression Analysis
and Computational Tools

Rupesh P. Amin, Hisham K. Hamadeh, J. Todd Auman, Lee Bennett,

Cynthia A. Afshari, Pierre R. Bushel, and Richard S. Paules
National Institute of Environmental Health Sciences, Research Triangle Park,
North Carolina, U.S.A.
J. Christopher Corton
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina, U.S.A.

INTRODUCTION

Technologies that probe genomic responses to chemical exposure are expected to

improve our understanding of molecular mechanisms of toxicity. Accelerated by
advances in high-throughput DNA sequencing, a number of animals that are impor-
tant as models of human disease (e.g., yeast, worm, fruit fly) have been completely
sequenced. A draft of the human genome released in February 2001 and a draft of
the mouse genome released in December 2002 have provided a wealth of genetic infor-
mation much sooner than initially anticipated (1–3). A high-quality draft covering
more than 90% of the Brown Norway rat genome has also been recently reported
(4). The availability of genomic sequences has provided revolutionary opportunities
to monitor gene expression changes across an entire genome. This flood of genomic
data has also led to the identification and quantitation of sequence differences between
individual humans or between animal strains in the form of deletions, insertions, and
single nucleotide polymorphisms. Dual sources of information about global gene
expression and genetic differences will facilitate an understanding of the molecular
basis of disease in general and, in particular, how chemicals interact with genes,
proteins, and other biological components to cause adverse or beneficial effects.
Changes in the abundance of messenger ribonucleic acid (mRNA) are some of
the earliest events that occur after chemical exposure and usually precede observable
perturbations in cellular and tissue homeostasis. Elucidating the molecular mechan-
ism(s) of chemical exposure using the old paradigm of hypothesis-driven research in
which a gene or group of related genes was examined for association with toxicity has
been difficult. Availability of complementary sequences derived from expressed genes
has allowed construction of tools that can be used to simultaneously interrogate the
expression changes in hundreds or thousands of genes upon chemical exposure

5
6 Amin et al.

(5–7). Transcript profiling has provided an avenue for simultaneously testing multiple
hypotheses of chemical toxicity, thereby accelerating the identification of pathways
mechanistically linked to toxicity. The use of these tools is facilitating novel insights
into genome-wide effects of environmentally relevant chemicals of concern to society
as well as new chemical entities of interest to the pharmaceutical industry. Recent
applications of expression profiling have provided insights into compound-specific
and mode of action-related gene expression changes relevant to xenobiotic-induced
organ system toxicity, such as hepatotoxicity (8–14) and nephrotoxicity (15–17).
Identification of specific gene networks perturbed by classes of xenobiotics holds
tremendous potential for understanding responses to toxicants and will be greatly
facilitated by the use of bioinformatics tools to allow for improved biological integra-
tion and interpretation of vast amounts of expression data (18–22).
This chapter provides an overview of microarray data analysis of gene expres-
sion applied to toxicology (toxicogenomics). We discuss the basics of performing a
microarray experiment, including statistical analysis of expression data and compu-
tational image analysis. We use as an example a study comparing gene expression
profiles between different classes of chemicals to illustrate the types of analyses that
can be useful in toxicogenomics studies.

MICROARRAY-BASED GENE EXPRESSION PROFILING

The National Center for Toxicogenomics (NCT) was created by National Institute
of Environmental Health Sciences (NIEHS) to foster the development and applica-
tions of toxicogenomics. One of the goals of the NCT is to develop a public database
that will be populated with toxicity data associated with exposures to a variety of
xenobiotics representing various classes of industrial chemicals, environmental tox-
icants, and pharmaceutical compounds, as well as the corresponding gene expression
profiles that result in multiple species and tissues (23). A database containing gene
expression information, along with classical toxicity parameters (e.g., clinical chem-
istry, urinalysis, histopathology, etc.), will create an opportunity to mine the data for
identifying novel mechanisms of toxicities or for associating traditional toxicity end-
points with hallmark gene profiles. This database would be generally useful for aca-
demic, industrial, and government toxicologists interested in understanding mode of
action and relevance of human exposure more quickly. The NCT has initiated a large
program to generate xenobiotic-induced gene expression profiles in appropriate tis-
sues, to house the profiles in a publicly available database, and to continue to
develop algorithms for the comparison of ‘‘unknowns’’ to the database. More infor-
mation about this effort can be obtained from the cited website (24). The usefulness
of the database will depend in part on the analysis tools used to facilitate the classi-
fication of unknown compounds and to predict their toxicity based on similarities to
profiles of well-studied compounds. Information derived from database queries (e.g.,
biomarkers of exposure) will provide opportunities to conduct exposure assessment
studies in workplace or environmental exposure scenarios. Identification of sub-
groups with genetic susceptibility to environmentally induced disease will help to
monitor and prevent disease in those individuals.
The NIEHS Microarray Group of the NCT developed a gene expression ana-
lysis platform based on the cDNA microarray approach of Brown and Botstein (6).
DNA sequences for the microarrays are generated by polymerase chain reaction
amplification from purified plasmid DNA encoding cloned cDNA inserts from
Toxicogenomics: Gene Expression Analysis and Computational Tools 7

expressed sequence tag (EST) libraries. ESTs are sequences of DNA derived from a
gene’s mRNA and are generated from an mRNA population isolated from a tissue
or cell population. Many ESTs are uncharacterized genes without an official name or
known function. Even ‘‘known’’ genes often have only inferred or, at best, partly
characterized functions. A computer-driven high-speed robotic arrayer is used to
create high-density spotted cDNA arrays printed on glass slides (5–7), thus allowing
for critical quality control aspects of manipulating thousands of bacterial clones
and DNA fragments to be controlled more easily. The arrayer selects predeter-
mined cDNAs from multiwell plates and spots the DNA onto poly-L-lysine–coated
glass slides at designated locations. The location of each sequence arrayed onto a
chip is computer tracked by a ‘‘gene in plate order’’ file. Presynthesized oligonucleo-
tides that are commercially available can also be printed on glass slides using a
similar technique with a few modifications. Other commercial methods for produ-
cing microarrays for gene expression studies include ink-jet fabrication (25) and
photolithography (26).
cDNA arrays representing several distinct genomes, i.e., human, mouse, rat,
and yeast (Saccharomyces cerevisiae) have been printed and used at the NCT for a
wide variety of toxicogenomics studies (27). Each chip contains between 1700 and
20,000 spots depending on the genome being represented. Having chips containing
an entire set of genes in a genome will help identify the most informative genes
for a particular biological response or process, possibly leading to the development
of smaller, focused microarrays specialized for monitoring an optimal subset of
genes involved in a biological pathway or mechanistic response.
The NIEHS microarray efforts began with the development and application of
the NIEHS Human ToxChip (28), a collection of clones representing approximately
2000 genes proposed by a variety of experts to be particularly relevant to cellular
responses linked to toxicity. The genes on the NIEHS Human ToxChip fall into a
number of categories, including genes implicated in apoptosis, DNA replication
and repair, cell cycle control, and oxidative stress, as well as oncogenes, tumor sup-
pressor genes, peroxisome proliferator, aryl hydrocarbon receptor-, and estrogen-
responsive genes, genes for transcription factors, receptors, kinases, phosphatases,
and heat shock proteins, and cytochrome P450 genes. The NIEHS rat and mouse
chips contain genes representative of these categories from their respective genomes.
In addition, the laboratory utilizes mouse and human oligo chips, each containing
approximately 17,000 oligonucleotides 70 nucleotides long, as well as commercial
oligonucleotide chips, with the goal to monitor as broad a representation of the
expressed genome (i.e., transcriptome) as possible.
An overview flowchart for a typical microarray study used by NCT is depicted
in Figure 1. The approach involves isolating total or polyAþ RNA from experimen-
tal samples (e.g., control or treated, normal or diseased) that will be compared for
changes in gene expression. Isolation of intact mRNA is critical to the success of
a microarray experiment, and optimized protocols for isolation of high-quality
RNA for gene expression experiments are available (29). An indication of the quality
of a RNA sample can be obtained from gel electrophoresis analysis, or from a Bio-
analyzer (Agilent Technologies, Palo Alto, California, U.S.A.). Results from the
Bioanalyzer are displayed in an electropherogram, which is a computer-generated
gel-like image of the sample that provides the 28S/18S ribosomal RNA ratios and
the RNA concentration in the sample. These methods help to determine whether
there is RNA degradation in a sample. Poor quality RNA can affect subsequent
labeling reactions and lead to erroneous results in expression profiling experiments.
8 Amin et al.

Figure 1 Experimental design and flowchart for microarray analysis. This illustration pro-
vides an overview of a toxicogenomics study, including experimental design. Rats (n ¼ 3 per
group) were treated with either a peroxisome proliferator (clofibrate, Wy-14,643, or gemfibro-
zil) or an enzyme inducer (phenobarbital) for 1 and 14 days with daily dosing. D-mannitol
served as a negative control in this study. Steps involved in a cDNA microarray experiment
and the data analysis approach are illustrated.

The identification and exclusion of poor quality RNA from microarray analysis will
increase the chances that a database is populated by high-quality data.
Isolated mRNA is converted to cDNA using reverse transcriptase (RT) in the
presence of one of two fluorescently tagged nucleotides, commonly Cy3-dUTP or
Cy5-dUTP. The two populations of labeled cDNAs are mixed in equal amounts
and hybridized for 18 to 24 hours onto glass slides containing the spotted cDNAs.
After washing off labeled cDNAs not specifically hybridized to the slide, the amount
of the two dyes on each spot is quantitated using a specialized laser scanner that cap-
tures the fluorescence intensity of hybridized dye-labeled cDNA at each spot on an
array. The Cy3 and Cy5 dyes exhibit different wavelengths for excitation (532 and
635 nm) and emission (575 and 675 nm), respectively. The images derived from the
Cy3 or Cy5 dyes are used for subsequent image analysis and data acquisition, mak-
ing it possible to detect the relative abundance of mRNA.
Analysis of microarray images is complex and can be subdivided into array tar-
get segmentation, background intensity extraction, target detection, target intensity
extraction, normalization and ratio analysis, and measurement quality assurance
(29). The NIEHS Microarray Group uses a variety of software tools, including the
ArraySuite image-processing software originally written by Chen et al. (30) at
the National Human Genome Research Institute (NHGRI). This software is avail-
able as MicroArray Suite (Scanalytics Inc., Fairfax, Virginia, U.S.A.) and carries
out a number of functions, including detection of targets, i.e., spots, and associates
these spots with identified genes on an x–y coordinate map of the chip. Another
program in MicroArray Suite processes mean pixel intensity for each spot as well
Toxicogenomics: Gene Expression Analysis and Computational Tools 9

as the intensity of the background region, the latter being used for local background
subtraction of the images (30). Once the background-subtracted intensities are deter-
mined, a ratio of the intensities at 635 and 532 nm (Cy5/Cy3) is calculated for each
spot. However, first the ratio intensities need to be normalized to correct for scanning
imbalances. To do this, the software fits all of the ratio intensity data from a chip to a
probability distribution and calculates an estimated normalization constant, which is
then used to normalize the chip-to-chip ratio intensities. Another program in Micro-
Array Suite facilitates superimposing the images generated by scanning the chip at
two different wavelengths, creating one composite overlaid image, called a pseudo-
image. The pseudoimage is a visual representation of the normalized ratio intensities,
with different colored spots indicating different levels of expression, ranging from red
(Cy5/Cy3 ratio >1) to green (Cy5/Cy3 ratio < 1). Thus, a red spot indicates greater
abundance of transcripts from the experimental sample (provided the experimental
RNA was tagged with Cy5), whereas a green spot indicates greater abundance of
transcripts from the control sample. A yellow spot indicates roughly equivalent tran-
script levels in control and experimental samples (Cy5/Cy3 ffi 1), whereas no spot
indicates that neither sample contains detectable amounts of that transcript.

