Interpretation of data from translational rodent nerve injury and repair models

Introduction

Preclinical studies using animal models are often necessary in developing new therapies or procedures. This review aims to guide and inform how rodent models, the most common preclinical models used to study nerve injury and regeneration, should be interpreted by clinicians to appropriately translate these findings to clinical application.

Comparing nerve anatomy and biology of rodents versus humans

Rodent models are frequently used to model human disease; however, there are key differences between the two species to consider in the context of nerve injury and repair. Anatomically and structurally, while rodent and human nerves share many similarities, there are important differences. In both species, nerves are comprised of unmyelinated and myelinated axons apposed by Schwann cells (SCs). These axons are immediately surrounded by extracellular matrix (ECM) forming an endoneurium, then further sorted into fascicles defined by a discrete connective tissue sheath perineurium. In humans, nerves in the proximal root-level brachial plexus region are generally mono-fascicular; however, even at the trunk-level motor and sensory fascicles are grouped separately. As the nerves continue distally into the arm, they become poly-fascicular and topography becomes more clearly defined.¹ Rodent nerves are generally considered mono-fascicular. However, as an example using the sciatic nerve, the location of the peroneal, tibial, and sural fascicles are still topographical even in the proximal thigh, and rodents maintain topographical information similar to humans with motor and sensory fiber grouping.² Despite these differences, the histology of the rodent nerve on a fascicular level is indistinguishable from human tissue. In rodents, there is generally less ECM surrounding the fascicles of the nerve. Humans have more connective tissue and accompanying stromal cells surrounding the fascicles of nerve, which could contribute to ECM remodeling and introduce differences to fibrous tissue formation after nerve injury,^3,4 As such, this difference may need consideration in an experimental design using rodents.

One obvious difference between rodents and humans is their physical size and, therefore, their nerve sizes. Due to their smaller size, the time to achieve end organ reinnervation after peripheral nerve injury is more rapid in rodents. Also, it is generally agreed upon that after nerve injury, axonal growth rates during regeneration are generally more rapid in rodents compared to humans. These factors are important considerations as downstream denervated nerve and muscle undergo extensive changes after injury with respect to time. In the setting of transection and acute repair, there will therefore be less denervation-induced atrophy in the distal nerve and muscle in rodents compared to humans. However, carefully selected rodent injury models, such as delayed repair and models of chronic denervation, can account for these changes (see next section).

Although the aforementioned anatomical, structural, and physical size differences in nerve exist between these species, there is extensive overall genomic homology between rodents and humans. Eighty percent (80%) of protein-coding genes in Mus musculus have a single identifiable orthologue in humans. Further, only approximately 300 genes seem to be unique to these species⁵. More specific to nerve injury and regeneration, rodents have been shown to be dependent on a pro-regenerative, supportive SC phenotype to promote overall nerve regeneration.⁶ Repair genes and transcription factors that mediate this SC phenotype in mice, such as c-Jun and p75NTR, have been confirmed to be upregulated following human nerve injury as well.⁷ Also, general mechanisms of nerve repair, such as the mechanism of Wallerian degeneration⁸, regeneration speed⁹, as well as trophic factor¹⁰, cytokine, and chemokine profiles¹¹ at the site of nerve regeneration all show significant similarities. However, important metabolic processes that could affect support of a regenerating nerve differ between rodents and humans. Rodents have higher rates of production of reactive oxygen species, have different adaptive immune system function, such as mechanisms for memory T cell population maintenance¹², and have completely different xenobiotic metabolism. However, systemic genomic changes in the rodent associated with states of short term inflammatory stress such as burn, trauma, or sepsis tend to mimic those of humans.¹³ These similarities argue for the utility of rodents as an animal model of systemic human disease. The difference in regenerative potential after traumatic injury between species may be in large part due to transcriptomic and lipidomic changes that accompany transition into a repair state in rodents, but are absent in humans. In an ex vivo nerve model, it was shown that differentiated SC markers such as Erbb2 and Egr2 were more efficiently downregulated in rodents, while repair genes such as Shh and Atf3 were induced more rapidly and strongly.¹⁴ Also, rodent SCs showed robust changes in lipid metabolism driven by PPARγ and S1P downregulation proposed to facilitate myelin removal and promotion of a repair phenotype more efficiently. Notably, human SCs did not show these lipidomic changes, but responded pharmacologically to PPARγ up- and downregulation, providing a potential target for modulating repair after peripheral nerve injuries.¹⁴ While these differences require consideration with regards to findings in rodent models, the many general similarities between rodents and humans demonstrate the utility of rodents as a useful first step to improve understanding of nerve injury, regeneration, and therapeutics to manage injuries.

