Research Article

Molecular Pain
Volume 15: 1–12
The upregulation of EGFR in the ! The Author(s) 2019
Article reuse guidelines:
dorsal root ganglion contributes to sagepub.com/journals-permissions
DOI: 10.1177/1744806919857297
chronic compression of dorsal root journals.sagepub.com/home/mpx

ganglions-induced neuropathic pain in rats

Shuo Wang1,2, Siyi Liu1,2, Linping Xu1,2, Xuan Zhu1,3,

Wanyuan Liu1, Lixia Tian1,2, Yu Chen1, Yuying Wang1,2,
Borra V Padma Nagendra1, Shushan Jia3, Lingli Liang1,2 , and
Fu-Quan Huo1,2

Abstract
The epidermal growth factor receptor (EGFR) located in dorsal root ganglion has been found as a new target for chronic
pain treatment. However, it is not clear whether the change of EGFR expression in the dorsal root ganglion contributes to
neuropathic pain development. In this study, we used a chronic compression of unilateral lumbar dorsal root ganglions
(CCD)-induced rat neuropathic pain model and found that CCD caused the upregulation of both phosphorylated EGFR and
total EGFR expression in compressed lumbar 4/5 (L4/L5) dorsal root ganglions by western blotting and immunohistochem-
istry methods. Either inhibition of EGFR activation by EGFR inhibitor or knockdown of EGFR expression by EGFR small
interference RNA (siRNA) relieved CCD-induced pain hypersensitivities to mechanical, thermal, and cold stimuli in rats.
Moreover, EGFR knockdown reversed CCD-induced the increase of intracellular mammalian target of rapamycin (mTOR)
expression as well as the activation of the satellite glial cells in the ipsilateral compressed L4/L5 dorsal root ganglions. These
findings suggest that not only activated EGFR but also total EGFR contribute to CCD-induced neuropathic pain by enhancing
intracellular mTOR signaling.

Keywords
EGFR, mTOR, dorsal root ganglion, neuropathic pain, chronic compression of dorsal root ganglions
Date Received: 26 February 2019; revised: 24 April 2019; accepted: 12 May 2019

Introduction 1
Department of Physiology and Pathophysiology, School of Basic Medical
Neuropathic pain arises from injuries or diseases affect- Sciences, Xi’an Jiaotong University Health Science Center, Xi’an, China
2
ing the somatosensory nervous system at any level of the Key Laboratory of Environment and Genes Related to Diseases, Xi’an
Jiaotong University, Ministry of Education, Beijing, China
peripheral or central nervous system,1 which is charac- 3
Department of Anesthesiology, Yantai Affiliated Hospital of Binzhou
terized by intermittent burning pain, allodynia, and Medical University, Yantai, China
hyperalgesia, as well as spontaneous ongoing pain.2 The first three authors contributed equally to this work.
Neuropathic pain is usually caused by trauma (e.g., Corresponding Authors:
peripheral nerve, dorsal root, dorsal root ganglion Lingli Liang, Department of Physiology and Pathophysiology, School of Basic
(DRG), spinal cord, or brain injury) and some disorders Medical Sciences, Xi’an Jiaotong University Health Science Center, 76 Yanta
(e.g., multiple sclerosis, stroke, human immunodeficien- Road West, Xi’an, Shaanxi 710061, China.
Email: [email protected]
cy virus-induced neuropathy, and diabetes) in clinic and
Fu-Quan Huo, Department of Physiology and Pathophysiology, School of
is best treated with a combination of multiple therapeu- Basic Medical Sciences, Xi’an Jiaotong University Health Science Center, 76
tic approaches. However, current therapeutic Yanta Road West, Xi’an, Shaanxi 710061, China.
approaches of neuropathic pain management although Email: [email protected]

Creative Commons Non Commercial CC BY-NC: This article is distributed under the terms of the Creative Commons Attribution-
NonCommercial 4.0 License (http://www.creativecommons.org/licenses/by-nc/4.0/) which permits non-commercial use, reproduction and dis-
tribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.
sagepub.com/en-us/nam/open-access-at-sage).
2 Molecular Pain

