Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Neuroscience. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 25 July 2021

Axon Regeneration in Peripheral Nervesfree

Axon Regeneration in Peripheral Nervesfree

  • Arthur EnglishArthur EnglishDepartment of Cell Biology, Emory University School of Medicine


Despite the intrinsically greater capacity for axons to regenerate in injured peripheral nerves than after injury to the central nervous system, functional recovery after most nerve injuries is very poor. A need for novel treatments that will enhance axon regeneration and improve recovery is substantial. Several such experimental treatments have been studied, each based on part of the stereotypical cellular responses that follow a nerve injury. Genetic manipulations of Schwann cells that have transformed from a myelinating to a repair phenotype that either increase their production of axon growth-promoting molecules, decrease production of inhibitors, or both result in enhanced regeneration. Local or systemic application of these molecules or small molecule mimetics of them also will promote regeneration. The success of treatments that stimulate axonal protein synthesis at the site of the nerve injury and in the growing axons, an early and important response to axon injury, is significant, as is that of manipulations of the types of immune cells that migrate into the injury site or peripheral ganglia. Modifications of the extracellular matrix through which the regenerating axons course, including the stimulation of new blood vessel formation, promotes the navigation of nascent regenerating neurites past the injury site, resulting in greater axon regeneration. Experimental induction of expression of regeneration associated gene activity in the cell bodies of the injured neurons is especially useful when regenerating axons must regenerate over long distances to reinnervate targets. The consistently most effective experimental approach to improving axon regeneration in peripheral nerves has been to increase the activity of the injured neurons, either through electrical, optical, or chemogenetic stimulation or through exercise. These activity-dependent experimental therapies show greatest promise for translation to use in patients.


  • Motor Systems

Axon Regeneration in Peripheral Nerves

Even though axons in peripheral nerves are said to be capable of complete reinnervation of peripheral targets after peripheral nerve injury (PNI), functional recovery from most injuries is very poor. Very few patients with nerve transection injuries report significant restoration of function (Scholz et al., 2009). For decades, this poor recovery was considered a surgical issue, but despite improvements in surgical techniques and materials, functional recovery has not improved. It was said that surgical approaches “have reached a plateau where surgical repair techniques cannot be refined any more” (Lundborg, 2002). In response, several novel experimental approaches to enhancing axon regeneration after PNI have been studied, each of which are based on discoveries in the cell biological responses found after nerve injury. Each of those post-injury changes in cells in peripheral nerves following PNI and the novel experimental therapies that emanate from them will be discussed. Many of these approaches have been subjects of review articles, and the reader is directed to them as appropriate.

Structure and Organization of Peripheral Nerves

Peripheral nerves are structures that connect the central nervous system (CNS), the brain and spinal cord, with targets. They provide a conduit for transmission of motor signals to skeletal muscles, for sensory feedback to the CNS from skin and deeper structures, and they provide for CNS control of blood vessels and glands. Spinal nerves arise from body segments defined by vertebrae and cranial nerves from the brainstem. The general structure of all nerves is similar. They consist of neural elements, axons, which are processes of nerve cells or neurons. The axons are surrounded by supportive Schwann cells. Individual medium to large size axons are surrounded by Schwann cells that form a thick insulating myelin sheath (Figure 1A: SC). Groups of small axons are bundled together within a single Schwann cell, known as Remak bundles (Figure 1A: SC Nucleus). Schwann cells in Remak bundles do not make myelin, so the small axons within them are considered unmyelinated.

Most peripheral nerves contain a mixture of different types of axons. Motor axons arise from motoneurons in the spinal cord ventral horn or the brainstem and provide innervation to skeletal muscles. Sensory axons arise as the peripheral processes of axons from dorsal root ganglion (DRG) neurons, in the case of spinal nerves, or from brainstem-associated ganglia for cranial nerves. They vary considerably in size and function. Large myelinated sensory axons innervate proprioceptive structures such as muscle spindles and Golgi tendon organs in muscle and low threshold mechanoreceptors in skin and deeper structures. Smaller sensory axons are poorly myelinated or unmyelinated and most are associated with higher threshold mechanoreceptive and nociceptive functions. In addition, most nerves contain unmyelinated axons of sympathetic ganglion neurons. These postganglionic axons innervate smooth muscle, especially in blood vessels and glands, but some have been shown to innervate neuromuscular junctions (Khan et al., 2016). Parasympathetic preganglionic axons are found in some cranial nerves and some sacral spinal nerves. Their targets are postganglionic neurons in or near the smooth muscles or glands that they innervate.

