Trigeminal Nerve Injury ErbB3/ErbB2 Promotes Mechanical Hypersensitivity

• Peripheral nerve injury results in demyelination and activation of Schwann and other support cells around peripheral nerves

• The role of this process in neuropathic pain remains obscure

• In rats, genetic disruption of neuregulin 1, which signals proliferation and activation of Schwann cells, delayed downstream receptor signaling and hypersensitivity to touch on the whisker pad after infraorbital nerve injury

• Lapatinab, an experimental drug in phase 3 trials for the treatment of breast cancer and which targets this downstream receptor signaling pathway, reduced hypersensitivity in rats with nerve injury

TRIGEMINAL neuropathic pain is one of three main neuropathic pain-related diagnoses. The orofacial pain condition is characterized by chronic aching, burning pain, and sudden excruciating, electric-like shooting pain caused by unintentional injury to the trigeminal system. There are few investigations on nociception mechanisms and analgesic trials in the trigeminal region. Thus, new analgesic strategies need to be explored before chronic trigeminal neuropathic pain can be relieved successfully.

Deterioration of the trigeminal nerve myelin sheaths is one causal mechanism for trigeminal neuropathic pain. Mechanical hypersensitivity accompanying trigeminal neuropathic pain involves spontaneous and low-threshold activity in injured myelinated fibers.1These pathologic changes cause ectopic discharge or impulse generation from sites along axons where damage has occurred, rather than just at sensory nerve endings.2Sciatic nerve constrictive injury induces demyelination and is a source of pathologic ectopic firing leading to mechanical allodynia and heat hyperalgesia.3Neuregulin 1 (Nrg1) is a peptide ligand signaling via  the receptor tyrosine kinase ErbB3-ErbB2 heterodimer.4The Nrg1-ErbB3-ErbB2 signaling complex plays a key role in axon-Schwann cell interactions promoting Schwann cell development and remyelination after nerve injury. Nrg1 binds catalytically inactive ErbB3, whereas ErbB2 contributes tyrosine kinase activity in Schwann cells.5,6Blockade of ErbB-receptor–initiated cell signaling in glial cells wrapping either myelinated or nonmyelinated nerves produces unique sensory dysfunctions.7Thermal and mechanical hypersensitivity and thin myelination of sciatic nerves are found in ErbB receptor mutant mice.8,9Up-regulation of ErbB3 messenger RNA is maintained in different types of median nerve regenerating axons after upper limb injury, whereas ErbB2 messenger RNA is unchanged.10Possible changes of ErbB after nerve injury or a role in generating trigeminal neuropathic pain have not been explored and are the focus of this study. The current experiments used Nrg1 transgenic rats (Nrg1Tg) to explore the behavioral changes in whisker pad and ErbB receptor concentrations in injured nerve after infraorbital nerve chronic construction injury (CCI-ION). Lapatinib (Tykerb), an ErbB2 inhibitor, was applied after CCI-ION to investigate its effect on the pain-related behavior. Lapatinib is used in ErbB2-positive breast cancer treatment and reportedly has reduced pain in a phase 2 clinical trial.11We also investigated the relationship of ErbB3 concentrations in injured nerves to the development of whisker pad mechanical allodynia after CCI-ION. We hypothesized that the ErbB receptor concentration increase at infraorbital nerve injury sites promotes mechanical hypersensitivity in the whisker pad.

Adequate measures were taken to minimize pain or discomfort in these studies. Experiments were carried out in accordance with the Guidelines of the National Institutes of Health regarding the care and use of animals for experimental procedures. Experiments were approved by the Institutional Animal Care and Use Committee at the University of Kentucky, Lexington, Kentucky. Male Fischer 344 wild type or Nrg1Tg rats were accommodated in ventilated animal housing with a reverse light–dark 12:12 hr cycle and weighed 200–250 g at the beginning of the experiments. A breeding pair of Fischer 344 Nrg1Tg rats generated by introduction of a Sleeping Beauty transposon with a bidirectional splice acceptor gene trap12,13was provided by the Knock Out Rat Consortium. A nonsense mutational insertion into intron 1 of the sequence encoding for Nrg1 domain resulted in ablation of all Nrg1 isoforms. The Nrg1Tg rats used in this study were bred in our animal facility, and the Nrg1^{Tn(sb-T2/Bart3)2.183Mcwi}deficient genotype was confirmed.

