Journal of the Peripheral Nervous System 6:67–72 (2001)
Etanercept reduces hyperalgesia in experimental painful neuropathy Claudia Sommer, Maria Schäfers, Martin Marziniak, and Klaus V. Toyka Department of Neurology, University of Würzburg, Germany
Abstract
Etanercept, a recombinant tumor necrosis factor receptor (p75)-Fc fusion protein competitively inhibits tumor necrosis factor-alpha (TNF). Etanercept has been successfully used in patients with rheumatoid arthritis, where it reduces pain and inflammation. Because locally produced proinflammatory cytokines play a role in pain after nerve injury, we investigated whether etanercept can reduce pain and hyperalgesia in an animal model of painful neuropathy, the chronic constriction injury of the sciatic nerve. C57BL/6 mice received etanercept or sham treatment by local near-nerve injection to the injured nerve or by systemic application. Treatment with etanercept reduced thermal hyperalgesia and mechanical allodynia significantly in both modes of application. The effect of etanercept was present in animals that were treated from the time of surgery and in those that were treated from day 6, when hyperalgesia was already present. These results suggest the potential of etanercept as a treatment option for patients with neuropathic pain. Key words: tumor necrosis factor-alpha, etanercept, chronic constriction injury, neuropathic pain
Introduction these studies, endogenous TNF is rapidly upregulated in the local microenvironment after nerve injury, and the anti-inflammatory cytokine IL-10 is downregulated (George et al., 1999; 2000). A recombinant tumor necrosis factor receptor (p75)-Fc fusion protein (TNFR:Fc) (Mohler et al., 1993) competitively inhibits TNF, binding to cell surface TNF receptors, and thus acts as a TNF antagonist (Mohler et al., 1993). TNFR:Fc has been successfully used in patients with severe rheumatoid arthritis, where it reduces pain and inflammation (Moreland et al., 1997 ). It is now approved in various countries for the treatment of rheumatoid arthritis under the name of etanercept. In this study we investigated whether TNFR:Fc reduces pain and hyperalgesia in an established mouse model of painful neuropathy, the chronic constriction injury of the sciatic nerve (CCI).
Treatment of neuropathic pain is often unsatisfactory. The drugs shown to be efficacious in clinical trials, such as antidepressants and anticonvulsants (McQuay et al., 1996; Sindrup and Jensen, 1999), are not successful in all patients and induce considerable side effects. New therapeutic strategies for neuropathic pain are therefore warranted. Proinflammatory cytokines have been increasingly recognized as cofactors in the development of pain after nerve lesion (Covey et al., 2000; George et al., 1999; Myers et al., 1999). If applied to the nerve, tumor necrosis factor-alpha (TNF) can induce pain and spontaneous activity (Sorkin and Doom, 2000; Sorkin et al., 1997; Wagner and Myers, 1996). In contrast, inhibition of TNF or interleukin-1 (IL-1) or administration of an anti-inflammatory cytokine can reduce pain related behavior in animal models of neuropathic pain (Sommer et al., 1998a, 1998b; 1999a; Wagner et al., 1998). In accordance with
Materials and Methods
Address correspondence to: Dr. Claudia Sommer, Neurologische Klinik der Universität, Josef-Schneider-Str. 11, 97080 Würzburg, Germany. Tel: 49 931 201 2621; Fax: 49 931 201 2697; E-mail:
[email protected] © 2001 Peripheral Nerve Society, Inc.
All animal experiments were performed according to the regulations of the local Animal Studies Commit67
Blackwell Science Publishers
Sommer et al.