provide symptomatic but still not unsatisfactory relief. water and food ad libitum. All the experimental proto-
More and more work is endeavoring to explore the new cols were approved by the Institutional Animal Ethics
potential targets or strategy for antinociception. Committee of the Xi’an Jiaotong University Health
The epidermal growth factor receptor (EGFR) Science Center (No. XJ20120117) and are consistent
belongs to the well-studied ErbB family of receptor tyro- with the ethical guidelines of International Association
sine kinases which regulate cellular growth, survival, for the Study of Pain. All the efforts were made to min-
proliferation, and differentiation of fibroblasts and hep- imize animal suffering and to reduce the number of ani-
atocytes.3–5 A recent study showed that the gene coding mals used.
EGFR had genetic association with temporomandibular
disorder, a human chronic pain syndrome.4 EGFR Neuropathic pain models
inhibitors, the first-line treatment for nonsmall cell
lung cancer, have been reported analgesic effect in neu- The neuropathic pain models including spinal nerve liga-
ropathic pain patients and preclinical neuropathic pain tion (SNL) and CCD in rats were developed based on
models.6,7 Furthermore, it has been found that phos- the previous publications. For SNL surgery,9,10 lumbar 5
phorylated EGFR was transiently activated in affected (L5) spinal nerve was exposed, ligated, and transected
DRGs of murine inflammatory pain or neuropathic pain under isoflurane anesthetization. Sham group underwent
models.4 EGFR inhibitors showed short analgesic effect the identical operation but had no ligation and
(less than 1 h) on these animal models.4 Furthermore, it transection.
has been demonstrated that phosphoinositide 3 kinase For CCD surgery,11,12 the L4 and L5 intervertebral
(PI3K)/Protein kinase B (Akt)/mammalian target of foramina were exposed and L-shaped stainless steel rods
rapamycin (mTOR) cellular signaling pathway is one (4 mm in length and 0.6 mm in diameter) were carefully
of the pathway mediating EGFR-activated pain behav- inserted into the L4 and L5 foramina to compress the
iors.4 Therefore, it seems that EGFR is a potential target DRGs. The surgical procedure in sham group was iden-
for pain management. tical to that in CCD group, except that the stainless steel
Chronic compression of unilateral lumbar DRGs rods were not inserted into the intervertebral foramina.
(CCD)-induced neuropathic pain model has been used
to mimics low back pain and radicular pain syndromes DRG cell culture and siRNA transfection
caused by multilevel nerve root compression and inter- EGFR siRNA (50 -GAUGGAGUCAGCAAGU
vertebral foramina stenosis in human.8 In this study, we GUATT-3 ; 5 -UACACUUGCUGACUCCAUCTT-30 )
0 0
observed the upregulation of EGFR expression in and negative control (NC) siRNA (50 -UUCUCC
injured DRGs of neuropathic pain models and further
GAACGUGUCACGUTT-30 ; 50 -ACGUGACACG
explored the effect of EGFR upregulation in CCD- 0
UUCGGAGAATT-3 ) were designed and synthesized
induced pain hypersensitivities. We found that unilateral
by Genepharma (Shanghai, China). The DRGs from
CCD induced the increase of phosphorylated EGFR (p-
three- to four-week-old SD rats were harvested and cul-
EGFR) and total EGFR (t-EGFR) expression in ipsilat-
tured for examining the knockdown efficiency of EGFR
eral compressed L4/L5 DRGs. Either inhibition of
siRNA. All harvested DRGs were then digested with
EGFR activation by its inhibitor or knockdown of
0.25% trypsin solution without EDTA (Beyotime
EGFR expression by EGFR small interference RNA
Biotechnology, Shanghai, China). Following trituration
(siRNA) relieved CCD-induced pain hypersensitivities
and centrifugation, dissociated cells were resuspended
to mechanical, thermal, and cold stimuli. EGFR knock-
and cultured in cold NeurobasalTM-A Medium (Gibco/
down also restored the expression of total EGFR and its
intracellular downstream target mTOR in compressed ThermoFisher Scientific, Waltham, MA) with 10% fetal
L4/L5 DRGs of CCD rats. Moreover, EGFR knock- bovine serum (JR Scientific, Woodland, CA), B-27TM
down reversed the increased expression of glutamine Supplement (1) (Gibco/ThermoFisher Scientific), 100
synthetase (GS), a marker for satellite glial cells, in com- units per ml penicillin, and 100 mg per ml streptomycin
pressed DRGs of CCD rats. (Beyotime Biotechnology) in a six-well plate precoated
with 50 mg per ml poly-D-lysine (Beyotime
Biotechnology). The cultured cells were incubated in
Materials and methods an incubator with 95% O2, 5% CO2, and at 37 7.
After 24 h incubation, EGFR siRNA (250 pmol) or
Animals equivalent NC siRNA was delivered by Lipo6000 trans-
Adult male Sprague Dawley (SD) rats (200–250 g) or fection reagent (Beyotime Biotechnology) into cultured
three- to four-week-old SD rats were kept in an environ- cells in six well plates. Two days later, the cultured
ment with adequate temperature (25
1 ) and ventila- cells were harvested into radio immunoprecipitation
tion in a light–dark cycle of 12 h–12 h. The animals had assay (RIPA) lysis buffer for western blotting.
Wang et al. 3

Intrathecal catheter implantation and drug delivery The cold hyperalgesia was measured with a model
ZH-6C cold plate (Zheng-Hua Biologic, Anhui,
For intrathecal drug delivery, a polyethylene (PE)-10
China). Each rat was placed in a cylindrical transparent
catheter was implanted into the subarachnoid space of
Plexiglas chamber (height: 28 cm, diameter: 20 cm) on a
the spinal cord under isoflurane anesthesia. The detailed
round cold plate (diameter: 19 cm). The temperature of
procedure has been described in our previous publica-
cold plate was set at 0 t. The positive response of rat
tion.13 Rats showing neurologic deficits postoperatively indicating cold hyperalgesia exhibited as a quick and
were excluded from the study; 10 ml of siRNAs, EGFR strong paw flinching. The latency of the first positive
inhibitor, or vehicles were administered intrathecally fol- response was recorded. Each trial was repeated three
lowed by flushing of 10 ml sterile normal saline. times at 10-min intervals. A cutoff time of 60 s was
EGFR siRNA, NC siRNA, or vehicle control (phos- used to avoid tissue damage of both hind paws.
phate buffer saline (PBS)) was administered intrathecally
from the third day after CCD surgery once daily for four RNA extraction, reverse transcription and quantitative
days. Lipo6000 transfection reagent (Beyotime
real-time polymerase chain reaction
Biotechnology) was used as a delivery vehicle for
siRNA to improve delivery efficiency and prevent degen- L4/L5 DRGs were collected and kept in RNAlater sta-
eration of siRNA.14,15 bilization solution (Thermo Scientific). DRGs were
Gefitinib (HY-50895, MedChem Express, Shanghai, homogenized in a cold tissue homogenizer (Shanghai
China) or vehicle (20% DMSO) was administered intra- Jingxin, China), and RNA was extracted using
thecally by single injection on day 7 after CCD surgery RNAeasyTM RNA extraction kit (Beyotime
to observe the acute effect of gefitinib or once daily for Biotechnology). RNA concentration was quantitated
seven days from the first day after CCD surgery to using a nanodrop spectrophotometer (Thermo
observe the chronic effect. Scientific); 500 ng RNA was reverse transcribed with
oligo (dT) primer using the RevertAid First strand
cDNA Synthesis Kit (Thermo Scientific) according to
Behavior tests
the manufacturer’s instructions. Quantitative polymerase
Paw withdrawal thresholds (PWTs) in response to chain reaction (qPCR) was performed on a 20 ml reaction
mechanical stimuli were measured with the up–down with 20 ng of cDNA, 250 nM forward and reverse pri-
testing paradigm.16–18 Briefly, the unrestrained rat was mers, and 10 ml of BeyoFastTM SYBR Green qPCR Mix
placed in a Plexiglass chamber on an elevated mesh (Beyotime Biotechnology) by using the Egfr primer
screen after adaptation to experimental environment. (Forward: 50 -ACAACACCCTGGTCTGGAAG-30 ;
Calibrated von Frey filaments in log increments of Reverse: 5 -GCCCTTCTGGTTGTTGACAT-30 ) or
0