Rats were genotyped by polymerase chain reaction using the following primers: 5′-AGGAACCAAAAAAGTAGACTCAGTGTG-3′, 5′-GTGAGCTTTTCTGTAAGGTGGTAACTT-3′, and transposon primer 5′-CTGACCTAAGACAGGGAATT-3′. Polymerase chain reaction was carried out at 94°C for 2 min, and then 94°C for 15 s, 60°C for 30 s, and 72°C for 90 s for 35 cycles, followed by a 5-min extension at 72°C by PTC-100 programmable thermal controller (MJ Research Inc., Waltham, MA). The expected 1,041 bp product for wild type allele, 1,041 and 440 bp products for heterozygous allele, and 440 bp product for the mutant allele were separated on a 1% agarose gel.

Rats were anesthetized with intraperitoneal sodium pentobarbital (50 mg/kg), and all surgeries were performed in sterile conditions with the use of a surgical microscope. The infraorbital nerve on one side was exposed using the surgical procedure described by Vos et al.  14The infraorbital nerve was dissected free at its most rostral extent in the orbital cavity and two chromic gut (5–0) ligatures were loosely tied around the infraorbital nerve (2 mm apart). Nerve exposure but no ligation was performed in the sham operation. To obtain the desired degree of constriction, a criterion formulated by Bennett and Xie15was applied (i.e. , ligations reduced the nerve diameter just noticeably). The ligation retarded but did not totally occlude circulation through the superficial vasculature.

Mechanical sensitivity of vibrissal whisker pad, which is the infraorbital nerve receptive field, was measured with eight von Frey fibers (0.4, 0.6, 1, 2, 4, 6, 8, 15 g; Stoelting, Wood Dale, IL) by modified up-and-down method with a default maximal threshold at 18.72 g.16Mechanical stimuli were applied within the infraorbital nerve territory, near the whisker pad centers, both ipsilateral and contralateral to the surgery site. Responses to von Frey filaments applied to the rat whisker pad determined the threshold required for 50% head withdrawals. Each filament was applied five times at intervals of a few seconds. If head withdrawal was observed at least three times after probing with a filament, the rat was considered responsive to that filament. When a positive response to a stimulus occurred, the next smaller von Frey filament was applied. Otherwise, the next higher filament was applied. Behavioral changes to mechanical stimuli were tested once a week for 10 weeks after surgery. Intraperitoneal injection of tyrosine kinase ErbB2 inhibitor Lapatinib (0, 0.2, 1, 5 mg/kg, in 300 μl dimethyl sulfoxide, Selleckchem, Houston, TX)17was administered to CCI-ION rats after mechanical allodynia was confirmed. Behavioral responses were determined for whisker pads on both sides at 30 min and 1, 3, and 6 h after Lapatinib administration.

Rats were anesthetized with intraperitoneal sodium pentobarbital (70 mg/kg) and perfused transcardially with heparinized saline followed by 4% ice-cold paraformaldehyde in 0.1 M phosphate buffer solution (pH 7.4).

Infraorbital nerves were dissected and placed in the same fixative solution at 4°C overnight. Samples were switched to ethanol (70%), dehydrated through graded ethanol, and embedded in paraffin. Infraorbital nerve tissue sections were cut (5 μm), mounted onto glass slides (Superfrost Plus; VWR, Radnor, PA), deparaffinized (Citrisolv; Fisher Scientific, Pittsburgh, PA), dehydrated with graded ethanol, and rinsed in ddH²O.

Slides were immersed in 0.1% hematoxylin, washed in tap water, immersed in 0.1% eosin, and washed in distilled water. Sections were dehydrated in ethanol and coverslipped (Permount, Fisher Scientific).

After block of nonspecific antigen sites with 3% normal goat serum (30 min), sections were incubated overnight (4°C) with rabbit antimyelin protein zero (1:1,000, Abcam, Cambridge, MA) and mouse antipan-axonal-neurofilament protein (1:1,000, Covance, Princeton, NJ) antibodies. Subsequently, sections were incubated with secondary antibodies, fluorescein isothiocyanate donkey antimouse antibody and Texas Red donkey antirabbit antibody (1:1,000, 1 h, Santa Cruz, Santa Cruz, CA). Sections were coverslipped with glycerol-based mounting media (Vector Laboratories, Burlingame, CA). Staining was visualized using a Nikon E1000 microscope (Nikon Instruments, Inc., Melville, NY) equipped with MetaVue and Act-1 Programs. Photomicrographs were taken under the same conditions. Fluorescent intensity was analyzed offline using an advanced image-analysis system MetaMorph (Molecular Devices, Sunnyvale, CA). The numbers of axons with intact myelin sheaths in the best five representative sections of three rats from each group were counted.18,19The person who counted the number of myelinated axons was blinded to experimental groups.