Journal of the Peripheral Nervous System 6:67–72 (2001)
tee and adhered to the guidelines for experiments with awake animals (Zimmermann, 1983). Female C57BL/6mice (16 g – 21 g, n 52) were used (Harlan Winkelmann, Germany). In 47 mice a CCI was performed in deep barbiturate anesthesia by placing 3 ligatures (7/0 prolene) around the sciatic nerve (Sommer et al., 1999b) with a contralateral sham operation. Five animals were not operated and served as untreated controls. These mice were used to control for the consistency of behavioral testing, but not for statistical comparisons. The operated mice were treated with different doses (50 g to 1 mg) of the tumor necrosis factor receptor (p75)-Fc fusion protein (TNFR:Fc) etanercept or with the corresponding amount and volume of human immunoglobulin (IgG, Sigma, Taufkirchen, Germany). For the first 2 experiments, TNFR:Fc was kindly provided by Immunex Corp. (Seattle, USA). For the later experiments, etanercept was purchased from Wyeth (Münster, Germany). Because these compounds are identical, the term ‘etanercept’ is used in the description of all experiments. In 10 animals, etanercept and IgG was given perioperatively in a volume of 35 L by application onto the exposed nerve after surgery and then daily at 6 a.m. under ether anesthesia. A 26 G needle was inserted cranially to the skin incision and inserted in a cranio-caudal direction until a loss of resistance was felt. We have previously shown that drugs can reach a damaged nerve with this method of application and that injection of normal saline in this way does not change animal behavior or nerve morphology (Sommer et al., 1998b). Other animals received etanercept by systemic intraperitoneal (i.p.) injection, either from the day of surgery (n 8) or from day 6 after surgery (n 4). Thermal nociceptive thresholds of the hindpaws were measured using a device (Hargreaves et al., 1988) purchased from Ugo Basile (Comerio, Italy) as described elsewhere (Sommer et al., 1999b). Paw withdrawal latencies (means of 5 tests) on the control side and on the operated side were automatically recorded, the reduction in latency giving a measure of thermal hyperalgesia. Mechanical allodynia was assessed with von Frey hairs using the up-and-down method of Dixon (1965 ) modified for mice (Sommer et al., 1998b). The 50% withdrawal threshold (force of the von Frey hair to which an animal reacts in 50% of the presentations) was recorded. The mice were sacrificed and tissue was harvested one day after the last test. Sciatic nerve segments (3-mm long) distal to the ligatures and corresponding segments from the controls were deep-frozen. A 2-mm segment distal to the previous one was processed for plastic embedding. Immunohistochemistry was performed on frozen sections with a monoclonal antibody to mouse macrophages (Mac-1, Serotec, Eching, Germany) with an ABC-system (Vector, Burlingame, USA) for detection. Toluidine-blue stained
semithin sections were used to control for the extent of the lesion. Statistical significance of differences between treatment groups was assessed with ANOVA followed by LSD test (Statistica software) for thermal behavioral data. Friedman’s test was used for statistical analysis of 50% thresholds to mechanical stimuli with Newman-Keul’s test for post hoc analysis. Significance was assumed below p values of 0.05. Values are given as mean standard deviation.
Results In the first experiment 10 mice received CCI. Five mice were treated with 50 g of etanercept, 5 with 50 g of human IgG by near-nerve epineurial injection daily from the day of surgery. The 2 mice unoperated and untreated were considered controls. Behavioral tests were performed daily from day 3 to 7. Thermal hyperalgesia was present in all mice compared with controls. Withdrawal latencies were higher in etanercept treated mice from day 3 to 7 (mean 5.6 0.3 s) than in IgG treated mice (mean 4.4 0.5 s) indicating a specific attenuation of hyperalgesia in etanercept treated mice (Fig. 1a). Mechanical thresholds were lower than preoperative values in both groups. Etanercept treatment did not change the thresholds significantly (Fig. 1b). Of an additional 12 mice, 5 received CCI and etanercept (87.5 g per day) in the same manner as above, 5 received CCI and 87.5 g per day of IgG, 2 were unoperated and untreated controls. Behavioral tests were performed on day 4 and 7. Again, thermal hyperalgesia was attenuated in etanercept treated mice with a significant difference on day 4 and 7 (mean 6.5 0.3 s vs 3.6 0.2 s in IgG treated mice, Fig. 1a). The reduction in mechanical thresholds (“mechanical allodynia”) was attenuated in etanercept treatment with a significant difference on day 4 and 7 (Fig. 1b). A further increase of the epineurial dose to 100 g of etanercept did not yield better results in 4 mice with CCI (data not shown). A major increase in the dose of etanercept (e.g. by a factor of 10) was precluded using this mode of application due to the volume of the drug. We therefore turned to systemic application. Four mice with CCI received either 100 g or 500 g of etanercept i.p. daily from the day of surgery, and 4 mice received 500 g of IgG. Three mice had CCI only, and one was a control. Thermal withdrawal latencies were higher in mice treated with etanercept in both doses on day 5 (4.7 0.6 s and 5.3 0.7 s) than in mice with CCI only (3.1 0.4 s); the difference to IgG treated mice (3.8 0.4 s) was not significant (Fig. 2a). However, on day 7 withdrawal latencies were significantly higher with both doses (5.1 0.7 s and 6.2 1.0 s) than with CCI only (3.0 0.4 s) and with IgG treatment (3.6 0.3 s, Fig. 2a). Mechanical thresh68
Sommer et al.
