Animal Models of Experimental Neuropathic Pain

SUMMARY

Neuropathic pain has an adverse impact on quality of life and remains an important unmet medical need. Consequently, many attempts have been made to “model” neuropathic pain in animals, each with its strengths and shortcomings. The observation that animals would attack a denervated limb (“autotomy”) following axotomy of a nerve led to the suggestion that such behavior reflected the presence of pain (i.e., a possible model of “anesthesia dolorosa”). Since other behavioral assessments were not possible following complete axotomy, later models focused on incomplete nerve injuries, including partial sciatic section, constriction of the sciatic nerve, and ligation of the L5 and L6 spinal nerves or distal branches of the sciatic nerve (i.e., the tibial and common peroneal nerves). Each of these injuries results in somewhat varying behavioral manifestations and may allow assessment of potentially different mechanisms that could contribute to a pathological pain state. Chronic constriction injury, for example, elicits robust thermal hyperalgesia and allodynia to cold stimuli. Ligation of the L5 and L6 spinal nerves (spinal nerve ligation model) produces enhanced responses to light dynamic or static touch, as well as to thermal stimuli. Recently, attempts have been made to model painful diabetic neuropathy and neuropathies resulting from cancer chemotherapeutic agents (e.g., vincristine, paclitaxel). The underlying mechanisms through which chemically induced neuropathic pain might occur remain uncertain and might be considerably different from those that drive traumatic injury–induced neuropathic pain. Although these preclinical approaches have been critical to the evolution of thinking on mechanisms of pain, the difficulty in evaluating spontaneous, or non-evoked, “pain” in animals has been an impediment to the field. Nevertheless, progress has been made and is continuing, a number of approaches to study spontaneous pain have been proposed, and some have begun to be validated. Some of these approaches appear to offer significant potential for the exploration of mechanisms that may help in ultimately increasing translational to novel therapy for human neuropathic states.

Introduction

Chronic pain is a significant health problem that has a negative impact on the quality of life of afflicted individuals, as well as on society in economic terms (more than $100 billion annually) ( ). Of the chronic pain states, neuropathic pain is perhaps the most perplexing. Most patients obtain inadequate relief with currently available pain medications. Neuropathic pain may arise from disease states, nerve trauma, exposure to chemotherapeutic agents, and other causes. The development of animal “models” that reflect such a diverse etiology has been difficult. Whether animal models of neuropathic pain are relevant to specific human conditions is not known. Patients with neuropathic pain often complain of multiple symptoms, including spontaneous “burning” pain, as well as pain from normally innocuous stimuli such as touch and cold (allodynia) and hyperalgesia to normally painful stimuli. Not all patients with neuropathic pain have these symptoms, and the symptoms that do occur may vary with time, thus presenting additional challenges in identifying relevant mechanisms and, by extension, appropriate treatments. It is important to emphasize that although terms such as allodynia and hyperalgesia have clear meaning in humans, application of these terms preclinically relies on assumptions of an animal’s perception in response to innocuous or noxious stimuli. Accordingly, we anticipate that a withdrawal reaction to a stimulus that does not normally evoke a response suggests “allodynia” and that an exaggerated response to a normally nociceptive stimulus suggests hyperalgesia.

Despite recent progress, our mechanistic understanding of neuropathic pain remains limited. The preliminary knowledge that we do have is related to the development of a number of animal models that mimic, at least in a limited way, some aspects of the pain that is reported clinically and that allow the generation of hypotheses that can lead to the testing of new potential therapies. This chapter summarizes information that has emerged from some of the models that have influenced our understanding of the mechanisms of neuropathic states.

