Somatosensory Systems, Pain NeuroReport 8, 1613–1618 (1997) 1 1 CHRONIC constriction injury (CCI) of the rat sciatic nerve, which within 3 days induces thermal and mechanical hyperalgesia and mechanical allodynia, is used as a model for pain resulting from nerve injury. Involvement of nerve growth factor (NGF) in the development of this hyperalgesia is suggested by the increase in the level of mRNA encoding NGF in cells in the injured area and in dorsal root ganglia at the level of the lesion and the greatly increased NGF levels (determined by ELISA) in the ganglia ipsilateral to the CCI. Application of antiserum to NGF at the site of CCI delayed the appearance of hyperalgesia, whereas pre-immune serum appeared to enhance it. These results are consistent with the view that NGF is an important factor in the appearance of hyperalgesia associated with unilateral mononeuropathy. Key words: Inflammation; Neuropathy; Pain Nerve Growth 1 1 p There is persuasive evidence that nerve growth factor (NGF) plays an important role in mediating inflammatory pain.1 Systemic administration of large doses of NGF also produces thermal and mechanical sensitivity. These investigators suggested that the mechanical hyperalgesia induced by NGF is probably centrally mediated, because it could be blocked by the non-competitive NMDA receptor blocker MK801.2 Other investigators reported that local injections of NGF act similarly and that administration of NGF antibodies prevents sensitivity to painful stimuli following intraplantar administration of complete Freund’s adjuvant (CFA). The antibody did not prevent paw swelling or erythema, suggesting a direct effect of the NGF on peripheral nerves as a pain promotor.3 The administration of TrkA–IgG fusion protein, which sequesters NGF, prevented the hyperalgesia that develops following carrageenaninduced inflammation.4 The role of NGF in pain produced by chronic nerve injury is less well defined. Nerve growth factor has been reported to alleviate painful peripheral neuropathy following chronic constrictive injury (CCI) of nerves in rats.5 After unilateral nerve crush, however, bilateral induction of NGF mRNA and NGF content in the lumbar dorsal root ganglia (DRG) has been demonstrated.6 Taken together with the evidence for a role of NGF in both peripheral and central components of inflammatory pain, it is © Rapid Science Publishers Uri Herzberg,CA Eli Eliav,1 Jill M. Dorsey, Richard H. Gracely1 and Irwin J. Kopin Clinical Neuroscience Branch, National Institute of Neurological Disorders and Stroke, 9000 Rockville Pike, Bethesda, MD 20892; 1Pain and Neurosensory Mechanisms, National Institute of Dental Research, Bethesda, MD, USA Factor; Introduction 1 NGF involvement in pain induced by chronic constriction injury of the rat sciatic nerve CA Corresponding Author possible that NGF is also involved in pain associated with nerve injury. To test this hypothesis, we examined the effects of application of a polyclonal antibody to NGF at the site of unilateral chronic constriction injury (CCI) of the sciatic nerve. We assessed the effects of the antibody on thermal and mechanical hyperalgesia, mechanical allodynia, and levels of NGF and NGF mRNA in the lumbar DRGs in animals subjected to CCI or sham surgery. Materials and Methods The experiments were performed as approved by the NIDR Animal Care and Use Committee, in accordance with federal law, the regulations of the National Institutes of Health and the guidelines of the International Association for the Study of Pain.7 Surgical procedures: Painful unilateral mononeuropathy was induced in male Sprague–Dawley rats (Taconic Farm, Germantown, NY) weighing 250– 350 g as previously described.8 Briefly, rats were anesthetized with pentobarbital (50 mg/kg, i.p), and four loose ligatures of 4-0 chromic gut were placed around the right sciatic nerve, 1 mm apart, ~2 mm proximal to the nerve trifucation. In sham-operated animals four pieces of ~5 mm 4-0 chromic gut were placed along the nerve trunk. To examine the effects of NGF antibody, the ligation area was wrapped with 2 3 8 mm bands of hemostatic oxidized cellulose (Oxycell, cotton type, Becton Dickinson, Franklin Vol 8 No 7 May 6 1997 1613 U. Herzberg et al. 1 11 11 11 11 11 1p Lakes, NJ) previously saturated with either 150 ml of sheep anti-NGF antibody (diluted 1:10), or the same dilution of pre-immune sheep serum (Accurate Antibodies, Westbury