force (0.41, 0.69, 1.20, 2.04, 3.63, 5.50, 8.51, 15.14, and Gapdh primer (Forward: 50 -TCGGTGTGAACGGA
26.00 g) were applied to the plantar surface of the hind TTTGGC-30 ; Rerverse: 50 -TCCCATTCTCGGC
0
paws of the rat. The 2.04-g stimulus was applied first. If CTTGACT-3 ) in a BIO-RAD CFX96 real-time PCR
a positive response occurred, the next smaller von Frey system (Bio-Rad Laboratories, Hercules, CA). The
filament was used; or not, the next larger filament was cycle parameters were an initial 3-min incubation at 95
used. The test stopped when (1) a negative response to 5, followed by 40 cycles of 95 5 for 10 s, 60 0 for 30 s,
the 26.00-g filament or (2) 3 stimuli after the first positive and 72 2 for 30 s. All data were normalized to glyceral-
response. The PWT was calculated with the formula pro- dehyde-3-phosphate dehydrogenase (GAPDH), an inter-
vided by Dixon by converting the pattern of positive and nal control. Ratios of mRNA levels in ipsilateral side to
negative responses to a 50% threshold value.17,18 contralateral side were calculated using the 䉭Ct
Paw withdrawal latencies (PWLs) to noxious heat method (2䉭Ct).
were measured with a Model 37370 Analgesic Meter
(UGO, Italy).10,16 Rats were placed in a Plexiglas cham- Western blotting
ber on a glass plate. A radiant heat was applied by a The L4/L5 DRGs or spinal dorsal horns were collected
light beam to the middle of the plantar surface of each after decapitation of rats under deep anesthetization by
hind paw. The light beam was turned off once the rat has chloral hydrate (400 mg/kg). The DRGs or spinal dorsal
a strong paw withdrawal response. The PWL was horns were homogenized with ice-cold RIPA lysis buffer
defined as the length of time between the start of the (50 mM Tris, 150 mM NaCl, 1% NP-40, 0.25% sodium
light beam and paw withdrawal response. Five values deoxycholate, 1 mM phenylmethylsulfonyl fluoride,
in each trial were recorded at 10-min intervals and aver- pH7.4) with protease and phosphatase inhibitor cocktail
aged as PWL value for each paw. A cutoff time of 20 s (Beyotime Biotechnology) using Shanghai Jingxin tissue
was used to avoid tissue damage to paw. homogenizer. After the crude homogenate was
4 Molecular Pain

centrifuged at 4 t for 15 min at 1000 g, the supernatants room temperature. The primary antiserum was omitted
were collected for cytosolic proteins. Protein concentra- for control DRG sections. The slides were coverslipped
tion was measured using a Bradford Protein Assay Kit with SouthernBiotech Fluoromount-G
(Beyotime Biotechnology). The equal amount samples (SouthernBiotech, Birmingham, AL). Pictures were cap-
were heated for 5 min at 99 9 and loaded onto a 8% tured by an Olympus BX53 fluorescence microscope. For
separating sodium dodecyl sulfate (SDS)-polyacrylamide colabeling of EGFR with DRG cell markers, DRG
gel electrophoresis gel. The proteins were then electro- sections were incubated with rabbit anti-EGFR anti-
phoretically transferred onto a nitrocellulose membrane body (1:100; Beyotime Biotechnology) and mouse
followed by blocking with 3% nonfat milk in Tris- antineurofilament-200 (NF200, 1:500; Sigma), mouse
buffered saline containing 0.1% Tween-20. The first anticalcitonin gene-related peptide (CGRP, 1:50;
antibodies included rabbit antiphosphor-EGFR Abcam), or mouse anti-GS (1:500; EMD Millipore) over-
(Tyr1068, p-EGFR, 1:1000; Beyotime Biotechnology), night at 4 t, followed by goat antirabbit antibody con-
rabbit anti-EGFR (1:1000; Beyotime Biotechnology), jugated to Alexa Fluor 488 (1:200; Abcam) and goat
rabbit anti-mTOR (1:1000; Cell Signaling Technology), antimouse antibody conjugated to Cy3 (1:200; Abcam)
mouse anti-GS (1:1000; EMD Millipore, Darmstadt, for 2 h at room temperature. Cy3-conjugated-Avidin
Germany), or mouse anti-b-actin (1:1000; Beyotime (1:200, Abcam) was used as the second antibody for
Biotechnology) antibodies. After incubating with the double labeling of EGFR with biotinylated-isolectin B4
first antibodies overnight, the membranes were incubated
(IB4, 1:100, Sigma). The slides were coverslipped with
in horseradish peroxidase-conjugated antirabbit or anti-
SouthernBiotech Fluoromount-G (SouthernBiotech) or
mouse secondary antibody (1:3000; EMD Millipore) for
antifade mounting medium with 40 ,6-diamidino-2-
2 h. The proteins on membrane were detected by western
phenylindole (DAPI) (Beyotime Biotechnology). All
peroxide reagent and luminol/enhancer reagent
colabeled pictures were viewed and captured using
(Immobilon Western Chemiluminescent HRP Substrate;
Nikon C2 confocal microscope (Nikon, Tokyo, Japan).
EMD Millipore) and visualized using the Champchemi
All the pictures were quantified by NIH Image J Software.
System with SageCapture software (Sagecreation Service
for Life Science, Beijing, China). The intensity of blots
was quantified by using NIH Image J software. Each Statistical analysis
blot from the targeted protein was normalized to the cor- The animals were distributed into various treatment
responding b-actin. The average value from the control groups randomly. The number of rats is more than six
groups was set as 100% after normalization. The relative in each group in behavior experiment and more than
levels of the targeted protein from time points or the three in western blotting or immunofluorescence experi-
treated groups were determined by dividing the normal- ments. All of the results were given as means
standard
ized values from these groups by the average value of the error of the mean. The data were statistically analyzed
control groups. with SigmaPlot 12.5 software (San Jose, CA) by two-
tailed, paired Student’s t test for two group comparison,
Immunofluorescence one-way analysis of variance (ANOVA) for more than
The rats were deeply anesthetized with chloral hydrate two-group comparison or two-way repeated measure
(400 mg/kg) and perfused transcardially through the left (RM) ANOVA for results from behavior tests. When
cardiac ventricle with 100 ml of perfusion buffer, fol- ANOVA showed a significant difference, post hoc
lowed by 250 ml of 4% paraformaldehyde in 0.1 M phos- Tukey method was used to perform pairwise compari-
phate buffer (PB), pH 7.4, at room temperature for sons between means. P < 0.05 was considered significant.
15 min. Subsequently, the L4/L5 DRGs was removed and
postfixed for 24 h at 4 t and then dehydrated with 30% Results
sucrose in 0.1 M PB for 48 h. DRGs were embedded in
optimum cutting temperature medium (Tissue Tek OCT; Increased EGFR expression in neuropathic
Sakura) and were cut at 20 mm using a Leica CM3050 S pain models
cryostat. DRG sections were placed directly on gelatin-
covered slides. After rinsing with PBS for 10 min, slides Previous RNA-seq analysis had shown a significant
with DRG sections were incubated with 10% normal increase of EGFR mRNA levels in injured DRGs of
goat serum for 1 h. For single labeling of EGFR, DRG SNL-induced neuropathic pain mice.19 Using quantita-
sections were incubated overnight at 4 t with rabbit anti- tive real time-PCR, we confirmed the increase of EGFR
EGFR (1:100, Beyotime Biotechnology) antibody. The mRNA level in injured DRGs in SNL rats on day 7 after
sections were then incubated in goat antirabbit antibody SNL surgery. EGFR mRNA level in SNL rats is 2.60-
conjugated to Alexa Fluor 488 (1:200, Abcam) for 2 h at folds of sham rats (P < 0.01, Figure 1(a)).
Wang et al. 5