Figure 1 (A). Electron micrograph of transverse sections of adult mouse nerve showing individual myelinated axons surrounded by Schwann cells (SC) and multiple unmyelinated axons in Remak bundles with a central Schwann cell nucleus. Red arrows indicate the endoneurium surrounding each (Fricker et al., 2009). (B) Connective tissue coverings of peripheral nerves

In addition to ensheathment by Schwann cells, axons are surrounded by a connective tissue covering known as the endoneurium (Figure 1B). The endoneurium is separated from the outer membrane of myelinating Schwann cells by a Schwann cell basal lamina. For unmyelinated axons, all the axons within a Remak bundle are so encased, but the Schwann cell basal lamina is thought to be absent (Griffin & Thompson, 2008). Groups of encased axons and their Schwann cells form fascicles within a nerve and these fascicles are surrounded by a perineurium. Whole nerves are composed of groups of fascicles and are surrounded by a substantial epineurium.

Peripheral Nerve Injuries

Injuries to peripheral nerves are common (Taylor et al., 2008). The severity of nerve injuries is usually defined by a scale of 1 to 5 that correlates with the extent of nerve structure damage and the potential for functional recovery (Sunderland, 1990). The least severe PNI (Stage 1, Neurapraxia) is characterized by a temporary loss of nerve conduction. The axons and connective tissues covering them remain intact, but sensory and/or motor difficulties are experienced distal to the injury. Recovery from these injuries is nearly complete. Stage 2 injuries (Axonotmesis) are more severe. Continuity of axons and ensheathing myelin is disrupted and anterograde (Wallerian) degeneration of the axon segment distal to the injury is found. Connective tissue coverings of the nerve remain intact, and axons can regenerate and remyelinate. Recovery is possible without surgical intervention. These are roughly comparable to nerve crush injuries performed experimentally. In Stage 3 injuries (Neurotmesis), continuity of the entire nerve fiber is lost. The axon, myelin, Schwann cells, and endoneurium are damaged, leading to anterograde degeneration of the axon and the potential for connective tissue scarring. The epineurium and perineurium remain intact. Surgical intervention may be required if any hope for axon regeneration and functional restoration is to occur. Stage 4 injuries are like Stage 3 injuries, except that only the epineurium remains intact. In Stage 5 injuries, a complete severance of the nerve is found. Surgical intervention is required if any functional recovery is expected.

Cellular Responses to Peripheral Nerve Injury

Following a PNI, especially those classified as Stages 2–5, a series of cellular events takes place that is associated with the process of axon regeneration. In the segment of the nerve distal to the injury, the axons themselves degenerate, beginning soon after the injury. The surrounding myelin disintegrates, along with the debris from degenerating axons. These products are cleared first by Schwann cells, using autophagy to break down their own cellular components, and later by immune cells that migrate to the injury site.

Formation of Repair Schwann Cells

The Schwann cells distal to a PNI undergo a process known as adaptive cellular reprogramming: a change in these differentiated cells in the injured nerves from a myelinating to a repair-supportive phenotype. A cessation of myelin gene expression by these repair Schwann cells, along with expression of proteins that were produced during development, are post-injury changes that have been used to suggest that the formation of repair Schwann cells is a process of dedifferentiation, but other changes in repair Schwann cells suggest that the changes involved are different. There is increased expression of proteins supporting neuronal survival and regeneration, as well as of cytokines and chemokines that attract immune cells to the injury site. Repair Schwann cells, including those in Remak bundles, also proliferate, elongate, and rearrange themselves within the endoneurium as cellular columns that are known as bands of Büngner. These are tracks along which regenerating axons grow (Gomez-Sanchez et al., 2017). Details of the signaling events in this transformation are nicely reviewed elsewhere (Jessen & Arthur-Farraj, 2019).

The role of the proliferation of Schwann cells distal to the nerve injury in axon regeneration has been of some interest. There is a clear increase in the number of Schwann cells in the distal segment of injured nerves, starting 3–4 days after injury, and this increase is blocked in mice deficient in the cell cycle regulator molecule, cyclin D1 (Yang et al., 2008). Whether the newly formed Schwann cells are similar to the reprogrammed cells described as repair Schwann cells is not clear, but no distinct phenotype has been described. What is clear is that the number of Schwann cells present after successful axon regeneration is not increased. It has been suggested that the new Schwann cells are somehow excluded from the regeneration pathway and die from apoptosis due to a lack of axonal contact (Murinson et al., 2005). Whether this is selective for newly generated Schwann cells is not certain. In cyclin D1 knockout mice, proliferation is blocked, but axon regeneration appears to be similar to that found in wild type controls (Yang et al., 2008). The importance of Schwann cell proliferation remains unknown.

