SCs play a central role in the early stages of WD by initiating repair responses that support regeneration. Myelinating SCs, which typically form the insulating myelin sheath, rapidly convert to a repair phenotype after nerve injury. This repair phenotype involves downregulation of myelin-associated genes, upregulation of regeneration-associated transcription factors, secretion of neurotrophic factors and chemokines, expression of innate immune receptors, and participation in myelin clearance and antigen presentation (56, 57). Because of early secretion of cytokines and chemokines, SCs recruit and activate macrophages, in turn triggering the cascade of degeneration-associated inflammation that is also essential for subsequent regeneration (58). Non-myelinating (Remak) SCs ensheathe small-diameter axons and exhibit distinct transcriptional programs compared with myelinating SCs. Remak SCs provide metabolic support to nociceptive and autonomic fibers, upregulate immune-related genes, and importantly can adopt repair-like states after injury that promote chemokine production and cross-talk with macrophages (59). Collectively, SCs orchestrate several important functions during the immediate response to nerve injury and subsequent regeneration. Although not the primary focus of this review, these functions have been discussed in greater detail elsewhere (60–63).

Although perineurial glia have been less extensively studied than SCs or macrophages, live-imaging studies in zebrafish models have demonstrated that they contribute to the resident cellular response after nerve transection injury (38, 64, 65). Perineurial glia facilitate two cardinal functions after nerve injury: (1) phagocytosis of debris and (2) formation of cellular bridges across injury gaps. Lewis and Kucenas observed that perineurial glia rapidly extend processes toward damaged sites and phagocytose debris immediately after laser-induced nerve transection injury (64). In addition, time-lapse imaging revealed that perineurial glia sequentially localize to the proximal and then distal degenerating stumps around the injury gap, whereas macrophages appeared to home to the injury zone specifically. The second finding identified that perineurial glia traverse the injury gap to form scaffold-like structures in advance of SCs and axonal regeneration (64). Notably, when the size of the gap is larger, perineurial bridge formation fails, and correspondingly, a lack of axonal regeneration is observed. In subsequent experiments, genetic inhibition of perineurial glia development resulted in an absence of bridge formation, with nearly half of nerves failing to regenerate axons after injury. Further mechanistic perturbations demonstrated that this bridging phenomenon is mediated by TGF-β signaling, with SCs acting synergistically with perineurial glia to facilitate bridge formation (65). A role for perineurial glia that is similar to regeneration has also been observed in a rodent models; after a nerve gap injury, perineurial glia migrate into conduits by 7 days after injury, followed by blood vessels, SCs, and axons in a collaborative effort to rebuild the nerve microenvironment (66). However, because these observations were based primarily on electron microscopy studies, further molecular characterization of perineurial glia is needed to confirm their identity and function.

To understand macrophage contributions to nerve injury and repair it is essential to first characterize the resident populations present during homeostasis. Although resident macrophages have previously been identified in the PNS (26, 85, 86), their developmental origins, homeostatic turnover, and function have remained largely unexplored.

To determine the developmental origins of PNS resident macrophages, lineage tracing studies using Cx3cr1CreERT2:Rosa26-YFP mice tagged at embryonic day 16.5 revealed that approximately 60% of sciatic nerve macrophages (snMacs) at postnatal day 4 were embryonically derived (71). However, by postnatal day 42, only half the resident macrophage population were positively labeled, suggesting a postnatal window of turnover. To test this theory, Ydens and colleagues used S100a4Cre:Rosa26-YFP mice to label hematopoietic stem cell–derived monocytes and found that snMac labeling doubled between 2 and 6 weeks postnatally, indicating that snMacs are established embryonically but are gradually replaced by hematopoietic stem cell-derived monocytes over the course of development. As in the adult CNS, resident macrophage turnover under homeostatic conditions was limited, with resident macrophage populations observed to be maintained through at least 36 weeks of age (71). Therefore, unlike CNS microglia or CNS-associated macrophages, resident PNS macrophages do not appear to be solely yolk sac-derived (83).