Animal Models of Neuropathic Pain

Animal models of nociceptive pain depend on nocifensive responses to an acute noxious stimulus and have been very important in elucidating the basic mechanisms and neuroanatomical substrates that underlie nociception and its modulation. These models have allowed understanding of the roles of neurotransmitters and neuropeptides in inhibition and facilitation of pain and have become standard tools in drug development, especially predicting the analgesic efficacy and potency of strong (e.g., opioid) analgesics against acute pain states. A shortcoming that became evident, however, was the fact that models of acute nociception are insufficient to account for abnormal pain states, such as the pain that can arise as a consequence of injury to nerves within the central nervous system or in the periphery. Treatments that have been consistently reliable for experimental and clinical acute pain states have been disappointing for chronic neuropathic pain states, thus highlighting the need for models that approximate the features and perhaps give insight into the mechanisms of human neuropathic states. A number of important models of neuropathic pain have been developed and characterized, all with advantages and disadvantages and all with limited ability to accurately reflect the clinical condition. Nevertheless, these models have provided a basis for the mechanistic understanding that we have achieved to date of neuropathic pain syndromes and are an important step in the discovery process for new pain medications.

Neuroma Model

One of the earliest explicit models of neuropathic pain, nerve axotomy, was developed from the observation that transection of the sciatic nerve results in self-mutilation, or autotomy, of the denervated paw ( ). Autotomy was originally thought to occur because the animal perceived the denervated limb as a foreign object and not as a part of its own body ( ). However, Wall and colleagues surmised that the neuroma formed at the site of nerve section may be the source of a noradrenergic-sensitive afferent discharge that might be perceived as pain referred to the anesthetized, denervated limb ( ). Based on these suppositions, the transection or neuroma model was developed to study anesthesia dolorosa and phantom limb pain. Nerve transection in mice and rats was performed by removing approximately 5 mm of the sciatic and saphenous nerves at the level of the mid-thigh ( ). Denervation, indicated by complete anesthesia, was confirmed by the complete absence of a nocifensive response to a strong pinch. The degree of autotomy was quantified by applying a scoring method reflecting the severity of damage to the paw ( ). Animals showed time-dependent damage to the paw during the first 5 weeks with no increased damage afterward ( ). Normal weight gain and behavioral activity were otherwise observed ( ). Interestingly, injury to the saphenous nerve (a purely sensory nerve), a crush injury of the sciatic nerve, or a combination of both manipulations produced little or no autotomy over a 9-week observation period ( ). Although crush produces wallerian degeneration, nerve regeneration and reinnervation of the hindpaw can occur to at least some degree and in this manner possibly attenuate the expression of autotomy ( ). In contrast, sciatic nerve ligation and transection produced time-dependent autotomy that was observed even when saphenous nerve section was performed more than 100 days after sciatic nerve section ( ). Although autotomy can also occur after dorsal root avulsion, the driving mechanism may differ since the spinal horn neurons lose afferent input, thereby possibly resulting in hyperactivity of these transmission pathways so critical to pain processing, and it may thus represent deafferentation pain ( ). After axotomy distal to the dorsal root ganglion (DRG), however, the proximal portions of the sectioned nerves continue to generate input to the spinal cord ( ).

Although this model is one of the first reported attempts to model clinical neuropathic pain states, specifically, anesthesia dolorosa and phantom limb pain, autotomy behavior has not been widely adopted. One reason may be the uncertainty whether autotomy behavior reflects pain (but see ). It has been suggested that autotomy is related to persistent, spontaneous ectopic discharges from the neuromas that result in ongoing pain ( ). Other suggestions, however, are that autotomy may be an attempt to remove an insensate appendage or that it occurs as a result of excessive grooming because of the absence of sensory feedback ( ). Objections have also been raised that injury-induced autotomy as a model of neuropathic pain should be discarded for several reasons, including the fact that it is “esthetically repugnant” and that other models should be considered ( ).