Axon Growth Promotion by Repair Schwann Cells

Among the proteins expressed by repair Schwann cells are axon guidance molecules that provided cues for elongating axons during development. Four families of axon guidance molecules are recognized: the netrins, semaphorins, slits, and ephrins (Kolodkin & Tessier-Lavigne, 2011). Some are growth promoting, like netrins, and some are chemorepulsive, like the slits. The most widely studied of these molecules is netrin-1. Its expression is increased dramatically by repair Schwann cells soon after PNI. By binding to one of its receptors, Deleted in Colorectal Carcinoma (DCC), netrin-1 acts as a chemoattractant and plays an important role in the guiding of regenerating axons across an injury site. Its roles in elongation and branching of regenerating axons are less well known. The roles of these axon guidance molecules in axon regeneration are reviewed (Dun & Parkinson, 2017). They remain a potential target for novel treatments for PNI that warrant further exploration.

Additionally, a number of growth-promoting molecules are synthesized by the repair Schwann cells and secreted into the Schwann cell basal lamina. These include glial cell-line derived neurotrophic factor (GDNF), artemin, brain derived neurotrophic factor (BDNF), neurotrophin-3 (NT3), nerve growth factor (NGF), vascular endothelial growth factor (VEGF), erythropoietin, fibroblast growth factors (FGFs), pleiotrophin, N-cadherin, and the common neurotrophin receptor p75NTR. These secreted molecules interact with regenerating axons and may contribute to axon regeneration. The roles of these molecules are reviewed elsewhere (Boyd & Gordon, 2003; Brushart et al., 2013).

It is known that different suites of these growth-promoting molecules are produced by repair Schwann cells associated with different types of axons (Höke et al., 2006). Repair Schwann cells in axotomized ventral roots and cutaneous nerves express different growth factor profiles, and the magnitudes of the differences vary with central-peripheral location (Brushart et al., 2013). This diversity is thought to define preferential or modality-specific axon regeneration first described by Brushart (Brushart, 1988, 1993, 2011). Indeed, application of modality-specific growth factors and extracellular matrix components to nerve conduits selectively promoted modality-specific regeneration (Santos et al., 2016).

One of these growth factors, BDNF, has been studied extensively as a focus for enhancing axon regeneration following PNI. Regeneration of both sensory and motor axons is slowed significantly if BDNF availability is reduced by function blocking antibodies (Zhang et al., 2000), by genetic manipulation of BDNF expression in repair Schwann cells (Wilhelm et al., 2012), or if its high affinity receptor, tropomyosin receptor kinase B (trkB), is knocked out selectively in neurons (English et al., 2013). At moderate doses, treatments with recombinant human BDNF (rhBDNF) promoted regeneration, but at higher doses there was no effect (Boyd & Gordon, 2003). This attenuation was linked to greater BDNF signaling through p75NTR (Boyd & Gordon, 2003). The role of p75NTR is complex, as blocking its activity promotes the early elongation of regenerating axons across the injury site (McGregor et al., 2021) but when it is eliminated from the repair Schwann cells, further elongation of regenerating axons is inhibited (Tomita et al., 2007).

The biological half-life of rhBDNF is relatively short (1–3 h) (Kishino et al., 2001), making its use as a novel therapeutic agent improbable. Using an in vitro cell survival assay, Ye and colleagues screened a large number of natural products and discovered two small molecules that bind the trkB receptor with high affinity, but not p75NTR, and generate the same downstream signaling cascades as BDNF (Jang, Liu, Chan, et al., 2010; Jang, Liu, Yepes, et al., 2010). When applied locally to injured mouse nerves or even administered systemically after PNI, both deoxygedunin and 7,8 dihydroxyflavone (7,8 DHF) enhanced axon regeneration significantly at very low doses (Figure 2) (English et al., 2013). Further use of these small molecule trkB agonists as a treatment for PNI deserves consideration.

Figure 2. Effect of treatments with small molecule trkB ligands on axon regeneration after PNI. Top: Common fibular nerves in mice expressing yellow fluorescent protein in axons were cut and repaired with segments of nerves from non-fluorescent littermates. Some were treated with 7,8 DHF, some with deoxygedunin, by daily intraperitoneal injection. Some were left untreated (Control). Repaired nerves were harvested two weeks later and lengths of fluorescent axon profiles measured. Bottom: Median axon profile lengths were significantly longer in the treated nerves.

Reprinted from English et al. (2013).