SELECTION OF DIFFERENTIALLY EXPRESSED GENES

A microarray experiment can monitor the expression changes of thousands of genes

simultaneously, yet only a fraction of the genes in any given experiment will exhibit a
statistically significant level of differential expression. The genes that remain, pre-
sumably, unchanged can thus be used to normalize the intensity ratios across the
entire chip to approximately one and then are frequently excluded from further ana-
lysis. As research into the analysis of microarray expression data has progressed,
many different methods have been developed for identifying differentially expressed
genes, ranging from simple fold-change cutoffs to sophisticated stochastic models.
Chen et al. (30) developed one of the first statistical treatments of microarray
expression-ratio data from two-color experiments. This method allows the researcher
to specify a confidence level for the analysis, and a confidence interval is calculated
for the ratio intensity values on the chip; genes having ratio values outside of this
interval are considered differentially expressed with a high degree of confidence.
The first version of this method used the distribution of ratio values from housekeep-
ing genes to produce fixed-width confidence intervals, but an extension of the
method allows for gene-specific adaptive confidence intervals that account for differ-
ences in the signal-to-noise ratio between spots (31). The ratio distribution models
developed by Chen et al. (30) are available in the MicroArray Suite package and
are one of the types of gene selection methods used by members of the NCT.
If the ratio model fits the data well and independent replicate hybridizations
are performed, a binomial probability distribution can be used to determine the
probability of a gene being detected as differentially expressed multiple times strictly
by chance. To diminish experimental error in microarray analysis, hybridizations can
be performed using technical replicates from individual samples; a binomial distribu-
tion can be used to model the results of the analyses at given confidence levels (32).
For example, scoring genes as differentially expressed at the 95% confidence level
(p ¼ 0.05) four or more times (k  4) out of nine replicate experiments (n ¼ 9) has
a binomial probability (p) of 0.00064 of being detected by chance. In addition, it
is advisable to reverse the fluorescent molecules tagged to the control and treated
10 Amin et al.

samples by tagging RNA from experimental samples with Cy3 instead of Cy5 and
control samples with Cy5 instead of Cy3, which in microarray parlance is called
fluor-flipping or dye-reversal. By utilizing fluor-flips in the experimental design, false
positives in gene selection due to dye biases can be avoided. The use of fluor-flipping
in microarray experiments including minimizing the number of replicates needed has
been recently addressed (33). An examination of all data will yield a list of genes that
are consistently identified as differentially expressed according to criteria specified by
the researcher generating a statistically validated list of differentially expressed genes.
Genes with highly variable expression changes across hybridizations are flagged
owing to a large coefficient of variation or a modified Z-score computation to detect
outliers (32,34). Thus, a simple calculation can be used to convey some measure of
confidence for the results of an entire experiment.
In addition to using fold-change models, a variety of statistical techniques have
been developed or adapted for the analysis of microarray data to detect differential
gene expression, including the use of linear models and analysis of variance
(ANOVA) (35,36). These models incorporate all of the replicate hybridization data
into one analysis, and provide a measure of statistical significance, a p-value, for each
individual gene. Additionally, models of this type allow for estimates of the sources
of experimental variation, such as effects due to the dyes, differences between chips,
and even spatial variability within chips. The mixed linear model method developed
by Wolfinger et al. (35) uses two separate ANOVA models to identify genes that are
significantly changed by a treatment. First, a model is fit to normalize the data across
arrays; systematic effects due to the dyes are identified and the variation due to chip-
to-chip differences is quantified. The difference between the observed data and the
fitted values given by this model are called the residual values, and these serve as
the input for the second set of ANOVA models, which are fit on a gene-by-gene
basis. These models are designed to assess effects due to the treatment while account-
ing for gene-specific dye biases as well as spot variation between arrays. In both these
models, the arrays are considered to be a ‘‘random effect’’; that is, it is assumed that
the chips used in the experiment were drawn from a large population of available
chips. Doing this allows us to make inferences about the treatments that are derived
from the analysis, to be generalized to a larger population of chips rather than
restricted to only the arrays in the experiment.
These ANOVA models rely on typical statistical assumptions for linear
models. First, it is assumed that the ‘‘error’’ terms in the individual gene models,
which measure the variation that the model cannot account for, are normally distrib-
uted with a mean of zero and constant variance. Second, the random effects due to
arrays are assumed to have a normal distribution with a mean of zero. The limited
number of observations for each gene in a microarray experiment can make it difficult
to assess the validity of these assumptions, but some diagnostic tools are available to
verify that the models fit the data adequately. The inference results are sometimes dis-
played in the form of a ‘‘volcano plot’’ (Fig. 2), which shows graphically the relation-
ship between fold-change estimates and statistical significance values, reported as the
negative log10(p-value). The horizontal line on the plot indicates the cutoff value for
statistical significance while the vertical lines represent a twofold change in gene
expression, both of which are predetermined by the researcher. Spots above the hor-
izontal line in the center area of this plot indicate genes that have small expression
changes but are nonetheless determined to be statistically significant by mixed linear
modeling. Owing to the multiple testing problem created by the large number of genes
on an array, multiple comparison correction procedures such as the Bonferonni
Toxicogenomics: Gene Expression Analysis and Computational Tools 11

Figure 2 Example of a volcano plot from mixed model analysis. Fold-change values are
compared to statistical significance values for each gene on a microarray. The dashed vertical
lines represent fourfold induction and repression, as measured on the x-axis. The horizontal
line at –log(p) ¼ 5 represents a p-value of 0.00001. Genes falling above this line exhibit differ-
ential expression that is statistically significant based on this p-value.

method are necessary to reduce false positives but are often considered to be overly
conservative (37).
Another type of method to select differentially expressed genes from microar-
ray data is the error model-based approach. Hughes et al. (38) utilized a model for
the uncertainties in individual array experiments that took into account both addi-
tive and multiplicative error components in the channels of a two-color experiment.
In addition, a p-value and the minimum-variance–weighted average used to compute
the mean log10 intensity ratio of each reported gene are computed to allow the ana-
lyst to select differentially expressed genes with more reliability. In the case where
the distribution of the gene expression data is not known or desired to be assumed
for analysis, the nonparametric gene selection of Callow et al. (39), which assumes no
specific parametric form for the distribution of the expression levels and employs a
permutation procedure to estimate the joint null distribution of the t-test statistic for
each gene, is a reasonable option. Significance analysis of microarrays (40) uses
permutations of repeated measures of microarray data as well to estimate the false
discovery rate—the percent of genes identified by chance.

FINDING EXPRESSION PATTERNS

Many researchers who use microarrays to monitor gene expression are interested in
identifying groups of genes that exhibit similar expression profiles. These genes may
provide insight into the biological mechanisms at work as cells or tissues respond to
chemical or physical perturbations. Cluster analysis, an exploratory data analysis
technique familiar to biologists because of its use in phylogenetic analysis, has
become a popular tool for finding patterns in gene expression data. Since its first
application to microarray expression data by Eisen et al. (41), clustering results have
12 Amin et al.

appeared regularly in the gene expression literature. The use of clustering by scien-
tists at the NCT Microarray Group is illustrated later in this chapter using gene
expression data generated from a rat in vivo study.
A popular type of cluster analysis is hierarchical cluster analysis, which pro-
duces output in the form of a ‘‘tree’’ or dendrogram (41). This type of analysis is
a form of unsupervised clustering, meaning that an informed classifier is not used;
also labels (categories) or prior information about the samples are not used for clas-
sification. A similarity measure, such as Pearson correlation or Euclidean distance, is
used to determine the level of dissimilarity or similarity between two clusters. A
method for assigning the distance between the clusters, termed ‘‘a linkage method,’’
must be selected by the researcher. Single, complete, and average linkage methods
define the distance between clusters as the minimum, maximum, and average pair-
wise distance between members of the clusters, respectively. The agglomerative pro-
cess of comparing and merging clusters continues until only one cluster remains, and
the resulting dendrogram illustrates the relationships between clusters at each step.
Another method is k-means clustering, which requires a priori knowledge of
the number of clusters in the data. k-means clustering iteratively assigns objects to
a cluster according to minimization of an objective function that measures the dis-
tance of the object to the mediod (center of a cluster). The process continues for a
large number of iterations, until there are no more movements (reassignments) of
objects to clusters or there is no more change in the objective functions after some
predetermined number of iterations. Self-organizing maps (SOMs), a technique
developed by Kohonen (42) in the early 1990s, is based on neural network theory
and was first used by Tamayo et al. (43) to reveal unsupervised nodes (clusters) of
genes that distinguish acute myeloid leukemia from acute lymphocytic leukemia bio-
logical samples.
Other clustering algorithms are known as supervised, where the labels of the
samples and/or prior information about the samples (learning set) are provided to
the classifier. For example, k-nearest neighbors (KNN) assign samples to the class
of the nearest neighbor to it by majority vote, and support vector machines (SVMs)
use the labels of the training samples near or at the decision boundaries separating
two classes to classify the unknown/unlabeled sample. Steiner et al. (44) used multi-
ple SVMs to separate classes and subclasses of toxicants, separate nonresponders
from hepatotoxicants, identify sets of genes that discriminate hepatotoxicants from
nonhepatotoxicants, and reveal how a predictive model built from one strain of rats
can be used to classify treatment of another strain of rats.
One application of toxicogenomics analysis is the comparison of different
treatment groups for classification of chemicals based on gene expression profiles.
Biological samples derived from toxicant or control-treated animals can be repre-
sented as gene expression patterns consisting of fingerprints or profiles. These
profiles are analyzed with automated pattern recognition analyses aimed at deter-
mining similarity between datasets rather than probing the genes for mechanistic
information. For example, one goal of an in vivo study by Hamadeh et al. (11)
was to use expression profiling to classify blinded samples derived from livers of
xenobiotic-treated rats. Linear discriminant analysis (LDA) is one classical approach
used to solve a task of this nature, but this method often relies on the assumption
that the data follows a multivariate normal distribution. For gene expression data,
Li et al. (45,46) have utilized an innovative nonparametric approach that combines
a genetic algorithm (GA) for gene selection with the KNN method for sample clas-
sification (47). Subsets of candidate genes, ‘‘chromosomes,’’ are generated and tested
Toxicogenomics: Gene Expression Analysis and Computational Tools 13

for their ability to separate known classes of samples using the KNN criterion.
Through probabilistic ‘‘mutations,’’ the best subsets of genes evolve as the process
goes through several iterations. When a large number of these subsets have been
selected, the frequency with which each gene appears in the subsets is tallied. Intui-
tively, the most frequently occurring genes should be the most informative in terms
of discriminating between the queried classes. It is important to note that, Ooi and
Tan (48), who developed the GA/maximum likelihood (GA/MLHD) method for
multiclass prediction of gene expression data, indicate that the GA/KNN method
may not be optimal for multiclass classification purposes because of limitations with
using KNN with a high dimensional dataset; the computational complexity of the
approach and the gene selection component of the algorithm is best suited for binary
classification purposes. In any case, the flexibility of combined algorithms for gene
selection and classification of microarray gene expression data is appealing because
different components of the algorithm can be replaced with more improved
approaches as they become available.
Principal component analysis (PCA) has also been used with microarray
expression data (49). The goal of this type of analysis is to reduce the complexity
of a dataset through the creation of new variables, called the principal components.
The principal components retain a large percentage of the information that is present
in the original variables, but are uncorrelated with each other. Depending on the
goal of the analysis, either the genes or the arrays can be considered the original vari-
ables in a microarray experiment. If genes are used, the resulting principal compo-
nents may provide insight into the particular features of the genes that explain the
experimental response (50). The visualization of high dimensional data in two- or
three-dimensional principal components space may reveal groups in the dataset. This
information can then be used as input for classification methods, such as clustering,
SOM or KNN. If there is significant separation between the clusters, PCA may also
be used to assist with classification problems. Unknown samples can be ‘‘scored’’
using the results of a PCA on known samples, and these scores may be used to place
the unknown into one of the previously identified clusters.

DATA MANAGEMENT

One of the primary challenges that has arisen with the development of microarray
technology is that of efficiently managing the large volumes of information that these
experiments produce. The NIEHS Microarray Group of the NCT has integrated two
web-based applications that allow users to track and analyze their experimental infor-
mation from start to finish—MicroArray Project System (MAPS), a laboratory infor-
mation management system, and ArrayDB, an analysis information management
system that allows users to store data and analyze individual array experiments.
The MAPS application, developed by Bushel et al. (32) at NIEHS, is a tool that
allows researchers to store information about projects, samples, hybridizations, and
quality control parameters. Originally designed as a Microsoft Access database with
a ColdFusion web interface, MAPS has since been moved to an Oracle database for
faster and more robust performance. A key feature of MAPS is the ability to store
information about genes that were identified as differentially expressed using the
confidence interval model described above, thus enabling the experimenter to ana-
lyze multiple hybridizations easily. For any combination of arrays, a list of differen-
tially expressed, statistically valid ‘‘signature’’ genes that appear in some specified
14 Amin et al.

number of those arrays can be produced. Additionally, MAPS provides some statis-
tical details about the genes that were chosen to give a measure of the quality of the
data. For example, genes that show large variation in expression level across hybri-
dizations can be flagged for further investigation or exclusion. In addition, MAPS
also tracks information on RNA quality, scanner details, and other relevant para-
meters that vary and affect the quality of microarray data.
ArrayDB, originally developed at the NHGRI, provides an interactive system
for the analysis of microarray experiments, and has been modified for use with an
Oracle database at NIEHS. Once the image analysis is performed on the scanned
array slide, the images and data are stored in the database for access via the web.
An experimenter can view the pseudocolor array image, check intensity and expres-
sion values for particular spots on the array or for particular genes, and perform a
differential expression analysis using the same method as the ArraySuite software
described above. Because ArrayDB stores data for every gene on an array instead
of just those that are differentially expressed, it provides a platform from which other
programs can query data for analysis. The Microarray Group is currently in the pro-
cess of interfacing several other analysis tools with ArrayDB to facilitate new analy-
sis of existing datasets without reformatting.