Expanding on studies of developmental origins and turnover dynamics, advances in single-cell technologies have revealed spatial and molecular heterogeneity within the resident macrophage populations. Notably, single-cell RNA sequencing (scRNA-seq) has permitted the identification of two distinct macrophage subsets: epineurial macrophages (Relmα⁺Mgl1⁺) and endoneurial macrophages (Relmα⁻Mgl1⁻) (71). Endoneurial macrophages share genetic signatures more closely aligned with microglia than with epineurial macrophages, suggesting that proximity to axons and SCs promotes a microglia-like phenotype. This initial distinction of two resident macrophage populations has been corroborated by additional single-cell sequencing studies (87, 88). Furthermore, these findings in the PNS parallel observations in other tissues, such as lung and brain, where local environmental niches appear to instruct macrophage identity and fate specification (89, 90).

Despite these findings, significant challenges remain in studying these resident macrophage populations. Notably, because identification currently relies on a combination of markers together with anatomical context, and because gene expression profiles and canonical markers are dynamically regulated over the course of injury and recovery, it remains challenging to definitively link specific macrophage populations to distinct injury-related functions. Furthermore, the cues that drive the fate specification of these distinct subsets, how these populations are maintained, and their functional roles during homeostasis remain poorly understood.

Early studies in the 1980s questioned whether recruited immune cells contribute to WD. Beuche and Friede addressed this question by placing injured sciatic or phrenic nerves into Millipore diffusion chambers with varying pore sizes to restrict cellular infiltration before implanting them into the peritoneal cavity for 8 weeks (91). In chambers that prevented leukocyte entry, no increase in cell density was observed and myelin was largely maintained. By contrast, in chambers permitting leukocyte infiltration, cell increases were detected by 3 days post-transplantation, with many cells appearing “myelinophagic” by 1 week. These explanted nerves recapitulated features of WD observed in vivo, with marked myelin reduction by 4 weeks after insertion (91). Follow-up studies using silica injections to physically block cellular infiltration aimed to characterize the cellular identity of the infiltrating cells and test whether these cells precipitate WD (86). Blocking cell entry significantly reduced both cell density and myelin debris clearance immediately after peripheral nerve injury. Furthermore, MAC-1 (CD11b) and F4/80 staining identified the infiltrating cells as myeloid cells and macrophages, and subsequent studies demonstrated that recruited monocytes differentiate into macrophages upon entering the PNS (69, 70, 85, 92, 93).

These experiments demonstrated that recruited immune cells substantially contribute to WD and were later expanded upon with technological advances. Notably, scRNA-seq identified a substantial macrophage population at 1 day after crush injury that was absent in naïve nerve tissue, suggesting robust early recruitment after injury (71). This finding was corroborated using a fate-mapping approach in which irradiated mice (with the surgical limb shielded) were transplanted with CD45.1 bone marrow. Resident macrophages were still detectable in the nerve, yet by 1 day after injury a large proportion of CD45.1⁺ Ly6C⁺ monocytes had infiltrated the endoneurium (71). Subsequently, these cells downregulated Ly6C and adopted macrophage phenotypes. Interestingly, similar to observations in the CNS (94), there appears to be potential for monocytes to repopulate the peripheral nerve after injury and contribute to homeostatic resident macrophage population thereafter (71); however, the cues that permit monocyte imprinting into the peripheral niche remain unexplored.

Deconstructing the roles of infiltrating vs. resident macrophages remains challenging because of limited lineage-specific markers. Once recruited to the injured nerve, monocytes downregulate markers such as CCR2 and Ly6C (bone marrow–derived monocyte markers) and acquire macrophage signatures, which complicates efforts to distinguish resident from peripherally derived macrophages during the distinct phases of nerve regeneration (71). In addition, studies that broadly deplete macrophages to assess impact on WD may present confounding results because of simultaneous impact on resident, recruited, and CNS macrophage populations. Although research efforts have elaborated important insights into the contributions of recruited macrophages to injury and repair, the distinct contributions of specific subtypes (e.g., recruited vs. resident) are likely more nuanced than is currently appreciated.

Macrophages in the peripheral nerve demonstrate robust phenotypic diversity across the course of nerve injury and regeneration. Macrophages have historically been categorized as predominantly M1 (pro-inflammatory) and M2 (anti-inflammatory); however, with improved sequencing technologies, it is clear that this binary paradigm is oversimplified. Macrophages instead exhibit marked heterogeneity and embody a spectrum of activation states that dynamically shift across physiologic conditions (28, 95).