Figure 1. Expression changes of EGFR mRNA and protein in the DRG and spinal dorsal horn in neuropathic pain models. (a) EGFR
mRNA increased in the ipsilateral L5 DRG on day 7 after SNL. N ¼ 5 rats per group. **P < 0.01 versus the corresponding sham group by
two-tailed unpaired Student’s t test. (b) EGFR mRNA increased in the ipsilateral L4/L5 DRG on days 3, 7, and 14 after CCD. N ¼ 3 or 4
rats/time point. One-way ANOVA (expression vs. time points) followed by post hoc Tukey test, *P < 0.05, ***P < 0.001 versus the
corresponding naive rats (day 0). (c) Phosphor-EGFR (p-EGFR) and total EGFR (t-EGFR) increased in the ipsilateral L4/L5 DRGs after CCD
surgery. N ¼ 3 or 4 rats/time point. One-way ANOVA (expression vs. time points) followed by post hoc Tukey test, *P < 0.05, ***P < 0.001
versus the corresponding naive rats (day 0). (d and e) No significant changes in p-EGFR and t-EGFR were seen in the contralateral L4/L5
DRGs (d) and the ipsilateral L4/L5 spinal dorsal horns (e) at all observed time points. N ¼ 3 or 4 rats/time point. One-way ANOVA
(expression vs. time points) followed by post hoc Tukey test. CCD: chronic compression of DRG; DRG: dorsal root ganglion; EGFR:
epidermal growth factor receptor; SNL: spinal nerve ligation.

CCD induces pain hypersensitivities in rats by causing did not affect the expression of p-EGFR and t-EGFR in
DRG injury via compressing nerve root and DRG,8,12 contralateral L4/L5 DRGs (P > 0.05, Figure 1(d)) and ipsi-
which is more close to clinical neuropathic pain conditions. lateral L4/L5 spinal dorsal horns (P > 0.05, Figure 1(e)) at
Therefore, we further examined EGFR expression using all observed time points. In comparison, sham surgery did
CCD-induced neuropathic pain model. As expected, not cause any change on EGFR mRNA and protein level
CCD increased EGFR mRNA and protein levels in com- (data not shown).
pressed DRGs. The mRNA levels of EGFR is 1.88, 2.92, We further observed EGFR expression in DRG using
and 5.97-folds of control group (0 day) on days 3, 7 and 14, immunofluorescence method. We first examined the spe-
respectively (F3,13 ¼ 29.10, P < 0.001, Figure 1(b)). The sig- cificity of EGFR antibody by omission of primary anti-
nificant changes were observed only on days 7 and 14 by a body (Figure 2(a)). No positive signals were detected only
one-way ANOVA followed by post hoc Tukey method. except DAPI showed the nuclei staining in control DRG
However, a significant increase on day 3 was also observed slices (Figure 2(a)). Significant EGFR immunoreactivity
if using a two-tailed, paired Student’s t test. Consistently, (EGFR-ir) was observed in both neuronal and nonneuro-
EGFR protein levels increased by 2.27 and 2.32-folds on nal cells as shown in Figure 2(b). Dr. Martin et al. has
days 7 and 14, respectively (F3,15 ¼ 23.15, P < 0.001, Figure reported that EGFR distributed averagely in DRG neu-
1(c)). EGFR total protein did not change significantly on rons by counting neuron numbers with EGFR-ir at differ-
day 3 after CCD by western blotting. By normalizing to ent sizes.4 However, this classification is not accurate to
b-actin, the activated p-EGFR significantly increased by distinct the expression of EGFR in different types of DRG
2.30, 4.79, and 3.00-folds on days 3, 7, and 14 neurons. In DRG, medium/large neurons with myelinated
(F3,11 ¼ 5.37, P < 0.05, Figure 1(c)), while the ratio of p- A fibers was usually labeled by NF200, small DRG pepti-
EGFR to total EGFR (t-EGFR) increased 1.81 and 2.11- dergic neurons by CGRP, and small nonpeptidergic DRG
folds on days 3 and 7 (F3,11 ¼ 45.77, P < 0.001, Figure 1(c)). neurons by IB4.20–22 By colocalization of EGFR with
The p-EGFR/t-EGFR ratio did not significantly change on NF200, CGRP, or IB4, we found that about 80.1% of
day 14 after surgery. These results indicate that CCD not EGFR-positive neurons were colocalized with NF200,
only causes the activation of EGFR but also leads to the 14.9% with CGRP, and only 5.5% with IB4 (Figure 2
upregulation of EGFR mRNA and protein expression in (c)). We also observed EGFR expression in satellite glial
injured DRGs. The increased p-EGFR is partially due to cells by colocalization of EGFR with GS, a marker for
the increase of total EGFR protein level. Unilateral CCD satellite glial cells (Figure 2(c)). The cellular nuclei were
6 Molecular Pain

Figure 2. Distribution of EGFR protein in the lumbar DRGs of rats. (a) No EGFR signals were detected in the control DRG slices without
EGFR antibody incubation. (b) EGFR was expressed in the cytosol of neurons and nonneuronal cells around cellular nuclei (labeled by
DAPI) in the DRG. (c) EGFR was colocalized with NF200, CGRP, IB4, or glutamine synthetase. (d) CCD increased the numbers of DRG
neurons with high EGFR-ir intensity in the ipsilateral L4/L5 DRGs. N ¼ 5 rats/group. **P < 0.01 versus the corresponding sham group by
two-tailed unpaired Student’s t test. Short arrows: DRG neurons with high EGFR-ir intensity; Long arrows: glial cells with EGFR-ir. Scale
bars: 50 mm. CCD: chronic compression of dorsal root ganglion; CGRP: calcitonin gene-related peptide; DAPI: 40 ,6-diamidino-2-phenyl-
indole; EGFR: epidermal growth factor receptor; EGFR-ir: EGFR immunoreactivity; IB4: isolectin B4; NF200, neurofilament-200.