BIOLOGICAL DATABASES USEFUL FOR TOXICOGENOMICS

Making biological sense of gene expression data generated by microarray analysis is

a major rate-limiting step, yet probably the most critical and exciting component of a
toxicogenomics study. Bioinformatics tools help visualize and dissect toxicant-
induced gene expression data; however, interpreting and making biological sense
of the coordinated regulation of the hundreds or thousands of genes requires tremen-
dous scientific input. Xenobiotics can affect multiple molecular pathways and it is
not uncommon to expect tissue-specific regulation of genes involved in various cel-
lular processes to be simultaneously modulated by a chemical. Understanding and
elucidating mechanism(s) of toxicity requires identification of coordinately regulated
genes in common pathways, developing hypotheses as to how changes occur and the
toxicological significance of the changes, and testing those hypotheses by conducting
additional mechanistic studies. Table 1 provides some genomics-related web res-
ources available for investigating the function of genes and gene products. These
resources include those that allow characterization of ESTs, identification of biolo-
gical and biochemical function of gene products as well as their cellular location, and
identification of orthologous genes in other species and methods to search promoter
sequences for common elements that may coordinately regulate gene expression.
New resources are rapidly evolving and coming online.

EXAMPLES FROM AN IN VIVO TOXICOGENOMICS STUDY

The utility of using microarrays to perform a toxicogenomics study can be illustrated

by describing the results of pioneering studies performed at the NCT (10,11). An
in vivo rat model system was used to investigate the hypothesis that treatment with
different xenobiotics results in chemical-specific patterns of altered gene expression.
This hypothesis was based on the premise that genes that altered following exposure
to different classes of chemicals can differentiate one class from another. Alterations
Toxicogenomics: Gene Expression Analysis and Computational Tools 15

Table 1 Resources for Toxicogenomics

Category Database Web locationa

Public EST sequences GenBank www.ncbi.nlm.nih.gov/Genbank/

index.html
DbEST www.ncbi.nlm.nih.gov/dbEST/index.html
cDNA databases UniGene www.ncbi.nlm.nih.gov/entrez/
query.fcgi?db¼unigene
TGIR gene indices www.tigr.org
DoTS www.allgenes.org
Eugene eugenes.org/
Gene annotation and DAVID http://david.niaid.nih.gov/david/
pathway(s)
Genecard www.genecards.org
DRAGON pevsnerlab.kennedykrieger.org/dragon.htm
S.O.U.R.C.E. http://source.stanford.edu
KEGG www.genome.jp/kegg/
OMIM www.ncbi.nlm.nih.gov/entrez/
query.fcgi?db¼OMIM
Ensembl-Human www.ensembl.org/
Genome Server
SPAD www.grt.kyushu-u.ac.jp/spad/index.html
Signal transduction stke.sciencemag.org/search/
maps searchpage.dtl?search_my¼all
Other ExPASy us.expasy.org
GO www.geneontology.org/
Pubmed www.ncbi.nlm.nih.gov/entrez
Mouse, rat, and www3.ncbi.nlm.nih.gov/Homology/
human comparative index.html
maps
Locuslink http://www.ncbi.nih.gov/entrez/
query.fcgi?db¼gene
Transcription element www.cbil.upenn.edu/tess/
search system
Links to mammalian www.genelynx.org/
genomes
Genome genome.ucsc.edu/
bioinformatics
a
Website addresses as of December, 2005.
Abbreviations: DAVID, database for annotation, visualization, and integrated discovery; DoTS, database
of transcribed sequences; DRAGON, database referencing of array genes online; EST, expressed sequence
tag; ExPASy, expert protein analysis system; GO, gene ontology; KEGG, kyoto encyclopedia of genes and
genomes; OMIM, online mendelian inheritance in man; SPAD, signalling pathways database.

in hepatic gene expression of Sprague-Dawley rats were studied 1 and 14 days after
treatment with nongenotoxic rodent hepatocarcinogens (Fig. 1) (10). The com-
pounds studied represent two classes of chemicals that alter gene expression through
nuclear receptors. The enzyme inducer, phenobarbital, activates the constitutive
androstane (or activated) receptor (51) and three known peroxisome proliferators
(clofibrate, Wy-14,643, and gemfibrozil) activate the peroxisome proliferator-
activated receptor a (PPARa) (52). D-Mannitol, which caused no detectable patho-
logical effects in the liver, was used as a negative control. Microarray analysis was
16 Amin et al.

performed using the NIEHS rat chip containing approximately 1700 genes, as
described above.
Transcript changes observed in rat livers after treatment were analyzed using a
number of computational approaches. Clustering analysis was useful for determining
similarities and differences in expression ratios between the genes and samples
being analyzed, as visualized by the dendrogram in Figure 3. The three peroxisome
proliferators clustered together, indicating a similar pattern of gene expression that
was different from the gene expression profile for phenobarbital. Many of the changes
in gene expression were in agreement with known responses to these agents. Genes
reported to be involved in peroxisomal or mitochondrial fatty acid b-oxidation
were upregulated by peroxisome proliferators, but suppressed by phenobarbital.
Genes upregulated by phenobarbital treatment, but not by peroxisome proliferator
treatment, included genes known to be involved in the detoxification and metabolism
of xenobiotics. Hamadeh et al. (10) also observed ESTs with unknown functions that
were coordinately regulated with named genes with known functions [Fig. 3 (node
II)]. Clustering of microarray data could be used in toxicological studies to associate
poorly characterized genes with expression profiles similar to genes involved in
well-defined pathways or in the mechanism(s) of toxicity of chemicals. This has been
termed ‘‘guilt by association’’ and will likely increase our understanding of gene
function in general (38,53).
Detailed analysis of gene expression patterns of each compound revealed time-
dependent and time-independent changes. Following phenobarbital treatment, the
expression of 57 and 81 genes were significantly altered at 1 and 14 days, respectively
(Fig. 4), with 38 genes altered at both time points, 19 genes specifically expressed at
one day and 43 genes specific at 14 days. Cdc2 was upregulated only one day after
phenobarbital treatment, while lipoprotein lipase was upregulated only at 14 days,
indicating that phenobarbital treatment resulted in specific early effects related to
hepatocyte hyperplasia and late time-dependent changes related to lipid metabolism.

Figure 3 Hierarchical clustering of validated genes. Hierarchical tree depicts the grouping of
correlated genes in specific nodes. I and II are nodes depicting classes of correlated genes.
Toxicogenomics: Gene Expression Analysis and Computational Tools 17

Figure 4 Gene alterations by phenobarbital and Wy-14,643. Validated gene outliers for
each compound at 1 and 14 days, obtained using a binomial distribution analysis at
95% confidence interval, were compared and the results are presented using a Venn diagram.
(A) Time-dependent and time-independent gene alterations by phenobarbital at 1 and 14 days,
(B) time-dependent and time-independent gene alterations by Wy-14,643 at 1 and 14 days, and
(C) partial overlap in genes regulated by phenobarbital or Wy-14,643 at one day are illustrated.

Phenobarbital-induced increases in lipoprotein lipase and hepatic triglyceride lipase

activities have been reported with concomitant increases in hepatic synthesis of
triglyceride, but lower serum concentration of triglyceride (54,55). Expression of
CYP2B2 and glutathione-S-transferase were elevated at both time points. Liu
et al. (56) demonstrated that a 163-bp fragment of the rat CYP2B2 gene contains
sequences that mediate phenobarbital responsiveness, and mutations within this
region reduced responsiveness to phenobarbital.
A similar analysis of gene expression altered by Wy-14,643 also revealed time-
dependent and time-independent patterns of gene expression, with a total of 75 and
199 genes significantly modulated in rat liver 1 and 14 days after treatment, respec-
tively (Fig. 4B). Of these, 58 genes were common to both time points, while 17 genes
were differentially expressed only at one day and 141 genes were only differentially
expressed at 14 days after treatment. Acyl-CoA oxidase and acyl-CoA dehydrogen-
ase, genes involved in the first and second steps of fatty acid b-oxidation, were 2 of
the 17 genes that were observed to be uniquely modulated by Wy-14,643 at one day.
One of the 141 genes modulated by Wy-14,643 at 14 days was lipid-binding protein.
Thiolase and stearyl-CoA desaturase exhibited increased expression at both time
points. All of these genes are known to be transcriptionally regulated by PPARa
(57,58). An understanding of early and late changes in gene expression may not only
be indicative of altered biological processes that arise from xenobiotic exposure, but
may also be useful for chemical classification.
Alterations in gene expression induced by compounds in these two different
chemical classes were compared one day after treatment. The data from this com-
parison showed that Wy-14,643 and phenobarbital significantly modulated the
18 Amin et al.

expression of 75 and 57 genes, respectively, of which only 15 genes were common to

both compounds (Fig. 4C). Therefore, Wy-14,643 and phenobarbital treatment
uniquely modulated 60 and 42 genes, respectively, suggesting that these genes may
help define chemical class-specific ‘‘gene signatures.’’ For example, the expression
of CYP2B2 was increased by phenobarbital but not by Wy-14,643, while the expres-
sion of thiolase was increased by Wy-14,643 but not by phenobarbital. In contrast,
both Wy-14,643 and phenobarbital increased transcript levels of uridine diphosphate
(UDP)–glucose dehydrogenase, suggesting that increasing the expression of this
enzyme is a common response to xenobiotic exposure because this enzyme furnishes
UDP-glucuronic acid for Phase II xenobiotic metabolism (59).
Comparison of the altered gene expression profiles after treatment with the
three peroxisome proliferators (clofibrate, Wy-14,643, and gemfibrozil) revealed 12
genes that were regulated in the same manner by all three compounds one day after
treatment and 13 genes commonly regulated 14 days after treatment (Fig. 5A and B).
One of the genes overexpressed by all three peroxisome proliferators at both time
points was rat liver stearyl-CoA desaturase. In addition to changes in gene expres-
sion shared by the three compounds, the peroxisome proliferators caused changes
in gene expression that were unique for each compound (Fig. 5A and B). Further
analyses of these observations will hopefully provide evidence that gene expression
studies can be used to identify subtle similarities and differences in mechanism(s)
of action of compounds within a chemical class, which may, in part, be attributed
to differences in chemical structure, receptor–ligand interactions, drug metabolism,
and/or gene targets.
As the first study was able to identify genes that distinguish two classes of com-
pounds, the next study by Hamadeh et al. (11) sought to determine if it was possible

Figure 5 Gene expression changes by peroxisome proliferators. Validated gene outliers for
each peroxisome proliferator compound (i.e., clofibrate, Wy-14,643, and gemfibrozil), at the
(A) one-day time point and (B) 14-day time points, obtained using a binomial distribution
analysis and 95% confidence interval are compared and the results are presented using a Venn
diagram.
Toxicogenomics: Gene Expression Analysis and Computational Tools 19

to assign an unknown compound to a specific chemical class based on changes in

gene expression elicited by treatment. Multiple approaches were used to find highly
discriminatory and informative genes whose expression pattern could distinguish
RNA samples derived from livers exposed to different chemicals. Two methods,
LDA and GA/KNN discussed above were useful in revealing genes that could sepa-
rate known samples based on the class of chemical involved in the exposure in a
time-independent manner. Using these procedures, 22 highly informative genes that
clearly exhibited different patterns of expression between enzyme inducers and per-
oxisome proliferators were identified. A blinded study was conducted using rat liver
RNA from animals treated with various compounds, comparing the gene expression
alterations in the blinded samples, using computational tools, to the gene expression
changes elicited by compounds from the two chemical classes, as well as the negative
control studied. The results were extremely encouraging because this approach made
it possible to successfully classify 22 out of 23 unknown compounds based on the
expression signatures of known compounds (11).
A number of other groups have published studies that categorize chemicals
based on gene expression patterns (Table 2). Most of these studies examined pattern
changes after chemical exposure in the rat liver because the liver is the primary
site for toxicity and much is known about mechanisms of chemical toxicity in this
organ, allowing for preliminary chemical classification based on mode of action.
In some studies, unsupervised methods were initially used for clustering the com-
pounds; success was limited because of the lack of reproducible gene responses
likely due to biological variability and limitations imposed by lack of replicates
(60) or due to incomplete understanding of the types of liver toxicity induced by
the queried compounds (63). Most studies used supervised methods that required
prior knowledge of the type of toxicity induced by chemical exposure. Although
these studies focused on high doses that led to clear toxicity as assessed by conven-
tional end-points, a major challenge will be to include in the model building process
expression profiles after treatment with doses of chemicals that do not induce the
conventional toxicology endpoints. For example, a recent study analyzing gene
expression changes in the livers of rats treated with acetaminophen showed that
gene expression changes predicted toxic effects that were not observed by conven-
tional endpoints at higher doses (14). Earlier time points should also be considered
to determine if gene expression changes are early predictors of developing toxicities
before standard tests could detect them. Together these findings indicate that it will
be possible to obtain compound-related gene expression signatures that are useful in
chemical class prediction.