The M1/M2 profile was originally defined from in vitro studies in which the M1 macrophage phenotype was described by stimulation with interferon gamma (IFN-γ)/lipopolysaccharide, resulting in expression of inducible nitic oxide synthase and TNF-α, also termed the “classical response”; the M2 phenotype was described by stimulation with IL-4/IL-13, resulting in expression of Arg1/CD206, also termed the “alternative response” (95). Despite this binary characterization, in vivo studies that assessed macrophage polarization demonstrated a mixed inflammatory response rather than clear temporal pattern in M1/M2 dichotomy (96). Interestingly, appreciable changes in M1 markers such as inducible nitic oxide synthase or IFN-γ were not observed in nerves for the time points profiled within the first 2 weeks of injury. By contrast, M2 markers such as Arg1, TREM2, and Ym1 increased, with each exhibiting a distinct temporal expression profile during the first 2 weeks after injury (96). Meanwhile, another study examining the temporal dynamics of macrophage polarization during the early injury phase (days 3–5) identified IFN-γ, a proinflammatory (M1-like) cytokine, as a key upstream regulator associated with differentially expressed genes such as Nos2, which were highly expressed at day 5 yet declined over time (97). Cells expressing Arg1 (an M2-like, anti-inflammatory marker) were also identified in which expression decreased from days 5 to 14 but showed a rebound by day 28 (97). Collectively, these initial studies demonstrate the multidimensional complexity of the inflammatory response to nerve injury. Although we maintain M1-like and M2-like terminology for consistency, this strict dichotomy likely does not appreciably represent the plasticity and dynamic states of macrophages (98).

Single-cell sequencing has greatly expanded our understanding of macrophage diversity during homeostasis, injury, and repair states. Notably, it has revealed that multiple macrophage populations coexpress M1/M2 markers across clusters and injury timepoints (71), further reinforcing that this strict dichotomy fails to capture the diverse transcriptional and functional states of macrophages after injury. For example, one study examining nerve crush injury at 1, 3, and 5 days after injury identified about 5 macrophage populations, with the largest attributed to monocyte-derived macrophages expressing Ly6c and Ccr2 (88). Notably, one-to-one matching of clusters across timepoints could not readily be distinguished, underscoring the dynamic macrophage plasticity in response to injury. Furthermore, as described above, multiple groups have now distinguished two resident populations within the nerve: endoneurial-associated macrophages and epineurial-associated macrophages (71, 87, 88, 99). Further examination of each population in response to injury revealed that epineurial macrophages were largely unresponsive to injury, whereas endoneurial macrophages upregulated multiple chemoattractants (Ccl6, Ccl7, Ccl8, Ccl12) between days 1 and 5 after injury (71). To further explore transcriptional diversity as a function of spatial location, Zhao et al. performed scRNA-seq comparing injured vs. distal nerve segments at 3 DPI and identified an Arg1⁺ cluster enriched at the crush injury site, whereas a CD38⁺ population was largely restricted to the distal nerve (88). Complementing this finding, Kalinski et al. employed an Arg1-YFP mouse reporter line and further demonstrated that Arg1-YFP⁺ cells were localized to the crush injury site at 7 DPI but were completely absent from the distal stump, highlighting spatial restriction of specific macrophage subsets (80). Collectively, these studies provide early insight into the spatial organization of macrophage functional states during homeostasis, nerve injury, and recovery.

After the identification that WD is mediated by recruited myeloid cells, critical questions emerged regarding the cues contributing to recruitment and sustained cellular activation. This initial inflammatory response largely involves danger-associated molecular pattern (DAMP) signaling and complement activation (Figure 2).

These respective mechanisms were elaborated by in vitro coculturing of nerves with macrophages in the presence of antibodies against complement receptor 3 or C3-deficient serum. Notably, both of these conditions prevented myelin opsonization and phagocytosis (49). In vivo complement depletion using cobra venom factor further corroborated these findings; cobra venom factor–treated animals demonstrated significantly lower macrophage numbers, reduced phagocytic activity, and increased myelin retention at 7 DPI (50). Similarly, mice deficient in TLR2, TLR4, or MYD88 had fewer CD68+ macrophages at 7 DPI with correspondingly increased myelin load, demonstrating that DAMP signaling largely drives macrophage recruitment and efferocytosis (47, 49, 50).