Partial Sciatic Nerve Section

A variant of the neuroma model was designed to avoid the complications inherent with complete denervation of the hindpaw and to more closely resemble trauma-induced nerve injury by tightly ligating approximately one-third to one-half of the sciatic nerve fascicle and sparing the remaining fibers from injury ( ). The rats displayed licking and guarding behavior of the hindpaw, suggestive of spontaneous pain, but no autotomy was observed ( ). Tactile allodynia, suggested by exaggerated withdrawal responses to light touch, was observed bilaterally ( ). Mechanical and thermal hyperalgesia was also observed ( ). Chemical sympathectomy with guanethidine abolished these behavioral responses, thus suggesting that these signs of neuropathic pain are probably sympathetically maintained ( ). The combination of an almost immediate onset and long duration of allodynia and hyperalgesia, the presence of mirror-image pain, and the dependence on sympathetic activity led to suggestions that this preparation may serve as a model for causalgia elicited by partial nerve injury ( ). Neonatal rats were treated with capsaicin to selectively obliterate capsaicin-sensitive C-fiber nociceptors in an attempt to differentiate the fiber types mediating these behavioral responses ( ). Tactile hyperesthesia and touch-evoked tactile allodynia developed in the capsaicin-treated rats after partial sciatic nerve ligation, but not thermal hyperalgesia. These results indicated that tactile hyperesthesia is probably mediated through large-diameter myelinated fibers whereas thermal hyperalgesia is mediated through thermal nociceptive C fibers as noted previously ( ). One shortcoming of the model is that it is difficult to accurately reproduce the same injury repeatedly. Consequently, the degree of neuropathic pain behavior might be affected by the degree of the injury. Moreover, there is considerable mixing of injured with uninjured fibers, and the uninjured fibers may express enhanced discharges through ephaptic transmission.

Chronic Constriction Injury

The chronic constriction injury (CCI) model was developed to inflict reproducible nerve injury without complete denervation ( ). A set of four chromic gut sutures were placed loosely around the common sciatic nerve at intervals of 1–2 mm so that they did not completely impede circulation through the epineurium ( ). During the first week after surgery, gait and posture were highly variable among the animals and stabilized for up to 2 months thereafter. The rats walked with a definitive limp; exhibited ventriflexion of the toes, which were held tightly together; and avoided placing weight on the affected hindpaw ( ). Guarding behavior, consisting of raising the limb and keeping it close to the flank, was observed. A striking feature of rats with CCI is thickening and elongation of the claws, to the point that in extreme cases they would curve back to the toe pads because of avoidance of grooming the injured hindlimb. Rats may show mutilation of the claw tips and roots but, unlike autotomy, no mutilation of the digits ( ). Thermal hyperalgesia may be evident within 2 days and last 2 months after CCI. These changes in thresholds appear much sooner than autotomy behavior following complete nerve transection. Compound electromyographic activity of the biceps femoris and semitendinosus muscles in response to a heated probe applied to the hindlimb was markedly greater and showed long-lasting afterdischarge activity in rats with CCI when compared with sham-operated animals ( ). Hypersensitivity to cold (4°C) and to light tactile stimuli, but not to noxious mechanical stimuli, was observed after CCI ( ). Gentle stroking of the affected hindlimb resulted in a marked increase in Fos expression in the superficial laminae of the spinal dorsal horn ( ). These behavioral observations may indicate that the CCI model might suggest spontaneous pain, as revealed by abnormal postures, guarding, and coincident grooming behavior, actions mimicking the signs of complex regional pain syndrome (CRPS). For example, some CRPS patients do not trim their nails because of the pain experienced ( ).

The CCI model may elicit both spontaneous and evoked neuropathic pain because some afferent fibers of the sciatic nerve survive after surgery. Within 1 day of CCI, intraneural edema caused by partial constriction of the vasculature of the epineurium results in translucent, demyelinated constrictions 25–75% of the original diameter of the sciatic trunk; the constrictions merge into a single area of uniform thinning with swelling proximal to the constricted area ( ). This is reminiscent of entrapment neuropathies (e.g., carpal tunnel syndrome) ( ). The proximal sciatic nerve appeared normal to within 0.5 cm of the constriction, where marked degeneration and endoneurial edema became apparent along with massive demyelination and nearly total loss of large-diameter Aα and Aβ fibers and a lesser loss of small-diameter myelinated Aδ fibers ( ). Within 3 days of CCI, 89% of Aβ, 87% of Aδ, and 32% of C fibers were lost, with progression to loss of 94% of myelinated fibers and 73% of unmyelinated fibers within 14 days ( ). Subsequently, there was regeneration of unmyelinated and small-diameter myelinated but not large-diameter myelinated fibers to nearly normal levels ( ). The hyperalgesia observed after CCI is probably not due to the loss of large-diameter mechanosensitive fibers since it was resolved 56 days after CCI whereas the large-diameter fibers remained absent ( ). It is possible that hyperalgesia may be mediated through sensitized Aδ and C fibers, some of which normally act as low-threshold mechanoreceptors ( ). Electrophysiological studies have shown that primary afferents, including large-diameter myelinated fibers, spontaneously discharge at ectopic foci proximal to the injury and that these abnormal discharges might play a role in spontaneous and evoked manifestations of neuropathic pain ( ).

