Earlier, we found that neonatal mice with conditional deletion of c-Jun in Schwann cells (c-Jun mutant mice) show delayed loss of myelin proteins and mRNA after nerve injury (Parkinson et al., 2008). This suggested that Schwann cell c-Jun might play an important role in specifying the phenotype of denervated Schwann cells. To test this comprehensively, we used Affymetrix whole-genome microarray to examine gene expression in the sciatic nerve of adult c-Jun mutant mice and control (WT) littermates and compared this with gene expression in denervated cells in the distal stump of transected nerves without regenerating axons, to avoid the complicating effects of axon-induced redifferentiation (Figure 1). We chose 7 days after injury since in regenerating mouse nerves this is near the mid-point of active axonal regrowth. Seven day denervated cells therefore represent the terrain that confronts regenerating axons in WT and mutant nerves.

Before injury, the nerves of adult c-Jun mutant mice were normal on the basis of a number of criteria. Thus, the numbers of myelinated and unmyelinated axons (see Figures 4E and 4F), myelinating Schwann cells and Remak bundles (see Table S1 available online), g-ratios (Figure S1), sciatic functional index (SFI) (see Figure 7E), motor performance in a rotarod test (unpublished), and responses to heat and light touch (see Figures 7B and 7C) were similar to WT controls. While c-Jun was excised from almost all Schwann cells (Parkinson et al., 2008), c-Jun expression in neurons, macrophages, and fibroblasts was normal, and the rate of axonal disintegration after cut was similar in WT and mutants (Figures S2 and S3).

The close similarity between WT and mutant nerves was confirmed by the Affymetrix screen (Figure 1), since only two genes (keratin 8 and desmoplakin) were differentially expressed. Furthermore, following injury, a comparable number of genes changed expression in WT and c-Jun mutants (Figure 1A). Importantly, however, comparison of the distal stumps of WT and c-Jun mutants revealed 172 significant differences in gene expression (Figure 1 and Tables S2 and S3). The differentially regulated genes included genes which have been implicated in regeneration and trophic support such as BDNF, GDNF, Artn, Shh, and GAP-43 that failed to upregulate after injury, together with genes that failed to downregulate normally after injury such as the myelin genes Mpz, Mbp, and Cdh1 (also known as E-cadherin). Gene ontology analysis indicated that known functions of these 172 genes were particularly related to neuronal growth and regeneration (Figure 1C).

We selected 32 of the 172 disregulated genes for further analysis by RT-QPCR. In every case this confirmed the disregulation shown by the microarray data (Figures 1D–1F and Table S3). Six of the thirty-two genes were then analyzed in purified Schwann cell cultures. Comparison of c-Jun mutant and WT cells confirmed the regulation seen in the distal stumps. Furthermore, as predicted, enforced c-Jun expression in c-Jun mutant cells, by adenoviral gene transfer, activated c-Jun, BDNF, and GDNF expression but suppressed Chd1, Mpz, and Mbp expression (Figure 2A). These results show directly that c-Jun regulates these genes in Schwann cells, demonstrates that this control is independent of the nerve environment, and confirms results obtained by microarray and RT-QPCR.

Lastly, we found that three proteins implicated in regeneration, N-cadherin, p75NTR, and NCAM, were disregulated in cut mutant nerves, although their mRNAs were normally expressed. Injured mutant nerves expressed strongly reduced N-cadherin and p75NTR but elevated levels of NCAM (Figures 2B and 2C). Sox2 protein, which, like c-Jun, is upregulated in WT Schwann cells of injured nerves (Parkinson et al., 2008), remained normally upregulated in injured nerves of c-Jun mutants (Figure S4).

Denervated Schwann cells in injured adult nerves are often considered similar to immature Schwann cells in developing nerves. However, the immature cells for instance do not share the axon guidance, myelin breakdown and macrophage recruitment functions of denervated cells, and these cells differ in molecular expression (Jessen and Mirsky, 2008). To explore the idea that the denervated cell represents a distinct Schwann cell phenotype regulated by c-Jun, we examined three genes, Olig1, Shh, and GDNF, which showed strong, c-Jun-dependent activation in denervated cells (Figure 1D). Using RT-QPCR and in situ hybririsization we confirmed strong expression of these genes in WT adult denervated cells, but found that they were not (Olig1 and Shh) or borderline (GDNF) detectable in immature Schwann cells (from WT embryo day 18 nerve). They were also essentially absent from uncut nerves (Figures 2D and 2E and Table S4). This supports the notion that denervated adult Schwann cells and immature Schwann cells in perinatal nerves represent distinct cell types. It shows also that c-Jun takes part in controlling the distinctive molecular profile of the adult denervated cell. The response of neonatal cells to injury remains to be determined.