Assessing regenerative outcomes using rodent versus human

The assessment of nerve regeneration after injury in rodents is frequently performed using histological and functional metrics, which can range from electrophysiological to behavioral analysis, at defined endpoints. In humans, semi-quantitative and qualitative techniques such as sensation, control of motor function and strength testing have traditionally been used instead.¹⁵ Critically, there is not a direct linear relationship between function and axonal regeneration or axon quantity.^16–18 (Fig. 1) The interpretation, sources of variability, and bias using rodent models can often be mitigated by experimental design and injury model, rational outcome metrics and endpoints,^19,20 mechanistic studies to provide explanation for conclusions drawn, and reproducibility of the proposed effect across different research groups.

Fig. 1.

Muscle function and percentage of motor innervation do not exhibit a linear relationship. Sunderland and Seddon Classifications are included and match up to varying levels of axonal loss. Note that this figure depicts the anticipated functional recovery after a nerve injury has stabilized, and not the function immediately following nerve injury. (Adapted with permission from Power et al., 2019.)

Experimental Design and Injury Models

When using rodent nerve models to assess nerve regeneration, it is critical to select a nerve injury and repair model with strong negative and positive control groups. These must be designed to show impaired regeneration in the negative control group that is statistically and reliably different than the positive control group. As an example, suppose a novel intervention (i.e. a novel graft) to repair a nerve gap injury is evaluated. A comparison to a silicone conduit (negative control) and an autograft (positive control) would be needed to provide context. It is well established that the silicone conduit will yield outcomes (histology, functional recovery) that are inferior to the autograft, which provides context to how the experimental intervention performs. Studies that fail to provide both controls when examining a novel intervention limit the ability to understand the impact of the proposed intervention.

The sciatic nerve model is still the most used for studying nerve injury and regeneration. The sciatic nerve is large and easily accessible for procedures and provides excellent histology, electrophysiology, and behavioral measurements. However, when considering sciatic nerve crush and cut-repair procedures, which represent the most commonly studied injury modalities, histological measures of axon regeneration will often match pre-injury levels given enough time for regeneration. Further, total recovery of motor function in the sciatic nerve distribution is not observed after a cut-repair procedure, whereas a crush injury is able to show almost complete recovery.^21,22 These outcomes emphasize the model’s validity as a starting point to evaluate therapies to improve regeneration.

As mentioned in the previous section, time is a major factor in clinical outcomes, as downstream denervated nerve and muscle undergo extensive changes after injury. The nerve and muscle distal to the injury, devoid of axons for periods of months and termed “chronic denervation,” becomes less hospitable and conducive to axon regeneration and reinnervation/recovery, respectively.^23,24 The rodent nerve cut-repair model with immediate repair does not adequately reflect the time-dependent denervation-induced changes in muscle and SC that occur in humans with immediate repair after proximal nerve injury. Instead, models that delay nerve repair after initial injury have been adopted as relevant to model changes in muscle and nerve that occur during long distances (and therefore regenerative time) for human nerve. These models have served to improve our understanding of a key challenge: denervation-induced atrophy of muscle and SCs is the major barrier to achieving satisfactory regeneration and functional recovery in the clinic.^25,26 Early work inducing chronic denervation in rodents highlighted the importance of SC support to promote axonal growth to reinnervate targets, correlating distal SC degeneration with decreased axonal regeneration to targets and placing importance on swift repair after injury.^27,28 Independently, it has also been shown that there is a rapid decrease in muscle mass and function if the repair of a transected rat tibial nerve was delayed for more than one month similar to Gordon’s data.²⁹ (Fig. 2A) In a model created to weigh the relative contributions of SC and muscle denervation to negative impact on functional recovery, it has been shown that the deleterious effects of muscle denervation likely has a greater impact.²⁶ Though models of chronic dennervation are not a perfect recapitulation of long-distance nerve regeneration in humans, the rodent model of chronic denervation has been invaluable to improving how to interpret regenerative findings from rodents to humans.