An extension of the CCI model is CCI of the infraorbital nerve, which has been used as a model of trigeminal neuropathic pain ( ). The edge of the orbit of anesthetized rats was exposed and the contents of the orbit gently pushed aside with a cotton-tipped swab to expose the infraorbital nerve just caudal to the infraorbital foramen ( ). A pair of chromic gut sutures were placed around the nerve without occluding circulation in the nerve sheath. Rats with constriction injury of the infraorbital nerve demonstrated a significant increase in facial grooming along with decreased exploratory behavior ( ). They also spent increased periods immobile in a “frozen” posture ( ). The rats expressed touch-evoked agitation and vigorous nocifensive responses to probing the periorbital area with von Frey filaments in the later (15–130 days) postoperative period but were hypoesthetic during the immediate postoperative period ( ). Moreover, the constriction injury resulted in increased expression of Fos in the medullary dorsal horn corresponding to the intensity of the touch stimulus ( ). Tactile allodynia was blocked by injections of triptans ( ), endothelin ET _B receptor antagonists ( ), and the antiepileptic lacosamide ( ). This model may lead to better understanding of trigeminal neuropathic pain states and to enhanced therapeutic regimens for this condition.

Intubation of the Sciatic Nerve

Intubation of the sciatic nerve with polyethylene cuffs was developed as an alternative to CCI performed with sutures in an attempt to standardize the degree of constriction of the sciatic nerve ( ). Polyethylene cuffs with internal diameters of approximately 0.75 mm produced behavioral signs of neuropathic pain and morphological changes indicative of peripheral nerve injury. A capsule of connective tissue formed over the cuff along with swelling of the nerve within 3 days ( ). The translucence and constriction of the nerves apparent after CCI did not develop after cuffing. Wallerian degeneration with large-scale depletion of large-diameter myelinated fibers was evident within 7 days and remained apparent throughout the 6-week observation period. In contrast, unmyelinated fibers and thinly myelinated, small-diameter fibers were initially diminished 7 days after cuff placement but appeared to rebound such that a net increase in the number of small-diameter axons was seen by 14 days after cuffing, possibly as a result of collateral sprouting ( ). From 28–42 days after cuffing, the number of small-diameter unmyelinated axons was not significantly different from that in controls, the number of myelinated small-diameter fibers was elevated, and the number of large-diameter myelinated fibers remained depressed ( ). The changes in morphology after placement of the polyethylene cuffs were found to be more consistent than after CCI ( ).

Cuffing of the sciatic nerve produced abnormal gait and defensive postures, including guarding of the injured hindlimb, uneven weight-bearing distribution, and hyperextension of the limb with ventriflexion of the toes ( ). Spontaneous elevation of the hindlimb was observed along with occasional biting of the claws on the affected limb. The rats with polyethylene cuffs demonstrated enhanced responses to mechanical stimuli, including vocalization and prolonged struggling after light paw compression; such behavior was not evident in control animals. In addition, light touch produced rapid and repeated flinching-type behavior ( ). Moreover, the afflicted animals also demonstrated definitive allodynia to cold stimuli. Interestingly, it was suggested that the enhanced nociceptive responses did not correlate with the degree of changes in nerve fiber morphology. The maximal behavioral signs of neuropathic pain were observed 2 weeks after the injury but had largely resolved by the fourth week after injury ( ).