shown by DAPI to better recognize the structure of satel- 40 min in chronic constriction injury of sciatic nerve or
lite glial cells. To be consistent with previous report,4 we spared nerve injury-induced-neuropathic pain models.4
analyzed the neurons with high EGFR-ir intensity. We further examined the effect of gefitinib on CCD-
Around 25.86 (
6.55) % of DRG neurons showed high induced mechanical pain hypersensitivity. Ipsilateral
EGFR-ir intensity in sham group. The high EGFR-ir PWTs to von Frey stimuli decreased significantly on
intensity neurons increased by 1.53-folds in CCD group day 7 post-CCD compared to baseline (***P < 0.001,
compared to sham group (P < 0.05, Figure 2(d)). vs. 1 day, Figure 3(a) and (b)). Intraperitoneal injec-
tion (i.p.) of gefitinib dose dependently increased PWTs
on the ipsilateral side at 20 min and 40 min after drug
Inhibition of EGFR activation by EGFR inhibitor
injection compared to vehicle treatment (Figure 3(a),
attenuates CCD-induced pain hypersensitivities F12,129 ¼ 3.98, #P < 0.05; ##P < 0.01, ###P < 0.001, vs.
It has been reported that EGFR inhibitors produced Veh). Intrathecal injection (i.t.) of gefitinib dose depen-
complete and dose-dependent reversal of allodynia for dently increased PWTs at 30 min and 60 min on the
Wang et al. 7

Figure 3. Effect of EGFR inhibitor gefitinib on CCD-induced nociceptive hypersensitivities. Single injection of gefitinib intraperitoneally
(i.p., a) or intrathecally (i.t., b) dose dependently reversed the decrease in PWTs to mechanical stimulation. All rats were performed CCD
surgery. N ¼ 7 or 8 rats/group. Two-way RM ANOVA (effect vs. group  time interaction) followed by post hoc Tukey test, ***P < 0.001
versus baseline value before CCD surgery. #P < 0.05, ##P < 0.01, ###P < 0.001, versus the corresponding time point in the vehicle (Veh,
20% DMSO) group or versus the value before drug administration. I.t. injection of gefitinib (once daily for 7 days) dose dependently
blocked the decrease of PWTs to mechanical stimulation (c), PWLs to thermal stimulation (d), or positive response latencies to cold
stimulation (e) on the ipsilateral side of CCD rats. The first injection starts at 30 min before CCD surgery. N ¼ 6 rats/group. Two-way RM
ANOVA (effect vs. group  time interaction) followed by post hoc Tukey test, ***P < 0.001 versus the corresponding time point in the
sham þ Veh group. #P < 0.05, ###P < 0.001 versus the corresponding time point in the CCD þ Veh group. CCD: chronic compression of
dorsal root ganglion; i.p.: intraperitoneal; i.t.: intrathecal; PWL: paw withdrawal latency; PWT: paw withdrawal threshold.

ipsilateral side compared to vehicle treatment (Figure 3 Figure 3(d): F20,179 ¼ 11.64, Figure 3(e): F20,179 ¼ 15.80,
(b), F12,159 ¼ 13.20, ##P < 0.01, ###P < 0.001, vs. Veh). ***P < 0.001, vs. Sham þ Veh, ###P < 0.001, vs.
Moreover, the analgesic effect of i.t. injection of 5 mg CCD þ Veh). Gefitinib at 0.05 mg did not significantly
gefitinib lasted for at least 2 h (Figure 3(b), affect CCD-induced pain hypersensitivities during initial
###
P < 0.001, vs. Veh). five days while finally lead to pain relief on day 7 post-
To determine the analgesic action of repeated admin- CCD (#P < 0.05, ###P < 0.001, vs. CCD þ Veh, Figure 3
istration of EGFR inhibitors on CCD-induced pain (c) to (e)). Gefitinib at the dose of 1 mg did not affect basal
hypersensitivities, we intrathecally injected gefitinib at paw withdrawal responses in sham rats (P > 0.05,
different doses (0.05, 0.5, 1 mg) starting on day 1 after Figure 3(c) to (e)). However, it should be noted that
CCD surgery once daily for seven days and observed repeated administration of the higher dose of gefitinib
paw withdrawal responses to mechanical, thermal, and (5 mg or 10 mg) caused the decrease of PWTs and PWLs
cold stimuli in rats. Increased pain hypersensitivities to in both sham and CCD rats (data not shown).
mechanical, thermal, or cold stimuli in CCD rats were
attenuated dose dependently by repeated injection of EGFR knockdown by EGFR siRNA attenuates
gefitinib (Figure 3(c) to (e)). Compared with vehicle treat-
ment, gefitinib at the doses of 1 mg and 0.5 mg completely
CCD-induced pain hypersensitivities
or partially reversed the decrease of PWTs to mechanical In order to determine the role of total EGFR in the
stimulation, PWLs to thermal or positive response laten- development of CCD-induced pain hypersensitivities,
cies to cold stimulation on the ipsilateral side from we tested the effect of i.t. injection of EGFR siRNA
days 3 to 7 post-CCD (Figure 3(c): F20,179 ¼ 9.37, on CCD-induced pain hypersensitivities. We first
8 Molecular Pain

Figure 4. Effect of EGFR siRNA (Si) on CCD-induced nociceptive hypersensitivities. (a) The expression level of EGFR protein was
markedly reduced by treating with EGFR siRNA (250 nM) in in vitro DRG cell culture. N ¼ 3 repeats (six wells from three rats) per
treatment. *P < 0.01 versus NC siRNA by two-tailed unpaired Student’s t test. (b) Intrathecal injection of EGFR siRNA (10 mM in 10 ml)
blocked the increase of EGFR induced by CCD and did not affect the basal expression of EGFR in sham group. The first injection starts on
day 3 post-CCD and once daily for four days. Ipsilateral L4/L5 DRGs were harvested on day 7 after surgery. N ¼ 3 rats/time point. One-
way ANOVA (effect vs. the treated groups) followed by post hoc Tukey test, **P < 0.01, ***P < 0.001 versus the sham þ Veh group.
##P < 0.01 or ###P < 0.001 versus the CCD þ Veh group. Intrathecal injection of EGFR siRNA reversed the decrease of PWTs to
mechanical stimulation (c), PWLs to thermal stimulation (d), or positive response latencies to cold stimulation (e) on the ipsilateral side in
CCD rats. N ¼ 6 rats/group. Two-way RM ANOVA (effect vs. group  time interaction) followed by post hoc Tukey test. ***P < 0.001
versus the corresponding time point in the sham þ Veh group. ###P < 0.001 versus the corresponding time point in the CCD þ Veh or the
CCD þ NC group. CCD: chronic compression of DRG; DRG: dorsal root ganglion; EGFR: epidermal growth factor receptor; NC:
negative control; PWL: paw withdrawal latency; PWT: paw withdrawal threshold.