FUTURE PROSPECTS FOR TOXICOGENOMICS

Advances in toxicogenomics are expected to facilitate pharmaceutical and industrial

lead compound development by identifying, much earlier than is presently possible,
which compounds have the propensity to cause human toxicity and, perhaps, to pre-
dict the target population for either the pharmacological or toxicological effect. An
era of genetic medicine in which therapeutic strategies will be tailored to the needs of
individuals with known genome sequence variations is likely to emerge in the future.
The potential for monitoring subtle gene expression changes will facilitate the deve-
lopment of biomarkers of exposure and effect, providing future opportunities to
screen ongoing molecular changes in accessible human tissues, i.e., blood, urine,
 20

Table 2 Chemical Classification Studies

Experimental parameters
(#cmpds; system; treatment
References time; dose) Discriminant method Supporting data Results

Burczynski et al., 100; HepG2 cells; Unsupervised; supervised None Discriminated between DNA
2000 (60) 24 hr; various doses computational algorithm damaging and anti-inflammatory
drugs
Waring et al., 15; rats; 3 days; known Unsupervised methods: hierarchical, Histopathology and Profiles correlated to clinical
2001 (8) hepatotoxic doses divisive hierarchical, k-means, clinical chemistry chemistry and pathology
and SOMs
Waring et al., 15; rat primary hepatocytes; Unsupervised unweighted pair-group None Compounds with similar toxic
2001 (61) 24 hr; 20 mM method with arithmetic mean mechanisms formed clusters
Hamadeh et al., 4; rats; 1 and 14 days; Supervised; LDA GA/KNN Histopathology, Correctly identified 22 out of 23
2002 (10,11) various doses RT-PCR blinded samples
De Longueville et al., 11; rat primary hepatocytes; Unsupervised classical agglomerative RT-PCR Steatosis inducers clustered together
2003 (62) 24 hr; below toxic dose hierarchical clustering
McMillian et al., 100; rats; 24 hr; MTD Unsupervised; supervised RT-PCR of cytokines Discriminated between macrophage
2004 (63) discriminant analysis and activators and peroxisome
cross-validation proliferators
Steiner et al., 31; rats; 6 hr to 14 days; Supervised SVMs with recursive Serum chemistry; Discriminated between hepatotoxic
2004 (44) various doses feature elimination histopathology and nonhepatotox; predicted class
in most cases

Abbreviations: cmpds, compounds; GA, genetic algorithm; KNN, k-nearest neighbors; LDA, linear discriminant analysis; MTD, maximum tolerated dose; RT-PCR, reverse
transcriptase-polymerase chain reaction; SOMs, self-organizing maps; SVMs, support vector machines.
Amin et al.
Toxicogenomics: Gene Expression Analysis and Computational Tools 21

buccal scrapings. Not only will environmental and occupational health physicians
and scientists be able to identify toxic compounds and their mechanism(s), but it
is also expected that genomic and technological advances will propel medical
advances and provide opportunities for physicians to intervene during disease devel-
opment and progression.

REFERENCES

1. Venter JC, Adams MD, Myers EW, et al. The sequence of the human genome. Science
2001; 291:1304–1351.
2. Lander ES, Linton LM, Birren B, et al. Initial sequencing and analysis of the human
genome. Nature 2001; 409:860–921.
3. Waterston RH, Lindblad-Toh K, Birney E, et al. (Mouse Genome Sequencing Consor-
tium) Initial sequencing and comparative analysis of the mouse genome. Nature 2002;
420:520–562.
4. Gibbs RA, Weinstock GM, Metzker ML, et al. (Rat Genome Sequening Project evolution)
Genome sequence of the Brown Norway rat yields insights into mammalian evolution.
Nature 2004; 428:493–521.
5. Duggan DJ, Bittner M, Chen Y, Meltzer P, Trent JM. Expression profiling using cDNA
microarrays. Nat Genet 1999; 21:10–14.
6. Brown PO, Botstein D. Exploring the new world of the genome with DNA microarrays.
Nat Genet 1999; 21:33–37.
7. Bowtell DD. Options available—from start to finish—for obtaining expression data by
microarray. Nat Genet 1999; 21:25–32.
8. Waring JF, Jolly RA, Ciurlionis R, et al. Clustering of hepatotoxins based on mechanism
of toxicity using gene expression profiles. Toxicol Appl Pharmacol 2001; 175:28–42.
9. Hamadeh HK, Knight BL, Haugen AC, et al. Methapyrilene toxicity: anchorage of
pathologic observations to gene expression alterations. Toxicol Pathol 2002; 30:470–482.
10. Hamadeh HK, Bushel PR, Jayadev S, et al. Gene expression analysis reveals chemical-
specific profiles. Toxicol Sci 2002; 67:219–231.
11. Hamadeh HK, Bushel PR, Jayadev S, et al. Prediction of compound signature using high
density gene expression profiling. Toxicol Sci 2002; 67:232–240.
12. Ruepp SU, Tonge RP, Shaw J, Wallis N, Pognan F. Genomics and proteomics analysis
of acetaminophen toxicity in mouse liver. Toxicol Sci 2002; 65:135–150.
13. Ueda A, Hamadeh HK, Webb HK, et al. Diverse roles of the nuclear orphan receptor
CAR in regulating hepatic genes in response to phenobarbital. Mol Pharmacol 2002;
61:1–6.
14. Heinloth AN, Irwin RD, Boorman GA, et al. Gene expression profiling of rat livers
reveals indicators of potential adverse effects. Toxicol Sci 2004; 80:193–202.
15. Huang Q, Dunn RT II, Jayadev S, et al. Assessment of cisplatin-induced nephrotoxicity
by microarray technology. Toxicol Sci 2001; 63:196–207.
16. Kramer JA, Pettit SD, Amin RP, et al. Overview on the application of transcription pro-
filing using selected nephrotoxicants for toxicology assessment. Environ Health Perspect
2004; 112:460–464.
17. Amin RP, Vickers AE, Sistare F, et al. Identification of putative gene based markers of
renal toxicity. Environ Health Perspect 2004; 112:465–479.
18. Henry CJ, Phillips R, Carpanini F, et al. Use of genomics in toxicology and epidemiol-
ogy: findings and recommendations of a workshop. Environ Health Perspect 2002;
110:1047–1050.
19. Amin RP, Hamadeh HK, Bushel PR, Bennett L, Afshari CA, Paules RS. Genomic inter-
rogation of mechanism(s) underlying cellular responses to toxicants. Toxicology 2002;
181–182:555–563.
22 Amin et al.

20. Gant TW. Application of toxicogenomics in drug development. Drug News Perspect
2003; 16:217–221.
21. Suter L, Babiss LE, Wheeldon EB. Toxicogenomics in predictive toxicology in drug
development. Chem Biol 2004; 11:161–171.
22. Irwin RD, Boorman GA, Cunningham ML, Heinloth AN, Malarkey DE, Paules RS.
Application of toxicogenomics to toxicology: basic concepts in the analysis of microarray
data. Toxicol Pathol 2004; 32(suppl 1):72–83.
23. Waters M, Boorman G, Bushel P, et al. Systems toxicology and the chemic effects in
biological systems (CEBS) knowledge base. EHP Toxicogenomics 2003; 111:15–28.
24. www.niehs.nih.gov/nct/.
25. Hughes TR, Mao M, Jones AR, et al. Expression profiling using microarrays fabricated
by an ink-jet oligonucleotide synthesizer. Nat Biotechnol 2001; 19:342–347.
26. Lipshutz RJ, Fodor SP, Gingeras TR, Lockhart DJ. High density synthetic oligonucleo-
tide arrays. Nat Genet 1999; 21:20–24.
27. www.dir.niehs.nih.gov/microarray/.
28. Nuwaysir EF, Bittner M, Trent J, Barrett JC, Afshari CA. Microarrays and toxicology:
the advent of toxicogenomics. Mol Carcin 1999; 24:153–159.
29. www.nhgri.nih.gov/DIR/Microarray.
30. Chen Y, Dougherty ER, Bittner ML. Ratio-based decisions and the quantitative analysis
of cDNA microarray images. J Biomed Opt 1997; 2:364–374.
31. Chen Y, Kamat V, Dougherty ER, Bittner ML, Meltzer PS, Trent JM. Ratio statistics of
gene expression levels and applications to microarray data analysis. Bioinformatics 2002;
18:1207–1215.
32. Bushel PR, Hamadeh H, Bennett L, et al. MAPS: a microarray project system for gene
expression experiment information and analysis. Bioinformatics 2001; 17:564–565.
33. Rosenzweig BA, Pine PS, Domon OE, Morris SM, Chen JJ, Sistare FD. Dye bias correc-
tion in dual-labeled cDNA microarray gene expression measurements. Environ Health
Perspect 2004; 112:480–487.
34. Bushel PR, Hamadeh HK, Bennett L, et al. Computational selection of distinct class- and
subclass-specific gene expression signatures. J Biomed Inform 2003; 35:160–170.
35. Wolfinger RD, Gibson G, Wolfinger ED, et al. Assessing gene significance from cDNA
microarray expression data via mixed models. J Comp Biol 2001; 8:625–637.
36. Kerr MK, Martin M, Churchill GA. Analysis of variance for gene expression microarray
data. J Comp Biol 2000; 7:819–837.
37. Lin DY. An efficient Monte Carlo approach to assessing statistical significance in geno-
mic studies. Bioinformatics 2004; 21(6):781–787.
38. Hughes TR, Marton MJ, Jones AR, et al. Functional discovery via a compendium of
expression profiles. Cell 2000; 102:109–126.
39. Callow MJ, Dudoit S, Gong EL, Speed TP, Rubin EM. Microarray expression profiling
identifies genes with altered expression in HDL-deficient mice. Genome Res 2000;
10:2022–2029.
40. Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the
ionizing radiation response. Proc Natl Acad Sci USA 2001; 98:5116–5121.
41. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and display of genome-
wide expression patterns. Proc Natl Acad Sci USA 1998; 95:14,863–14,868.
42. Kohonen T. Self-organizing maps. In: Kohonen T, ed. Springer Series in Information
Sciences. Vol. 30. Berlin: Springer, 2001:501.
43. Tamayo P, Slonim D, Mesirov J, et al. Interpreting patterns of gene expression with self-
organizing maps: methods and application to hematopoietic differentiation. Proc Natl
Acad Sci USA 1999; 96:2907–2912.
44. Steiner G, Suter L, Boess F, et al. Discriminating different classes of toxicants by tran-
script profiling. Environ Health Perspect 2004; 112:1236–1248.
Toxicogenomics: Gene Expression Analysis and Computational Tools 23