Spinal Nerve Ligation

The spinal nerve ligation (SNL) model of traumatic nerve injury has become one of the most commonly used and studied models ( ). The primary impetus for development of the SNL model derived from the belief that it was not possible to adequately control the numbers and types of primary afferent fibers that were injured with the previous models ( ). The L5 and L6 spinal nerve branches of the sciatic nerve are carefully identified, isolated, and tightly ligated with 3–0 silk suture between the trifurcation of the sciatic nerve and distal to the DRGs. Removal of the L6 spinal transverse process is required for unobstructed contact with this spinal nerve ( ). The hindpaws of rats with L5/L6 SNL are slightly everted and the toes held together, and the rats shift weight away from the injured hindpaw and limp with an uncoordinated gait ( ). After SNL, rats would frequently suddenly withdraw the injured hindpaw and lick or gently bite the claws of the hindlimb ( ). Most importantly, no signs of autotomy were observed in these rats ( ). Critically, ligation of only the L4 spinal nerve caused extreme motor deficits and dragging of the hindlimb because of denervation of the proximal muscles of the leg ( ). The many subsequent experiments performed with the SNL model indicate that limited motor deficits are seen with SNL.

Robust enhanced responses to gentle tactile and noxious thermal stimuli were present within 1–2 days after SNL and persisted throughout the observation period; they were interpreted as evidence of tactile allodynia and thermal hyperalgesia lasting longer than 10 weeks ( ). Other studies using the paw withdrawal threshold, which involved the application of a series of von Frey filaments of increasing and decreasing strength, demonstrated that tactile allodynia developed within 1–2 days after SNL and remained throughout the observation period of more than 50 days ( ). Hyperalgesia in response to noxious radiant heat had a similar time course as tactile allodynia did ( ). The cold allodynia after SNL is much less robust than that seen after CCI ( ). Application of acetone drops to the hindpaw elicits brisk withdrawal responses or increased dorsal horn unit activity ( ). However, interpretation of such cold allodynia may be adulterated by the temperature of the hindpaw, the acetone, the rate of evaporation, and the tactile stimulation produced by the drop itself ( ). Behavioral responses to a 5°C cold plate were significantly less pronounced in SNL rats than in CCI rats ( ). Electrophysiological studies of convergent dorsal horn units showed an increase in the number of responsive neurons, an increase in response frequencies, and an increase in the slope of the stimulus–response function of the dorsal horn units in response to light touch, brush, or heat ( ). A presumptive cooling sensation after a drop of acetone onto the hindpaw also produced enhanced responses of spinal dorsal horn units after SNL ( ). Interestingly, the response thresholds or response frequencies of spinal dorsal horn neurons to electrical stimulation of Aβ or C fibers applied to their receptive fields did not differ significantly among naïve, sham-operated, or SNL rats ( ).

Whether expression of the behavioral signs of neuropathic pain requires input from the injured nerves has been controversial. Interruption of neuronal communication between the DRG of the ligated nerves and the spinal cord through either dorsal rhizotomy or the local application of bupivacaine reversed the hyperesthetic responses to thermal, tactile, and cool stimuli ( ), thus suggesting the importance of the injured nerves. Rhizotomy of the L5/L6 spinal nerve roots did not produce any behavioral signs of neuropathic pain, and dorsal rhizotomy of the L3 and L4 spinal nerve roots, in addition to L5/L6 SNL, produced complete denervation of the hindlimb along with loss of motor activity and autotomy, in effect mimicking sciatic nerve section ( ). Bupivacaine applied onto the L3 and L4 spinal nerve roots of rats with L5/L6 SNL blocked tactile but not thermal or cold hyperesthetic responses ( ). It was proposed that evoked pain requires input from both injured and uninjured afferent fibers whereas spontaneous pain may be mediated through injured afferent input ( ). However, other studies have questioned this conclusion. For example, reported that section of the L5 spinal nerve followed by L5 dorsal rhizotomy did not block enhanced evoked responses. In contrast, L5 spinal nerve section followed by L4 dorsal rhizotomy blocked evoked input, which suggests that the hypersensitivity was mediated via uninjured fibers ( ), although this conclusion may have been confounded by significant denervation of the hindpaw under these conditions. Recent studies evaluating non-evoked pain (see later) have supported a role for injured fibers ( ), consistent with the clinical observations ( ). However, this study could not exclude an additional role for uninjured fibers. Notably, a recent microneurographic study found that individuals with small-fiber neuropathic pain, as well as animals with five different types of experimental nerve pathologies, demonstrated markedly increased spontaneous activity of intact C-fiber nociceptors, thus providing evidence that the spontaneous activity of nociceptors may additionally contribute to spontaneous pain ( ).