It has long been thought that the elongation of regenerating axons follows a gradient of growth-promoting molecules such as BDNF. Upon contact with repair Schwann cells, a process of remyelination is initiated. Schwann cells increase expression of proteins related to myelination and decrease expression of growth-related proteins (Jessen & Mirsky, 2016). Although direct in vivo evidence for the presence of these gradients is not known, the importance of this short lifetime of the repair Schwann cells is perhaps shown by attempts to promote axon regeneration using nerve grafts containing Schwann cells that had been engineered to express growth-promoting molecules. This approach stimulated the early regeneration of axons into the grafts, but the prolonged delivery of these molecules in the pathway of the regenerating axons, and a lack of any proximal-to-distal gradient in the distribution of growth factors, resulted in a trapping of motor axons and failure to leave the graft (Hoyng et al., 2014; Santosa et al., 2013; Tannemaat et al., 2008). More carefully controlling the timing of expression of growth factors had a much more positive effect on functional recovery (Eggers, de Winter, Arkenaar, et al., 2019; Eggers, de Winter, Hoyng, et al., 2019).

Similarly, if injured nerves are repaired after a significant delay, as is common clinically, or if regenerating axons must elongate over considerable distances to reinnervate targets, leaving repair Schwann cells without axonal contact, regeneration is significantly impaired (Gordon et al., 2011). These findings have stimulated efforts to prolong the expression of growth-promoting molecules by repair Schwann cells using both surgical and gene therapy approaches (Hoyng et al., 2015; Sulaiman & Gordon, 2018).

Immune Cells

The blood-nerve barrier is broken down after PNI and immune cells are attracted to the site of PNIs by chemokines and cytokines released by the repair Schwann cells, such as monocyte chemoattractant protein (CCL2), acting on the CCR2 receptor on the monocytes. The CX3CR1 ligand (CX3CL1 or fractalkine) also has been implicated in this attraction (Leonard et al., 1991), although no increase in the mRNA for this cytokine was found in the distal sciatic nerve after axotomy (Lindborg et al., 2017). Macrophage infiltration is first observed at the site of injury 2–3 days post-injury and peaks at 7 days post-injury (Taskinen & Roytta, 1997). Recruited monocytes are heterogeneous but can be divided generally into inflammatory and anti-inflammatory (patrolling and reparative) subsets. Inflammatory monocytes are thought to play a key role in the removal of axonal and myelin debris. The anti-inflammatory monocytes are thought to promote tissue regeneration. Once recruited to the site of the nerve injury, these two classes of monocytes can differentiate into two classes of macrophages. Type 1 or M1 macrophages arise from inflammatory monocytes, produce pro-inflammatory cytokines, and are associated with phagocytic processes. Anti-inflammatory monocytes give rise to tissue repair (M2a, M2c), and regulatory (M2b) macrophage subtypes. The M2a and M2c cells release anti-inflammatory cytokines and secrete growth factors. Experimental manipulations that favor their expression result in enhanced axon regeneration (Mokarram et al., 2017).

The removal of debris related to anterograde degeneration following PNI is vital to the success of subsequent axon regeneration and it has been thought to be the function of repair Schwann cells and infiltrating monocyte-derived M1 type macrophages (Gaudet et al., 2011). However, work on the role of neutrophils has questioned this view and introduced a potential new therapeutic target. Knockout of the CCR2 gene reduced the appearance of macrophages at the site of a PNI, but it did not reduce the clearance of myelin debris from the injury site (Niemi et al., 2013). These authors show that infiltration by neutrophils is the primary source of compensation for the loss of CCR2+ macrophages in these knockout mice, but neutrophil depletion in wild type mice also leads to decreased removal of myelin debris (Lindborg et al., 2017).

In addition to the accumulation of macrophages in the distal nerve segment, these cells accumulate in sensory and sympathetic ganglia after PNI, attracted there by expression of CCL2 (and perhaps CX3CR1) in the axotomized neurons, especially by medium to large size cells. The macrophages are concentrated around the cell bodies of axotomized, but not intact, ganglion cells. Unlike macrophages at the site of nerve injury, that are associated with anterograde degeneration phenomena, macrophages in ganglia may stimulate axonal growth. The molecular mechanism of this growth promotion is not well-established. This infiltration of macrophages into sensory ganglia also has been reported to contribute to the development of neuropathic pain, and may do so differently in males and females (Yu et al., 2020). Neuroinflammation within the DRG has been shown to increase the expression of certain regeneration-associated genes, but this requires further investigation. More details of the role of macrophages in peripheral ganglia can be found in the excellent review Zigmond and Echevarria (2019).