Fig. 2.

A) Gastrocnemius muscle mass to body mass ratio in rat tibial nerve cut-repair (grey) and transection with no repair (black) at varying denervation times. Similar to humans undergoing denervation, there was a period (by 3 months) when denervation resulted in sub-optimal muscle mass recovery even after repair. Adapted with permission from Kobayashi et al. 1997. B) Sciatic Functional Index (SFI), Tibial Functional Index (TFI), and Peroneal Functional Index (PFI) after transection and repair (cut-repair) of each respective nerve. Sciatic nerve repair resulted in some return of function between 8 and 24 weeks. Tibial nerve repair recovered by 4 weeks and continued to show improved function. Peroneal nerve repair never yielded complete loss of function (−100 PFI) and values were comparable to control by 8 weeks. (Adapted with permission from Hare et al. 1992.)

In instances where a nerve gap repair model is selected, the use of a critical nerve gap is preferred to strengthen conclusions. However, there is no broad consensus as to the extent at which critical nerve gaps, a gap long enough at which axon regeneration across the gap is limited without intervention, differ between species.^30,31 Critical nerve gaps in rats have originally been reported as ~1 cm, whereas in humans it is considered ~3 cm.³¹ However, this rodent critical gap length was based upon the limit to axon regeneration across an empty conduit (i.e. >1 cm). As such, rodents are frequently used to model experimental therapies to repair nerve gap lengths of 1–1.5 cm based on this assumed critical nerve gap. To this point, there is evidence to suggest that the nerve gap differences between species are more similar than different. In rodents, the critical gap length with the nerve graft alternative, acellular nerve allografts (ANAs), is around 3 cm, the same gap length recommended by most surgeons for similar use in humans.³² Additionally, complete resection of the rat sciatic nerve has shown that axons can regenerate up to ~2.5 cm without any structure to support axon regeneration from the proximal nerve, which is more in-line with the critical gap length in humans.³³ This similarity emphasizes the ability to extrapolate preliminary results in rodent nerve grafts to human nerve grafts. Also, axon regeneration across grafts is independent of graft geometry,³⁴ and therefore it is possible to expand physiologic gap lengths through tortuous grafts to model much sought-after longer gaps of clinical relevance using rodents. This has been accomplished by multiple groups within the last decade.^35,36

Finally, translation of data from hind limb models can be hindered by differences in neural plasticity in the rodent driven in large part by differences in regenerative profile and spinal cord circuit involvement that is not present in humans.^12,14,37,38 Using alternative models such as forelimb or isolated tibial nerve injuries may mitigate some of these differences, however. Muscle force testing is one way to measure functional outcomes while attempting to mitigate behavioral input (discussed more in latter sections).^38,39,40As most human peripheral nerve surgeries are performed in the upper extremity, the use of the upper extremity in rodent models could improve translation.⁴¹

Outcome Metrics and Endpoints: Histological

Regeneration in rodents is frequently measured via direct postmortem histological analysis of regenerated axons. The advent of transgenic rodents with fluorescent reporters has made identifying and tracing nerve regeneration possible without sacrifice of the animal.⁴² But as briefly touched upon in the previous section, histological measurements of regeneration, in particular axon regeneration, require careful consideration of both the model and endpoint. For example, in evaluation of myelinated axon regeneration using histomorphometry, error in interpretation of outcome can be made by not following these principles. In the rodent crush and cut-repair models, there exists a small timeframe to assess superior axonal regeneration between experimental variables tested. This phenomenon of limited yet critical endpoints for outcome measures has been coined the “blow-through” effect, in which regenerating axons overcome experimental defects given sufficient time.²⁰ For example, in the sciatic nerve cut-repair model, there is a significant increase in fibers distal to the repair site matching or exceeding pre-injury values by 4 weeks, in part due to axonal sprouting with subsequent pruning and return to baseline after 24 months.⁴³ If an intervention was assessed in this model, the interpretation of improved axon regeneration would be missed if the histomorphometric assessment was not performed prior to 4 weeks.