45. Li L, Darden TA, Weinberg CR, Levine AJ, Pedersen LG. Gene assessment and sample
classification for gene expression data using a genetic algorithm/k-nearest neighbor
method. Comb Chem High Throughput Screen 2001; 4:727–739.
46. Li L, Weinberg CR, Darden TA, Pedersen LG. Gene selection for sample classification
based on gene expression data: study of sensitivity to choice of parameters of the GA/
KNN method. Bioinformatics 2001; 17:1131–1142.
47. Pei M, Goodman ED, Punch WF. Feature extraction using genetic algorithms. Proceed-
ing of International Symposium on Intelligent Data Engineering and Learning’ 98
IDEAL October 1998:371–384.
48. Ooi CH, Tan P. Genetic algorithms applied to multi-class prediction for the analysis of
gene expression data. Bioinformatics 2003; 19:37–44.
49. Hilsenbeck SG, Friedrichs WE, Schiff R, et al. Statistical analysis of array expression
data as applied to the problem of tamoxifen resistance. J Natl Cancer Inst 1999;
91:453–459.
50. Raychaudhuri S, Stuart JM, Altman RB. Principal components analysis to summarize
microarray experiments: applications to sporulation time series. Pac Symp Biocomput
2000:455–466.
51. Kakizaki S, Yamamoto Y, Ueda A, Moore R, Sueyoshi T, Negishi M. Phenobarbital
induction of drug/steroid-metabolizing enzymes and nuclear receptor CAR. Biochim
Biophys Acta 2003; 1619:239–242.
52. Corton JC, Anderson SP, Stauber A. Central role of peroxisome proliferator-activated
receptors in the actions of peroxisome proliferators. Annu Rev Pharmacol Toxicol
2000; 40:491–518.
53. Stuart JM, Segal E, Koller D, Kim SK. A gene-coexpression network for global discov-
ery of conserved genetic modules. Science 2003; 302:249–255.
54. Goldberg DM, Parkes JG. Lipolytic enzymes as markers of induction and differentia-
tion. Clin Biochem 1987; 20:405–413.
55. Goldberg DM, Roomi MW, Yu A. Modulation by phenobarbital of lipolytic activity in
postheparin plasma and tissues of the rat. Can J Biochem 1982; 60:1077–1083.
56. Liu S, Rivera-Rivera I, Bredemeyer AJ, Kemper B. Functional analysis of the phenobar-
bital-responsive unit in rat CYP2B2. Biochem Pharmacol 2001; 62:21–28.
57. Anderson SP, Dunn C, Laughter A, et al. Overlapping transcriptional programs regu-
lated by the nuclear receptors peroxisome proliferator-activated receptor alpha, retinoid
X receptor and liver X receptor in mouse liver. Mol Pharmacol 2004; 66(6):1440–1452.
58. Anderson SP, Howroyd P, Liu J, et al. The transcriptional response to a peroxisome pro-
liferator-activated receptor alpha (PPAR alpha) agonist includes increased expression of
proteome maintenance genes. J Biol Chem 2004; 279:52,390–52,398.
59. Sheweita SA. Drug-metabolizing enzymes: mechanisms and functions. Curr Drug Metab
2000; 1:107–132.
60. Burczynski ME, McMillian M, Ciervo J, et al. Toxicogenomics-based discrimination of
toxic mechanism in HepG2 human hepatoma cells. Toxicol Sci 2000; 58:399–415.
61. Waring JF, Ciurlionis R, Jolly RA, Heindel M, Ulrich RG. Microarray analysis of hepa-
totoxins in vitro reveals a correlation between gene expression profiles and mechanisms
of toxicity. Toxicol Lett 2001; 120:359–368.
62. de Longueville F, Atienzar FA, Marcq L, et al. Use of a low-density microarray for
studying gene expression patterns induced by hepatotoxicants on primary cultures of
rat hepatocytes. Toxicol Sci 2003; 75:378–392.
63. McMillian M, Nie AY, Parker JB, et al. Inverse gene expression patterns for macrophage
activating hepatotoxicants and peroxisome proliferators in rat liver. Biochem Pharmacol
2004; 67:2141–2165.
3
Proteomics

Frank A. Witzmann
Department of Cellular and Integrative Physiology, Biotechnology Research and
Training Center, Indiana University School of Medicine, Indianapolis,
Indiana, U.S.A.

INTRODUCTION

It is well known that xenobiotics exert their biological effects via the alteration of
protein expression, through up- and down-regulation, alteration of protein synthesis
and degradation rates, or chemically modifying existing proteins. Either way, chemi-
cal toxicants can have very complicated cellular effects—those that can be studied by
proteomics. Proteomics measures the quantitative and qualitative changes in cellular
or tissue protein expression and explores protein–protein and protein–ligand interac-
tions. Toxicoproteomics is an approach for the identification and characterization of
proteins whose expression is altered by chemical exposure, and it is complementary
to toxicogenomics.
There seems to be general agreement among toxicologists that differential pro-
tein expression information is a critically important component of a comprehensive
biomolecular (panomics) approach in characterizing, understanding, and even pre-
dicting chemical toxicity (1–3). Exactly how one should obtain such proteomic infor-
mation remains a matter of diverse opinion, given the unique design of toxicological
experiments and the often complicated responses exhibited by cells and tissues to che-
mical exposure. Nonetheless, the assumption that changes in protein expression may
provide evidence of specific mechanism of toxic effect, by increased expression,
decreased expression, or even subtle post-translational modifications, is based on
years of published observation. The prospect of using protein expression alterations
as diagnostic markers of exposure or effect, often in such hugely complex samples as
serum, plasma, cerebrospinal fluid, or urine is an additional driving force empowering
the application of proteomic methodologies in toxicity studies.
Notwithstanding the logical combination of proteomics and toxicology, this
technological union seems to be underrepresented in the scientific literature. As
Figure 1 illustrates, despite nearly 30 years of quasi global protein expression analysis
(initially by two-dimensional electrophoresis or 2-DE) and well over 20,000 published
citations, relatively few papers combining 2-DE and tox
have been published (app-
roximately 2.5%). The same is true for the more recent appellation ‘‘proteomics,’’ e.g.,
the global analysis of the proteins encoded (and many not encoded) by the genome.

25
26 Witzmann

Figure 1 Comparison of published papers cited in PubMed since 1975 using search terms
tox
, two-dimensional electrophor
, 2-DE, 2-DE, 2-D gel, pharm
, proteom
in various
combinations, and searching all fields. The results suggest that 2-DE, and other proteomic
techniques have not been exploited fully in toxicology-related studies. Abbreviation: 2-DE,
two-dimensional electrophoresis.

Since the proteomics explosion began during the last decade (4), this combination of
various protein analytical techniques has been cited in PubMed nearly 7500 times,
while its combination with tox
appears in only approximately 4% of those papers.
Whatever the underlying reasons for this discontinuity, the comparatively small num-
ber of tox-related citations may be due to the difficulty toxicologists encounter in
applying the various proteomic technologies in their experiments.

Unique Challenges for Toxicoproteomics

The bulk of toxicity testing is conducted using animal models with subsequent extra-
polation to humans. Therefore, to best exploit the power of toxicoproteomics, it is
necessary to establish and fully characterize the unique target tissue proteomes of
various species (5), including normal and sensitive human populations. This also
applies to their response to intoxication. Furthermore, proteomic toxicity tests fre-
quently take place in the context of other ‘‘omics’’ (notably functional genomics
and metabolomics), the dose–response, and the time–course. The latter two place
a unique burden on the proteomic technology selected to study differential protein
expression, and its implementation in toxicology has proven to be challenging (6).
Perhaps the most significant difficulty associated with toxicoproteomics relates
to the nature of the typical toxicologic experiment. As alluded to above, the design
often includes the dose–response relationship, a paradigm made even more compli-
cated by the single versus repeated–dose response, the time–course, interaction
Proteomics 27

assessment, complex synergisms or antagonisms, potentiation, metabolic biotrans-

formation reactions, or various combinations of these. Moreover, a toxicologic
investigation frequently includes various classes of tests—acute, subacute, subchro-
nic, chronic, and even mutagenic. Consequently, the typical experimental design is
complicated and includes large numbers of samples per experiment. This poses tre-
mendous challenges to proteomic application, regardless of which strategy or plat-
form one employs.

The Goal of Toxicoproteomics

Ultimately, the question that must be asked is ‘‘What is it that we really want to
know about protein expression in response to chemical exposure?’’ Typically, the
answer lies in globally assessing quantitative and qualitative changes in the pro-
teome. This necessitates examining increases or decreases in protein expression with
exposure, requiring relative quantitation at least and absolute quantitation at best.
Secondly, abnormal protein posttranslational modification (PTM) must be exam-
ined, requiring the assessment of the extent, chemical nature, and specific location
of the chemical modification. Any alterations in protein expression noted by the
above must be related to relevant toxic end points, rendering these as ‘‘toxico-
logically relevant’’ proteins. Finally, one must ascertain that the observed protein
changes occur in both the animal model and the human, requiring the investigation
of homologous systems. The greatest challenge to toxicologists lies in selecting a pro-
teomic platform that best addresses these issues and provides proteomic information
with interspecies relevance.
In a toxicologic target cell or tissue, each expressed protein is unique and the
final, fully functional protein product rarely resembles the gene(s) from which it
was coded. Therefore, a comprehensive approach that includes a variety of proteo-
mic techniques is necessary to detect, quantify, and identify individual proteins, to
elucidate protein activities, and to determine protein–protein interactions. Unlike
mRNA analysis, protein analysis is severely limited by the tremendous dynamic
range of protein expression and the large number of relevant proteins whose cellular
expression or abundance in body fluids is minuscule.
This chapter will address several of the core technologies that comprise con-
temporary differential expression proteomics most relevant to toxicology, focusing
on those whose protein analyses are quantitative and where they have been applied
successfully in toxicologic investigations.

DIFFERENTIAL EXPRESSION PROTEOMIC TECHNOLOGIES

2-DE
It is still considered by many to be the most powerful and readily applicable of the
proteomic techniques, 2-DE involves the polyacrylamide-based separation of com-
plex protein mixtures first by protein charge (i.e., isoelectric point or pI) via isoelectric
focusing (IEF) in a pH gradient, followed by mass separation in the presence of
sodium dodecyl sulfate (7). In 2-DE, resolved proteins are visualized by any of a vari-
ety of stains or dyes, and generally, comparative quantitation of protein expression is
conducted by image and statistical analyses. Distinctively, this technique is simulta-
neously analytical and preparative. Separated protein spots can be cut from the
gel, digested proteolytically, and the resulting peptides analyzed further by mass
28 Witzmann

spectrometry (MS) (8). Online peptide-mass or -sequence database comparisons then

identify and characterize the proteins [see section on MS and Tandem MS (MS/MS)].
PTMs are frequently accompanied by changes in protein charge and are easily
observed by their altered pI. In 2-DE, protein phosphorylation, glycosylation,
and chemical adduct formation—three of the most relevant PTM in toxicologic
studies—frequently generate multiple spots of differing pI so that individual proteins
may be represented numerous times on a 2-D gel. Consequently, a gel depicting 2000
detected protein spots should not be misinterpreted as having resolved 2000 unique
proteins. Instead, a significant number of spots are charge isoforms representing far
fewer actual gene products. Nonetheless, 2-DE is one of few proteomic approaches
capable of readily detecting PTM, quantitatively comparing the extent of PTM, and
enabling the determination of the PTMs’ chemical nature.
The combination of 2-DE and MS constitutes a powerful platform for protein
expression analysis. Unfortunately, 2-DE has several shortcomings that have
spawned a number of chromatographic and mass spectrometric approaches designed
to overcome these deficiencies and render protein expression analysis truly global.
The major weaknesses of the 2-DE approach include an inability to resolve very
hydrophobic proteins and those with extremes of pI (particularly basic proteins).
The hydrophobicity issue is difficult to overcome, as it is in all proteomic
approaches, while the pI issues continue to be addressed by technical developments.
Another major problem that has limited 2-DE’s applicability is its poor dynamic
range. This is due to a combination of factors, including limited physical space for
protein separation (gel-format) and protein detection (stain or dye sensitivity). If
the range of the putatively expressed > 100,000 different cellular protein forms
resembles that postulated for plasma, greater than nine orders of magnitude (9),
researchers using this approach will remain incapable of analyzing physiologically
relevant, low abundance proteins altered by chemical exposures. Finally, toxico-
logists who choose 2-DE to make relative quantitative comparisons between numer-
ous individual groups of protein samples quickly realize that gel-to-gel variation is
an issue. While the underlying reasons for this are numerous, this and other difficul-
ties can be overcome.