Extracellular Matrix

A rich extracellular matrix (ECM) component in nerves is found in association with the epineurium, perineurium, and endoneurium (Muir, 2010). It contains numerous growth-promoting elements, such as laminin, but it also contains growth inhibitory elements, most notably chondroitin sulfate proteoglycans (CSPGs) (Zuo et al., 1998). These consist of large core glycoproteins and associated side chains of glycosaminoglycan (GAG) polymers of chondroitin sulfate (Figure 3A). The CSPGs associate with laminins and inhibit their growth-promoting potential, primarily through these GAGs. The chondroitin monomers in the GAGs can be differentially sulfated at the 4 and/or 6 positions. Single 6 sulfation is least inhibitory to axon elongation, followed by 4 sulfation, with 4,6 dual sulfation being most inhibitory (Gilbert et al., 2005).

Following PNI, CSPGs are upregulated and contribute to poor axon regeneration (Braunewell et al., 1995; Morgenstern et al., 2003; Zuo et al., 1998). The CSPGs with more inhibitory 4 sulfated GAGs are specifically found surrounding the endoneurium, where they may play a role in guidance of regenerating axons into endoneurial tubes, but also contribute to inhibiting or slowing this entrance to a regeneration pathway. A single treatment with the bacterial enzyme, chondroitinase ABC, at the time of repair of a cut nerve cleaved the GAG side chains from CSPGs, irrespective of their sulfation state, and resulted in more robust axon regeneration (Figure 3B) (Groves et al., 2005; Zuo et al., 2002). These treatments also result in faster restoration of neuromuscular connectivity and improved functional recovery (Sabatier et al., 2012).

Figure 3. A. Structure of chondroitin 4,6 sulfate elements of proteoglycan side chains. The arrow points to the site of cleavage by chondroitinase ABC (Chase ABC). B. A single treatment of cut and repaired common fibular nerves with Chase ABC resulted in a striking enhancement of axon elongation one week later. Nerves of mice in which a subset of axons were marked completely with yellow fluorescent protein (YFP) (A) were cut and repaired with a graft harvested from a wild type littermate (B). One week later, lengths of YFP+ axons in Chase ABC–treated nerves measured in the grafts were more than three times longer than controls.

The Schwann cell basal lamina is the preferred pathway for elongation of regenerating axons (Witzel et al., 2005). It is known that growth-promoting, and growth-inhibiting molecules can be secreted into this ECM by repair Schwann cells and the regenerating axons, but the contributions from mesenchymal cells have been overlooked. Using an unbiased systems biology approach, multiple ligands were identified in genetically marked mesenchymal cells derived from injured nerves (Toma et al., 2020). Some of these molecules are also produced by repair Schwann cells, but many are unique to mesenchymal cells. Regenerating axons expressed receptors for many of these ligands. It will be interesting to see if this analysis leads to novel targets for promoting axon regeneration after PNI.

An additional component of the ECM comes into play in response to Class 3–5 PNIs, when both axon and endoneurial continuity is disrupted. After these injuries, the proximal and distal nerve segments withdraw, creating a gap at the injury site, and a new fibrous bridge forms in it. Blood vessels form in this gap, stimulated by release of VEGF-A from hypoxia-sensitive macrophages, and their recruitment of endothelial cells from both the distal and proximal nerve stumps. Repair Schwann cells migrate into the gap and align along these vessels. These Schwann cells, surrounded by fibroblasts, form cords that then attract regenerating axons and serve to guide them across the bridge (Cattin et al., 2015).

Some axons are not able to follow the cords across the bridge and never regenerate successfully. Additionally, the association of some axons with the cords has been described as a “tentative and constrained penetration” into the bridge (McDonald et al., 2006), meaning that their ability to enter a regeneration pathway is delayed. A consequence of this temporally staggered regeneration is that different regenerating axons may enter regeneration pathways where the extent of available support by repair Schwann cells may vary considerably.

The navigation through the fibrous bridge clearly represents a critical early stage of axon regeneration following PNI. The ultimate reliance on macrophage VEGF-A for this process might explain the successful enhancement of axon regeneration by manipulation of VEGF expression in two different manners. Direct application of VEGF-A, either as recombinant protein (Hobson et al., 2000) or using genetic approaches (Pereira Lopes et al., 2011; Zor et al., 2014), accelerated axon regeneration. Blocking VEGF expression retarded blood vessel formation and axon regeneration (Nishida et al., 2018). Macrophages are known to respond to hypoxia with increased activity of the transcription factor, hypoxia inducible factor 1 (HIF1). HIF1 is a heterodimetric complex consisting of a constitutively expressed β‎ subunit and a low-oxygen (or hypoxia) inducible, α‎ subunit. The HIF1 complex (α‎ and β‎ subunits) is primarily regulated by the stability of the alpha subunit (HIF1α‎), which is tightly controlled by enzymes called prolyl hydroxylases and is highly sensitive to cellular oxygen states. Under normoxic conditions, HIF1α‎ has a very short half-life and is degraded in the proteasomal pathway. Under low oxygen states, HIF1α‎ is stabilized and rapidly translocated to the nucleus, where it dimerizes with the β‎ subunit to target HIF1 response elements. One prominent downstream target of HIF1 is VEGF-A (Ramakrishnan et al., 2014). Treatments with repeated intermittent hypoxia resulted in an elevation of VEGF expression and improved axon regeneration (Zhou et al., 2018). Similarly, Ward and colleagues (Hassan et al., 2019) treated mice with a HIF prolyl hydoxylase inhibitor, Roxidustat, and found that a single treatment following sciatic nerve transection and repair both increased HIF1 stability and produced a substantial enhancement of axon regeneration and functional recovery.