NY). Eight rats were treated with NGF antibody and eight with pre-immune serum. Behavioral assessment and analysis: Heathyperalgesia was assessed as described elsewhere.9 The radiant heat stimulus intensity was adjusted to induce a pre-surgical latency of ~10 s. Decrements (in seconds) in the time to withdrawal (paw withdrawal latency, PWL) were used as an index of hyperalgesia, whereas increments in PWL indicated hypoalgesia. Mechanical allodynia is expressed as the log ratio of the force (in mg) required to elicit a withdrawal response from the operated side divided by the unoperated side. Increments in duration of paw withdrawal (in seconds) after a pin prick or cold stimulus were used as indices of mechanical and cold hyperalgesia, respectively. Cold stimulus was induced by applying 200 ml acetone to the plantar surface of the hind paw with a blunt needle as previously described.10,11 NGF ELISA: NGF was assayed with a two site enzyme-linked immunoassay (ELISA) using antibodies and reagents purchased from Boehringer Mannheim (Indianapolis IN). Corning #25801 plates were incubated for 2 h at 37°C with 150 ml of 0.5 mg/ml anti-NGF monoclonal antibody (BMB #1008218) in 0.05 M carbonate buffer, pH 9.6. Nonspecific binding to the plates was then blocked by incubation for 30 min at 37°C with 0.5% bovine serum albumin in coating buffer (0.05 M carbonate buffer, pH 9.6) followd by three washes in 50 mM Tris–HCl, 0.2% Triton X-100, pH 7.4. Eight rats (for each time point) which had been subjected to CCI or sham operation were deeply anesthetized with sodium pentobarbital and perfused transcardially with ice cold saline before harvesting DRGs from lumbar regions L2–L6. Individual DRGs were weighed and homogenized (Kontes, Micro ultrasonic cell disrupter, Vineland NJ) in 350 ml buffer (100 mM Tris–HCl pH 7.4) containing 0.2% Triton X-100, 2% bovine serum albumin, 1% PMSF, 7 mg/ml aprotinin (ICN, Aurora OH) and 4 mM EDTA. After pulse centrifugation for 1–2 min (Sorval MC 12V), 100 ml of supernatant was applied to well, in triplicate, together with standards (mouse NGF-b) and incubated at 4°C overnight. Plates were washed three times as described above and 100 ml anti-NGF antibody (0.2 U/ml) conjugated to b-galactosidase (BMB #1008234) was applied to each well. After incubation for 6 h at 37°C, plates were again washed three times and then incubated for 6 h at 37°C with 150 ml substrate solu1614 Vol 8 No 7 6 May 1997 FIG. 1. CCI-induced thermal hyperalgesia. Thermal hyperalgesia is expressed as decrements in paw withdrawal latency (in seconds) of the right hind paw (CCI or sham operated side) compared to the left, unoperated side upon exposure of the plantar side of the hindpaw to a radiant heat.* p < 0.05 v sham-operated animals. tion containing 2 mg/ml chlorphenol red–b-D-galactopyranoside, 100 mM HEPES, 150 mM NaCl, 2 mM MgCl2 and 1% bovine serum albumin. Absorbance was read at 570 nm (Molecular Devices microplate reader). Assay sensitivity was ~10 pg/ml. Immunohistochemistry and in situ hybridization: Accumulation of NGF in DRG soma was demonstrated by immunohistochemistry for NGF protein; in situ hybridization for the NGF mRNA was used to determine sites of NGF synthesis. DRGs, sciatic nerve trunks and spinal cords from four CCI and four sham-operated animals were examined. For immunohistochemistry, we used a modification of the previously described procedure.12 Deeply anesthetized rats were perfused with 250 ml ice cold saline (pH 11.0) followed by 4% paraformaldehyde in phosphate buffered saline (PBS) (pH 11.0). DRGs, spinal cord and sciatic nerves were harvested and cryprotected by overnight fixation in 30% sucrose in PBS. Frozen sections (20 mm) were cut in a cryostat (Leica 2080) and mounted on gelatin-coated slides. Slides were incubated with pre-immune rabbit serum (4% in PBS) for 1 h at room temperature, followed by an overnight incubation at 4°C with a primary antibody solution consisting of 1:1000 dilution of sheep anti-NGF (Accurate Antibodies, Westbury NY), 4% normal rabbit serum, 1% bovine serum albumin and 0.5% Triton X-100. Control sections in which the primary antibody was omitted were also included. After three washes in PBS, the slides were incubated with rabbit anti-sheep IgG (1:200 dilution) conjugated to tetramethyl rhodamine (Jackson Laboratories, West Grove PA), washed three times NGF involvement in a rat model of neuropathic pain 1 FIG. 2. CCI-induced mechanical allodynia. Mechanical allodynia expressed as log ratio (injured/non-injured) of the threshold force required to elicit hind paw withdrawal from graded mechanical stimuli (von Frey filaments).