Together these results show that c-Jun controls the molecular reprogramming that transforms mature Schwann cells to the denervated cell phenotype following injury. This includes the regulation of genes that differentiate denervated from immature cells and extends to the posttranscriptional control of protein expression.

Denervated Schwann cells form cellular columns that replace the axon-Schwann cell units of intact nerves and serve as substrate for growing axons. We examined these structures by electron microscopy in the distal stump 4 weeks after cut. Because these cells have been without axonal contact for 4 weeks they are comparable to the cells encountered by growing axons in distal parts of crushed nerves in the c-Jun mutant where regeneration is delayed beyond the normal 3–4 week period, while at this time WT nerves have just reached their targets. We found that the structure of these regeneration tracks is strikingly abnormal in c-Jun mutants (Figure 3A). There are many fewer cell profiles per column, indicating reduced process formation (Figure 3B) and the cells have flattened as confirmed by reduction in the roundness index of cell profiles in vivo (Figure 3C). Flattening and paucity of processes are also seen even in c-Jun^−/− cells from neonatal nerves in vitro (Figures 3D and 3E). Therefore, this is a robust phenotype that does not depend on long term denervation in vivo.

Thus, c-Jun is an cell-intrinsic determinant of Schwann cell morphology that controls the structure of the essential regeneration tracks that guide growing axons back to correct targets.

c-Jun specification of gene expression and morphology of denervated cells suggested that Schwann cell c-Jun might exert a decisive control over nerve repair. Because survival of injured neurons is the basis for repair, we measured the survival of small and large dorsal root sensory (DRG) neurons following sciatic nerve crush at the sciatic notch. We counted axons in L4 dorsal roots (Coggeshall et al., 1997) and the tibial nerve, and neuronal somas and nucleoli in DRGs. Comparable results were obtained using all methods.

Axon counts in WT dorsal roots showed that 20%–25% of the unmyelinated axons were lost following crush, as expected (Coggeshall et al., 1997). In contrast, 55%–60% of these axons were lost in c-Jun mutants, showing increased death of small DRG neurons in the mutant. This was confirmed by corrected (Abercrombie, 1946) counts of B neuron profiles in DRGs, showing 25%–30% loss in WT but 45%–65% loss in the mutants (Figures 4A and 4B).

The number of myelinated axons in dorsal roots remained unchanged in injured WT mice as expected (Coggeshall et al., 1997). But surprisingly, in the mutants the number of myelinated axons was reduced by 30%–35%, indicating death of large DRG cells. In confirmation, the corrected number of large A cell profiles in DRGs was reduced by about 40% in the mutants. The number of these profiles did not change significantly in injured WT (Figures 4C and 4D). We also carried out counts on DRG sections using nucleoli as the counted entity, an approach that theoretically provides increased accuracy. Nucleoli in A type DRG neurons from uncut WT (n = 3), 10 week cut WT (n = 3), and 10 week cut c-Jun mutant (n = 3) mice were counted, corrected (Abercrombie, 1946), and expressed as percentage of uncut WT. This showed a 12% reduction in cut WT, (not significant; p > 0.40) but a 50% reduction in the c-Jun mutant (highly significant; p < 0.017). This provides a third line of evidence (in addition to counts of myelinated axons in dorsal roots and cell profiles) for the notion that nerve injury results in the loss of A type DRG neurons in mice that selectively lack c-Jun in Schwann cells.

Consistent with neuronal death, the number of myelinated axons in the mutant tibial nerve 10 weeks after crush was reduced by about 35% and unmyelinated axons were reduced by about 65%, both compared to crushed WT controls (Figures 4E and 4F).

These experiments show that without Schwann cell c-Jun, small, unmyelinated DRG neurons are about twice as likely to die following axonal damage. Significantly, about a third of the large, myelinated DRG neurons also die in crushed c-Jun mutants, although none die in injured WT controls, and in other studies these cells are resistant to death following axonal damage (Tandrup et al., 2000). These experiments establish that a key function of denervated Schwann cells is to prevent the death of injured neurons and that this rescue depends on c-Jun activation.

The number of myelinated axons in ventral roots of both WT and mutant mice remained unchanged following sciatic nerve crush (Table S6). Therefore, unlike DRG neurons, survival of injured ventral horn motoneurons is independent of Schwann cell c-Jun. Nevertheless, the corrected (Abercrombie, 1946) counts of motoneurons that reconnected with the target muscle showed a large reduction in the mutant, even as late as 10 weeks after injury, reaching only about 55% of that in controls, judged by motoneuron backfilling (Figures 5A and 5B). This indicates that in the mutants, axonal regeneration by surviving neurons is severely and permanently compromised.

