Elsevier

Journal of Neuroscience Methods

Volume 227, 30 April 2014, Pages 166-180
Journal of Neuroscience Methods

Basic Neuroscience
Invited review
Sciatic nerve injury: A simple and subtle model for investigating many aspects of nervous system damage and recovery

https://doi.org/10.1016/j.jneumeth.2014.01.020Get rights and content

Highlights

  • Sciatic nerve injury models many different nervous system pathologies.

  • Different functional components of the sciatic nerve are segregated in its branches.

  • Selective injury to the nerve components results in neuropathic pain states.

  • Technologies exist to accurately reproduce different degrees of injury.

  • Many methodologies to assess damage and promote recovery are described.

Abstract

Sciatic nerve injury has been used for over a century to investigate the process of nerve damage, to assess the absolute and relative capacity of the central and peripheral nervous systems to recover after axotomy, and to understand the development of chronic pain in many pathologies. Here we provide a historical review of the contributions of this experimental model to our current understanding of fundamental questions in the neurosciences, and an assessment of its continuing capacity to address these and future problems. We describe the different degrees of nerve injury – neurapraxia, axonotmesis, neurotmesis – together with the consequences of selective damage to the different functional and anatomic components of this nerve. The varied techniques used to model different degrees of nerve injury and their relationship to the development of neuropathic pain states are considered. We also provide a detailed anatomical description of the sciatic nerve from the spinal cord to the peripheral branches in the leg. A standardized protocol for carrying out sciatic nerve axotomy is proposed, with guides to assist in the accurate and reliable dissection of the peripheral and central branches of the nerve. Functional, histological, and biochemical criteria for the validation of the injury are described. Thus, this paper provides a review of the principal features of sciatic nerve injury, presents detailed neuroanatomical descriptions of the rat's inferior limb and spine, compares different modes of injury, offers material for training purposes, and summarizes the immediate and longterm consequences of damage to the sciatic nerve.

Introduction

The use of sciatic nerve axotomy (SNA)2 as an experimental model in neuroscience dates back at least to the turn of the 20th Century, when Santiago Ramón y Cajal described its use in his 1906 Nobel prize acceptance speech (Ramón y Cajal, 1906). His seminal contribution to the study of nervous system injury, the two volume Estudios sobre la Degeneración y Regeneración del Sistema Nervioso (Ramón y Cajal, 1928) published in 1913 and 1914, made frequent reference to the use of sciatic nerve lesions by others, suggesting that the technique was already widely practiced by neurophysiologists and neuroanatomists in the mid to late 19th Century (Anderson, 1902, Fraidakis, 2010, Garcia-Poblete et al., 2003, Lobato, 2008). Through clever application of the model and rigorous description of the results, Cajal made substantial contributions to the defense of neurotropic theory and nerve continuity in regeneration.

Cajal experimented with degrees of injury to the sciatic nerve, but it was not until the end of the Second World War that Sir Herbert Seddon published his ternary classification of nerve damage severity based on his observation of hundreds of trauma cases (Seddon, 1942, Seddon, 1943). In increasing order of damage, he defined neurapraxia, a rapidly reversible compression injury; axonotmesis, loss of axon continuity with preservation of the nerve sheath; and neurotmesis, where the nerve itself is transected. It is of no small significance that this classification is still current 70 years later.

Subsequently, Sunderland refined this system into five categories based on the histopathology, rather than the degree of injury, and added electrodiagnostic and clinical criteria that related the categories to the possibility of regeneration with or without surgical intervention (Sunderland, 1951). In Sunderland's classification the first and fifth degrees of injury correspond to Seddon's neurapraxia and neurotmesis, respectively. Thus the principal refinement was to subcategorize axonotmesis into three degrees of injury with progressively worse prognosis for spontaneous recovery; surgical intervention was considered by the third and recommended by the fourth degree. In 1988, Mackinnon and Dellon proposed a sixth degree (Mackinnon and Dellon, 1988) for cases where different parts of the nerve had suffered a combination of the previous five grades of injuries, resulting in a mixed syndrome.

The idea that the degree of injury might be of profound clinical importance had been mooted by the physicians Weir Mitchell, Morehouse and Keen (Richards, 1967a) during the American Civil War of 1861–1865. They described the development of an intractable burning pain after gunshot injuries to nerves and blood vessels (Weir Mitchell et al., 1864) that they later termed causalgia, from the Greeks words for heat and pain (Richards, 1967b). Weir Mitchell, who had studied in Paris under the renowned physiologist Claude Bernard, subsequently experimented on the sciatic nerve of rabbits (and human cadavers) in an attempt to better understand the damage caused by trauma (Weir Mitchell, 1872), showing that the tough sheath of the sciatic nerve made it particularly resistant to injury. His observations also led him to conclude that the nerve maintained functional and anatomical segregation of the axons:

“The toughness and general elasticity of nerve trunks sometimes serve a useful purpose in cases of ball wound, and I have repeatedly seen nerves escape total destruction from missiles simply because they were thrust aside, instead of being divided. … On the other hand, injuries of nerves in connection with bone or near to joints are likely to be severe and lasting, because at these points and in these positions the nerve trunk is more firmly anchored than elsewhere…

… When a spinal nerve emerges from the intervertebral canal it is motor and sensory, by the union of the anterior and posterior roots, which represent motion and sensation respectively. Whether or not these fibres become at once scattered so that every part of the area of the nerve contains an equal share of the nerve tubes, both of sense and motion, is not at present very clear. Such, however, is the popular medical belief, though there is a good deal of reason to think that the nerve filaments of either function remain in bundles; because, as we shall see later, it is very common to find that a nerve trunk, injured by a missile, has suffered in its sensory or motor functions alone, which could scarcely be accounted for upon any other supposition than that last mentioned. Any other explanation must presuppose some greater susceptibility to injury in one set of fibres than in another.” (Weir Mitchell, 1872) pp. 25–26.

Thus, by the end of the 19th Century, the SNA had been established as a fundamental experimental model in the two major fields that would continue to motivate its use over the following century – nerve regeneration and neuropathic pain. In both fields, recognition of the importance of the degree of nerve damage has driven the development of multiple technologies to replicate the consequences of different clinically relevant injuries. In the last two decades of the 20th Century, several important innovations were introduced to reliably induce neuropathic pain (see Section 1.3) and the technologies to evaluate and ameliorate the consequences of nerve damage are becoming ever more sophisticated.

The SNA has found widespread use not only in neuroscience, but also in endocrine (Pan et al., 2009, Sahenk et al., 2005), immunological (Cullheim and Thams, 2007, Vallejo et al., 2010) and translational (Choi et al., 2012, Eaton et al., 2002, Gu et al., 2005, Kato et al., 2003, Merle et al., 1994, Sahenk et al., 2005, West et al., 2007) research into nerve injury. Much recent work has focused on the role of cytokines (Abbadie et al., 2009, Jeon et al., 2009, Strong et al., 2012) and immune cell infiltration (Moalem and Tracey, 2006 and see Fig. 4) in the development and maintenance of neuropathic pain in a range of clinical conditions (Blackbeard et al., 2012, Calvo and Bennett, 2012, Maratou et al., 2009). The use of various kinds of grafts (Wood et al., 2011), both biological and artificial, or in combination (IJkema-Paassen et al., 2004, Martins et al., 2005a, Seo et al., 2013), to promote regeneration after peripheral nerve section is a promising field of research.

Increasingly sophisticated methodologies have ensured that the latest experimental techniques can be applied to SNA. In many clinical fields, the extended local application of drugs has been an important tool for understanding and treating pathological alterations. In the sciatic nerve, prolonged application of exogenous drugs or macromolecules can be achieved through release from osmotic minipumps implanted in the nerve (Lever et al., 2007), the DRG3 (Zhou et al., 1999) or the spinal cord (Geremia et al., 2010). Permeable artificial or natural polymer implants are a popular alternative (Maratou et al., 2009, Pu et al., 1999, Smith and Skene, 1997, Van der Zee et al., 1988) with several advantages, cost not being the least.

Genetic models and interventions are increasingly common in the biological sciences. While transgenic lines of many species, including the rat, are available, the ability to manipulate gene expression locally and temporally is a vital modern tool. The use of both siRNA and antisense oligonucleotides to knockdown protein expression (Jankowski et al., 2012, Lee et al., 2009, Tsantoulas et al., 2012) and viral vectors to transfect proteins (Gu et al., 2005, Hermens et al., 1997, Maratou et al., 2009, Tsai et al., 2010) have been described in conjunction with nerve crush or transection in the context of regeneration and neuropathy. An additional method for gene transfer into the sciatic nerve by muscle injection of an inactivated virus has been described (Kato et al., 2003). The powerful Cre/loxP system for targeted gene manipulation, widely used in the mouse (Eijkelkamp et al., 2010b, Nijboer et al., 2010), has recently been adapted for the rat nervous system (Schonig et al., 2012) along with other transgenic technologies (Huang et al., 2011, Tong et al., 2011). Immune cell infiltration into the nerve after axotomy has been measured by transplantation of GFP-labeled precursors (Pan et al., 2009), a field of particular interest given the use of precursor cell transplants to promote and support nerve regeneration. Thus, cell transplantation procedures have been used to create bridges for regeneration (Ladak et al., 2011), investigate the origin and development of chronic pain (Radtke et al., 2010), and post-injury behavior through the secretion of monoclonal antibodies from hybridoma cells (Cui et al., 2004).