Addressing 2-DE’s Limitations

2-D DIGE and Multiplexing. To enable the separation and differential expres-
sion analysis of two or more samples on a single 2-D gel, fluorescent two-dimensional
difference gel electrophoresis (2-D DIGE) was developed (10). Despite having its
own limitations (11), this approach overcomes most gel variability issues and can
be useful in small experiments with limited numbers of samples for comparison. In
DIGE (Fig. 2), complex protein samples are labeled with fluorescent cyanine dyes
Cy2, Cy3, and Cy5 and the samples mixed so that they can be examined in a single
2-D gel with each dye giving an independent channel of measurement. For each spot
detected in the gel pattern, the intensities in the respective dye channels are com-
pared, and ratios are calculated from these intensities to indicate the extent of differ-
ential protein expression. With DIGE, the uncertainty of coordinate gel matching
across two or more gels is overcome, for the most part. With the advent of saturation
labeling (12,13), 2-D DIGE has demonstrated unprecedented sensitivity. For
instance, Stühler et al. (14) recently used this approach to detect approximately
2400 protein spots on 2-D gels loaded with protein from approximately 5000 cells
(~2 to 3 mg) microdissected from precursor lesions of pancreatic adenocarcinoma.
Proteomics 29

Figure 2 Schematic of a 2-D DIGE experiment with an internal pooled standard and three
fluorescent dyes. Samples A and B are labeled with either Cy3 or Cy5, and a pooled internal
standard is also constructed from equal amounts of all the samples in the experiment and
labeled with Cy2. After mixing these protein samples and performing 2-DE, the protein spot
patterns can be visualized by illuminating the gel with the specific excitation wavelengths.
Samples A and B protein spot intensities are then normalized by dividing each by the corre-
sponding spot intensity of the pooled internal standard. Analyzing the normalized spot inten-
sities enables the detection of subtle differences in protein expression levels with a higher
statistical confidence. Abbreviations: 2-D DIGE, 2-D gel, fluorescent two-dimensional differ-
ence gel electrophoresis; 2-DE, two-dimensional electrophoresis. Source: From Ref. 11.
30 Witzmann

DIGE has been used effectively in several toxicology studies. The mechanism
of acetaminophen hepatotoxicity has been characterized and the DIGE approach
optimized (15,16), and preclinical studies resulted in the discovery of predictive bio-
markers of compounds with a propensity to induce liver steatosis (17). Taking
advantage of running liver samples from hydrazine-treated and control rats on a sin-
gle 2-D gel, Kleno et al. (18) applied multivariate analysis to discover new biomar-
kers of hydrazine hepatotoxicity. Finally, combining the power of DIGE and rodent
airway epithelial cell isolation, Wheelock et al. (19) improved the applicability of
2-DE-based proteomics to respiratory toxicology by demonstrating the enrichment
(by 36%) of epithelial cell-specific proteins and resolving 2365 detectable spots.
A related approach avoids Cy dyes but incorporates the multiplexing strategy
by means of multiple, repetitive staining of individual gels to detect all proteins in
the 2-D gel pattern, along with phosphoproteins (20) and glycoproteins (21). This
methodology enables parallel determination of altered glycosylation and phosphory-
lation patterns and protein expression level changes without running three gels.
Large-Scale, Highly Parallel 2-DE. An alternative to multiplexing samples in
2-DE by DIGE is found in highly parallel 2-DE separations. While current maximal
capacities of commercially available 2-DE apparatus are limited to 12 gels per run, the
ISO-DALT System (22,23), in which 20 large-format 2-D gels can be cast and run
simultaneously, is uniquely well suited to overcome gel-to-gel variability. The separa-
tion of up to 20 samples per run (or 10 replicates per run) directly addresses the
demands placed on 2-D gel–based proteomics by complicated toxicologic experimen-
tal design. For example, recently, we have analyzed the effect of various concentra-
tions of hydrazine and cadmium exposure in vitro in rat hepatocyte primary culture
(Witzmann, unpublished data) using multiple runs of the 20-gel ISO-DALT System.
Using PDQuest 2-D Gel Analysis Software (BioRad), 144 individual sample gels
representing 144 individual wells from 24 six-well culture plates were analyzed. Figure 3
illustrates the composition of this matchset and the remarkable pattern reproducibility
achieved. An average of 1100 proteins were matched in this gel set, 415 of them in
every single pattern. Whether the toxicological experiment is larger or smaller, using a
highly parallel platform, technical limitations now seem to rest in detection sensitivity
and image analysis capacity, not in the reproducibility of electrophoretic separation.
Increasing Analytical ‘‘Depth of Field’’—–Sample Prefractionation. With the
ability to run literally hundreds of gels with acceptable consistency, 2-DE analysis
in toxicoproteomics is poised to take advantage of another trend in proteomics—
the reduction of sample complexity. It is clear that global protein analysis
across the range of protein expression is currently impossible, no matter which
proteomics platform one applies. Consequently, the strategy of examining subsets
of the proteome has gained significant momentum, and methods to reduce sample
complexity in all manner of proteomic studies are becoming routine (25).
In toxicoproteomics, reducing sample complexity generally involves decreasing
sample heterogeneity and improving the analytical ‘‘depth of field’’ by digging dee-
per into the proteome. For example, rather than analyzing protein expression in
whole liver, kidney, or brain samples where numerous cell types reside, one uses care-
fully microdissected regions obtained by laser capture (26) or cells isolated by tradi-
tional collagenase treatment and centrifugation. Increased depth of field is achieved
through the analysis of cell organelles, specific membrane fractions (27), or multiple
subproteomes (28). One can also study specifically enriched proteomic subsets gen-
erated by depletion of highly abundant proteins, as in serum or plasma (29,30) and
urine (31), by preparative solution phase sIEF (32–34) or by chromatography (35).
 Proteomics

Figure 3 Screen-capture of PDQuest montage illustrating 99 of 144 rat hepatocyte 2-D gel images (plus 1 reference pattern, first frame) analyzed in a single
matchset (100 frames/display window is maximum). Hepatocytes were exposed to hydrazine or cadmium in a range of exposures in six-well plates, solu-
bilized in situ after removal of media, separated by 2-DE using the ISO-DALT System (20 gels/run), and stained with colloidal Coomassie blue, as referred
to in the text. Each frame illustrates the same region of the 2-DE gel pattern, containing calreticulin precursor (upper left in each frame) and hsp60
31

(middle right), among others. Approximately 1100 protein spots were matched in each pattern, with 415 spots matched in every pattern. Abbreviation:
2-DE, two-dimensional electrophoresis. Source: From Ref. 24.
32 Witzmann

In photography, the phrase ‘‘depth of field’’ can be defined as ‘‘the distance

range between the nearest and farthest objects that appear in acceptably sharp
focus.’’ Its use (or misuse) here is intended as a point of emphasis with respect to
the vast dynamic range of protein expression and our desire to ‘‘sharply focus’’ on
and accurately analyze expression of the least abundantly expressed protein along-
side the most abundant protein, and everything in between. To improve analytical
depth for 2-DE, sIEF provides perhaps the best and most attractive remedy. This
approach subdivides a complex protein mixture into well-resolved fractions based
on the pI of each protein in the mixture. This is accomplished in a conveniently
low-volume platform (32,36) using series of chambers connected in tandem and sepa-
rated by thin membranes that contain covalently attached buffers of defined pH
(immobilines). The protein sample is loaded into the chambers separated by these
disks and spacers and subjected to sIEF. The result is a significant reduction in ori-
ginal sample complexity in the form of a set of five highly resolved protein fractions
suitable for further analysis by 2-DE.
In our laboratory’s 2-D gel–based toxicoproteomics efforts, we have begun to
address this issue in various proof-of-concept studies by prefractionating samples
using a combination of subcellular fractionation (by differential centrifugation) com-
bined with further fractionation using sIEF. We hypothesize that by combining the
two fractionation techniques, a significantly greater portion of a cell’s or tissue’s pro-
tein complement can be analyzed. Figure 4 illustrates the components of this
approach and emphasizes the need for a 2-DE platform capable of handling the large
number of samples. The successful application of this approach is heavily dependent
on reproducibility in the fractionation technique as well as in 2-DE. Using a large
capacity, highly parallel gel system makes this a viable approach. It is unlikely that
a more limited-capacity platform or multiplexing strategy would be sufficient.
In a preliminary study comparing baseline protein expression in rat hippocam-
pus and nucleus accumbens in various strains of rats determined by 2-DE (37), it was
readily apparent that substantive differences between these proteomes occurring at
low levels of protein expression were not detectable, mainly because of a lack of
depth of field. To rectify this limitation, we have proposed the approach illustrated
in Figure 5. Recent results in cerebral synaptosomes (38) and sIEF fractions of
nucleus accumbens (Bai et al., unpublished results) support the final protein spot
numbers speculated in Figure 4, where 20,000 to 30,000 protein spots are putatively
resolved from a single brain region sample (albeit on 20 individual gels per sample).
For example, in an initial experiment, 200 mg synaptosomal and cerebral cyto-
sol protein loading resulted in the detection of roughly 1000 and 1500 protein spots
on broad range 2-DE, respectively. Heavier gel loading (approximately 1 mg) would
have increased that total significantly. In a separate analysis using the commercial
rendition of Zuo and Speicher’s (33) original concept of sIEF fractionation, the
Zoom1 IEF Fractionator (Invitrogen), a rat brain cerebral protein fraction pH
4.6 to 5.4 resolved by narrow-range 2-DE (24 cm, pH 4.5–5.5), yielded 1250 protein
spots. By combining the two fractionations and loading significant protein amounts,
an optimistic estimate of 20,000 to 30,000 resolvable protein spots per brain region
seems plausible, corresponding to a significant improvement in depth of field. Studies
are currently underway to substantiate these predictions.
In another preliminary assessment of the prefractionation strategy, in the pro-
karyote Escherichia coli, we subjected cell lysates to sIEF, and separated the resulting
fractions by large-format, narrow-range 2-DE (Fig. 6) (24). Consistent with previous
observations in E. coli 2-DE in the literature, broad range IEF (pH 3–10) resulted in
Proteomics 33

Figure 4 Proposed application of a multi-step sample prefractionation approach for com-

plexity reduction and increased proteomic depth of field in assessing the toxic effect of alcohol
exposure in the brain. The approach includes (1) brain region dissection NA, HIP, AMG,
STR, and PFC, (2) subcellular fractionation by differential centrifugation, and (3) solution
phase isoelectric focusing. Consequently, the proteome of each brain region is represented
on 20 individual 2-D gel patterns. Based on preliminary results, we anticipate the resolution
and analysis of 20,000 to 30,000 protein spots—a significant improvement over single sample
analysis. Abbreviations: 2-D, two-dimensional; NA, nucleus accumbens; HIP, hippocampus;
AMG, amygdala; STR, striatum; PFC, prefrontal cortex. Source: From Ref. 24.

the detection of 1577 protein spots. In contrast, the sum of five sIEF fraction gels
totaled 5525 proteins, another improvement in depth of field. These kinds of studies
suggest that 2-DE, when applied to eukaryotic systems in the manner just described,
continues to have powerful utility in toxicoproteomics.