Infraorbital nerve fragments from the ligation site were dissected and homogenized in radioimmunoprecipitation assay buffer. Total protein samples (50 μg) were loaded onto SDS-PAGE (Bio-Rad, Hercules, CA), transferred to polyvinylidene fluoride membrane, blocked overnight, and probed with rabbit anti-ErbB3 (1:200) and mouse anti–β-actin (1:10,000, Santa Cruz, Santa Cruz, CA) (2 h). The membrane was incubated with antirabbit and antimouse peroxidase-conjugated secondary antibody (1:10,000; GE Healthcare, Piscataway, NJ) (1 h). Immunoreactive proteins were detected by enhanced chemiluminescence (Amersham Biosciences, Pittsburgh, PA). Bands recognized by the primary antibody were visualized by X-ray film exposure and analyzed with Image J (National Institutes of Health, Baltimore, MD). The intensities of the ErbB3 bands were normalized to β-actin intensities, and the relative intensities of ErbB3 from each group were compared.

All data were expressed as mean ± SD, analyzed using the Prism 4 statistical program (Graph Pad Software, Inc., La Jolla, CA). The behavioral changes after nerve injury among four groups (Nrg1Tg and wild type rats with or without nerve ligation) for ipsilateral or contralateral sides were analyzed on a weekly basis by two-way ANOVA followed by Bonferroni posttests. The odds ratio for the effect in week 2 for wild type versus  transgenic rats was calculated using McNemar's test with a 95% confidence interval. The mechanical threshold changes on both sides of four groups (Nrg1Tg and wild type rats with or without nerve ligation) over 10 weeks and the time-course data from 0.5, 1, 3, and 6 h after Lapatinib administration were analyzed by repeated-measures, one-way ANOVA followed by Tukey multiple comparison post hoc  tests. For the dose-dependent responses and the western blot data from each group and groups compared with initial concentration, comparisons were executed by independent t test with two-tailed P  values. The same comparisons were done for the contralateral side. P ⩽ 0.05 was considered significant. There were no missing data from any of the experimental results.

Behavioral testing to determine mechanical threshold of the whisker pad to von Frey fiber stimuli was performed once a week for 10 weeks after CCI-ION in both Nrg1Tg and wild type rats. Behavioral alterations observed after nerve injury were indicative of severe sensory disturbances within the injured nerve receptive territory (i.e. , whisker pad). After constrictive nerve injury, rats exhibited responses to mechanical stimuli in two phases. First phase: During the first 2 weeks after CCI-ION there was no change in responses to mechanical stimuli applied to whisker pad with any of the eight test filaments. Second phase (mechanical allodynia): Two weeks or more after CCI-ION surgery, low intensity stimuli applied to the injured nerve territory evoked a sensitized response in wild type rats. Wild type rats showed mechanical allodynia (3.11 ± 5.93 g, n = 9) in week 2 after nerve injury that persisted at least 8 weeks (i.e. , a statistically significant decrease in mechanical threshold compared with responses of the sham group [18.72 ± 0.00 g, n = 4; P < 0.001; fig. 1A]).

There was no mechanical threshold difference between wild type and Nrg1Tg rats before surgery or week 1 after CCI-ION. In Nrg1Tg rats, development of second phase hypersensitive responses (4.39 ± 3.51 g, n = 8) (reduced threshold) occurred with a statistically significant delay (week 6) compared with wild type rats (i.e. , the first phase lasted for 5 weeks rather than 1 week). Thus, differences in mechanical thresholds on the whisker pads were statistically significant between wild type and Nrg1Tg rats in weeks 2–5 (3.44 ± 4.60 g vs.  18.72 ± 0.00 g, P < 0.001) and in week 6 (3.05 ± 4.73 g vs.  9.42 ± 7.74 g, P < 0.01) after surgery. Second phase hypersensitivity to mechanical stimuli persisted for 5 weeks in Nrg1Tg rats, from week 6 to week 10 (P < 0.01 in week 6 and P < 0.001 in weeks 7–10 compared with sham group of Nrg1Tg rats, n = 4). Mechanical thresholds on the contralateral whisker pads were decreased compared with thresholds of the contralateral sides for both sham groups (wild type and Nrg1Tg rats) (11.54 ± 4.70 g, 11.01 ± 2.49 g, 10.46 ± 5.10 g, P < 0.05 in weeks 5, 8, 9 in wild type rats and 11.47 ± 6.65 g, 10.89 ± 6.87 g, P < 0.05 in weeks 8 and 10 in Nrg1Tg rats) (fig. 1B).