examined the knockdown efficiency of EGFR siRNA by PWLs to thermal and positive response latencies to
using in vitro DRG cell culture (Figure 4(a)). cold stimulation on the ipsilateral side on days 5 and 7
Transfection of EGFR siRNA (250 pmol) significantly after CCD surgery (Figure 4(c): F16,149 ¼ 10.98, Figure 4
reduced total EGFR expression compared to NC siRNA (d): F16,149 ¼ 19.30, Figure 4(e): F16,149 ¼ 19.48,
(P < 0.01, Figure 4(a)). In in vivo experiment, we intra- ###
P < 0.001, vs. CCD þ Veh or CCD þ NC). NC
thecally injected siRNAs once daily for four days start- siRNA had no effect on CCD-induced mechanical allo-
ing on day 3 post-CCD and found that EGFR siRNA dynia, thermal, or cold hyperalgesia (P > 0.05, Figure 4
(10 mM/10 ml), but not NC siRNA (10 mM/10 ml), dimin- (c) to (e)). Basal mechanical, thermal, or cold responses
ished the CCD-induced increase in the level of EGFR on the ipsilateral sides of sham rats were not affected by
protein in the ipsilateral L4/L5 DRG (F4,14 ¼ 12.51, i.t. injection of either siRNA (Figure 4(c) to (e)).
**P < 0.01 vs. Sham þ Veh, ##P < 0.01, vs. CCD þ Veh;
Figure 4(b)). No significant changes in the basal level of
mTOR signaling pathway mediates the effect of EGFR
EGFR protein were seen from the EGFR siRNA plus
sham surgery group (Figure 4(b)), which may due to the It has been demonstrated that activation of EGFR
low basal level of EGFR expression. Three days after enhanced nociception through a mechanism involving
surgery, CCD induced the significant decrease in PWTs mTOR signaling pathway. We further confirmed that
to mechanical stimulation, PWLs to thermal or positive mTOR was a downstream effector of EGFR by exam-
response latencies to cold stimulation in all groups with ining the effect of EGFR knockdown on mTOR expres-
CCD operation (***P < 0.001, vs. Sham þ Veh, Figure 4 sion. In in vitro DRG cell culture, transfection of EGFR
(c) to (e)). EGFR siRNA dramatically blocked CCD- siRNA markedly reduced mTOR expression compared
induced increase of EGFR expression and thereby to transfection with NC siRNA (P < 0.01 for mTOR;
increased PWTs to von Frey filaments stimulation, Figure 5(a)). In in vivo experiments, repeated i.t.
Wang et al. 9

Figure 5. Intracellular signaling molecule mTOR mediated nociceptive effect of EGFR. The expression level of mTOR (a) or GS (c) was
markedly reduced by treating with EGFR siRNA (Si, 250 nM) in in vitro DRG cell culture. N ¼ 3 repeats (six wells from three rats) per
treatment. **P < 0.01, ***P < 0.001 versus NC siRNA treatment by two-tailed unpaired Student’s t test. Intrathecal injection of EGFR
siRNA (10 mM in 10 ml vehicle solution) blocked the increase of mTOR (b) or GS (d) induced by CCD and did not affect their basal
expressions in sham group. The first injection was administered on day 3 post-CCD and once daily for four days. Ipsilateral L4/L5 DRGs
were harvested on day 7 after surgery. N ¼ 3 rats/time point. One-way ANOVA (effect vs. the treated groups) followed by post hoc Tukey
test, *P < 0.05 versus the sham þ Veh group. ##P < 0.01 versus the CCD þ Veh group. CCD: chronic compression of DRG; DRG: dorsal
root ganglion; GS: glutamine synthetase; mTOR: mammalian target of rapamycin; NC: negative control.

injection of EGFR siRNA for four days reversed satellite glial cells activation. In in vitro DRG cell
CCD-induced the upregulation of mTOR expression culture, which included DRG neuron and nonneuronal
(F4,14 ¼ 9.09, *P < 0.05 vs. Sham þ Veh, ##P < 0.01, cells, EGFR siRNA largely reduced the expression of
vs. CCD þ Veh; Figure 5(b)). EGFR siRNA did not GS, a marker for satellite glial cells (P < 0.05, Figure 5
alter the basal level of mTOR in the ipsilateral L4/L5 (c)). Consistently, EGFR knockdown by i.t. injection
DRG of sham rats (Figure 5(b)). These results sug- of EGFR siRNA reversed the increase of GS expres-
gested that mTOR might mediate the pro-nociceptive sion in compressed DRGs of CCD rats (F4,14 ¼ 8.32,
effect of EGFR under the chronic condition of neuro- *P < 0.05 vs. Sham þ Veh, ##P < 0.01, vs. CCD þ Veh;
pathic pain. Figure 5(d)). These results suggested glial EGFR
As EGFR was expressed in DRG nonneuronal might also contribute to neuropathic pain
cells, we further tested if EGFR knockdown affected development.
10 Molecular Pain