MS and Tandem MS
Gel-separated protein spots can be identified following proteolytic digestion and ana-
lysis by any of several MS methods. For instance, peptides resulting from a tryptic
digest can be analyzed by matrix-assisted laser desorption ionization time-of-flight
(MALDI-TOF) MS, a process referred to as peptide mass fingerprinting (PMF)
(39). The measured and optimized monisotopic mass data are then compared with
theoretically derived peptide mass databases generated by applying specific enzyma-
tic cleavage rules to predicted or known protein sequences. MALDI-TOF–based
PMF enables high-throughput, accurate, and sensitive mass detection. For unambig-
uous identification of 2-D gel–separated proteins, peptide fragment spectra generated
by collision-induced dissociation (CID) and detected as MS/MS spectra can be com-
pared to spectra predicted from sequence databases using search engines, such as
Mascot, (40) and algorithms, such as SEQUEST (41). With the advent of the
TOF/TOF (42,43) and FT-ICR (44) MS, protein quantitation, identification, and
Exploring the Variety of Random
Documents with Different Content
 "Bashful or not, they were lots of fun. Just wonder what we
did that scared them away?" "Maybe," Jan who was younger but
wiser, answered, "maybe it was something that we didn't do!"
"Well," Barbara replied, "if they don't come back, they'll never know
what they’re missing . . . tomorrow is artother night! cc UJ 0. < u 65
 Always the outdoor girl, sweet Susan — a flower in her own
right — spent years growing rare tea geraniums. So much tike her
are these pink flowers with fragrant, lobed leaves, they almost
describe our Sue. However, S.S. likes all things that grow and
develop — plant, animal — and especially man. So, if you are o man
and want to spend a pleasant moment or two, drop in and take a
peek at Susan's flower bed— you won't be sorry you did.
 Susan obviously doesn't spend all her time in the hot house
— although the men who know her will tell you that she radiates a
heat all her own. So, between botany and boys Susan is kept well
occupied — ond derives a great deal of pleasure from each. 67
 (£ < O OVER 2 MILLION COPIES SOLD! FOR THE FIRST
TIME, HOW AVAILABLE IH AMERICAN BOOKSTORES! OVER 420
PHOTOGRAPHS! OVER 50 DIAGRAMS A RENOWNED DOCTOR
WRITES A FRANK, HONEST BOOK ABOUT FULFILLING MARITAL
RELATIONSHIPS! A HAPPIER SEX LIFE STDR.SHAKOKKEN Dr. Sha
Kokken's unique, photographic approach to marital happine.'is
skyrocketed this book into sales of 1.700.000 copies in Japan alone.
First published in Japan six years ago. the book is currently enjoying
tremendous success in Germany. China. England and the
Scandinavian countries. And note, this book is oim'fa6/e l
 “But I wanted to. Can’t you tell?” “All right.” “You say all
right to everything?” “It depends.” She gave a throaty chuckle,
moved up close, laid one thigh against his leg, and said, “1 want you
to love me. How about that?” He grinned at her. “I should be at the
shop. Suppose your husband drops by?” “I don’t give a damn.” “You
always do this to him?” “No, this is the first time.” She moved closer,
put one arm around his neck, and kissed him on the mouth. She
leaned into it, then drew away, “C’mon,” she said. “There’s nobody
here.” He could see their reflections in the pink mirrors. He looked at
her, grinning. She was out to get what she was certain she could
have. She worked all day long, running to dog shows, cat shows, gin
teas, cocktail mauls. Jose was becoming tiresome, and her natural
vibrance excused her suddenly wanton eagerness. He suddenly
didn’t give a damn. He knew she' didn’t. He reached for her, and
kissed her. She gave a sigh and slipped her tongue into his mouth,
moving her legs slightly apart. The straw purse struck the floor.
“Strip me,” she breathed. He begaii peeling off that black bikini, his
knuckles digging into the soft, lush, scented flesh. When she was
naked, he marveled at that body. She was panting, now, her large
breasts undulant, the provocative curve of her hips an eager
promise. She writhed against him with a wild concern, an avidity so
sexual he was immediately swamped. She thrust against him,
working her hips. “Oh, Jim,” she whispered. Pulling him over to the
bed, she sprawled down on the crimson spread. Immediately very
soft music began playing, and he realized the mattress held a
switch. He heard a gentle humming. As he stretched out beside her,
long, nervous, silver-tipped fingers tore and twisted at his belt,
quickly unzipped his pants. “Do it good— I can’t wait — do it fast,”
she said. Naked, they came together, rubbing, cleaving in abrupt,
wild passion. Her hands were all over him, seeking. Her mouth
sucked his throat, her tongue darling, and she moaned deep down,
and there was an untame, curiously frightened look in those slanted
blue eyes. Jim Richards loved her the way she obviously wanted,
needed. The lush mattress oozed beneath their furious bodies, and
he had never had a woman so crazy for love. Little throaty moans
began coming from her, her eyes rolling in her head, and then she
began to [continued on page 7 1 1 SECRETS OF SCANDINAVIAN
SEXUAL POW^: an illustrated book of 150 rare marital position^ „
DYNAMIC SEX An Exciting Sexual Breakthrough Crammed to the
Brim, Page after Page 300 Explicit Illustrations. Now discover the
most exquisite, intimate details of technique and sex, as performed
in Scandinavia. Read only a few pages at random and you will see
why DYNAMIC SEX by Karl Jacobsen could never have been written
by an American. Explore these sensational pages and learn what
exotic adventures await you. It is then you will discover what others
have: Scandinavian uninhibited technique draws forth your sexual
powers to their very fullest, sometimes even beyond, and brings out
in any man or woman more than you -'could ever imagine, sweeping
away every inhibition and restraint. I SEX FOOD by Ffin FcUfssii
PRACTICALLY EVERY POSSIBLE WAY IN WHICH THE HUMAN BODY
CAN BE SEXUALLY AROUSED IS INCLUDED— WITH DARING
PICTURES. Have you ever tried ‘Riding the Stallion', 'The Panther's
Kiss’, and 'The Coital Boomerang’? If you haven't, you haven’t really
lived! They’re all here. Africa and Asia. The sensual pleasures
revealed in DYNAMIC SEX are so enormous, so staggeringly varied,
one delighted connoiseur has called it a 'Sexual Smorgasbord!'
LEARN WHAT IT IS LIKE TO REACH SEXUAL FULFILLMENT AGAIN
AND AGAIN AND AGAIN AND AGAIN! Learn for yourself those
legendary techniques for unleashing sexual power . . . pure, raw,
glorious love power. Power to open up an exciting new world of
erotic marital pleasure, leading you into unexplored areas of ecstasy.
Engulf yourself in the teachings of DYNAMIC SEX. Saturate yourself
with the pleasures of this brand of supercharged sex! SEX FOOD by
Fritz Peterson Europeans through the centuries, to this very day,
have regarded certain foods, drinks and recipes as the fuel of love—
able to stimulate the user to unusual heights of sexual power and
body energy. Now, on an exclusive basis you will receive FREE, a
copy of SEX FOOD by Fritz Petersen, when you purchase DYNAMIC
SEX. A SAMPLE OF WHAT'S WAITING FOR YOU!: • Numerous ways
of harnessing your sexual potential! • The art and science of making
love in the nude! • Intriguing sex games to play! • Highly
unconventional ways to stimulate a woman with your lips! • How to
use ice to. obtain super-charged coition! • Original methods to drive
an experienced woman to new, unimagined heights of frenzy! •
Rough but effective ways to break down a woman’s inhibitions! •
Complete guide to genital twitching, for strange new sexual
sensations! • Specific ways to work your wife to new peaks of almost
unendurable passion! • Invigorating, erotic types of massage— can
work wonders for any man! • New, proven techniques to combat
premature ejaculation! • Unique, tantalizing ways a woman can
stimulate you! ■ OFFER AVAILABLE TO MATURE ADULTS OVER 21 ■
NOVEL PRESS Dept CP ic 31 Second Ave., New York, NY 10003
Please rush the following in plain sealed wrapper: Q Dynamic Sex —
only $4.95 -|-50d PP [ ] p'lus a free edition of ‘The Art of Love’ and
SEX FOOD - - .'2nd BONUS! ‘ART OF LOVE' by R. BurtonN ’ orjg.
published at $5.00! The amazing book on the secret love techniques
of the Afro-Asian world, which describes the INCREDIBLE ‘SPINNING
TOP vTECHNIQUE’. Handsome hard-cover edition. enclosed find: I
OcashQ] check [j money order I Q 1 am over 21 j City _ -State. .Zip.
(C LU 0. < U 69
 $3 Brings you SO ft. of my . film 2nd free reveelins L
brochures. SUPER ujomeo^GiRtS OPENLY POSED t you like them -in
INTIMATE, ' NUDE Bedroom Scenes} e 200 Ft. Color Feature Film
l.Wa5S5»
 SWING WITH ME [continued from oage 69] squeal, digging
fingernails into his back. Suddenly, she burst into uncontrollable
laughter, those eyes streaming tears, panting, saying, “You bastard—
you bastard, love me good— oh, you wonderful big bastard— you’re
a stallion — don’t stop!” Everything about her collapsed except her
hips. They still pressed, as if searching for a last twinge of ecstasy,
and then suddenly she fell back with a satiated groan. She stared at
him, and he could actually see the slow, lazy return of that natural
disdain. He marveled at it. At first he could not believe it, but then it
was obvious. The only thing was, she was forcing it. She knelt on
the bed, still eyeing him, then climbed off the bed and looked down
at him. “Thanks,” she said. “Think nothing of it.” He didn’t tr>’ to
figure her. It had been too good. She was everything she'd
promised, and more. He would not forget her. She moved, picked up
the flimsy bikini, dressed quickly, ran fingers through that thick
wealth of hair, rubbed her thigh with the palm of one hand. “All
through, now?’’ he asked. She lifted her chin. "You're very good.” “So
are you.” “I’d rather you'd forget it.” All the cool quality had returned
now. She was a poised, aloof woman. “Baby," he said. "You slay me.
Really. You came and got me, brought me here, begging for it — ”
“Stop!” “You practically raped me—” “Stop, I tell you!” “You loved
every blessed minute, and you won't even admit it to yourself.” She
gave a little toss of that beautiful head. He got off the bed, reached
out and grabbed her arm. “You want it again?” he asked. “That
what's the matter?” She began to weep, silently, standing there. Fat
tears worked from her eyes, and she watched him through a mist.
He let go her arm. Her voice was suddenly soft. “When 1 said
thanks,’ I really meant it.” She covered her mouth with her hand,
then took her hand aw'ay, and smiled. “Honest, Jim— you’re the
best.” She paused. “Come with me.” She took his hand, tugged.
They moved to the mirrored far w’all. She pressed a button and the
mirrors folded back, revealing a shadowed, small room. “Inside,
Jim.” [continued on page 7 2] WHAT HER "TEACHERS" DIDN'T
KNOW ABOUT THE EXOTIC... SHE TAUGHT THEM! Neighbors in the
Southern California city called them the Gay Sisters — the pair in the
secluded house down the block. But Deborah and Sandy were
beyond all care as they shared their most intimate desires within
their sensual world, devotees to the pursuit of physical sensations.
When their advertisement brought the delicate beauty named Elicia
to their door, they couldn't believe the shy teenager's innocent desire
to share their sensate world, to be slave to their unbridled demands.
But their "student" became their teacher as this wanton nymphet
became passion's mistress, enslaving the sisters to her own sensual
demands. ORIGINAL UN-RETOUCHED VERSION WITH
UNRESTRICTED FULL VIEWS 8MM BLACK A WHITE 219 FT- ORIG
$25.00. NOW ONLY $9.95 j HOVEL PRESS, dept. cp502 I 31 Second
Ave. New York' N.Y. 1IM3 I Please rush me the gay sisters [ I
enclose $9.95 Cash □ Check □ [ M.O, plus 50( P.P. ! □ I am over 21 !
NAME e { ADDRESS ! CITY STATE ZIP
^mmmmmmmmmmmmmmmmmmmmmmmmmrn formbu ^IVIEil
tt707 BRIEF ZIP This Brief' zipper front. Soft Iwo-woy stretch Nylon
Heloncp with □ dual purpose. They double os senscjtlonal sr-ioi
trunks! White, Block. Nude, (Word of Warning, becomes traosporentl
| $3.99 Save! TWO FOR $7.50 REGENCY SQUARE • Dept. L6224 1
800 North Highland Hollywood. Calif. 90028 exciting Mve a privite
55? a letter from • me tool ' Rubi £ Jan Suite i‘t75 234 5th ave. N.Y..
N.Y. 10001 "FRENCH PHOTOS fRlia Guaranteed to surprise you! To
cover mail costs only send 25
 uncommon products fOR„ I Our business is the se(AARRIED