* p < 0.05 v pre-surgical baseline values. 1 FIG. 4. Total NGF contents in dorsal root ganglia ipsilateral (A, right) and contralateral (B, left) to the sciatic nerve subjected to CCI. The control NGF levels in gangllia of sham operated or unoperated animals were 100–300 pg/ganglia (shaded area). *p < 0.05 v control values. 1 FIG. 3. CCI-induced hyperalgesia to pin prick and cold. Hyperalgesia is expresses as the increment in duration (seconds) of hindpaw withdrawal following application of the noxious stimulus.* p < 0.05 v presurgical baseline values. 1 1 p in PBS and coverslipped using Flourmount embedding media (Gurr, UK). For in situ hybridization, deeply anesthetized animals were perfused with 250 ml ice cold saline (pH 7.4) followed by 4% paraformaldehyde and 0.5% glutaraldehyde in PBS (pH 7.4). Spinal cord, DRG and sciatic nerves were harvested and cryoprotected with overnight fixation in 30% sucrose (RNAse free) in autoclaved PBS. Frozen sections (20 mm) were cut in a cryostat (Leica 2080) and mounted on poly-Llysine coated slides. Pre-hybridization and hybridization were performed in sealed plastic chambers. Hybribuffer ISH (Biognostics, Germany) was used as buffer for probe dilution as well as for prehybridization. A purified probe of deprotected DNA with 30 bases complimentary to a target region of bases 304–333 of the coding region for NGF, conjugated to digoxigenin, as well as a random control sequence consisting of the same bases also conjugated to digoxigenin were used (Biognostics Germany). Probe sequence was 59-GTCAAGGGAATGCTGAAGTTTAGTCCAGTG-39, the reverse complement to the target gene.13 After prehybridization at 30°C for 4 h, sections were hybridized overnight at 30°C with a solution containing 2.8 pmol/ml purified probe or the random sequence control (15–30 ml/section). Following four washes with 0.13 SCC at 40°C sections were reacted with a 1:50 dilution of the Fab fragment of sheep anti-digoxigenin IgG, conjugated to fluorescein (Boehringer Mannheim, Indianapolis IN) for 2 h at room temperature, washed three times in PBS, dehydrated and coverslipped using Flourmount embedding media (Gurr, UK). Results Animals subjected to unilateral sciatic nerve CCI showed increased sensitivity to heat (Fig. 1), mechanVol 8 No 7 6 May 1997 1615 U. Herzberg et al. 1 11 11 11 11 11 1p FIG. 5. Effect of local application of NGF antibody or pre-immune serum (see methods) on sciatic nerve CCI-induced thermal hyperalgesia. Thermal hyperalgesia is expressed as decrements in paw withdrawal latency (in seconds) of the right hind paw (CCI side) compared with the left, unoperated side, upon exposure of the plantar side of the hindpaw to a radiant heat. Hypoalgesia is indicated by negative values. *p < 0.05 v pre-immune serum treated nerves. ical stimuli (Fig. 2) and cold and pin prick stimuli (Fig. 3) in the ipsilateral paw. Sham-operated animals showed no hyperalgesia or allodynia, and no significant changes in the sensitivity to heat, cold, pin prick stimuli or threshold to a mechanical stimulus observed in the hind limbs contralateral to the ligated sciatic nerve. In animals subjected to unilateral sciatic nerve CCI, DRG NGF immunoreactivity was increased bilaterally, localized mainly in the cell bodies. Both large and small diameter neuronal cell bodies demonstrated increased NGF immunoreactivity (Fig. 4a,c). Quantification using an ELISA revealed that the rise was greatest in the DRGs ipsilateral to the injured nerve in DRGs corresponding to lumbar regions L4–5 (Fig. 5a,b). NGF content in DRGs of sham-operated animals remained between 100 and 300 pg (shaded areas in Fig. 5a,b). Weights of the DRGs from CCI animals were not significantly different from those from sham-operated animals, nor were there significant differences between the weights of left and right DRGs in sham-operated or CCI animals. NGF mRNA was detected in both left and right DRGs following right sciatic nerve CCI, as well as in cells surrounding the injured nerve and cells within the injured nerve trunk (Fig. 4d,e,f). No NGF mRNA was detected in DRGs or sciatic nerve of sham-operated animals or in the unoperated sciatic nerve of CCI animals. Treatment of the injury site with antibodies to