The effect of the ErbB2 inhibitor Lapatinib on mechanical allodynia was tested at different doses comparing 0.2, 1, and 5 mg/kg to vehicle treatment. Lapatinib was given intraperitoneally to wild type rats with CCI-ION in week 4 after surgery. Mechanical allodynia was confirmed before drug administration. Mechanical thresholds to von Frey filaments were tested on whisker pads at 30 min and 1, 3, and 6 h after drug injection. Lapatinib dose-dependently alleviated mechanical allodynia in wild type rats with CCI-ION at 3 h after drug administration (fig. 2A). Mechanical threshold increased ipsilaterally at 30 min and 1 and 3 h in response to 5 mg/kg Lapatinib. The effect reached a peak at 3 h (15 ± 3.89 g vs.  2.45 ± 1.13 g, P < 0.001, n = 6) and diminished at 6 h (fig. 2B). There was no difference in mechanical threshold on the contralateral sides before or after Lapatinib administration (dimethyl sulfoxide vehicle, 0.2, 1, and 5 mg/kg).

Hematoxylin and eosin staining was performed on cross sections obtained from infraorbital nerve injury sites and equivalent nerve segments from the sham group 5 weeks after surgery. Sections from the sham group showed the normal peripheral nerve structure with dense labeling of axons and paler staining of myelin sheaths (fig. 3A inset). In injured nerves, axons were swollen, myelin was diminished, and normal structure was disrupted (fig. 3B inset ). However, myelin disruption was more evident with immunofluorescence identifying myelin protein zero (red , fig. 3A). CCI-ION–damaged myelin sheaths were evident when stained for myelin protein zero and counter staining of axons with pan-axonal-neurofilament antibody (green ) (fig. 3B). Most myelinated axons were damaged in the injured nerves, and few remained with intact myelin sheaths (fig. 3C, table 1). Large myelinated axons (5–10 μM) were absent.

The ErbB3 receptor protein concentrations at the CCI-ION site were examined using western blots normalized to β-actin. The ErbB3 receptor concentrations in injured infraorbital nerves from wild type (in week 1) and Nrg1Tg rats (in weeks 1 and 5) were similar those of rats with sham surgery. ErbB3 concentration increases were statistically significant in weeks 5 and 10 relative to sham wild type rats and rats with CCI-ION in week 1 (P = 0.038 in week 5 and P = 0.013 in week 10, respectively, P = 0.041, 0.014 vs.  CCI group in week 1, n = 3) (fig. 4A, B). ErbB3 concentration increases were statistically significant in week 10 (P = 0.03 vs.  sham group in week 10, the CCI group in weeks 1 P = 0.034 and 5 P = 0.016, n = 3) in Nrg1Tg rats but not in week 5 (P = 0.26 vs.  sham group in week 5, P = 0.59 vs.  the CCI group in week 1) (fig. 4, C and D). Statistically significant increases in ErbB3 receptor concentrations were concurrent with the behavioral hypersensitivity (i.e. , ErbB3 was increased in week 2 after injury in wild type rats and in week 5 in Nrg1Tg rats). These differences were not observed in the contralateral nerves (fig. 5).

The current study characterizes behavioral, structural, and neurochemical changes after CCI-ION in both wild type and Nrg1Tg rats. Mechanical allodynia persisting at least 8 weeks in wild type rats in our study is consistent with the 120-day CCI-ION model time course reported by Vos et al.  14Myelin sheaths were heavily damaged or destroyed by ligation in both wild type and Nrg1Tg rats. Trigeminal root biopsies from patients with neuropathic pain exhibit pathologic changes such as axonal loss and demyelination.20,21Constrictive sciatic nerve injury induces demyelination as sources of pathologic ectopic firing accompanying mechanical allodynia and heat hyperalgesia.3Nrg1 is involved in peripheral Schwann cell-axon communication, growth, migration, and myelination.22Mice deficient in Bace  1, which cleaves the extracellular domain of Nrg1, show myelin impairment and nociceptive hypersensitivity.23The current study provided data in support of a role for the Nrg1-ErbB3-ErbB2 signaling complex in axon-Schwann cell interactions promoting mechanical allodynia after nerve injury.