To analyze regeneration, we examined sciatic nerves 4 days after crush, using the nerve pinch test and by quantifying the number and length of axons in longitudinal sections immunolabeled by CGRP or galanin antibodies to label regenerating DRG and motoneurons. This showed a strong decrease in axon outgrowth in the mutants compared to WT (Figures 5C–5H).

Regeneration in the mouse sciatic nerve is independent of Schwann cell proliferation (Kim et al., 2000; Yang et al., 2008). Nevertheless, because c-Jun contributes to proliferation in vitro (Parkinson et al., 2008), we counted Schwann cells in WT and mutant distal stumps (Table S7). In crushed, actively regenerating nerves (14 days after injury) Schwann cell numbers were not significantly different between WT and mutants; both were elevated 5- to 6-fold compared to uncut nerves. Four days after crush, cell numbers were higher in WT nerves, while 7 days after cut, again the difference between mutants and WT was not significant. The tendency toward lower Schwann cell numbers in the mutants is in line with the involvement of c-Jun in proliferation (Parkinson et al., 2004).

Together this shows that in the absence of Schwann cell c-Jun, the regeneration of axons from surviving neurons is severely reduced, leading to a permanent deficit in the number of neurons that reconnect with denervated targets. The observation that that Schwann cell numbers in regenerating mutant nerves are elevated up to 5-fold compared to uninjured nerves, together with the independence of regeneration from elevated Schwann cell numbers (Kim et al., 2000; Yang et al., 2008) and analysis using microfluidic chambers (following section), shows that that regeneration failure in c-Jun mutants is not caused by lack of Schwann cells.

A nerve is a complex cell community. We therefore used microfluidic chambers containing neurons and purified Schwann cells to test whether the poor axon growth in mutants was caused by disturbance of direct axon-Schwann cell interactions or whether the effect depended on other cells.

Axon regeneration by axotomized, adult WT DRG neurons was strongly stimulated by control Schwann cells relative to laminin substrate alone, as expected (Figures 5I–5K). The c-Jun mutant cells, however, were ineffective, the number and area of axons extending on their surface falling to only 40%–50% of that seen on WT cells. Importantly, reactivation of c-Jun in mutant cells by adenoviral gene transfer, fully restored axon number and length to WT levels.

These experiments show that injury-activated Schwann cell c-Jun controls direct communication between Schwann cells and growing neurites.

We have shown that c-Jun controls three important functions of denervated Schwann cells, formation of regeneration tracks, support of neuronal survival, and promotion of axon regrowth. A fourth major role classically ascribed to these cells is removal of myelin and associated growth inhibitors, a task they accomplish by breaking down myelin early after injury and indirectly by instructing macrophages to complete myelin clearance (Hirata and Kawabuchi, 2002).

We found that myelin clearance was substantially delayed in mutants. Four weeks after sciatic nerve transection (without regeneration), the distal stump of WT nerves was translucent, while mutant nerves remained gray/white (Figure 6A). Osmium stained lipid debris occupied about 10-fold larger area in the mutant than WT nerves (Figure 6B).

Electron microscopy revealed that although transected mutant nerves did not contain intact myelin, many Schwann cells contained lipid droplets, a late product of myelin breakdown (Figure 6C). This was not seen in 4 week transected WT controls. We therefore tested whether myelin breakdown was impaired in mutant Schwann cells. First, in cut adult nerves, the loss of myelin sheaths was delayed in the mutants (Figure 6D). This was not due to infiltrating macrophages, because the difference between WT and mutants was fully maintained when the cut nerves were maintained in vitro (Figure 6E). Second, this delay was confirmed by slower breakdown of myelin basic protein (MBP) in vivo (Figures S5A and S5B). Third, when myelinating cells from postnatal day 8 nerves were cultured, myelin proteins were broken down slowly by c-Jun mutant cells compared to WT, and mutant cultures contained many Schwann cells bloated with myelin debris (Figures 6F–6H). Both types of culture contained similar numbers of F4/80⁺ macrophages (5.6+/−1.8% and 5.8+/−1.7%, n = 4, in WT and mutants, respectively, at 3 days in vitro), and no F4/80⁺ cells containing myelin proteins were seen, suggesting macrophages are not significantly involved in myelin breakdown in these experiments.

These findings show that c-Jun mutant Schwann cells are deficient in their ability to break down myelin.