NGF immediately following nerve ligation delayed the onset of both thermal hyperalgesia and mechan1616 Vol 8 No 7 6 May 1997 FIG. 6. Effect of local application of NGF antibody or pre-immune serum (see methods) on CCI-induced mechanical allodynia. Mechanical alloydynia expressed as log ratio (injured/non-injured) of the threshold force required to elicit hind paw withdrawal from graded mechanical stimuli (von Frey filaments). *p < 0.05 v preimmune serum treated nerves. FIG. 7 Effect of local application of NGF antibody or pre-immune serum (see methods) on CCI-induced hyperalgesia to pin prick and cold. Hyperalgesia is expressed as the increment in duration (seconds) of hindpaw withdrawal following application of the noxiuos stimulus. *p < 0.05 v pre-immune serum treated nerves. ical allodynia (Figs 6,7) and significantly decreased the degree of cold and mechanical hyperalgesia (Fig. 8), but the differences became insignificant 7–9 days after surgery. Discussion Experimental inflammatory pain is induced by the injection of carrageenan or CFA into the hind paw or the tibio-tarsal joint of rats or mice. Following injection of carrageenan into the joint, pain behavior NGF involvement in a rat model of neuropathic pain 1 1 1 1 1 p FIG. 8. (a, b, c) NGF immunoreactivity in dorsal root ganglion (L4) ipsilateral to a ligated sciatic nerve (a), or sham operated animal (b), or contralateral to the ligated nerve (c), 9 days following nerve injury. NGF immunoreactivity is predominantly localized to cell soma. Bar (a and b)=50mm, or 150 mm (c.) (d, e, f) NGF mRNA detected using in situ hybridization (see methods); (d) injured nerve trunk, arrowheads point to inflammatory cells expressing NGF mRNA outside the nerve trunk, arrows point to BGF mRNA expressed inside the nerve trunk, bar = 75mm (e) and (f) NGF mRNA in dorsal root ganglia L4 ipsilateral (e) and contralateral (f) to an injured sciatic nerve. Arrowheads point to digoxigenin immunoflourescence attributed to NGF mRNA, bar =15mm. lasts up to 3 days,14 whereas following CFA injection into the joint the pain subsides after 7–21 days and sensitivity returns to normal at 4 weeks.15 In a recently described model of neuritis, Surgicell containing CFA was applied to the sciatic nerve trunk. The induced pain lasts up to 5 days and there is minimal nerve damage.16 The role of NGF in mediating the hyperalgesia associated with inflammation has been reviewed.1 Anti-NGF neutralizing antibodies have been shown to prevent the development of hyperalgesia that results from intraplantar injec- tion of complete Freund’s adjuvant and to block upregulation of substance P and calcitonin gene related peptide in primary afferent sensory neurons.3 In these models of inflammation the duration of pain behavior is shorter than that induced by nerve injury: pain behavior usually lasts at least 2 months after loose ligation of the sciatic nerve.8,17 In another model of chronic pain induced by nerve injury, in which spinal nerves L5 and L6 are tightly ligated, mechanical allodynia and heat hyperalgesia last more than 10 weeks and cold allodynia lasts at least 16 weeks.11 The role Vol 8 No 7 6 May 1997 1617 U. Herzberg et al. 1 11 11 11 11 11 of NGF produced at the nerve injury site or in the DRG ipsilateral to the injured nerve (see Fig. 5) has not been determined. Our results, demonstrating a sustained expression of NGF at the nerve injury site and at DRGs L4–L5 and a prolonged increase in NGF immunoreactivity in DRGs L4–L5 of CCI animals point to a possible mechanism mediating the prolonged pain following nerve injury. Although there is a bilateral increase in NGF mRNA in the dorsal root ganglia after sciatic nerve crush,6, in animal models of nerve injury the behavioral changes associated with pain are ipsilateral to the injured nerve.8,18,19 In the present study we found that although increased bilaterally, the NGF increase is more pronounced in ipsilateral DRGs L4 and L5. The delay in the onset of pain behavior produced by the direct application of NGF antibodies at the injury site suggests that, at least during their the initiation, hyperalgesia and allodynia are mediated by production of NGF at the site of injury. Our results contrast with previously published data that demonstrated that NGF treatment protects against streptozocininduced diabetic neuropathy in rats.20 The difference presumably stems from different mechanisms leading to the