There was no initial difference in infraorbital nerve ErbB3 expression from baseline in Nrg1Tg and wild type CCI-ION rats during the first phase. Mechanical allodynia develops immediately after CCI injury in sciatic nerve,24but infraorbital nerve constrictive injury-induced mechanical hypersensitivity typically is evident several days later.14Although the first phase is speculated to be a residual effect of anesthetics or increases of endogenous opioids,25differences in the regions affected (orofacial vs.  somatic) and the mechanisms involved might be relevant. The increases in mechanical allodynia and ErbB3 were both statistically significant in week 5 after infraorbital nerve injury in wild type rats. However, development of mechanical allodynia was delayed in Nrg1Tg rats after CCI-ION. Our results showed that Nrg1Tg rats have a prolonged first phase during which there is no ErbB3 expression increase and no mechanical allodynia. Thus, increases in ErbB3 receptor concentrations after nerve injury were concurrent with whisker mechanical allodynia, suggesting ErbB3 receptor activation is involved integrally in mechanical sensitization. Increases in ErbB3 phosphorylation and expression concentrations have been reported in distal nerves after transection of sciatic nerves.26,27 

ErbB3 activation is a crucial time-dependent event among nerve injury signaling in wild type and Nrg1Tg rats, suggesting Nrg1 involvement in pain-related behavioral changes after injury. Nrg1 has been shown to increase ErbB3 and olfactory glial proliferation in culture.28Nrg1 delivered by virus in transected nerves promoted axonal and Schwann cell growth and increased responses to nociceptive stimuli after sciatic nerve axotomy.29Low Nrg1 in Nrg1Tg rats likely provides minimal ErbB3 activation after nerve injury. It is still unclear what compensatory mechanisms promote increased ErbB3 and sensitized behavior in week 10 in Ngr1Tg rats. It is possible that the Nrg1-ErbB response is slowed even more in the mutant Nrg1Tg rats during recovery from nerve trauma.

ErbB3 signals through its active tyrosine kinase heterodimer ErbB2 to trigger downstream cascades. However, in a study of ErbB2 mutant mice with sciatic nerve transection, ErbB2 was found to be nonessential for Schwann cell proliferation and survival.30Other potential ErbB mechanisms that might be involved include nerve ligation-induced microglial proliferation and activation initiated by Nrg1-ErbB and ERK1/2/p38 mediation of downstream signaling pathways.31,32 

Lapatinib is used in ErbB2 (HER2)-positive breast cancer therapy and in a phase 2 clinical trial reportedly improved patient functioning, symptom relief, and pain.11In the current study, the tyrosine kinase inhibitor Lapatinib provides significant alleviation of mechanical allodynia, indicating ErbB3-ErbB2 involvement in whisker pad mechanical allodynia. Blocking ErbB3 downstream signaling with Lapatinib dose-dependently ameliorated the behavioral allodynia in wild type rats. These results also suggest a relationship between whisker pad mechanical allodynia and infraorbital nerve injury-induced peripheral glial activation through ErbB3-ErbB2. The blunting of the behavioral pain-related behavior only on the ipsilateral side suggests a direct effect of the Lapatinib on the injured nerve firing and/or peripheral glial activation, while having no effect on ongoing central sensitization events at the time points tested. On the contralateral side that was uninjured, there was no difference in infraorbital nerve ErbB3 receptor expression comparing wild type and Nrg1Tg rats in weeks 1, 5, and 10 after the nerve injury. This finding serves as an internal control confirming that the ErbB3 increases in response to the injury.

One important factor contributing to the pathogenesis of trigeminal neuropathic pain is an abnormal compression of the trigeminal nerve by a vein or artery leading to deterioration of myelin sheaths.20The investigation of peripheral glial ErbB3-ErbB2 mechanisms in the current study using the constrictive nerve injury model provides implications for potential clinical therapies to alleviate the leading cause of trigeminal neuropathic pain. Reduction of the behavioral hypersensitivity by the ErbB2 inhibitor and the increase in ErbB3 protein after nerve injury suggest enhanced ErbB3-ErbB2 complex signaling after infraorbital nerve injury mediates mechanical allodynia in rat whisker pad.

Interventions addressing peripheral ErbB3-ErbB2 mechanisms potentially could be better choices as nonsurgical therapy, instead of microvascular depression surgery, which is sometimes an effective treatment for trigeminal neuropathic pain. Similarly, other nerve constrictive injuries may be alleviated by an ErbB2 inhibitor.

The authors thank Ryan Nesemeier, B.S., Technician, Department of Physiology, University of Kentucky, Lexington, Kentucky, for technical assistance, and Robert Danaher, Ph.D., Research Assistant Professor, and Craig Miller, D.M.D., M.S., Professor (tenured), College of Dentistry, University of Kentucky, for comments on the manuscript.