An open question in the field of nerve injury is the nature of the signaling mechanisms that couple axonal damage to somatic responses in neurons (Barron, 2004, Mandolesi et al., 2004, Smith and Skene, 1997). The retrograde transport of signaling complexes can be detected as a distal accumulation after ligation (Johanson et al., 1995), or polymer cuffs releasing colchicine to block transport or tetrodotoxin to block electrical activity (Smith and Skene, 1997), while multiple retrogradely transported fluorescent markers can be used to evaluate survival and regeneration and central remodeling after SNA (Welin et al., 2009). The signaling complexes associated with detergent-resistant “lipid rafts” have been isolated from DRGs, cultured DRG neurons, and from sciatic nerve (Pristerá et al., 2012). Crushed sciatic nerve has been used to prepare cryosections as substrates for regeneration in culture (Pettigrew et al., 2001) which are readily amenable to environmental manipulation and offer an attractive alternative to live animal experimentation. Such models have been used to identify components of the extracellular matrix (Golding et al., 1996), such as chondroitin sulfates (Groves et al., 2005, Jungnickel et al., 2009) and polysialylated adhesion molecules (Galtrey et al., 2007), that inhibit the regeneration of damaged axons.

There is growing interest in the use of different forms of magnetic resonance imaging (MRI) to evaluate injury response non-invasively and in real time (West et al., 2007). Both gross degeneration and regeneration (including vascularization) have been evaluated by MRI (Bendszus et al., 2004, Wessig et al., 2004), while diffusion tensor imaging (DTI) can be used to selectively detect axon-regrowth and myelination (Lehmann et al., 2010, Morisaki et al., 2011). Alterations in the DRG and spinal terminals can be visualized by including manganese infusion in the peripheral injury (Matsuda et al., 2010).

In short, well over a century after Weir Mitchell and Cajal's groundbreaking work with sciatic nerve injury scientists continue to find ingenious new applications for this model in many fields.

There is a minor cottage industry in inventing custom-made instruments to reproduce the characteristics of neurapraxia, axonotmesis and neurotmesis. Ameroid rings – steel rings with an inner layer of casein that swells with time – have been deployed to gradually compress the nerve (Tzabazis et al., 2004), mimicking neurapraxia. An alternative non-surgical method of much relevance to positional compression injuries uses an external inflatable cuff placed around the leg, to cause a long-lasting functional deficit associated with alterations in the node of Ranvier without axon transection (Ochoa et al., 1972).

The nerve-crush model, which reliably produces axonotmesis (Beer et al., 2001, Chen et al., 1992, Ronchi et al., 2010, Varejao et al., 2004a), appears to be robust with regards to tools, pressure, time or repetitions in the histological responses and subsequent regenerative capacity (Bridge et al., 1994, Chen et al., 1992, Chen et al., 1993, Mosconi and Kruger, 1996, Ronchi et al., 2009), although some authors have raised doubts about this (Beer et al., 2001, Mazzer et al., 2008). Different forms of neurotmesis (neurectomy) have been described and can have substantially different outcomes (de Medinaceli, 1995, de Ruiter et al., 2008, Malusht et al., 2004, Yao et al., 1998). Even a model for generating ischemic injury limited to the sciatic nerve has been described, using local laser activation of a systemically administered thrombotic agent (Kupers et al., 1998).

In 1965, Melzack and Wall put forward the now classic gate theory of pain (Melzack and Wall, 1965). While crush and transection injuries were used extensively in the 1970s and 1980s to develop theories of pain (Wall and Devor, 1983, Wall et al., 1979, Wall and Gutnick, 1974, Wall and Woolf, 1986), they did not reliably reproduce the characteristics of causalgia, now called Complex Regional Pain Syndrome, that researchers in the field required (Bridges et al., 2001). Thus, from the late 1980s on, new models were developed to address this need; the principle methods currently being used for this are the chronic constriction injury (Bennett and Xie, 1988), partial nerve ligation (Seltzer et al., 1990) and spinal nerve ligation (Kim and Chung, 1992). These were largely forms of axonotmesis – sometimes combined with neurotmesis – that deliberately sought to cause a partial injury to the nerve. The intention was to spare axons, either sensory, sympathetic or motor, that appear important for the development of neuropathic pain (Kim et al., 1997).

An additional model relevant to the SNA is the spared nerve injury – transection of two of the three branches of the sciatic nerve after their trifurcation (Decosterd and Woolf, 2000) leaving the predominantly sensory sural branch intact (see Section 1.4), in contrast to the spinal nerve ligation model where the sensory and motor components are lesioned at a specific spinal segmental level close to the DRG. Both heating and cooling of the sciatic nerve have been reported to produce some form of neuropathy (Willenbring et al., 1995), but these techniques have received less attention within the research communities.