Although MCP-1 (CCL2) had long been implicated as a potential macrophage recruitment cue, its role had not been directly tested until work by Siebert et al. (100). In this study, both CCR2 and CCR5 knockout animals were examined to explore the consequences on macrophage recruitment and the sequalae of WD. Notably, CCR2 knockout mice demonstrated lower recruitment and delayed myelin clearance compared with CCR5 knockouts (100). Next, to test the necessity of CCL2 for macrophage recruitment, Siebert and colleagues examined both global and inducible CCL2 knockouts. Surprisingly, both genotypes demonstrated normal macrophage accumulation, foamy macrophage morphology, and effective myelin clearance, suggesting that CCL2 is not uniquely required for recruitment. As CCR2 binds multiple ligands, including CCL7 and CCL12, which were upregulated in the absence of CCL2, these chemokines likely compensate as alternative recruitment cues. Collectively, DAMPs, complement, and chemokines are critical mediators establishing the acute inflammatory microenvironment that drives monocyte/macrophage recruitment during WD (Figure 2). However, further investigation is warranted to elucidate the redundancy and specificity of compensatory ligands and receptors.

Although macrophages were demonstrated to have a central role in WD by phagocytosing myelin and axonal debris, it was not clear whether macrophages also participated in promoting regeneration. Comparative studies between the PNS and CNS (optic nerve) revealed lower macrophage recruitment, scarce evidence of foamy macrophages, and delayed myelin clearance in the CNS, suggesting a fundamental difference in regenerative processes (69). To directly test the necessity of macrophages to PNS regeneration, studies using liposome-mediated macrophage depletion demonstrated reduced macrophage numbers corresponding to delayed myelin degradation (70). Whole-body irradiation experiments to deplete bone marrow myeloid precursors yielded similar results, with reduced myeloid cell recruitment and impaired myelin clearance into the nerve after injury, although potential off-target effects should be noted as a confounding factor (101). An even more targeted approach using CD11b-TK mt-30 mice and ganciclovir administration demonstrated that early myeloid cell ablation severely impaired myelin clearance, neurotrophin synthesis, and vasculogenesis at the injury site (26). Furthermore, by then terminating ganciclovir administration at day 7, there was partial recovery over time, with functional recovery achieved by day 49. Notably, even without early WD, reintroduction of myeloid cells partially restored regenerative processes, although distal myelin debris remained elevated (26).

Macrophages also play a central role in orchestrating the post-injury microenvironment through the secretion of cytokines, chemokines, and trophic factors (Figure 2). Pro-inflammatory cytokines such as TNF, IL-1, and IL-1β recruit additional immune cells and enhance phagocytic function. These cytokines act in an autocrine manner to induce increased macrophage phagocytosis (43) and can instigate neurotrophin production such as nerve growth factor (NGF) to exert positive regenerative effects (102). Furthermore, over the course of WD, there is increased production of anti-inflammatory cytokines (IL-10, IL-6, granulocyte macrophage colony-stimulating factor), with M2-like macrophages serving as a key source of IL-10 and IL-6 (34).

Profiling peripheral nerves after injury has also demonstrated the upregulation of many neurotrophic factors known to promote axonal growth and neuronal survival in both PNS and CNS (Figure 2). Preliminary work demonstrated that addition of neurotrophins to sciatic nerve after injury potentiates axon regeneration, myelination, and functional recovery (103). The cellular origins of NGF were then elaborated by experiments demonstrating that NGF mRNA levels in cultured injured sciatic nerve decreased over time but were rescued with the addition of activated macrophages (103). Furthermore, Barrette et al. (26) examined how depleting CD11b cells impacts neurotrophin synthesis and found that these factors were negligibly detected in the absence of CD11b+ (myeloid) cells. However, this does not rule out the possibility that myeloid cells may communicate with SCs for neurotrophin secretion rather than directly secreting these factors themselves. For example, one study identified that macrophage-derived IL-1β leads to upregulation of NGF from SCs (102). Collectively, macrophages actively contribute to sculpting the post-injury and recovery microenvironment through their secretory functions (26, 103).