Axon Responses to Peripheral Nerve Injuries

After nerve injuries involving axotomy, calcium is increased in the distal segment of the injured nerve, and this initiates the process of anterograde (Wallerian) axon degeneration (Conforti et al., 2014). Calcium concentration is also elevated in the proximal segment of injured axons, entering through voltage-gated calcium channels (Nehrt et al., 2007), and this provides a rapid retrograde injury signal to the neuron. It also contributes to the local generation of cyclic adenosine monophosphate (cAMP) (Smith et al., 2020). The increased calcium concentration also stimulates translation of resident axonal mRNAs. Some of these locally synthesized proteins form a second retrogradely transported injury signal that includes importin-β‎1, a part of a mechanism that regulates axon size (Rishal & Fainzilber, 2014). The transcription factor STAT3, which promotes survival of at least some dorsal root ganglion neurons after PNI (Ben-Yaakov et al., 2012), is also among these locally synthesized and retrogradely transported proteins. The calcium-driven local protein synthesis has been implicated in the initial fusion of axonal membranes at the injury site as well as in the formation of growth cones from rapidly forming regenerative sprouts (He & Jin, 2016). Locally synthesized proteins were first implicated, because sprouts and growth cones can form from isolated segments of nerves when placed in culture (Kato & Ide, 1994; Shaw & Bray, 1977). Stimulus for elongation of these sprouts is likely in response to a disruption of axonal transport by the injury that neurons use to regulate axonal length, which also depends on local protein synthesis in the axons (Perry et al., 2016). The role of this local protein synthesis in axon regeneration is nicely reviewed elsewhere (Costa & Willis, 2018; Koley et al., 2019; Smith et al., 2020).

At least some of the proteins involved in longer term elements of regeneration of axons after PNI are also synthesized locally, from mRNA stored in axons (Costa & Willis, 2018; Terenzio et al., 2018). Twiss and colleagues (Sahoo et al., 2018) found small particles described as stress granules in intact axons of peripheral nerves. These particles consist of translationally repressed mRNAs, bound to the RasGAP SH3 domain binding protein 1, G3BP1. Upon peripheral axotomy, the stress granules were found to disperse, release bound mRNAs, and generate the proteins they encode. The disassembly of granules was regulated by phosphorylation of G3BP1via a locally expressed G3BP1 kinase. Blocking the function of G3BP1 in assembly of the stress granules increased axonal protein synthesis and enhanced axon regeneration and functional recovery after PNI (Sahoo et al., 2018). Future manipulations of axonal protein synthesis as a potential therapeutic target for novel treatments for PNI are strongly supported by these findings.

However, if regenerating axons need to elongate over long distances to reach their targets, the distances involved will require that an abundance of raw materials be supplied to the growing tips via transport from their cell bodies. Indeed, the steady-state rate of elongation of regenerating axons after PNI is similar to that of the slow component b of kinesin-based axonal transport (Lasek et al., 1984), which is characterized by intermittent pausing of these cargoes during transport (Roy et al., 2007). A well-established cell body increase in mRNA and protein expression following PNI, along with the identification of specific axonal proteins, such as GAP-43, led to the identification of components of the cell body response termed regeneration-associated genes (RAGs). The introduction of unbiased large-scale profiling methods has expanded the list of RAGs associated with axon regeneration after PNI greatly. Neuronal RAGs are now recognized by their organization into a network of hub proteins, consisting mainly of transcription factors, each of which controls the transcription of multiple growth-associated terminal RAGs. Expression of the terminal RAGs can be influenced by multiple hub transcription factors. Products of terminal RAGs include but are not limited to cytoskeletal and membrane proteins needed for elongating axons, signaling molecules, growth factors and their receptors, and ion channels. However, RAGs also encode some micro-RNAs, small single-stranded molecules that regulate multiple other genes, and these can function to enhance regeneration (Strickland et al., 2011). Additionally, enzymes related to epigenetic regulation of gene expression are found among RAGs (Palmisano & Di Giovanni, 2018). An excellent review of RAGs can be found at (Ma & Willis, 2015).