In the context of animal well-being, the relationship between neuropathic pain and the occurrence of self-injury (autotomy) remains nebulous (Koplovitch et al., 2012). This behavior occurs more frequently in protocols using forms of ligation or transection related to neuropathy than in the crush injury described in Section 2 (Martins et al., 2005b, Minert et al., 2007, Mosconi and Kruger, 1996, Obata et al., 2003). Here, again, substantial strain-specific differences have been reported, although autotomy does covary with sensibility to neuropathic pain after axotomy within strains (Carr et al., 1992, Persson et al., 2009, Shir et al., 2001, Ziv-Sefer et al., 2009). Regenerative processes involved in the formation of the neuroma have been implicated in the development of both neuropathic pain and autotomy (Foltán et al., 2008, Radtke et al., 2010, Small et al., 1990).

With respect to the central conditioning effect (described in Section 1.4) of peripheral nerve injury, it bears mentioning that crush, ligation and transection have all been used as conditioning lesions for spinal sensory and motor axons (Jacob and McQuarrie, 1993, Lankford et al., 1998).

In designing experiments with any form of sciatic nerve injury, it is essential to consider the effects of cell death on the outcome. It is estimated that about a third of DRG neurons and up to half of spinal motoneurons die after sciatic nerve transection in the rat (Himes and Tessler, 1989, Pu et al., 1999, Tandrup et al., 2000), with differential effects on anatomical (Mazzer et al., 2008, Tandrup et al., 2000, Vestergaard et al., 1997) and functional (Vanden Noven et al., 1993, Welin et al., 2008) subpopulations, although the origin of this phenomenon remains unclear (Devor et al., 1985). Neuronal death is scarce but visible at a day post-transection (McKay Hart et al., 2002, Whiteside et al., 1998) and continues for 6–8 months after the injury (McKay Hart et al., 2002, Tandrup et al., 2000, Welin et al., 2008). It has been reported that neuronal death is significantly greater in young animals (Schmalbruch, 1987), that also show a greater capacity to regenerate. Chromatolysis (Barron, 2004), Wallerian degeneration (Bridge et al., 1994) and apoptosis (Groves et al., 1997) all contribute to this process, which is not confined to neuronal cells (McKay Hart et al., 2002). Surgical repair of a transected nerve (see Section 2.4) to permit regeneration has been reported to reduce the amount of neuronal cell death (McKay Hart et al., 2002). There are also indications that adult neurogenesis may occur after SNA (Devor and Govrin-Lippmann, 1985, Devor et al., 1985, Groves et al., 2003).

Not surprisingly, given the long history and wide use of SNA to study regeneration, there are many indices of recovery (Martins et al., 2005a, Raivich and Makwana, 2007, Wood et al., 2011). Axonal regrowth occurs at 3–4 mm/day, starting within ∼1/2 a day in the rat although longer latencies have been reported depending on neuron and injury type (Forman and Berenberg, 1978, Jacob et al., 2000, Lozeron et al., 2004). The signals that initiate the processes of regeneration are still an active field of research (see Section 1.2), although it is widely agreed that they involve a combination of locally generated and retrogradely transported factors (Johanson et al., 1995, Liu et al., 2011, Michaelevski et al., 2010, Perlson et al., 2004, Raivich and Makwana, 2007, Smith and Skene, 1997). Differential regeneration into the peripheral targets of the L4 and L5 DRGs has been reported (Puigdellivol-Sanchez et al., 2005), while full sensory and motor recovery after SNA has been described in as little as three weeks in rats (Vogelaar et al., 2004). Stress appears to affect several parameters of recovery (van Meeteren et al., 1997), and the process is clearly responsive to sex-specific steroids (Roglio et al., 2008) and age (Kang and Lichtman, 2013).

When performed purposefully and without neuropathic addenda the SNA leaves few functional sequelae, complicating behavioral validation of the lesion. Within hours of recovery from the anesthesia, the animals are observed to move, feed and groom without problems. Positive confirmation of nerve damage is often taken from the presence of a limp or using more sophisticated footprint analysis techniques (Baptista et al., 2007, Bervar, 2000), but these may also arise from muscle injury during surgery or post-surgical inflammation in the absence of effective SNA. A simple observational technique is based on the presence of characteristic cutaneous lesions (see Fig. 2) that accompany altered posture and foot-dragging (Baptista et al., 2007, Bozkurt et al., 2008) (and see Section 1.4), which allow an early and midterm (up to ∼10 days) evaluation of the injury. Longer-term evaluation of motor and sensory alterations require specific tests (Nichols et al., 2005, Varejao et al., 2004b), usually involving specialized equipment, but which have been reported to give ambiguous results in the short term (Monte-Raso et al., 2006, Monte-Raso et al., 2008). It should also be borne in mind that the time for full sensorimotor recovery, often quoted as 2–3 months post-SNA, varies with the test used (Luis et al., 2007), and may even represent recovery mediated through other nerves (Rupp et al., 2007a, Rupp et al., 2007b).