Macrophages also actively influence BNB permeability and promote revascularization during nerve regeneration (Figure 2) (104). Assessment of BNB integrity after compression injury revealed greater permeability by 2 weeks, followed by progressive disintegration (105). Depletion of macrophages before injury largely prevented these changes, indicating that macrophages contribute to BNB disruption, although the precise mechanisms remain unknown.

Macrophages also orchestrate post-injury angiogenesis (Figure 2). Early studies observed that axonal regeneration failed in the absence of macrophages, and myeloid cell depletion also reduced blood vessel formation in the distal stump (26). Later work by Cattin et al. (27) demonstrated that revascularization between proximal and distal stumps begins as early as 3 days after injury, forming a bridge across the gap. Co-staining with the SC marker S100 revealed that SCs closely associate with newly forming vessels, migrating along them to extend across the bridge. To further interrogate the mechanisms and cellular species orchestrating angiogenesis, the authors leveraged the knowledge that hypoxia promotes vascular endothelial growth factor (VEGF) secretion, which in turn promotes angiogenesis (27). By using a hypoxia indicator and co-staining with IBA1 to label macrophages, they identified that these cells can sense hypoxic regions, suggesting they secrete VEGF to guide endothelial cells across the bridge. To test this hypothesis, VEGF was selectively depleted in macrophages. Although macrophage numbers at 5 DPI were unchanged, vascularization of the bridge was severely impaired, accompanied by reduced SC migration into the bridge (27). Rescue experiments were also performed in which injection of VEGF into the bridge of knockout animals restored vasculogenesis and enhanced SC migration into the bridge along the newly formed vasculature.

Although inflammation is critical to initiate the early injury sequela and promote WD after peripheral nerve injury, persistent inflammation drives pathophysiologic consequences and promotes neurodegeneration (Figure 1). Notably, macrophages participate in inflammatory resolution by efferocytosis (Figure 2) (106, 107). This permits removal of inflammatory molecules such as DAMPs and cytokines from the local environment, transitioning the pro-inflammatory response toward a more anti-inflammatory niche (33, 108). The role of efferocytosis in facilitating WD and favorable regeneration has emerged as an important component of the regenerative process. After nerve injury, macrophages localized to the injury site demonstrate higher expression of efferocytosis-associated genes including MERTK and Gas6 and apoptotic corpse clearance receptors such as TREM2 (80). Notably, a recent study demonstrated that the proportion of macrophages expressing MERTK decreased over time after injury. By using a myeloid cell–specific knockout of MERTK, Pandey et al. observed a reduction of efferocytosis coupled with exacerbated neuropathologic outcomes, suggesting that impaired efferocytosis negatively impacts regenerative potential (109). Furthermore, the type of cell death also influences inflammatory resolution. In sciatic nerve transection, a dysregulated form of macrophage cell death known as pyroptosis skews the cytokine profile toward a more pro-inflammatory signature. By employing gasdermin D knockout mice, a key pyroptosis regulator, this proinflammatory signature was reduced, resulting in improved axonal regeneration and functional outcomes (107). This suggests that when apoptotic cell death is supplanted by pyroptosis, this sustained inflammation adversely impacts the regenerative environment.

Given that the type of cell death impacts inflammatory resolution, it is critical to understand macrophage fate and egress upon WD termination. If the large influx of myeloid cells after injury fails to exit, they may contribute to persistent inflammation and skew regeneration in a pathophysiologic direction. However, the mechanisms underpinning macrophage egress after injury have yielded varied results across injury models (33, 104). Early studies demonstrated that approximately 2%–4% of macrophages undergo apoptosis about 15 days after transection, coinciding with the termination of WD (110, 111). Another study using a transection model and grafting nerves between male and female mice (and vice versa) to distinguish recruited from endogenous monocytes/macrophages revealed that recruited macrophages egress from the graft into perineurial blood vessels, draining lymph nodes, and spleen (110), suggesting most recruited macrophages leave the nerve as inflammation resolves. However, whether the remaining macrophages undergo a different form of cell death remains uninvestigated. Critically, it remains unknown whether egress is maintained in pathophysiologic contexts or if macrophages become trapped and die in failed regenerative environments (33, 104). Therefore, therapeutic interventions promoting efferocytosis and inflammatory resolution present as a viable strategy to promote successful nerve regeneration.