Journal of the Peripheral Nervous System 6:67–72 (2001)
Figure 1. (a) Withdrawal latencies to heat of the operated hindpaws of mice treated with etanercept (50 g or 87.5 g) or IgG epineurially daily from the day of surgery. Treatment with etanercept elevates the withdrawal latencies compared with IgG treatment. (b) Withdrawal thresholds to von Frey hair stimulation of the operated hindpaws of mice treated with etanercept (50 g and 87.5 g) and IgG epineurially daily from the day of surgery. Thresholds are elevated on days 4 and 7 only with the higher dose of etanercept. Mean SEM. *p 0.05.
Figure 2. (a) Withdrawal latencies to heat of the operated hindpaws of mice treated with etanercept (100 g or 500 g) or IgG i.p. daily from the day of surgery. Note that both doses elevate latencies compared with untreated mice on day 5 and 7; only the higher dose elevated latencies significantly vs IgG treatment on day 7. (b) Withdrawal thresholds to von Frey hair stimulation of the operated hindpaws of mice treated with etanercept (100 g or 500 g) or IgG i.p. daily from the day of surgery. Thresholds are significantly elevated on day 7 with both doses of etanercept. Mean SEM. *p 0.05.
olds were slightly but significantly higher on day 7 with both doses of etanercept than with IgG-treated mice or with CCI-only mice (Fig. 2b). Having established an effect of etanercept when given daily from the day of surgery, we proceeded to investigate whether an effect could be obtained in already hyperalgesic animals. Eight mice received CCI and were tested on day 6, when all had developed thermal hyperalgesia (from a baseline of 8.3 s to 5.5 1.8 s and 5.2 1.1 s) and mechanical allodynia (from a baseline of 0.4 0.5 g to 0.01 0.01 g). Eight mice were treated with either 1 mg of etanercept i.p. or with the corresponding dose of IgG on days 6, 7, 8, 9, 12, and 14. Lowered thermal withdrawal latencies remained unchanged in the IgG treated mice but were elevated significantly from day 10 in the etanercept treated mice (latencies of 6.7 s to 7.0 s). Lowered mechanical thresholds also remained unchanged in the IgG-treated mice and had a trend to-
ward significance starting on day 10 in the etanercepttreated mice, with a significant difference from IgG on day 15 (0.2 0.12 g vs 0.03 0.02 g, Fig. 3b). Plastic embedded, toluidine-blue stained sections of the injured sciatic nerve and frozen sections immunoreacted for macrophages from all animals were studied by a blinded investigator. Morphologically, no difference could be seen between etanercept-treated and sham-treated mice (Figs. 4a, b). The amount of endoneurial and epineurial macrophages was not changed by treatment (data not shown).
Discussion The main finding of this study is that the recombinant tumor necrosis factor receptor (p75)-Fc fusion pro69
Sommer et al.
Journal of the Peripheral Nervous System 6:67–72 (2001)
Figure 3. (a) Withdrawal latencies to heat of the operated hindpaws of mice treated with etanercept (1 mg) or IgG i.p. daily from day 6 after surgery. Latencies are elevated starting on day 10 in etanercept treated mice compared to IgG treatment. (b) Withdrawal thresholds to von Frey hair stimulation of the operated hindpaws of mice treated with etanercept (1 mg) or IgG i.p. daily from day 6 after surgery. Thresholds are significantly elevated on day 15 in etanercept-treated mice compared with IgG-treated mice. Mean SEM. *p 0.05.
Figure 4. Photomicrographs of toluidine-blue stained 1-m plastic sections of nerves with CCI treated with etanercept (a) or IgG (b) by local application, day 8 after surgery. Note that the typical pathology of CCI is present independent of treatment. Bar 20 m.