Many studies on nerve injury have focused on the upper limb, due to the gravity of loss of the ability to manipulate objects. Nonetheless, the SNA model is anatomically simpler and more accessible than the corresponding upper limb injury (Bontioti et al., 2003), and causes less inconvenience and thus distress to the operated animal, a criticism that can also be made of the facial nerve model (Moran and Graeber, 2004). The lumbar femoral nerve, or its or its mostly sensory saphenous branch, have been proposed as simpler models (Irintchev, 2011, Kingery et al., 1993) but lack the popularity of the sciatic model, perhaps because they are not as surgically accessible or carry additional problems related to the closely associated vascaluture (Zimmermann et al., 2009).

The popularity of the sciatic model for investigating nerve injury is indisputable, and likely arises from the surgical accessibility of the sciatic nerve at the mid-leg level, and its well characterized central (Decosterd and Woolf, 2000) and peripheral (Greene, 1935) projections (see below). As noted, some forms of injury to the sciatic nerve can result in chronic pain conditions and a number of variations in the type of lesion have been developed to produce reliable experimental models of neuropathic pain. These models are considered in detail in Section 1.3.

The sciatic nerve is a mixed nerve containing sensory, and somatic and autonomic motor axons (Schmalbruch, 1986), originating predominantly from the 4th and 5th lumbar segments (Swett et al., 1991) (and see Suppl. Figs. 2–5) of the spinal cord and associated DRG in the rat (summarized in Suppl. Table 1). There is controversy in the literature concerning the contributions from the 3rd and 6th lumbar segments (Asato et al., 2000, Puigdellivol-Sanchez et al., 1998, Puigdellivol-Sanchez et al., 2000, Shehab et al., 2008), although it is generally agreed that these components are relatively minor. It should nonetheless be borne in mind that the neuroanatomy varies between species (Rigaud et al., 2008), strains and individuals (Puigdellivol-Sanchez et al., 1998) and may even be non-symmetric within a given animal (Asato et al., 2000). Failure to allow for this variability can contribute to confusion in the interpretation of results obtained with the model, and is the motivation for the detailed anatomic exploration recommended in Section 2 and the anatomical guides in the Supplementary Information.

As speculated by Weir Mitchell (1872), the peripheral extension of the sciatic nerve is not a homogenous mix of different functional subtypes of axon, but reflects to a significant extent their anatomical segregation according to the final targets of each nerve branch (see Fig. 1 and Suppl. Table 1). The partial nerve ligation (Bennett and Xie, 1988) and spared nerve injury (Decosterd and Woolf, 2000) models of neuropathic pain exploit this characteristic. In the former, multiple loose ligatures of the sciatic nerve damage some but not all of the axons, causing a variable proportion (Bridges et al., 2001) of the animals to develop hyperalgesia and/or allodynia. The spared nerve model was developed to reduce the variability in the injury by lesioning the terminal branches after their trifurcation, thus controlling the number and type of axons lesioned.

In Section 2, we describe the lesion of the sciatic nerve beyond the radiation of the pelvic nerve and nerves to muscles of the upper leg, but before the trifurcation of the tibial, sural and common peroneal branches (Rupp et al., 2007c). These latter nerves have also been subjected to individual lesions (Isaacs et al., 2013, Lozeron et al., 2004, Malusht et al., 2004, Povlsen and Hildebrand, 1993) to produce more restricted motor or sensory deficits (Inserra et al., 1998) but, as discussed in Section 1.3, such partial interventions substantially increase the likelihood of neuropathic complications.

After the trifurcation of the sciatic nerve at the distal thigh, the different muscular, articular and superficial bundles separate and radiate rapidly throughout the leg (Suppl. Figs. 6, 7 and 9) in a complex net of interconnections. The tibial nerve soon divides to produce the largely sensory sural branch, then further divides into the gastrocnemius, popliteus, soleus and plantar branches, innervating the eponymous muscle groups. There is also an articular branch to the knee (Povlsen and Hildebrand, 1993), The common peroneal (or fibial) branch soon bifurcates into the superficial and deep peroneal nerves which innervate muscles of the lower leg and foot, with an additional branch that contributes to the sural nerve. Damage to the tibial nerve causes loss of dorsoflexion of the paw and extension of the toes (see Fig. 2b).