Retrograde labeling of neurons allows for better characterization of axon outgrowth, specificity, and modality of regenerating axons.^19,44 Neurons that regenerate their axons can be labeled by applying dye(s) to the site of the regenerated axons, where the dyes are endocytosed and transported back to the cell bodies. This technique thus counts the number of motor and sensory neurons that regenerate their axons, rather than counting axons. Because several axon sprouts are emitted from parent axons, axon counts as just described do not necessarily reflect the numbers of neurons that regenerate axons. The limitation of the blow-through effect holds in some instances for retrograde labeling models, which in crush and acute cut-repair models given sufficient time yield results that match pre-injury levels.⁴⁴ Importantly, models of chronic denervation or a critical nerve gap help to mitigate the blow-through effect, as measuring the quantity of neurons regenerating axons at later timepoints can be effectively modeled without equilibration to controls.^45,46

Downstream of axon regeneration, histological evaluation of reinnervation is valuable data to improve interpretation of functional recovery. At a basic level, measurement of relative muscle weight in affected muscles provides a useful metric of recovery that is directly correlated with muscle force production.¹⁹ Additionally, muscle recovery vs atrophy can be characterized by weight but also muscle fiber cross-sectional area. Histologic analysis of the neuromuscular junction (NMJ) can further characterize reinnervation and recovery. However, visualization and counts of endplates with motor nerves have limitations akin to myelinated axon counts (i.e. sprouting). Each regenerated axon can sprout to encompass and reinnervate motor units by as much as 3–5-fold.⁴⁷ Skin reinnervation via measuring axon density is much less common, but fraught with interpretive challenges, as collateral sprouting from adjacent nerve(s) can confound evaluation.

Additionally, there has been recent interest in the terminal Schwann cell and its role in plasticity in nerve injury and repair.⁴⁸ In a homeostatic state, these cells contribute to development, maintenance and remodeling of the NMJ. In a regenerative state, however, they extend cytoplasmic processes to allow for axonal regeneration into end targets. Investigating not only nerve and muscle, but supporting cells improves understanding of regenerative processes and could yield future biomarkers.

Outcome Metrics and Endpoints: Functional Recovery

Although histological data in the rodent can be well-representative of humans, functional recovery in rodents can be a more powerful assessment tool, as multiple assays can be monitored non-invasively and serially. Electrophysiology (compound neural and muscle action potentials), muscle contractile forces (ITFT and sGST), and behavioral assessments (i.e. walking track and gait analysis) encompass functional recovery. Functional recovery is arguably the most important outcome measure and should be considered the primary outcome measure when the goal of the study is to develop a novel therapy to improve overall nerve regeneration.

Measurement of compound muscle action potentials reflect and may parallel recordings of muscle force in reinnervated muscles. A major strength of either measurement is their consistency and robustness (reproducibility) compared to behavioral assessment, as both avoid capturing cortical activation, coordination of motion, and other factors involved in generating a behavior that is measured. While isometric tetanic force testing (ITFT) requires evaluation at a terminal endpoint, use of compound muscle action potential or the more recently developed stimulated grip strength test (sGST) can be performed serially. A novel stimulated grip strength test (sGST) using the rodent median nerve has been shown to be more reliable than previous upper extremity grip strength testing and limits some problems inherent in the rodent hind limb model, such as the complex circuitry involved in gait, replacing it instead with stimulated muscle contractions in vivo.⁴⁰ The use of either metric allows for serial measurements allowing for better characterization of muscle force (and thereby reinnervation) over time. However, caution must be exercised in using muscle recordings, such as in a sciatic nerve model, as the muscle chosen can influence the measurement. Misdirection of regenerating axons can lead to random reinnervation of end targets, such as a greater degree regeneration into the downstream tibial compared to the common peroneal.^38,39 Furthermore, in any nerve injury model, regenerated axons can reinnervate as many as 4–5 times the normal number of muscle fibers to compensate for reduced numbers of axons that succeed in reinnervating the denervated muscle.⁴⁷