Surprisingly, abnormalities in myelin breakdown extended to the macrophage compartment, although the macrophages are genetically normal (Figure S2). Four weeks after cut, macrophages in the mutants contained large amounts of myelin debris, and counts of lipid droplets per macrophage showed that they were about 7 times more numerous than in WT (Figures 6I and 6J). Furthermore, the number of very large (>150 μm²) foamy macrophages bloated with debris was elevated 3-fold in mutant nerves at this time point.

Macrophage numbers in the distal stump were strongly elevated in both WT and c-Jun mutant nerves after injury. Three days after cut, their number close to the injury site was significantly higher in WT mice, and a migration assay using Boyden chambers showed that WT nerves attracted more macrophages than mutant nerves (Figures S5C and S5D). But at 1 and 6 weeks after cut, the number of macrophages was similar in WT and mutants, and at 4 and 14 days after crush, macrophage numbers were not significantly different (Table S7). RT-QPCR of cytokines in the distal stumps 36 hr after cut did not reveal significant differences between WT and mutants (Figure S5E). This indicates that c-Jun mutants do not suffer from a major failure in macrophage recruitment. Reduced numbers shortly after injury in the mutant might relate to lower Schwann cell numbers rather than to significant disturbance in the expression of macrophage attractants by individual cells.

These results show that injured c-Jun mutant nerves develop substantial problems with myelin clearance. This is evident not only in Schwann cells but also in macrophages, an observation that suggests a role for Schwann cells in the control of macrophage activation and myelin degradation.

Previous sections show that injury-activated Schwann cell c-Jun controls the generation of the denervated Schwann cell, and controls key cellular interactions during Wallerian degeneration and nerve repair. The end result of this process is functional recovery. This is remarkably effective in rodents, where full recovery is seen 3–4 weeks after crush. We found that this essential feature of peripheral nerves was abolished or strikingly compromised in c-Jun mutant mice.

To measure sensory function, we used the following: (1) toe pinching (pressure), (2) Von Frey hairs (light touch), and (3) the Hargreaves test (temperature).

In WT and c-Jun mutants, sciatic nerve crush abolished the response to toe pinching and decreased the responses to light touch and heat to the minimum measurable by these assays. Control mice recovered in 3–4 weeks as expected, using the toe-pinching test. c-Jun mutants, however, showed minimal recovery, even after extensive (up to 70 day) periods (Figure 7A). Similarly, even 70 days after crush mutant mice showed no recovery in the Hargreaves and Von Frey tests, although control mice recovered fully (Figures 7B and 7C).

To measure motor function we used the toe-spreading reflex. We found that toe extension, abolished by nerve crush, recovered on schedule in WT controls but failed to recover even at 70 days in mutants (Figure 7D).

To measure sensory motor coordination we measured the sciatic functional index (SFI; Inserra et al., 1998). We found the expected reduction in SFI following sciatic nerve crush in WT and c-Jun mutants, but a permanent failure of recovery in mutants only (Figure 7E). These experiments show that Schwann cell c-Jun is a necessary driver of functional recovery of injured peripheral nerves.

Having shown that c-Jun activation is necessary for the conversion of injured nerves to an environment that supports repair, we tested whether c-Jun activation alone was sufficient for this critical transformation.

We took advantage of Wld^s mice, in which axons degenerate slowly after cut, and Schwann cells therefore remain differentiated and unsupportive of repair (Coleman and Freeman, 2010). We confirmed delayed regeneration in these mice compared to WT controls using the nerve pinch test and counting galanin-positive fibers distal to nerve crush (Figures 7F and 7H). Previously we showed that Schwann cells in Wld^s nerves fail to activate c-Jun after injury (Jessen and Mirsky, 2008). We therefore used adenoviral infection to enforce c-Jun expression in crushed Wld^s nerves (Figure S6). Remarkably, this converted the inhospitable Wld^s nerves to a terrain that supported regeneration as effectively as WT nerves (Figures 7F–7H).

This shows that c-Jun is not only necessary but also sufficient for the generation of a growth supporting environment in injured nerves. This observation also confirms that regeneration failure in Wld^s mice is caused by a failure of timely Schwann cell injury response in these animals.

The Schwann cell response to injury is commonly referred to as dedifferentiation, implying that adult denervated cells revert to an earlier stage resembling the immature Schwann cells of perinatal nerves (Harrisingh et al., 2004; Jessen and Mirsky, 2005; Chen et al., 2007; Woodhoo et al., 2009). It is becoming clear, however, that this view is incomplete. These cells have a different structure, molecular profile, and function. Therefore, the immature cells, generated from Schwann cell precursors during development, and Bungner cells generated in response to adult nerve injury, represent two distinct differentiation states.