The sural nerve is predominantly cutaneous sensory and sympathetic, with little or no somatic motor component (Suppl. Table 1). The peroneal and tibial branches contain a mixture of cutaneous and motor sensory complements, and should not be considered principally motor nerves (Swett et al., 1991). Given the interconnectedness of the lumbar plexus and associated nerves (Suppl. Figs. 4 and 5), it is not surprising that there is considerable overlap in the sensory dermatomes reported for the different lumbar nerves (Bajrovic and Sketelj, 1998, Pinter and Szolcsanyi, 1995, Sheth et al., 2002, Takahashi and Nakajima, 1996). Indeed, the dermatomes corresponding to the three sciatic branches (see Fig. 4 in Takahashi and Nakajima, 1996) are principally associated with the L4 root, but with considerable input from the L3 and L5 DRG (Bajrovic and Sketelj, 1998), while the epicritical nociceptors of the sural branch may not have an autonomous territory at all. These distributions are largely derived from studies using C-fiber stimulated extravasation, and may not be representative of other fiber types (Bajrovic and Sketelj, 1998). It should also be borne in mind that the somatic dermatomes converge centrally with visceral innervation (Pinter and Szolcsanyi, 1995), which might contribute to referred pain (Shin and Eisenach, 2004), and that selective damage to the L5 roots and nerve induce hyperalgesia in the L4 dermatome (Sheth et al., 2002).

Procedures have been described to selectively remove the sympathetic component of the sciatic nerve (Nakamura et al., 2003, Povlsen and Hildebrand, 1993, Shir and Seltzer, 1991, Willenbring et al., 1995), both to investigate the contribution of sympathetic nerves to neuropathic pain as proposed by Weir Mitchell, and the possibility of alleviating it. Peripheral sympathetic lesions are complicated by the number of branches that arise from the lumbar plexus (see Suppl. Figs. 4 and 5) and the frequency with which they run on the surface of vascular structures. While these selective lesions are conceptually useful, they can produce different indices of recovery depending on the test used (Nichols et al., 2005) and the central sequelae of these lesions become progressively more difficult to interpret as the peripheral lesion becomes more selective.

While many studies have focused on the loss of motor function after nerve injury and the possibility of its recovery, the sensory aspect of SNA has an important characteristic that makes it of particular interest. The pseudounipolar sensory neurons of the DRGs emit a single axon that bifurcates locally to innervate both the peripheral target via the spinal nerves and the central nervous system via the dorsal roots. Within the spinal cord, the central projections synapse locally and form ascending and descending tracts, including the long dorsal column tracts that connect with the brain stem. As Cajal observed, the peripheral axons regenerate when cut, but the central projections of these same neurons re-grow only as far as the dorsal root entry zone (or transition zone) of the spinal cord, where specific regeneration fails. Contrasting the behavior of the two branches of the DRG neurons has made the SNA the canonical system for investigating the origins of successful and failed regeneration in the nervous system (Abe and Cavalli, 2008).

A further consequence of this unusual anatomic distribution of the sensory axons is that injury to the peripheral axon induces similar regenerative changes in both central and peripheral branches (Abe and Cavalli, 2008, Liu et al., 2011, Richardson and Issa, 1984). Thus the SNA can be used as a conditioning lesion to promote and investigate the limits of central regeneration (Neumann and Woolf, 1999) without resorting to the much more surgically invasive lesion of the central roots within the vertebral column. The procedure has even been proposed as a possible intervention to promote spinal cord regeneration after injury (Neumann et al., 2005, Silver, 2009).

Sciatic nerve anatomy, morphology and physiology vary not only between species (Rigaud et al., 2008) but also between strains, genders, and individuals of different ages (Barron, 2004, Chakrabarty et al., 2008, Roglio et al., 2008, Tehranipour and Moghimi, 2010, Ziv-Sefer et al., 2009). The anatomy is very similar between Sprague-Dawley, Wistar-Han, Lewis and Nude rats, but Fischer 344 rats evince early separation of the tibial and peroneal branches, potentially complicating the protocol for lesion of the entire sciatic nerve (Rupp et al., 2007c). Different strains have also been reported to vary in their response to SNA, particularly in their tendency to develop neuropathic pain (Rigaud et al., 2008) and engage in autotomy (Shir et al., 2001) (discussed in Section 1.3). In consequence, while literature citations indicate that Sprague-Dawley rats remain the most published model, it should not be assumed that results from this species and strain are quantitatively comparable to others.

It is standard to use homologous structures on the contralateral side of the animal as a control for the effects of unilateral SNA – with or without sham surgery (Kingery et al., 1993) – although it is clear that systemic effects after SNA do occur (Serarslan et al., 2009). Astrocyte proliferation in the spinal cord (Pavic et al., 2008, Zhao et al., 2008), development of neuropathic pain (Shir and Seltzer, 1991), upregulation of neurotrophins in the sciatic nerve (Shakhbazau et al., 2012), and apoptosis in the DRG (Whiteside et al., 1998) have all been described on the side contralateral to injury. It is also possible that there is some anatomic and functional coupling between sciatic nerve injury and effects mediated by adjacent spinal nerves (Shehab et al., 2008, Sheth et al., 2002), such as the saphenous nerve (Kingery et al., 1993), so caution should be exercised in using rostral or caudal segments as controls. Furthermore, it has been reported that selective lesion of the L5 motoneurons in the ventral root exerts a conditioning effect on the corresponding ipsilateral sensory neurons (Li et al., 2009), implying that there exists a mechanism of retrograde motor-sensory coupling in the response to axotomy. These reports should be borne in mind when paradoxical effects with respect to the control are observed after SNA.