Similar to humans, behavioral functional assessments, such as walking track measurement or grid walking analysis, do not normalize with time in rodent models.⁴⁹ Increased functional demands, distance from end target, and complexity of sensory and motor topography all contribute to varying degrees of functional recovery in each nerve. Subsequently, behavioral assessments tend to be inherently more variable compared to electrophysiology or muscle force measurements. However, a strength of behavioral assessment is the similarity it shares to clinical evaluation of outcome. We will highlight walking track measurements in our examples, and refer the reader to an extensive recent review on behavioral assessments in rodent nerve injury models.⁵⁰

Across injury models, the sciatic nerve model remains an adequate model for measuring behavioral recovery. Similar to humans, the distance to target and the complexity of the sensory and motor topography of the rodent nerve predicts recovery. For example, even at 1 year based on walking track measurements, a sciatic nerve cut-repair recovers little function (41% of the pre-injury value), whereas tibial nerve recovers some function (54%) and the common peroneal nerve, with the simple requirement of dorsiflexion recovers, almost normal function (100%) (Fig. 2B). With such a low baseline level of functional recovery²², the complete sciatic nerve transection model provides a paradigm in which functional recovery can be assessed and compared between experimental groups only if a therapeutic offers significant benefit. For this reason, the common peroneal and tibial nerve models offer more sensitivity to assess recovery, as gait is less impaired and strategies that would enhance nerve regeneration to some degree (i.e. Tacrolimus (FK506), Electrical Stimulation; see next section) could be assessed. By contrast, if there was ever a “miracle” therapeutic (i.e. theoretically PEG fusion), then functional measurements with a sciatic nerve transection remains a dependable model.^51,52

The combination of histology and functional measures from rodent data further provides a regenerative profile after nerve injury and repair. It is possible for small differences between potential treatments to be highlighted by one modality, without being detected by other modalities. For example, in a recent article highlighting the use of Tacrolimus (FK506) and Electrical Stimulation in the same animal to promote regeneration, two behavioral measures, terminal muscle force analysis (ITFT), muscle weight, and histomorphometric analyses of myelinated axons in nerve were all used in conjunction to draw conclusions about differences in regeneration.⁵³ In this case, muscle force values and histomorphometric differences between experimental groups were not significantly different, while behavioral outcomes highlighted the difference in effect of each experimental group. If only one outcome measure were chosen in this case, important aspects of the regenerative profile would have been left unseen.

Examples of translation to the clinical realm

Tacrolimus (FK506)

A clinical therapy now used in nerve surgery credited to basic science experiments in rodents is the use of Tacrolimus (FK506) to promote nerve regeneration after injury. FK506 is a strong immunosuppressant that is currently approved for anti-rejection therapy after tissue allotransplantation such as kidney or liver, as well as tissue in the face, hand, abdominal wall, penis, and other areas where reinnervation is critical.⁵⁴ However, FK506 has also been shown to increase the rate of axon regeneration. These capabilities were first demonstrated in a rodent sciatic nerve model,^55,56 while later experiments revealed mechanism. Importantly, our group has modelled nerve injuries from neurapraxia to axonotmesis to neurotmetic injuries in the rodent and have used FK506 is all those models and demonstrated the ability to enhance recovery. FK506 can bind to immunophilin receptors on injured neurons, leading to upregulation of heat-shock proteins and, likewise, upregulation of growth-associated proteins (GAPs) leading to a regenerative phenotype (Fig. 3A).^57,58 It has also been found that FK506 given before nerve injury can “prime” this regenerative phenotype.⁵⁹ In a clinical setting, especially within nerve transfer surgery where many surgeries are elective in nature, pre-surgical dosing of FK506 could facilitate better outcomes. Further, even FK506 given at sub-immunosuppressive doses can promote accelerated axon growth, mitigating some negative systemic effects of immunosuppression.^60,61

Fig. 3.