tein etanercept can attenuate pain-related behavior in mice with an experimental neuropathy both after prophylactic and after therapeutic application. The treatment effect was seen with local and with systemic application. We have previously shown that blocking the production, the processing or the biological activity of TNF or IL-1 can reduce pain related behavior in mice and rats (Sommer et al., 1997; 1998a; 1998b). Whereas inhibition of TNF production by thalidomide only had an effect on hyperalgesia when administered from the day of surgery (Sommer et al., 1998a), neutralizing antibodies to TNF were also efficient when given at a later time point (Sommer et al., 1999a). TNF is upregulated in the injured peripheral nerve with a peak at 12 hours after CCI (George et al., 1999), which may be a reason
TNF release needs to be inhibited very early in this model. However, TNF-receptors, particularly TNF-R2, are upregulated with a slower time-course after nerve injury (George and Sommer, 1999), which might argue for positive effects of an anti-TNF treatment even at a later stage, as seen here with etanercept. The question remains whether an anti-TNF treatment can also reduce pain-related behavior in a more chronic neuropathy model and studies with such a model, for example in diabetic or in vincristine neuropathy (Aley et al., 1996), are warranted. Morphological investigation of the sciatic nerve did not reveal a difference in nerve de- and regeneration and 70
Sommer et al.
Journal of the Peripheral Nervous System 6:67–72 (2001)
in macrophage influx after etanercept treatment. This indicates that etanercept does not induce inflammatory changes when applied locally, and, at least in the time frame studied here, does not impair nerve regeneration. Based on the data from this study, it is not possible to determine whether the effect of etanercept on neuropathic pain was due to a local or to a systemic site of action. The effect on mechanical allodynia seems much more pronounced after the local application of 87.5 g of etanercept than after the systemic application of 100 g (compare Figs. 1b and 2b). It is thus very likely that a high local concentration of the TNF inhibitor is important for the treatment effect, consistent with a high local increase in TNF levels (George et al., 1999). However, additional systemic effects cannot be excluded. TNF mRNA is induced in the DRG after axotomy (Murphy et al., 1995). Lumbar spinal TNF-like immunoreactivity increases after CCI (DeLeo et al., 1997) and mice overexpressing TNF in astrocytes have increased mechanical allodynia after L5 spinal nerve transection (DeLeo et al., 2000). After CCI, TNF-bioactivity is increased in certain areas of the central nervous system, where it is supposed to antagonize the endogenous pain suppressing system (Covey et al., 2000; Ignatowski et al., 1999). Thus, a TNF inhibitor might act at various levels of the peripheral and central nervous system. Etanercept has not been used in neuropathic conditions in humans, but is known to reduce pain and inflammation rapidly in rheumatoid arthritis (Moreland et al., 1997). We postulate that a subgroup of patients with peripheral neuropathic pain associated with increased TNF production might benefit from TNF inhibition. The cytokine production in Schwann cells differs greatly between individuals (Rutkowski et al., 1999), such that it is likely that neuropathies with and without increased TNF production exist. In conclusion, inhibition of TNF by etanercept is efficient in reducing thermal hyperalgesia and mechanical allodynia in a prophylactic as well as in a therapeutic application. Clinical treatment trials are warranted to analyze the potential of etanercept for the treatment of pain in human peripheral neuropathies.