In the following section, we present a step-by-step description, illustrated by photographs and videos, of how to consistently perform axonotmesis on the sciatic nerve using the minimally invasive double crush injury at the mid-thigh level. This reduces the chances of neuropathic pain and maximizes the possibility of axon regeneration, and has given us reliable results for over twenty years (Hess et al., 1993, Patterson and Skene, 1994). Furthermore, we present several alternative criteria for the validation of the nerve injury model. Finally, we provide a review of the principal problematic features of the experimental design that can affect the results; and we describe the surgical anatomy of the sciatic nerve from its trifurcation in the lower leg right up to its origin in the lumbar spinal cord to aid in efficient surgery and accurate post-surgical harvesting of the peripheral and central branches of the sciatic nerve. We expect that this protocol will allow researchers to organize knowledge into a cohesive framework in those areas where SNA is applied.

Section snippets

Experimental animals

Historically, the rat has been the preferred model for investigating peripheral nerve injury and regeneration (Tos et al., 2009), having similar microscopic structure and injury response to human nerves (Croteau et al., 2005, Mackinnon et al., 1985); and additionally has analogous central responses to pain (Becerra et al., 2013). Larger animals – rabbits, chickens, cats or sheep – have been used and have particular advantages, but are more difficult to house and maintain and may be subject to

Assessment of injury

Direct confirmation of the SNA can of course be obtained anatomically, by identification of the injury and neuroma in dissection; however, the formation and duration of the neuroma is variable in both time and size (Foltán et al., 2008). A major source of error in evaluating the consequences of SNA is a poor understanding of the associated anatomy leading to misidentification of the structures to be investigated. As discussed in Sections 1.4 Functional neuroanatomy of the sciatic nerve, 2.1

Animal subjects

The procedures described here were assessed and approved by the Institutional Committee for the Care and Use of Laboratory Animals (Comité Institucional para el Cuidado y Uso de Animales de Laboratorio, CICUAL), in accordance with national law and international standards. The Universidad Nacional de Cuyo has PHS Approved Animal Welfare Assurance (registry # A5780-01).

Conflict of interest statement

The authors declare no competing financial interests.

Acknowledgements

L.E.S., M.R.F. and J.A.R. were supported by Introduction to Basic Research Fellowships from the Facultad de Ciencias Médicas, and L.E.S. and M.R.F. by Research Fellowships from the Universidad Nacional de Cuyo (UNC). J.A.R. was recipient of a Scientific Vocational Award from the Consejo Inter-Universitario. S.R.L. was supported by a Research Formation Fellowship from the Instituto Nacional de Cancer. V.G.P. was recipient of a Graduate Fellowship from the Consejo Nacional de Investigación en

References (226)

  • M.M. Bradford

    A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding

    Anal Biochem

    (1976)
  • P.M. Bridge et al.

    Nerve crush injuries – a model for axonotmesis

    Exp Neurol

    (1994)
  • D. Bridges et al.

    Mechanisms of neuropathic pain

    Br J Anaesth

    (2001)
  • M. Calvo et al.

    The mechanisms of microgliosis and pain following peripheral nerve injury

    Exp Neurol

    (2012)
  • S. Cullheim et al.

    The microglial networks of the brain and their role in neuronal network plasticity after lesion

    Brain Res Rev

    (2007)
  • L. de Medinaceli

    Interpreting nerve morphometry data after experimental traumatic lesions

    J Neurosci Methods

    (1995)
  • G.C. de Ruiter et al.

    Misdirection of regenerating motor axons after nerve injury and repair in the rat sciatic nerve model

    Exp Neurol

    (2008)
  • I. Decosterd et al.

    Spared nerve injury: an animal model of persistent peripheral neuropathic pain

    Pain

    (2000)
  • M. Devor et al.

    Neurogenesis in adult rat dorsal root ganglia

    Neurosci Lett

    (1985)
  • M.J. Eaton et al.

    Immortalized chromaffin cells disimmortalized with Cre/lox site-directed recombination for use in cell therapy for pain after partial nerve injury

    Exp Neurol

    (2002)
  • R. Foltán et al.

    Mechanism of traumatic neuroma development

    Medical Hypotheses

    (2008)
  • D.S. Forman et al.

    Regeneration of motor axons in the rat sciatic nerve studied by labeling with axonally transported radioactive proteins

    Brain Res

    (1978)
  • N.M. Geremia et al.

    Endogenous BDNF regulates induction of intrinsic neuronal growth programs in injured sensory neurons

    Exp Neurol

    (2010)
  • S. Geuna et al.

    Histology of the peripheral nerve and changes occurring during nerve regeneration

    Int Rev Neurobiol

    (2009)
  • J.P. Golding et al.