Changes in Neuronal Properties After PNI

Significant changes in the properties of neurons in the CNS are found after PNI. Axotomized motoneurons become more excitable, as noted by decreased rheobase or action potential threshold. Glutamatergic excitatory synapses, and to a lesser extent, GABA-ergic inhibitory synapses are withdrawn from their somata and proximal dendrites, a process termed “synaptic stripping” (Figure 4A–C) (Blinzinger & Kreutzberg, 1968). This shedding is transient if motor axons regenerate and reinnervate muscle targets (Brannstrom & Kellerth, 1999), but over time there is a restoration of GABA-ergic synapses whether or not injured axons reinnervate peripheral targets (Alvarez et al., 2011). If axotomized motoneurons do not reinnervate targets until after a significant delay, excitatory inputs are lost permanently, especially those arising from primary afferent neurons, as their terminal branches withdraw from the ventral horn (Rotterman et al., 2014). This withdrawal of excitatory synapses has been thought of as a way of decreasing motoneuron activity to aid in axon regeneration, but this view has been challenged.

The potassium-chloride co-transporter 2 (KCC2), an important regulator chloride concentration within motoneurons, disappears rapidly from axotomized motoneurons (Figure 4D) (Akhter et al., 2019), resulting in a decrease or reversal of the Cl-gradient used by ionotropic GABA-A receptors to hyperpolarize the motoneurons and thus a decreased inhibition produced by the remaining GABA-ergic synapses (Takata & Nagahama, 1983). This non-inhibitory effect on GABA-ergic synapses is remarkably similar to that found during axon elongation by developing motoneurons, when KCC2 is also absent from the cell membranes (Wilhelm & Wenner, 2008). If this change in synaptic function is placed in context with the increase in excitability observed in motoneurons after PNI, a very different role for synaptic stripping has emerged. Rather than serving to decrease motoneuronal activity, the shift in the ratio of glutamtergic to GABA-ergic synapses produced by the greater withdrawal of excitatory synapses, combined with the functional changes of the GABA-ergic synapses on axotomized motoneurons resulting from the loss of KCC2, may contribute to promoting the regeneration of their axons by increasing their activation. This hypothesis is developed in more detail elsewhere (Alvarez et al., 2020).

Figure 4. A, B. Examples of NeuN-labeled spinal motoneurons in intact mice (left) and mice 2 weeks after PNI. Immunoreactivity to excitatory synaptic inputs (VGLUT2, A) and inhibitory inputs (VGAT, B), to show the reduction in synaptic coverage after PNI (Reprinted from Alvarez et al., 2011). C. Quantitative reduction of VGLUT1+ excitatory inputs and GAD-67+ inhibitory contacts (Data from Park et al., 2019). D. KCC2 immunoreactivity surrounding the somata of fast blue (FB)-labeled motoneurons (arrows on left) is lost 2 weeks after PNI.

After PNI, the peripheral process of sensory axons is injured and results in increased excitability through increased membrane input resistance, decreased rheobase, and changes in ion channel expression (Abdulla & Smith, 2001; Zhang et al., 1997). Adult DRG neurons do not express KCC2 but regulate intracellular chloride via the Na-K-Cl cotransporter, NKCC1 (Alvarez-Leefmans et al., 2001), which is increased in DRG neuron membranes after PNI (Pieraut et al., 2007). This change results in a doubling of intracellular chloride concentration. At least for medium sized to large DRG neurons, those with myelinated axons, the increased chloride concentration is correlated with neurite outgrowth and blockade of NKCC1 blocked axon regeneration (Modol et al., 2015). Thus, increased excitability may also be a feature of regenerating sensory axons. These findings are reviewed in more detail elsewhere (Akhter et al., 2020).