Anti-inflammatory treatments focus on reducing the acute and chronic inflammatory milieu that impairs axonal regrowth in the injured sciatic nerve (Figure 3). For example, FK506 reduces levels of pro-inflammatory cytokines such as IL-1β and TNF-α, modulates macrophage phenotype from pro-inflammatory to anti-inflammatory, and inhibits ECM remodeling enzymes like MMP13 (136). These effects collectively decrease nerve edema and macrophage infiltration, thereby fostering a more permissive microenvironment for regeneration. Other anti-inflammatory agents, including nonsteroidal anti-inflammatory drugs and minocycline, also demonstrate efficacy in reducing macrophage-mediated neurotoxicity and scar formation, although their specific effects in peripheral nerve injury models require further detailed studies (137, 138).

Cytokines and growth factors serve as critical mediators in the neuroimmune crosstalk during peripheral nerve regeneration (Figure 3). Therapeutic administration of cytokines such as IL-33 and IL-4 promotes the polarization of macrophages toward a M2-like phenotype, characterized by secretion of anti-inflammatory cytokines like IL-10 and enhanced production of neurotrophic factors including NGF, brain-derived neurotrophic factor (BDNF), and VEGF (136). This shift attenuates inflammation while supporting SC activity and axonal regrowth. Growth factors such as TGF-β1 further regulate immune responses via Smad-dependent signaling and contribute to ECM remodeling, myelination, and reduction of neuropathic pain (139).

Decellularized xenogeneic hydrogels have demonstrated immunomodulatory effects across multiple injury models. In a 15-mm nerve gap transection model with silicon conduits filled with decellularized xenogeneic hydrogels, macrophages were found to dominate the graft margins and migrate further into the conduit compared with controls without hydrogel (155). Although spatial distribution patterns of each were not fully described, there also appeared to be greater recruitment of M2 macrophages in the hydrogel-infused conduit, resulting in a greater M2:M1 ratio. In a shorter gap model at 5 days after injury, the hydrogel treatment showed decreased expression of genes associated with inflammation and scarring, such as collagens and fibroblast growth factor receptor, whereas genes associated with an anti-inflammatory profile were increased compared with conduit-only controls (156). At 10 days after injury, the hydrogel treatment group also demonstrated a decrease in M1-like macrophages but no observable differences in M2-like macrophages, resulting in a higher M2:M1 ratio. Furthermore, conduits with hydrogel matrix showed significantly more neuronal cell bodies labeled with retrograde tracer, demonstrating increased axonal crossing associated with the higher M2:M1 ratio (156). Finally, in a nerve crush-injured model, injection of a xenogeneic decellularized nerve matrix reduced M1-like macrophages at both 3 and 7 days after injury, corroborated by decreased pro-inflammatory cytokine expression (TNF-α and IL-1β) (154). Conversely, more M2-like macrophages and anti-inflammatory cytokines (TGF-β and IL-10) were observed in the hydrogel-treated condition. Western blot revealed higher expression of TLR4/MyD88/NF-κB axis proteins in the untreated group, suggesting mechanistic involvement of this signaling pathway in the anti-inflammatory effects.

Although decellularized xenograft hydrogels are the most common strategy employed, hydrogels with alternative compositions have been developed (Figure 3). For example, silk fibroin derived from silkworm cocoons offers low immunogenicity, presenting as a potential biocompatible and natural biomaterial for hydrogel composition. By enriching hydrogels with Mg²⁺ and silk fibroin, an increased M2:M1 ratio was observed compared with untreated NH4 commercial chitosan conduit groups at both 7 and 14 days after injury (157). Others have attempted to harness bioelectrical properties in conjunction with decellularized ECM to create conductive, biodegradable scaffolds to enhance nerve regeneration. In a 4-week post-transection gap model, scaffolds incorporating polydopamine-modified silicon phosphorus nanosheets to enhance conductivity exhibited reduced pro-inflammatory and increased anti-inflammatory macrophage densities (158). Collectively, these studies demonstrate that hydrogel composition can modulate macrophage phenotype and improve axon regeneration. Further investigation into the effects on other glial types such as perineurial glia and SCs, recruited immune cells, revascularization, and ECM remodeling, as well as comprehensive evaluation of functional outcomes, will be critical for clinical translation.