    An in vitro model of the rat dorsal root entry zone reveals developmental changes in the extent of sensory axon growth into the spinal cord

    Mol Cell Neurosci

    (1996)
  • J.W. Griffin et al.

    Measuring nerve regeneration in the mouse

    Exp Neurol

    (2010)
  • M.L. Groves et al.

    Axon regeneration in peripheral nerves is enhanced by proteoglycan degradation

    Exp Neurol

    (2005)
  • Y. Gu et al.

    Remote nerve injection of mu opioid receptor adeno-associated viral vector increases antinociception of intrathecal morphine

    J Pain

    (2005)
  • W.T. Hermens et al.

    Transient gene transfer to neurons and glia: analysis of adenoviral vector performance in the CNS and PNS

    J Neurosci Methods

    (1997)
  • J. IJkema-Paassen et al.

    Transection of peripheral nerves, bridging strategies and effect evaluation

    Biomaterials

    (2004)
  • A. Irintchev

    Potentials and limitations of peripheral nerve injury models in rodents with particular reference to the femoral nerve

    Ann Anat

    (2011)
  • J. Isaacs et al.

    A rodent model of partial muscle re-innervation

    J Neurosci Methods

    (2013)
  • S.M. Jeon et al.

    Expression of monocyte chemoattractant protein-1 in rat dorsal root ganglia and spinal cord in experimental models of neuropathic pain

    Brain Res

    (2009)
  • C.R. Jimenez et al.

    Proteomics of the injured rat sciatic nerve reveals protein expression dynamics during regeneration

    Mol Cell Proteomics

    (2005)
  • S.O. Johanson et al.

    Retrograde axonal transport of signal transduction proteins in rat sciatic nerve

    Brain Res

    (1995)
  • J. Jungnickel et al.

    Level and localization of polysialic acid is critical for early peripheral nerve regeneration

    Mol Cell Neurosci

    (2009)
  • W.S. Kingery et al.

    A loose ligature-induced mononeuropathy produces hyperalgesias mediated by both the injured sciatic nerve and the adjacent saphenous nerve

    Pain

    (1993)
  • P. Koplovitch et al.

    Spontaneous pain in partial nerve injury models of neuropathy and the role of nociceptive sensory cover

    Exp Neurol

    (2012)
  • R. Kupers et al.

    Photochemically-induced ischemia of the rat sciatic nerve produces a dose-dependent and highly reproducible mechanical, heat and cold allodynia, and signs of spontaneous pain

    Pain

    (1998)
  • A. Ladak et al.

    Differentiation of mesenchymal stem cells to support peripheral nerve regeneration in a rat model

    Exp Neurol

    (2011)
  • K.M. Lee et al.

    Tumor necrosis factor receptor 1 induces interleukin-6 upregulation through NF-kappaB in a rat neuropathic pain model

    Eur J Pain

    (2009)
  • H.C. Lehmann et al.

    Diffusion tensor imaging to assess axonal regeneration in peripheral nerves

    Exp Neurol

    (2010)
  • I.K. Lo et al.

    Arthroscopic knots: determining the optimal balance of loop security and knot security

    Arthroscopy

    (2004)
  • R.D. Lobato

    Historical vignette of Cajal's work “Degeneration and regeneration of the nervous system” with a reflection of the author

    Neurocirugia (Astur)

    (2008)
  • C.M. Loss et al.

    Ketamine reduces neuronal degeneration and anxiety levels when administered during early life-induced status epilepticus in rats

    Brain Res

    (2012)
  • P. Lozeron et al.

    Regeneration of unmyelinated and myelinated sensory nerve fibres studied by a retrograde tracer method

    J Neurosci Methods

    (2004)
  • A.L. Luis et al.

    Long-term functional and morphological assessment of a standardized rat sciatic nerve crush injury with a non-serrated clamp

    J Neurosci Methods

    (2007)
  • H.K. Anderson

    The nature of the lesions which hinder the development of nerve-cells and their processes

    J Physiol

    (1902)
  • F. Asato et al.

    Variation in rat sciatic nerve anatomy: implications for a rat model of neuropathic pain

    J Peripher Nerv Syst

    (2000)
  • L. Becerra et al.

    Analogous responses in the nucleus accumbens and cingulate cortex to pain onset (aversion) and offset (relief) in rats and humans

    J Neurophysiol

    (2013)
  • Cited by (87)

    • Slow progression of sciatic nerve degeneration and regeneration after loose ligation through microglial activation and decreased KCC2 levels in the mouse spinal cord ventral horn

      2022, Neuroscience Research
      Citation Excerpt :

      Peripheral nerve injury affects both motor and sensory functions (Gordon, 2020; Riccio et al., 2019; Savastano et al., 2014).

    View all citing articles on Scopus
    1

    Current address: Department of Neurosurgery, University of Michigan, Ann Arbor, MI, USA.

    View full text