Discussion in the development of neuropathic pain. As CCD did

not cause any change of EGFR in spinal cord, it is not
It has been reported that activated EGFR contributes to
likely that EGFR in spinal cord contributes to the anal-
neuropathic pain via mTOR and other signaling path-
gesic effect of EGFR inhibitor or EGFR siRNA.
way.4,7 In this study, we found that not only activated
The downstream effects of EGFR were mediated by a
EGFR but also total EGFR involved in peripheral
number of important signaling pathways, including
mechanism of neuropathic pain. EGFR are not only
mitogen-activated protein kinase (MAPK) and PI3K/
expressed in DRG neurons but also in satellite glial
AKT/mTOR.4,24 It has been demonstrated that epiregu-
cells in DRG. Most of EGFR positive neurons colocal- lin, one of the EGFR ligands, activated EGFR and then
ized with NF200-positive large myelinated DRG neu- potentiated pain behaviors through enhanced PI3K/
rons, a few with CGRP-positive DRG neurons, and AKT/mTOR signaling but not extracellular signal-regu-
very few with IB4-positive DRG neurons. Both p- lated kinase (ERK) signaling.4 Our results showed that
EGFR and t-EGFR increased in injured DRGs in DRG EGFR knockdown not only reversed the increase
CCD-induced neuropathic pain rats. Either EGFR inhi- of EGFR expression but also reversed the increase of
bition by its inhibitor gefitinib or EGFR knockdown by mTOR expression in CCD rats. Therefore, mTOR, the
EGFR siRNA gradually relieved pain hypersensitivities. downstream of EGFR, may also mediate the chronic
EGFR knockdown reversed the increase of EGFR, as effect of EGFR on nociceptive behaviors in CCD rats.
well as mTOR expression, induced by CCD. Moreover, Not only mTOR signaling pathway, EGFR also
CCD also caused the activation of DRG satellite glial affects opioid receptors,25 b‑adrenergic receptors,26 can-
cells which can be deactivated by EGFR knockdown. nabinoid type 1, and transient receptor potential vanil-
These findings suggest that not only activated EGFR loid 1 (TRPV1) receptors,27 which are important for pain
but also total EGFR contribute to the development of processing. A recent study has reported that the upregu-
neuropathic pain. lation of epiregulin in the blood activated EGFRs on
CCD mechanically deformed DRG so that mimics low DRG neurons to induce hypersensitivity through trans-
back pain and radicular pain syndromes caused by multi- activation of TRPV1 and the mTOR signaling pathway,
level nerve root compression and intervertebral foramina which increases MMP-9 translation.4 In addition, DRG
stenosis radiculopathies and tumors in human.8 Several satellite glial cells have been demonstrated an important
studies have shown that unilateral CCD surgery caused role in the development and maintenance of neuropathic
pain hypersensitivities not only on the ipsilateral side but pain.20,28 Consistent with previous studies,4,29,30 EGFR
also on the contralateral side.11,12,23 However, not all of was not only expressed in DRG neurons but also in sat-
animals exhibited pain hypersensitivities on the contralat- ellite glial cells, which were activated in compressed
eral side in our preliminary experiments. Moreover, we did DRGs by CCD surgery. EGFR knockdown deactivated
not find any significant change of p-EGFR and t-EGFR satellite glial cells in compressed DRGs of CCD rats. The
in contralateral DRGs. We therefore focused on the ipsi- deactivated satellite glial cells may due to the knockdown
lateral side to study involvement of EGFR in CCD- effect of EGFR siRNA on EGFR in satellite glial cells.
induced pain hypersensitivities. The increase of EGFR As activated glial cells release pro-inflammatory cyto-
mRNA expression in injured DRGs of neuropathic pain kines, such as interleukin (IL)-1, IL-6, and tumor necro-
mice has been screened based on a RNA-seq analysis for sis factor-a, leading to further activation of glia and act
DRGs from SNL mice and sham mice.19 Our current as a secondary stimulus,31,32 the decreased EGFR
study further confirmed the upregulation of both EGFR expression level in satellite cells may reduce the release
mRNA and protein in injured DRGs of CCD-induced proinflammatory cytokines and then relieve CCD-
neuropathic pain model. induced pain hypersensitivities.
Previous studies have demonstrated the analgesic In conclusions, we demonstrate that not only activat-
effect of EGFR inhibitor possibly by blocking EGFR ed EGFR but also total EGFR expression contribute to
activation.4,7 Our study also showed that EGFR inhib- neuropathic pain development by regulating mTOR sig-
itor gefitinib by i.p. injection or i.t. injection reversed naling. Both EGFR inhibition and EGFR knockdown
CCD-induced mechanical allodynia dose dependenly. are potential ways for neuropathic pain treatment.
Moreover, i.t. injection of gefitinib prolonged analgesic However, the expression of EGFR in both DRG neu-
time on CCD-induced mechanical allodynia. It sug- rons and glial cells make the antinociceptive effect of
gested that i.t. administration may be more effective EGFR inhibitors and EGFR siRNA complicated.
for pain relief. Furthermore, repeated i.t. injection of
gefitinib or EGFR siRNA produced the accumulating Author Contributions
analgesic effect on CCD-induced pain hypersensitivities. L. Liang conceived of the project and designed experiments,
The effect of EGFR siRNA on pain behaviors and analyzed the data, and wrote the manuscript. F.-Q. Huo and
EGFR expression indicates a key role of total EGFR S. Jia provided advices on the project. S. Wang, S. Liu, W. Liu,
Wang et al. 11