As in other biomedical modalities, species-specific differences may explain challenges to translation in peripheral nerve repair (165, 166). For example, most experimental models use rodents, whose nerves regenerate faster than human nerves and are uni- or bifascicular rather than polyfascicular, as in humans (167). In addition, most research studies are performed on male animals. Although this mirrors the human demographic in which these injuries predominately occur, there likely exist sex-specific differences in regeneration outcomes that could provide valuable insight for the development of therapeutic interventions (168). Furthermore, peripheral nerve injuries often lead to chronic neuropathic pain, and sex-specific differences in pain processing and pathophysiology are well documented (169). Finally, most preclinical studies fail to approximate the complex, highly variable, real-world clinical scenario. For example, animal studies commonly utilize short-gap injuries with immediate intervention, whereas clinical repairs often involve more complex injuries with delayed treatment timelines, creating temporal mismatches that complicate direct translation (170, 171). Another critical challenge is the difficulty of translating basic science outcome measures to clinically meaningful endpoints. It is not clear how certain outcomes indicative of recovery in preclinical models, such as the number of axons regenerating, amount of reinnervation, or electrophysiologic outcomes, directly correlate to neurologic improvement in patients (172). Establishing clinically meaningful endpoints that predict functional recovery remains a significant barrier to assessing therapeutic efficacy.

For drug-based therapies, a critical challenge is cell-specific targeting. Challenges with drug-based application of agents such as growth factors or immunosuppressive agents include ideal dosing strategies and temporal regimes (165). Currently, many interventions target receptors or pathways with inherent multiplicity across multiple cell types, and beneficial outcomes may be offset by negative side effects (173). Cellular-based strategies face challenges including immunogenicity, poor cell survival rates that can exacerbate local inflammation, and costs associated with patient-specific manufacturing that limit accessibility (140, 174). Finally, given the dynamic microenvironment of peripheral nerve regeneration, we still lack understanding of temporal and spatial changes to cell phenotypes and their synergistic interactions. Global strategies to alter regeneration are likely insufficient, and we need improved temporal and spatial understanding of favorable and failed regeneration to develop targeted interventions (28).

In the realm of electrical stimulation, one of the critical challenges is the optimization of stimulation protocols from rodent to human, or for specific neuronal types; however, there is promising evidence of therapeutic benefits in several preclinical models. Notably, electrical stimulation has been shown to improve regeneration through increased expression of growth factors associated with regeneration, such as NGF and BDNF, which are associated with improved axon regeneration and motor reinnervation (146, 175). Despite these benefits, not much is known about how electrical stimulation may modulate the immune microenvironment. Further interrogation of the impacts of electrical stimulation not only on nerves themselves but also on other cells such as fibroblasts, glia, regenerating vasculature, and infiltrating immune cells will be critical to optimize protocols for temporal and potentially broader improvement in the local microenvironment.

In addition, there is need for more studies in humans to establish standard protocols, outcome measures, and demonstrate safety for broader applicability (176). Although promising results exist in both rodent and human studies across diverse injury mechanisms and nerves, the expansion of this approach remains limited. This is largely a consequence of substantial variability among studies, including stimulation protocols, injury paradigms, timing of intervention after injury, and the use of direct nerve stimulation vs. transcutaneous paradigms during recovery (176), as well as marginal and inconsistent outcomes. Electrical stimulation likely follows a dose-response curve, but there are numerous factors that determine dosage. As of yet, there is no clear determination of optimal dosage.

For conduits, a critical challenge is the identification of materials that minimize immunogenicity while providing ideal mechanical properties to match native nerve tissue (162, 164). Furthermore, the gap size in mice and humans differs markedly. Whereas most preclinical studies use short gaps, clinical gaps often exceed these distances, and vascularization and cellular infiltration kinetics differ fundamentally (177, 178).

Despite the wide variety and novel insights into how conduits and hydrogels not only facilitate reapproximation but can also be leveraged with molecules that facilitate regeneration or bias infiltrating or local immune and glial cells toward pro-regenerative phenotypes, very few grafts have been approved for clinical use (165). Moreover, several technical challenges still exist, such as prolific axonal branching at the site of repair and increased axonal sprouting at axon terminals, resulting in polyinnervation of muscles (165). No conduit or hydrogel system has successfully bridged large gaps with functional outcomes comparable with autograft (171).