A) Schematic of Tacrolimus (FK506) versus electrical stimulation (E-stim) effects on neurons promoting nerve regeneration. FK506 (left) signaling upregulation of heat shock proteins and growth associated proteins (GAPs). E-stim (right) upregulates neurotrophic factors including BDNF, as well as regeneration associated genes (RAGs). Adapted with permission from Marsh et al. 2023. B) Functional recovery, as measured via Tibial Functional Index (TFI), for rats which had undergone tibial nerve injury and repair with E-stim (ES), FK506, or repair alone. Both ES and FK506 had increased TFI values compared to repair alone by day 25. C) ,^: functional nadir for FK506 and FK506+E-stim; v: functional nadir for E-stim and repair alone. Recovery compared with functional nadir for E-stim (a), FK506 (b), FK506+E-stim (c), and repair alone (d). Figures and data for B reproduced with permission from Jo et al. 2019. Figures and data for C reproduced with permission from Marsh et al. 2023.

After the robust validation of the capabilities of FK506 across rodent and then large animal studies, FK506 was investigated for its potential as a neuroregenerative immunosuppressant in a human nerve allograft.⁶² As in animals, FK506 administration after nerve repair with allograft showed increased regeneration and earlier return of function.⁶³ However, despite its success, a major issue with systemic administration of FK506 is the potential side effects, including central nervous system toxicity and nephrotoxicity.⁶⁴ The possibility of serious side effects makes systemic administration of FK506 unappealing and marks it as a boundary to overcome in the future via local methods of administration. However, this work with FK506 in nerve injury paved the way for its use with vascularized composite allograft (VCA) where the benefits of improved function merits its use. Recent work has also shown success in restricting FK506 administration to a particular nerve injury site in rodents.⁶⁵ Locally administered FK506 via intra-operative fibrin gel with a delivery system containing the FK506 produced a marked increase in regeneration of sensory and motor axons in a rat sciatic nerve cut-and-repair model.⁶⁶ Further, it was shown that a fibrin gel-based method of delivery greatly reduced bio-distribution of FK506, circumventing the issue of systemic immunosuppression and toxicity.⁶⁶

Therapeutic Electrical Stimulation

Another advancement is the development of intraoperative, therapeutic, electrical stimulation (E-stim) protocols, as a means to promote axon regeneration after nerve injuries.⁶⁷ First using rodent models, E-stim of surgically-repaired nerve was shown to promote a large portion of axons, rather than pioneer axons first and others later, to grow across a repair site in a rodent nerve cut-repair model.⁴⁴ Further studies corroborated these results and isolated key variables, low-frequency current (20Hz) and time (1 hour), for E-stim application to treat repaired nerve injuries.⁶⁸ A mechanism generally accepted for the effects of E-stim is an upregulation of neurotrophic factors and their receptors, as well as other regeneration associated genes (RAGs), within neurons in order to promote earlier axon growth at the injury site (Fig. 3A).^44,69 Rodent research also revealed the potential for E-stim to enable guidance of regenerating axons to appropriate targets, whether motor or sensory end-organs, which also accelerates recovery of function.⁴⁴ Further, the use of E-stim has been compared to FK506 in cut-repair models, where FK506 served as a positive control given its previous validation as a neuro-therapeutic. In these studies, both E-Stim and Tacrolimus improved nerve regeneration compared to controls (Fig. 3B&C).⁷⁰ These findings are especially intriguing given that FK506 promotes a generally similar nerve regeneration phenotype when compared to E-stim,⁷⁰ yet they share different mechanisms of action. When FK506 and E-stim were administered in the same animal using the same nerve injury model, overt synergy was not observed, but variance in behavioral outcomes was reduced. Translationally, the use of multiple therapeutics targeting different pathways could be helpful in order to better predict clinical recovery after injury.⁵³

In the first randomized controlled clinical trial of intraoperative E-stim in the context of severe carpal tunnel syndrome, E-stim coupled with carpal tunnel release surgery accelerated axonal outgrowth and target reinnervation in patients compared to those who received surgery alone.⁷¹ A 20Hz, 1-hour administration of E-stim was used, the same as in rodents, and accelerated recovery in these patients, including reinnervation of thenar muscles after 12 months compared with no recovery in the control group by this time frame.⁷¹ The use of E-stim has further been highlighted in recent clinical trials involving different sites of injury (Ulnar nerve, spinal accessory nerve, median nerve) and different modalities (motor and sensory), which have all concluded that patients receiving E-stim have better recovery than controls.^72–74 It is important to recognize the success in translation from rodents to humans of this procedure.