A variation of the L5/L6 SNL model is to ligate or cut the L5 spinal nerve only. This procedure produced a similar behavioral profile, but of lesser magnitude, perhaps because of the involvement of fewer nerve fibers ( ). Another adaptation that has been reported, though not widely used, called for ligation of the sacral rather than the lumbar spinal nerves to induce tactile and thermal hyperesthesia of the rat tail ( ). An interesting feature of ligation of the sacral afferents is the production of bilateral tactile allodynia and thermal hyperalgesia of the hindpaws ( ). This model allows injury at a given spinal segment but hyperalgesia revealed through evoked input at a different level. Moreover, clear, robust signs of tactile allodynia and thermal hyperalgesia are produced within a few days and last for months, which allows extensive examination of the progression of changes after nerve injury. A unique advantage of ligation or section of the spinal nerve is that this process allows independent examination of damaged and spared nerves and their associated DRG neurons. This permits examination of the changes in expression of multiple potential mediators and neuromarkers in injured and adjacent, uninjured primary afferent nerves.

Spared Nerve Injury

The spared nerve injury (SNI) model is created by tight ligation and subsequent resection of 2–4 mm of the common peroneal and tibial nerves while leaving the sural nerve intact ( ). These nerves represent the three distal branches of the sciatic nerve. Behavioral signs of neuropathic pain are evident within 1 day after SNI and are maintained for more than 9 weeks ( ). Spontaneous pain is suggested by avoidance of weight bearing on the injured hindpaw and eversion of the paw and rapid hindpaw flexion on contact ( ). Animals with SNI demonstrated enhanced responses to light and noxious tactile stimuli and to acetone drops, thus suggesting tactile, thermal, and cold hyperesthesia ( ). Notably, thermal hyperalgesia was not demonstrated by decreased latency to noxious heat; rather, the duration of the withdrawal response was increased. Evidence of neuropathic pain behavior occurred in response to stimuli applied to regions innervated by the sural nerve and the uninjured saphenous territories. The level of responses appeared to be greater when the stimuli were applied to the receptive field of the sural nerve than when applied to the saphenous nerve ( ).

With this model, the territories of injured and uninjured nerves can be examined independently ( ). Furthermore, the effects of nerve injury on an uninjured nerve (sural nerve) that shares a common nerve trunk with the injured nerve may be compared with an adjacent uninjured nerve that is anatomically isolated (saphenous nerve) ( ). Therefore, it is possible to study changes that occur in DRG neurons whose axons have minimal commingling with injured fibers but whose terminals overlap the territory of injured nerves ( ).

Distal Nerve Injury

CRPS type 1 was thought to occur in the absence of verifiable nerve injury ( ). This condition may be precipitated by a “noxious event” that might include fractures, joint sprains, strains, thoracic surgery, soft tissue injury, and cardiac ischemia and can be of short duration or continue long after the original injury has healed ( ). Idiopathic CRPS-1 occurs after noxious events that are so trivial that they may not even be recalled by the patient ( ; ; ; ). Such injuries would include venipuncture, lacerations, and other types of minor trauma ( ). Because of difficulty determining the presence of verified nerve injury, inclusion of CPRS-1 as a neuropathic pain state has been challenged ( ). However, evidence is emerging that CRPS-1 is associated with nerve injury but that technologies have been unable to detect such injury until recently ( ).