Covey WC, Ignatowski TA, Knight PR, Spengler RN (2000). Brain-derived TNFalpha: involvement in neuroplastic changes implicated in the conscious perception of persistent pain. Brain Res 859:113–122. DeLeo JA, Colburn RW, Rickman AJ (1997). Cytokine and growth factor immunohistochemical spinal profiles in two animal models of mononeuropathy. Brain Res 759:50–57. DeLeo JA, Rutkowski MD, Stalder AK, Campbell IL (2000). Transgenic expression of TNF by astrocytes increases mechanical allodynia in a mouse neuropathy model. Neuroreport 11:599–602. Dixon W (1965). The up-and-down method for small samples. J Am Statist Assoc 60:967–978. George A, Marziniak M, Schafers M, Toyka KV, Sommer C (2000). Thalidomide treatment in chronic constrictive neuropathy decreases endoneurial tumor necrosis factor-alpha, increases interleukin-10 and has long-term effects on spinal cord dorsal horn met-enkephalin. Pain 88:267–275. George A, Schmidt C, Weishaupt A, Toyka KV, Sommer C (1999). Serial determination of tumor necrosis factor-alpha content in rat sciatic nerve after chronic constriction injury. Exp Neurol 160:124–132. George A, Sommer C (1999). Upregulation of the tumor necrosis factor receptor 1 and 2 in mouse sciatic nerve after chronic constrictive injury. J Neurol 246 (Suppl. 1):115. Hargreaves K, Dubner R, Brown F, Flores C, Joris J (1988). A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia. Pain 32:77–88. Ignatowski TA, Covey WC, Knight PR, Severin CM, Nickola TJ, Spengler RN (1999). Brain-derived TNFalpha mediates neuropathic pain. Brain Res 841:70–77. McQuay HJ, Tramer M, Nye BA, Carroll D, Wiffen PJ, Moore RA (1996). A systematic review of antidepressants in neuropathic pain. Pain 68:217–227. Mohler KM, Torrance DS, Smith CA, Goodwin RG, Stremler KE, Fung VP, Madani H, Widmer MB (1993). Soluble tumor necrosis factor (TNF) receptors are effective therapeutic agents in lethal endotoxemia and function simultaneously as both TNF carriers and TNF antagonists. J Immunol 151:1548–1561. Moreland LW, Baumgartner SW, Schiff MH, Tindall EA, Fleischmann RM, Weaver AL, Ettlinger RE, et al. (1997). Treatment of rheumatoid arthritis with a recombinant human tumor necrosis factor receptor (p75)-Fc fusion protein. N Engl J Med 337:141–147. Murphy PG, Grondin J, Altares M, Richardson PM (1995). Induction of interleukin-6 in axotomized sensory neurons. J Neurosci 15:5130–5138. Myers RR, Wagner R, Sorkin LS (1999). Hyperalgesia actions of cytokines on the peripheral nervous system. In: Cytokines and Pain: Progress in Inflammation Research. L Watkins (Ed). Birkhauser Verlag, Basal, pp 133–158. Rutkowski JL, Tuite GF, Lincoln PM, Boyer PJ, Tennekoon GI, Kunkel SL (1999). Signals for proinflammatory cytokine secretion by human Schwann cells. J Neuroimmunol 101:47–60. Sindrup SH, Jensen TS (1999). Efficacy of pharmacological treatments of neuropathic pain: an update and effect related to mechanism of drug action. Pain 83:389–400. Sommer C, Schmidt C, George A, Toyka KV (1997). A metalloprotease-inhibitor reduces pain associated behavior in mice with experimental neuropathy. Neurosci Lett 237:45–48. Sommer C, Marziniak M, Myers RR (1998a). The effect of thalidomide treatment on vascular pathology and hyperalgesia
Acknowledgements We would like to thank B. Dekant and L. Biko for technical assistance. For part of the experiments, TNFR: Fc was kindly provided by Immunex Corp. Supported by Deutsche Forschungsgemeinschaft SFB 353.
References Aley KO, Reichling DB, Levine JD (1996). Vincristine hyperalgesia in the rat: a model of painful vincristine neuropathy in humans. Neuroscience 73:259–265.
71
Sommer et al.
Journal of the Peripheral Nervous System 6:67–72 (2001)
caused by chronic constriction injury of rat nerve. Pain 74: 83–91. Sommer C, Schmidt C, George A (1998b). Hyperalgesia in experimental neuropathy is dependent on the TNF receptor 1. Exp Neurol 151:138–142. Sommer C, Lindenlaub T, Teuteberg P, Hartung T, Toyka K (1999a). Anti-TNF neutralizing antibodies reduce pain-related behavior in two models of painful mononeuropathy. J Peripher Nerv Syst 4:151. Sommer C, Petrausch S, Lindenlaub T, Toyka KV (1999b). Neutralizing antibodies to interleukin 1-receptor reduce pain associated behavior in mice with experimental neuropathy. Neurosci Lett 270:25–28. Sorkin LS, Doom CM (2000). Epineurial application of TNF elic-
its an acute mechanical hyperalgesia in the awake rat. J Peripher Nerv Syst 5:96–100. Sorkin LS, Xiao WH, Wagner R, Myers RR (1997). Tumour necrosis factor-alpha induces ectopic activity in nociceptive primary afferent fibres. Neuroscience 81:255–262. Wagner R, Janjigian M, Myers RR (1998). Anti-inflammatory interleukin-10 therapy in CCI neuropathy decreases thermal hyperalgesia, macrophage recruitment, and endoneurial TNF-alpha expression. Pain 74:35–42. Wagner R, Myers RR (1996). Endoneurial injection of TNF-alpha produces neuropathic pain behaviors. Neuroreport 7:2897– 2901. Zimmermann M (1983). Ethical guidelines for investigations of experimental pain in conscious animals. Pain 16:109–110.
72