Nerve Gap Repair: Nerve Autograft Alternatives

Rodent models have predicted challenges in repairing large and long nerve gap injuries using autograft alternatives, such as conduits and ANAs. Problems first identified in rodent nerve models, such as the concept of a critical nerve gap, have provided insight for the clinic. During evaluation of nerve gap repair, rodent models revealed that nerve graft alternatives were more prone to reduced axon regeneration across the repaired gap based upon gap length than autografts (Fig. 4A)³⁵. This regenerative limitation was first found in nerve guidance conduits, which have been mirrored in clinical settings.⁷⁵ Furthermore, when ANA was used to repair a sciatic nerve gap in rats, axon regeneration across the ANA was successful until ANA length exceeded 3 cm (Fig. 4A).³⁵ Conversely, nerve isografts (an animal model of nerve autograft) successfully promoted axon regeneration across nerve gaps as long as 6 cm.^35,76 Overall, it was not the rodents’ inability to regenerate axons across long nerve gaps.⁷⁷ Instead, the lack of axon regeneration across longer gaps was due to the grafting material used to repair the nerve. The failure to regenerate across longer ANAs (>3 cm) were attributed to an increased accumulation of senescent SCs, slower and delayed angiogenesis, and disrupted inflammatory signaling within the graft resulting in an environment that limited axon regeneration across the ANA.^76–82 (Fig. 4B&C) Thus, the unmet clinical need for therapies to repair long nerve gaps are indeed accurately reflected using a rat nerve gap injury model,³¹ also showing caution should be exercised when translating new technologies to humans. While rodents are useful in providing a model to create and test alternatives to current problems that must be overcome in the field, they also provide utility in uncovering limitations in potential alternatives.

Fig. 4.

A) Axon regeneration across isografts or acellular nerve allografts repairing a sciatic nerve gap using Thy1-GFP rats. When graft length was increased to 4 cm, only the isograft was able to support axonal regrowth (GFP+ axons) across the graft. At 6 cm, even the isograft showed less robust regeneration than at shorter graft lengths. Images and data reproduced with permission from Saheb-Al-Zamani et al. 2013. B-C) Schematics of the cellular differences within short (B) and long (C) ANAs. C) Long ANAs were repopulated with cells expressing markers of senescence, changes to T cells and other inflammatory cells, and differences in blood vessels compared to short ANAs. Adapted with permission from Pan et al. 2019.

Summary

Experimentation on novel interventions necessitates the use of animal models, especially rodent. Rodents provide a model that is relatively low cost, as adequate animal numbers can be tested using specific injury models and standardized outcome measures with regenerative profiles sharing similarity to human outcomes. It would be challenging to progress the field of nerve surgery without the use of rodents before larger animals to show proof-of-principle results. However, the therapeutic interventions (i.e. FK506) described have limited data demonstrating validity in large animal models and have not been directly validated in randomized-controlled clinical trials. These clinical trials are challenging to design given the heterogeneity in nerve injury patients, as well as the variability that can result from clinical outcome measures. Therefore, there is increasingly high value in confirming rat model findings in larger animal models prior to clinical testing or translation. For example, large animals can address the vast difference in size and regenerative distance moving from rodent to human. Confirmation of rodent results in a large animal model would greatly improve the robustness and validity of therapeutic translation given the limits of randomized-controlled clinical trials in nerve injury patient population.

Key Points.

• Evaluating novel interventions necessitates the use of rodent models to provide adequate testing to establish proof-of-principle results.

• There are more neurobiological similarities between rodents and humans than differences.

• Interpretation of rodent data requires consideration of experimental design and injury model and rational outcome metrics and endpoints.

Synopsis.

• This review highlights the use of rodents as preclinical models to evaluate the management of nerve injuries, describing the pitfalls and value from rodent nerve injury and regeneration outcomes, as well as treatments derived from these rodent models. The anatomic structure, size, and cellular and molecular differences and similarities between rodent and human nerves is summarized. Specific examples of success and failure when assessing outcome metrics are presented for context. Evidence for translation to clinical practice include the topics of electrical stimulation, Tacrolimus (FK506), and acellular nerve allografts (ANAs).