Minimally invasive techniques to visualize intracutaneous axons in skin biopsy specimens now exist ( ). Rice and colleagues ( ) found reduced innervation of the epidermis, sweat glands, and vasculature by small fibers in skin sections obtained from the amputated legs of patients with CRPS-1. Oaklander and co-workers found a 29% reduction in intraepidermal neurites in biopsy samples taken from an affected site when compared with an unaffected site ( ). Since venipuncture is a cause of CRPS-1 in susceptible individuals, needlestick of the left tibial nerve of rats (i.e., distal nerve injury [DNI]) was used as a model of this condition ( ). The left tibial nerve of anesthetized Sprague-Dawley rats was exposed and a flat wooden platform inserted under the nerve. The nerve was pierced through with either a 30-, 22-, or 18-gauge needle and the wound closed ( ). Paw withdrawal responses to probing with von Frey filaments or a pinprick along with hindpaw position, color, and edema were measured at various times after the injury. Seven days after DNI, 67, 88, and 89% of rats with tibial nerves pierced by the 18-, 22-, and 30-gauge needles, respectively, showed a 50% or greater reduction in the paw withdrawal threshold ( ). Importantly, the development of mechanical hypersensitivity did not correlate with the size of the lesion. Moreover, 57% of the rats also showed sensitivity of the contralateral hindpaw, a model of the “mirror pain” seen in CRPS patients ( ). The prevalence of hypersensitivity to pinprick or cold was very low, but 14% of the rats showed abnormal posture as indicated by elevation of the lateral hindpaw along with paw eversion or plantar flexion of all digits ( ). Evidence of wallerian degeneration was seen in tibial nerve stumps taken 14 days after DNI ( ).

Ischemic Peripheral Nerve Injury

Injury to the sciatic nerve of mice or rats has been produced by photochemically induced ischemia. The photosensitive dye erythrosin B is injected intravenously, and the exposed sciatic nerve is irradiated with an argon laser with emission at a wavelength of 514 nm ( ). It was found that exposure lasting 30 seconds would selectively injure myelinated fibers and an exposure duration of 2 minutes would cause injury to both myelinated and unmyelinated fibers ( ). Within 1 day of irradiation, the blood vessels of the epineurium and within the fascicles were occluded. Axons demonstrated signs of initiation of degeneration at the site of irradiation ( ). Inflammatory and fibrotic tissue, wallerian degeneration, edema, and demyelination were evident within 7 days. Although the nerves remained thinner than normal, there was evidence of reinnervation after 3 months. The unmyelinated axons appeared to have normal morphology, whereas the myelinated axons were smaller with a thin myelin sheath ( ). Hyperesthesia to light touch and to cold was maximal 7 days after injury and resolved within 3 months. Interestingly, tactile and cold hyperesthesia developed only when both myelinated and unmyelinated fibers were injured. Damage to myelinated axons only was insufficient to produce signs of neuropathic pain ( ). Behavioral signs of spontaneous pain were not evident in this model.

Postischemic Pain

In addition to injury, CRPS-1 may depend partly on localized tissue ischemia ( ). Studies have demonstrated diminished blood oxygenation in skin capillaries ( ), along with biochemical evidence of anaerobic metabolism ( ). Tissue obtained from the amputated limbs of patients with CRPS-1 showed signs of oxidative stress and ischemic conditions ( ). These observations led to the development of an animal model of CRPS-1 based on ischemia and reperfusion ( ). Male Long-Evans rats were maintained under pentobarbital anesthesia for a period of 3 hours. A nitrile O-ring, 5.5 mm in diameter, was placed on the left hindlimb proximal to the ankle joint and remained there for 3 hours. This produced the equivalent of a tourniquet inflated to a pressure of 350 mm Hg ( ). The ring was removed, anesthesia was terminated, and the rats recovered during reperfusion ( ). Sham rats were prepared in an identical fashion but with a loosely fitting O-ring. Rats showed elevated paw temperature 5 minutes after reperfusion that lasted 2 hours and extravasation lasting from 2–12 hours after reperfusion ( ). Rats with postischemic pain demonstrated behavioral hypersensitivity to light touch with von Frey filaments, to pinprick, and to cold stimuli in the form of an acetone drop that lasted up to 4 weeks after the procedure ( ). Thermal hyperalgesia was not observed. Administration of the free radical scavengers N -acetylcysteine and Tempol blocked the behavioral signs of enhanced pain, which is consistent with the role of free oxygen radicals in reperfusion injury and suggests a possible role for these radicals in CRPS-1 ( ).

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here