painful state in the two animal models. Whereas in the CCI model pain may be related to NGF produced as a result of nerve injury, in diabetic neuropathy fewer sensory neurons are immunoreactive to substance P (SP) or calcitonin gene related peptide.21 NGF administration has been reported to restore the levels of these peptides in sensory neurons.22 Our results are also in apparent contrast to previously reported results5 showing that NGF application at the injury site abolished thermal hyperalgesia and reduced the mechanical allodynia produced by CCI of the sciatic nerve. The apparent contradiction could stem from the mode of application; whereas in the mentioned report NGF was administered via a slow minipump infusion, we applied NGF antiserum in a cellulose matrix. As the kinetics of antibody release from such a matrix are not known, it is difficult to speculate about the amount of antibody reaching the injury site. A biphasic response to NGF could also account for the apparent contrast in the results. Although the results of this study suggest that endogenous NGF has a role in the development of chronic pain after CCI, it cannot be determined whether the effects of NGF result from its trophic effects on the sympathetic nerves,23,24 at the dorsal root ganglia or in the spinal cord or by its actions in sensitizing primary afferents of pain fibers.25 Conclusion We demonstrate a role for NGF in the initiation of pain following unilateral constriction injury as well as an increase in NGF immunoreactivity in the dorsal root ganglia associated with the injured nerve in the first 2 weeks after surgery. The prolonged increase in NGF immunoreactivity correlates well with the prolonged pain behavior displayed in this particular pain model. NGF mRNA was detected at the site of nerve injury and in the DRG’s receiving fibers from both the injured and the non-injured nerve. References 1. Levi-Montalcini R, Dal Toso R, della Valle F et al. J Neurol Sci 130, 119–127 (1995). 2. Lewin GR, Rueff A and Mendell LM. Eur J Neurosci 6, 1903–1912 (1994). 3. Woolf CJ, Safieh-Garabedian B, Ma QP et al. Neuroscience 62, 327–331 (1994). 4. McMahon SB, Bennett DL, Priestly JV and Shelton DL. Nature Med 1, 740–741 (1995). 5. Ren K, Thomas DA and Dubner R. Brain Res 699, 286–292 (1995). 6. Wells MR, Vaidya U and Schwartz JP. Exp Brain Res 101, 53–58 (1994). 7. Zimmermann M. Pain 16, 109–110 (1983). 8. Bennett GJ and Xie Y-K. Pain 33, 87–107 (1988). 9. Hargreaves K, Dubner R, Brown F et al. Pain 32, 77–88 (1988). 10. Tal M and Bennett GJ. Pain 57, 375–382 (1994). 11. Choi Y, Yoon YW, Na HS et al. Pain 59, 369–376 (1994). 12. Zhou XF, Zettler C and Rush RA. J Neurosci Methods 54, 95–102 (1994). 13. Borsani G, Pizzuti A, Rugarli EI et al. Nucleic Acids Res 18, 4020 (1990). 14. Schott E, Berge OG, Angeby–Moller K et al. J Pharmacol Toxicol Methods 31, 79–83 (1994). 15. Malcangio M and Bowery NG. Br J Pharmacol 113, 1561–1566 (1994). 16. Eliav E, Ruda MA and Bennett GJ. Soc Neurosci Abstr 26, 865 (1996). 17. Guilbaud G, Gautron M, Jazat F et al. Pain 53, 147–158 (1993). 18. Seltzer Z, Dubner R and Shir Y. Pain 43, 205–218 (1990). 19. Carlton SM, Lekan AH, Kim SH and Chung JM. Pain 56, 155–166 (1994). 20. Apfel SC, Arezzo JC, Brownlee M et al. Brain Res 634, 7–12 (1994). 21. Sango K, Verdes JM, Hikawa N et al. J Neurol Sci 126, 1–5 (1994). 22. Schmidt Y, Unger JW, Bartke I and Reiter R. Exp Neurol 132, 16–23 (1995). 23. Conner JM and Varon S. Exp Neurol 130, 15–23 (1994). 24. Davis BM, Albers KM, Seroogy KB and Katz DM. J Comp Neurol 349, 464–474 (1994). 25. Dmitrieva N and McMahon SB. Pain 66, 87–97 (1996). Received 10 February 1997; accepted 6 March 1997 General Summary Nerve growth factor (NGF) is known to play a role in pain following inflammatory response, such as that induced by the injection of killed mycobateria. In this study we demonstrated its role in the pain that follows nerve damage. A rise in the NGF content in dorsal root ganglia (DRG) L4 and L5 was demonstrated in rats subjected to chronic constriction injury of the sciatic nerve. The rise was more pronounced in the DRGs ipsilateral to the injured nerve than in contralateral DRGs. Moreover, application of NGF antibodies, but not pre-immune serum, at the site of nerve injury immediately following nerve ligation postponed the intiation of pain behavior by 3 days, and decreased it up to day 9 post-surgery. 1p 1618 Vol 8 No 7 6 May 1997