Activity-Dependent Experimental Therapies

The changes in the excitability of axotomized neurons are almost certainly related to the most successful experimental therapies to enhance axon regeneration following PNI. Both application of 1 hour of supramaximal electrical stimulation (ES) and 2 weeks of moderate daily treadmill exercise have been shown repeatedly to promote axon regeneration and increased functional recovery (Figure 5) (Gordon & English, 2016). The initial use of exercise as an experimental treatment for PNI (Sabatier et al., 2008) was based on its effectiveness in spinal cord injury models and it has now been validated by others (Udina et al., 2011). The effectiveness of ES and exercise have been shown to require both increasing the activity of the axotomized neurons and increased neuronal expression of BDNF and its trkB receptor. Experiments using optogenetics (Ward et al., 2016, 2018), chemogenetics (Jaiswal & English, 2017; Jaiswal et al., 2020), or bioluminescent optogenetics (Jaiswal et al., 2020; Mistretta et al., 2019) have shown that the extent of increase in neuronal activity needed may be modest, but that it is both necessary and sufficient to promote axon regeneration. If the increased excitability of motor and sensory neurons is considered, these activity-based experimental therapies might be viewed as exploiting an intrinsic cell biological response to the nerve injury.

Figure 5. Effects of brief electrical stimulation and treadmill exercise on axon regeneration after PNI. A. Common fibular nerves from mice expressing yellow fluorescent protein in axons after transection and repair with a segment of nerve harvested from a non-fluorescent littermate. Longer axons are found in the nerve from the mouse treated with 1 hour of 20 Hz ES. B. Cumulative histograms of axon profile lengths in mice treated with 1 hour of ES or treated with 2 weeks of daily treadmill exercise. Dashed lines indicate medians. C. Results of retrograde labeling to mark the motoneurons whose axons had regenerated successfully and reinnervated muscles 2 and 4 weeks later in control and two different groups of exercised animals. Note the complete reinnervation found in the exercised mice.

Reprinted from Gordon and English (2016)

In contrast to these studies using short applications of moderate increases in activity, others have shown that more prolonged increases in the activity of cultured sensory neurons leads to markedly diminished neurite outgrowth (Enes et al., 2010). They concluded that this increased electrical activity is a signal that suppresses axon growth through L-type Ca2+ channels. Because calcium influx in this manner is an important part of the initial cellular response to axon injury (see “Axon Responses to Peripheral Nerve Injuries”), any applications of activity-dependent experimental therapies, such as ES or exercise, need to be done with moderation.

Hormonal Effects on Axon Regeneration

Striking and unexpected sex differences were found when using different exercise protocols to enhance regeneration (Figure 6A) (Wood et al., 2012). These findings lead to evidence of a requirement for androgen receptor activation for the success of both ES and exercise in enhancing regeneration in both sexes (Figure 6B, C) (Thompson et al., 2014). More recently, a similar requirement was discovered for estrogens (Acosta et al., 2017), at least for the effectiveness of exercise. Administration of androgens has been known for some time to promote axon regeneration, especially in cranial nerves (Brown et al., 2013). The success of activity-dependent experimental therapies thus may be a physical way of mobilizing these hormonal effects on axon regeneration.

Figure 6. Sex and hormone dependence on the effectiveness of exercise. A. Images of nerves from mice in which YFP is expressed in axons 2 weeks after PNI. Continuous treadmill training for 2 weeks (1 hr/day, 5 days per week) is effective in males only. Interval training (4 × 2 minutes of intense running separated by 5 min. rest periods, 5 days/wk) over the same time period is effective only in females. B Enhancing effects of sex-appropriate exercise are blocked in males by castration (C), or treatment with the androgen receptor blocker, flutamide. Implantation of a blank capsule (B) had no effect. C. In females flutamide also blocked the effects of interval training.

Reprinted from Thompson et al (2014).

Summary and Future Directions

In response to Lundborg’s call for new and innovative treatments for PNI, research on the cellular biology of PNI has expanded the list of potentially novel approaches considerably. Strikingly, movement of these experimental therapies toward clinical use has been less robust. In many ways, this lack of translation should not be surprising. Many approaches will require considerably more testing before they could be considered safe and effective enough to be used with patients. Most of these studies have used one of a very few nerve injury models, so that adapting novel experimental therapies to the diversity of nerve injuries that present clinically will be difficult. Further, many of the outcome measures used in these studies are specific to experimental models, usually in rodents, and it is just difficult to assess whether they would have a similar effect in humans. In formulating novel experimental therapies, preclinical researchers should be encouraged to develop and utilize at least some outcome measures that are like ones used clinically, so that this barrier to translation will be lessened.

A shining exception to this lack of translation of basic research to clinical practice is the use of ES. Gordon and colleagues first reported the effectiveness of this treatment around the same time as Lundborg’s admonition to develop novel therapies. They explored the use of ES in a limited manner in patients undergoing carpal tunnel release surgery (Gordon et al., 2010) and have reported on expanded clinical use of ES (Zuo et al., 2020). As reviewed (Ransom et al., 2020), ES is poised to be applied to a number of nerve injury types. It is hoped that some of the other therapies discussed might be given a similar consideration, applied either alone or along with ES.