\ curing of UNIQUE PERNtiN J SONAL ITEMS for married men ONLY.
Now available . . . complete selection of hard-to-find products.
Married men, send today for ILLUSTRATED pictorial catalogue and
future descriptive mailings. Enclose 25c for handling. 0«pi.C269 80S
South Robtrtson Blvd. Los Angeles. CeM. 90035 FILMS Mil OFFBEAT
I Sold to adulte oxOj’* Swinging fllM that are real different and
offbeat* Sai^le fLln atslp I (enoogb to reallx know) |1*| or , fell
action aaiq>l6 $^both sent I with illnslrated catalog* Order now*
Ton mst be 21 or orer* OF, Sui*el475 234 FIFTH AVENUE, NEW
YORK, N. Y. 10001 MEN ONLY! We hove the most unusual items and
novelties for men ever offered. Sample ossortments, only $2.00.
Cotolog only 25c, refunded on first order. ARTCO MF6. CO.. Dept.
C269 BOB S. Robertson Blvd., Los Angelas, Calif. 90035 ^ey s/fU
Buy my pictures I’m anxious to sell my collection of privately posed
pictures, taken under intimate circumstances (all nude). Must sell —
Need money quick. Sample set and personal letter ST.OO. ADULT
MEN ONLY! Jack! Joy Apt. #1475 234 Fifth Ave.. N. Y., N. Y. 10001
MOVIES (Omm, Super 8, Color, B&W, Extra Lengtli) The greatest
entertainment movies ever produced for the eclectically inclined
urban male (who appreciates unadorned feminine beauty} are
available now — along with 3Smm and 3-D slides. Satisfaction
guaranteed. Write to: INTERCONTINENTAL DISTRIBUTING CO. F.O.
Box 2149 Hollywood, Calif. 90028 UDCommon PHOTOS All types
available women & • men, also groups. Sample set and • excitingly
illustrated catalog * $2.00 Swingers only (21 & over) « Write to . . •
HI-LITE Dept.1475 • -^234 5th ave. N.Y.. N;y. 10001 J ^-
jMagaiSnes-^ from DENMARK Imported from DENMARK. The kind
that are not published in this country. Not nudist, but intimately
posed m-door photos completely un-retouched, and from the
country where censorshjp is un-known. Sample magazine and
ILLUSTRATED BROCHURE $3.00. ADULTS ONLY! Dansk Sales Dept#
1475 234 5th Ave, New York, N.Y. 10001 7T976 LONG LINE #977
SHORT LINE Lycra Power Knit support briefs that stretch with you,
breath with you. “Magic-Grip" power panels trim Inches & give solid
masculine support. Sizes S (28-32), Med. (3336), Lg. (37-40), XL
(41-44), . Order «976 LONG or #977 SHORT EdCll $6.99 For C.O.D,
RUSH ME THE FOLLOWIN6: #976 □ #977 □ Add 50c for mailing.
Enclosed Is $ REGENCY SQUARE Suite X6224
^^M^^J^^fighlandj^^foM^wood^^Colif^OW^^ have all types of
‘privately" printed magazines . . . For per• EXFERENT imported
Scandinavian imports and "selected" domestic publications that
cannot be bought elsewhere ! ! ! N»uraily illustrated catalog 25 4
Adults only. HAMILTON HOUSE, Dept 1475 234 Fifth Ave. N.Y., N.Y.
10001 ] Brings you THREE BOLD OARING ADULT BOOKS over 500
pages • SWAPPING SAMPLES < ANAL EROTICISM -ORAL LOVE also
receive free Adult Biwks & Film Brochures PRIVATE COLLECTORS
dept, cio ’ P.O. Box 46608, Los Angeles, Calif. 90046 SWING WITH
ME {continued from page 71] She lit a light. He saw a movie camera
fastened to the inside wall. It hummed. She turned it off. It had
obviously been taking film of what happened in the bedroom, its
lens at a sm^l aperture in the big oil painting of the nude man and
woman. “Look,” she said. She pointed out another slot in the wall,
through the painting. He pressed his eye to it. He saw the entire
bed. An easy chair stood in the room, just high enough for the peep
hole. “This is Joses real life,” she said, her voice flat, “You mean, he
forces you to . . . ?” “Not me, darling. He hires them, takes pictures,
watches. His kicks, understand. He has reels and reels and he makes
me look at them, nights.” She paused. “No. Not me.” She smiled at
Jim Richards. “I’m special, you see? I’m a vase on the mantel, a
lamp beside a chair, a picture on the wall.” “You mean— he can’t —
?” “That’s right, Jim. He can’t. A toughie, huh?” “But the humming.
The camera, just now.” “I took pictures of us, Jim. I’ll have
something to watch in private, and to remember.” Her blue eyes
were suddenly blacker, hot. “Jim,” she said, stepping close. She
pressed against him. “Oh, Jim— please— ” It was immediately as
strong with him as with her. In seconds they were half on the easy
chair, half on the floor, and her body strained, her hips working, and
she moaned and laughed and cried again. It was savage, and it
lasted a long, groping, thrusting, animal-like time, and he knew he
would have to travel a long way to find the hkes of her again. She
drove him back to the shop on Redington, and for a moment he
hesitated before leaving the car. “Now you know why I said ‘thanks,’”
she told him. “I really meant it, Jim Richards.” He got out. She
reached, touched his hand, smiled gently. The gleaming Lincoln’s
engine furred, and the car drew swiftly away. The next day he saw
her in the car, sitting there as Jos^ Martine drove past the shop. Jim
Richards was out by the highway. Jose ignored him, but as the car
flashed past, her head turned and one eye winked. Jose Martine was
smoking a long black cigar. Jim Richards never saw her again,
because two days later he dropped everything and left for California.
On the plane, munching a steak, he suddenly remembered that he
didn't even know her first name. He shook his head and grinned,
signalled the pretty blonde stewardess for more coffee. Life was hke
that. • 72
 PARKER [continued from page 9] his hand. He turned and.
nodded. Soon he lay beside her. They proceeded to drink up. “It
doesn’t pay to be literary, I guess,” he said. “What? Oh, that. Schiller.
No, I’ve become very basic in approach.” Brusquely — “Do you like
my nudes?” He looked at the sketches before answering. “It’s the
same chick,” he said. “What do you see?” she asked. “She’s getting
old. Fleshy in a Renoir way. There’s something frantic in her
expression. Lips and eyes. She seems to be” — he stopped. “What?”
“Nothing.” “Tell me.” “It’s really nothing. Honest.” She waited,
thinking he might change his mind. “The others couldn’t articulate
about the sketches,” she said finally. They finished the glass in
silence, she watching the wine contemplatively, he not taking his
eyes off her. He wanted to talk to her, but was stuck for what to say.
Once she shifted her eyes onto him, but turned away. The silence
lengthened. They drank another quickly. She nudged him and made
a puffing gesture with her lips. He found the joint and matches by
feeling along the night table. He lit up, took the first drag and
handed the joint to her. They passed it back and forth, taking a long
suck and letting it fog their heads. He did not touch her, but he could
feel the warmth of her thigh against his, smell the expensive
perfume she was wearing, and it made him stretch his body lazily in
pleasant anticipation. She did not move, only listened to a ticking
clock on the night table. Then the phone again, jarring her from her
smoky calm. “Parker,” again testing its sound, as she talked through
the ringing. “The artisan,” he said. “Artisan. Quite nice. Parker the
artisan. He makes things.” “Damn fine things,” glancing at the
phone. “You’re high, aren’t you, baby?” “I suspect I am.” “Yoii must
be. Regular loquacious you are.” “I don’t know your name,” he said.
The phone stopped ringing at [continued on page 74] brown BABY
DOLL... A little girl in a big, beautiful, voluptuous body (40-25-37).
with skin like liquid honey! Come drink your fill. 8 soul stirring
COLOR pics all shot from unusual angles. $3.00 with excitingboldly-
illustrated catalog. Only adults may write to*«« MISS COCOA
BROWN 234 5th AVE. RM. 1475 N.Y..N.Y. 10001 THE ONLY
UNRETOUCHED MALE m NUDIST • Publication of its Kind! , § \ SEND
$5.00 Fnr Ynur Cnpy Nnw! Sn ff Send S1.no for Full Color Brochure
(Refundable on first nrder) tn: WYNGATE & BEYINS, INC. Dept. CIO
6311 Yucca Street, Hollywood, Calif. 90028 00 jo innnrinnrc" i hhatiy
iPrintea ^GIRLIE MAGS The kind you can't buy any where else!! °
Exciting sample selection $3.00 FREE % with order illustrated
brochure. <> H. I. P. SALES, 234 5th AVE. oept 1475“ N.Y.,N.Y.
10001 Sex Parties of the "Swinging Set" Openly Exposed in this New
8mm Motion Picture! \V RITTI-S' GVARAS'TEE with each order PART
1 200 FT , $20 00 PART 1 plus PHOTO SET $25.00 PART 1 & 2 400
FT $30.00 PART 1 & 2 plus PHOTO SET . . $35.00 PHOTO SET — 20
GLOSSIES $10.00 SAMPLE COLOR FILM $ 1.00 VIEWS, 5332
SUNSET BLVD., DEPT.^P? HOLLYWOOD. CALIFORNIA 90027
INTRODUCTORY^ OFFER FULL SHOW MOVIES $2.00 □ □ , MAN &
WOMAN WOMAN & WOMAN WEDEN FILMS • U.S. DELIVERY BOX
942. NATIONAL CITY, QAL. 92050 THIS WOMAIV SUBMITTED TO A
SHOCKIlUG SEX TEST 8MM FILM NOW AVAILABLE FOR QUICK
SALESl.OO FOR ACTUAL PHOTOS AND COMPLETE INFORMATION
269 LESLIE, P.O. BOX 61. GLENDALE, CALIF. 91209 i SHARE OUR
SCENE We're a married couple and we know what's happening. We
offer photos our activities. We are 30 and 24. Drop us a note and
send $1.00 for film/ BDM&& A uAi samples. Indicate DKUliWatVAL
your choice. 6311 Yucca St., Dept. 269 Hollywood, Cal. 90028 □ □
''ON FILM ANd^ iNJOY IT! COUPLE PLAYS WITH FRIENDS AND
OTHER COUPLES $1.00 FOR 8MM — 25# FOR 4x5 ■ PHOTO JIM &
LOIS 5334 SUNSET BLVD., HOLLYWOOD, CA. 90027 FILM For
sophisticated -couples & singles everywhere who value discretion
and confidence: A modern club devoted to unusual, exciting
correspondence with friends whose interests are similar. fhek::
Details and List of Sample Ads SELECT CLUB BOX 889, DEPT 39
CAMDEN. N. J. SHARE 1 .'’l^fli^EXPERIENCES ^ ¥ 1 P 1 CTU R E S
ET $!.□□ 1 1 IEXPERIENCES Dept. 269j MYi \ |63ll
mCAST.LDSANGELES.CALSDDai 1 FREE SAMPLE PHOTOS OF 1
COUPLES -IN -ACTION SENT ■ to all who send us a self-addressed 1
stamped envelope. Absolutely none sent ■ to minors. TRAVEL-ARTS,
P.O. Box 14110. 1 Houston, Texas 77021. Do You Own A Movie
Projector If You Do Just send your name and address to Diane &
Jean, Box 157v,'p Union City, N,J.. and we will write to you
personally. Must be over 21. 1 72 UNCENSORED d .00
CANOIDPHOTOsl. An ORIGINAL collection. UNCENSOREO and
UNRETOUCHEO. o1 72 wallet size CANDID PHOTOS. (f2 leit pf 6
photos) every peiillon sharp and clear, as you like them, sent prepaid
In a sealed envelope upon receipt of $1.00. No checks, no C.O.O.
orders, no samples. Rush your $1.00 today to livt WtRi NOVfITIil.
P.O. BOX « m lest Bseedwav Dept ru .N.T., N.T. 18M2 GUYS^GALS —
For information on these movies that cannot be described in this
issue, send 25c. If you would like me to include a special film,
enclose $1.00. AMATEUR FILMS DEPT. 269 P.O. BOX 2171, LOS
ANGELES, CALIF. 90053 cc LU 0. < y 73
 CAPER UHCEHSORED! UNRETOUCHED! UNCLOTHED! nvpri
A A ‘Scx-Sationar and over mil revealing action photos of Nude
Bronze Beauties! Belles en Co our EELLES E is ths exciting French
name for the most revealing ' nude magazine in America today! It is
3 nudist photobook that goes far beyond the volleyball court, the
swimming pool, or the beach into vivid new regions that will startle
you with their daring. Page after exciting page, photograph after
revealing photograph, an incredible array of fantastic bronzed
nudists follow each other, each seemingly bent on proving that she is
truly the most completely free, Onashamed and bold nudist ever
photographed. This is the most enticing collection of bronze beauties
ever to grace the pages of a magazine! 72 pages of glorious color
and snappy black and white, including an exhilarating full color
center spread, suitable for framing. These beauties will daszle and
thrill you like never before, as they allow themselves to be
photographed from every conceivable angle in the freedom and
abandon of any activity. uriknnvniLMDLb lu mni unt nuuLiguvLn&i
LINGERIE CATALOG City state Zip EXCLUSIVE MAIL-ORDER OFFER.
AVAILABLE ONLY THRU GREENWICH VILLAGE PRESS! FUTWIM
INTIMATE the revealing A Intimate apparel women wear in the
privacy of the bedroom. Your personal undress fashion show free.
Sea exciting models in lacy, unusual bras, clinging see-thru
negligees, baby dolls, penolres, panties, bikini brief sets and more.
Everything the seductive woman wears for the man In mind. Send
today for gig Free Cataloa. Just rush name, address and 25C to
cover postage A handllng:.VOSUE sRon, 0tptl475234 Fifth Avi., H.V.
N.Y. 1MQ1 .Street. GREENWICH VILLAGE BOOKS Dept. CP402 ' P.O.
Box 222, Cooper Station, NY, NY 10003 Please rush the following in
plain sealed wrapper: Q BELLES EN COLOUR: Specify ^ g2](^ only
$3 ea. + 50d pp Q SPECIAL: Any 2 different issues, #( ) & #( ) — 2
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookname.com