and B. V. P. Nagendra did behavior tests. L. Xu, X. Zhu, and Contribution of the suppressor of variegation 3-9 homolog
L. Tian performed western blotting and immunofluorescence 1 in dorsal root ganglia and spinal cord dorsal horn to
experiments. Y. Chen harvested tissues. Y. Wang did CCD nerve injury-induced nociceptive hypersensitivity.
surgery. S. Wang, S. Liu, and L. Xu equally contributed to Anesthesiology 2016; 125: 765–778. Oct
this work. 10. Zhao JY, Liang L, Gu X, Li Z, Wu S, Sun L, Atianjoh FE,
Feng J, Mo K, Jia S, Lutz BM, Bekker A, Nestler EJ, Tao
Declaration of Conflicting Interests YX. DNA methyltransferase DNMT3a contributes to neu-
ropathic pain by repressing Kcna2 in primary afferent neu-
The author(s) declared no potential conflicts of interest with
rons. Nat Commun 2017; 8: 14712.
respect to the research, authorship, and/or publication of 11. Hu SJ, Xing JL. An experimental model for chronic com-
this article. pression of dorsal root ganglion produced by interverte-
bral foramen stenosis in the rat. Pain 1998; 77: 15–23. Jul
Funding 12. Xie Y-B, Zhao H, Wang Y, Song K, Zhang M, Meng F-C,
The author(s) disclosed receipt of the following financial sup- Yang Y-J, He Y-S, Kuang F, You S-W, You H-J, Xu H.
Bilateral neuropathy of primary sensory neurons by the
port for the research, authorship, and/or publication of this
chronic compression of multiple unilateral DRGs. Neural
article: The authors thank for the support from the National
Plast 2016; 2016: 2130901.
Natural Sciences Foundation of China (81701112, 31871065 to 13. Liang L, Fan L, Tao B, Yaster M, Tao YX. Protein kinase
L. Liang; 81870886 to F.-Q. Huo) and the China Postdoctoral B/Akt is required for complete Freund’s adjuvant-induced
Science Foundation Grant (2018M633527) to L. Liang. upregulation of Nav1.7 and Nav1.8 in primary sensory
neurons. J Pain 2013; 14: 638–647.
ORCID iD 14. Li Z, Mao Y, Liang L, Wu S, Yuan J, Mo K, Cai W, Mao
Lingli Liang https://orcid.org/0000-0001-7594-387X Q, Cao J, Bekker A, Zhang W, Tao YX. The transcription
factor C/EBPbeta in the dorsal root ganglion contributes
to peripheral nerve trauma-induced nociceptive hypersen-
References sitivity. Sci Signal 2017; 10: eaam5345.
1. Jones RC III, Lawson E, Backonja M. Managing neuro- 15. Otsuka S, Adamson C, Sankar V, Gibbs K M, Kane-
pathic pain. Med Clin North Am 2016; 100: 151–167. Goldsmith N, Ayer J, Babiarz J, Kalinski H, Ashush H,
2. Woolf CJ. Central sensitization: uncovering the relation Alpert E, Lahav R, Feinstein E, Grumet M. Delayed intra-
between pain and plasticity. Anesthesiology 2007; thecal delivery of RhoA siRNA to the contused spinal cord
106: 864–867. inhibits allodynia, preserves white matter, and increases
3. Pastore S, Mascia F, Mariani V, Girolomoni G. The epi- serotonergic fiber growth. J Neurotrauma 2011;
dermal growth factor receptor system in skin repair and 28: 1063–1076.
inflammation. J Invest Dermatol 2008; 128: 1365–1374. Jun 16. Li Z, Gu X, Sun L, Wu S, Liang L, Cao J, Lutz B M,
4. Martin LJ, Smith SB, Khoutorsky A, Magnussen CA, Bekker A, Zhang W, Tao Y-X. Dorsal root ganglion mye-
Samoshkin A, Sorge RE, Cho C, Yosefpour N, loid zinc finger protein 1 contributes to neuropathic pain
Sivaselvachandran S, Tohyama S, Cole T, Khuong TM, after peripheral nerve trauma. Pain 2015; 156: 711–721.
Mir E, Gibson DG, Wieskopf JS, Sotocinal SG, Austin JS, 17. Dixon WJ. Efficient analysis of experimental observations.
Meloto CB, Gitt JH, Gkogkas C, Sonenberg N, Greenspan Annu Rev Pharmacol Toxicol 1980; 20: 441–462.
JD, Fillingim RB, Ohrbach R, Slade GD, Knott C, 18. Chaplan SR, Bach FW, Pogrel JW, Chung JM, Yaksh TL.
Dubner R, Nackley AG, Ribeiro-da-Silva A, Neely GG, Quantitative assessment of tactile allodynia in the rat paw.
Maixner W, Zaykin DV, Mogil JS, Diatchenko L. J Neurosci Methods 1994; 53: 55–63. Jul
Epiregulin and EGFR interactions are involved in pain 19. Wu S, Marie Lutz B, Miao X, Liang L, Mo K, Chang YJ,
processing. J Clin Invest 2017; 127: 3353–3366. Du P, Soteropoulos P, Tian B, Kaufman AG, Bekker A,
5. Citri A, Yarden Y. EGF-ERBB s: towards the systems Hu Y, Tao YX. Dorsal root ganglion transcriptome anal-
level. Nat Rev Mol Cell Biol 2006; 7: 505–516. ysis following peripheral nerve injury in mice. Mol Pain
6. Kersten C, Cameron MG. Cetuximab alleviates neuro- 2016; 12: 1744806916629048.
pathic pain despite tumour progression. BMJ Case Rep 20. Liang L, Wang Z, Lu N, Yang J, Zhang Y, Zhao Z.
2012; 2012: 1–6. doi: 10.1136/bcr.12.2011.5374 Involvement of nerve injury and activation of peripheral
7. Kersten C, Cameron MG, Laird B, Mjaland S. Epidermal glial cells in tetanic sciatic stimulation-induced persistent
growth factor receptor-inhibition (EGFR-I) in the treat- pain in rats. J Neurosci Res 2010; 88: 2899–2910.
ment of neuropathic pain. Br J Anaesth 2015; 21. Luo H, Cheng J, Han JS, Wan Y. Change of vanilloid
115: 761–767. receptor 1 expression in dorsal root ganglion and spinal
8. Briggs CA, Chandraraj S. Variations in the lumbosacral dorsal horn during inflammatory nociception induced by
ligament and associated changes in the lumbosacral region complete Freund’s adjuvant in rats. Neuroreport 2004; 15:
resulting in compression of the fifth dorsal root ganglion 655–658. 22
and spinal nerve. Clin Anat 1995; 8: 339–346. 22. Yu L, Yang F, Luo H, Liu FY, Han JS, Xing GG, Wan Y.
9. Zhang J, Liang L, Miao X, Wu S, Cao J, Tao B, Mao Q, The role of TRPV1 in different subtypes of dorsal root
Mo K, Xiong M, Lutz B M, Bekker A, Tao Y-X. ganglion neurons in rat chronic inflammatory nociception
12 Molecular Pain

induced by complete Freund’s adjuvant. Mol Pain 2008; responses in corneal epithelial cells. Exp Eye Res 2010;
4: 61. 91: 462–471.
23. Hatashita S, Sekiguchi M, Kobayashi H, Konno S, 28. Liu F-Y, Sun Y-N, Wang F-T, Li Q, Su L, Zhao Z-F,
Kikuchi S. Contralateral neuropathic pain and neuropa- Meng X-L, Zhao H, Wu X, Sun Q, Xing G-G, Wan Y.
thology in dorsal root ganglion and spinal cord following Activation of satellite glial cells in lumbar dorsal root gan-
hemilateral nerve injury in rats. Spine (Phila Pa 1976) glia contributes to neuropathic pain after spinal nerve liga-
2008; 33: 1344–1351. tion. Brain Res 2012; 1427: 65–77.
24. Zhang Y, Zheng L, Zhang J, Dai B, Wang N, Chen Y, He 29. Huerta JJ, Diaz-Trelles R, Naves FJ, Llamosas MM, Del
L. Antitumor activity of taspine by modulating the EGFR Valle ME, Vega JA. Epidermal growth factor receptor in
signaling pathway of Erk1/2 and Akt in vitro and in vivo. adult human dorsal root ganglia. Anat Embryol 1996;
Planta Med 2011; 77: 1774–1781. 194: 253–257.
25. Chen Y, Long H, Wu Z, Jiang X, Ma L. EGF transregu- 30. Ahmed Z, Read ML, Berry M, Logan A. Satellite glia not
lates opioid receptors through EGFR-mediated GRK2 DRG neurons constitutively activate EGFR but EGFR
phosphorylation and activation. Mol Biol Cell 2008; inactivation is not correlated with axon regeneration.
19: 2973–2983. Neurobiol Dis 2010; 39: 292–300.
26. Liu X, Wu WKK, Yu L, Li ZJ, Sung JJY, Zhang ST, 31. Koltzenburg M, Bennett DL, Shelton DL, McMahon SB.
Cho CH. Epidermal growth factor-induced esophageal Neutralization of endogenous NGF prevents the sensitiza-
cancer cell proliferation requires transactivation of tion of nociceptors supplying inflamed skin. Eur J Neurosci
beta-adrenoceptors. J Pharmacol Exp Ther 2008; 1999; 11: 1698–1704.
326: 69–75. 32. Cheng C-F, Cheng J-K, Chen C-Y, Lien C-C, Chu D,
27. Yang H, Wang Z, Capo-Aponte JE, Zhang F, Pan Z, Wang S-Y, Tsaur M-L. Mirror-image pain is mediated
Reinach PS. Epidermal growth factor receptor transacti- by nerve growth factor produced from tumor necrosis
vation by the cannabinoid receptor (CB1) and transient factor alpha-activated satellite glia after peripheral nerve
receptor potential vanilloid 1 (TRPV1) induces differential injury. Pain 2014; 155: 906–920.