In injured nerves, myelinating Schwann cells, that are specialized to support fast conduction of action potentials, transform to Bungner cells that are specialized for the unrelated task of organizing nerve repair. This represents an unambiguous change of function, brought about by the combination of dedifferentiation and activation of an alternative differentiation program, the c-Jun dependent Schwann cell repair program. Transitions that share this set of features have been described in other systems, where they are generally referred to as transdifferentiation (Jopling et al., 2011).

The regeneration defects in the c-Jun mutant are substantially more severe than those reported for other mouse mutants, in spite of the fact that the genetic defect is restricted to Schwann cells. The likely reason is the number and diversity of the molecules controlled by this single transcription factor. Among the 172 molecules that are abnormally expressed in the mutant are growth factors, adhesion molecules, growth-associated proteins, and transcription factors. This allows c-Jun to integrate a broad collection of functions that support nerve regeneration, and therefore to act as a global regulator of the Schwann cell repair program. This program involves regulation of molecules that have been directly implicated in repair such as the surface proteins N-cadherin, p75NTR, and NCAM, and the signaling molecules GDNF, artemin, sonic hedgehog, and BDNF. It also includes the morphogenetic processes that change myelinating cells to process bearing cells forming regeneration tracks, and the conversion of Schwann cells to cells equipped to rapidly clear myelin from injured nerves (Stoll et al., 2002; Chen et al., 2007; Vargas and Barres, 2007; Wang et al., 2008; Gordon et al., 2009; Höke and Brushart, 2010; Angeloni et al., 2011).

The exceptional repair potential of peripheral nerves is likely due to the coordinated functions of the repair program. Yet individual factors can also be presumed to play a prominent role, as exemplified by the enhanced regeneration seen when GDNF and artemin levels are increased in c-Jun mutant facial nerves (Fontana et al., 2012).

c-Jun is absent from Schwann cell precursors, expressed in immature cells in vivo and in cultured Schwann cells, suppressed by Krox-20 on myelination, but rapidly re-expressed at high levels in Schwann cells of injured nerves (Parkinson et al., 2004, 2008; D.K.W., unpublished).

Among potential intracellular activators of c-Jun is the AP-1 transcription complex, of which c-Jun is a key component. AP-1 activity, in turn, is controlled by numerous signals, including the major MAPK pathways Erk1/2, JNK, and p38. These are all activated in injured nerves and therefore potential upstream regulators of c-Jun (Sheu et al., 2000; Myers et al., 2003; Harrisingh et al., 2004; Jessen and Mirsky, 2008: Parkinson et al., 2008; Napoli et al., 2012; Yang et al., 2012). Genetically, the transcription factor Sox2 is not downstream of c-Jun, since Sox2 remains normally upregulated in injured c-Jun mutant nerves (Figure S4).

We described previously that c-Jun shows cross-inhibitory interactions with the pro-myelin transcription factor Krox20 (Parkinson et al., 2008). Mirroring the function of c-Jun in denervated cells, Krox20 is involved in the regulation of 100–200 genes in myelinating Schwann cells (P. Topilko, personal communication) and is required for the normal activation of the myelin program. We therefore suggest that Krox20/c-Jun are central components of a cross-inhibitory switch that regulates cell fate in injured and regenerating nerves.

The long term persistence of Schwann cell lipid droplets and large multivacuolated (foamy) macrophages in transected mutant nerves suggests problems with lipid clearance and macrophage activation and exit. Recent evidence indicates that failure of lipid breakdown may delay regeneration (Winzeler et al., 2011).

The reduced macrophage numbers in the mutant early after injury is unlikely to contribute substantially to the regeneration problems, a conclusion supported by the microfluidic chamber experiments, where axon growth fails in the presence of mutant Schwann cells, even in the absence of macrophages. Even severe depletion of invading macrophages has no effect on the number of myelinated axons in dorsal roots following nerve injury (Barrette et al., 2008).

The disregulated mutant Bungner cell not only fails to support axon regeneration, but also fails to rescue injured neurons from death. In the mutants, injured type B DRG neurons are about twice as likely to die as in WT mice. Even more notable is the death of about a third of type A neurons, because we find no death of these cells in WT animals, in agreement with previous work in mice and other species (Jiang and Jakobsen, 2004). The majority of facial motoneurons also die after facial nerve injury in the mutant (Fontana et al., 2012).

The observation that denervated adult Schwann cells acquire the ability to generate melanocytes, a property of Schwann cell precursors but not of immature Schwann cells (Adameyko et al., 2009), raises an intriguing possibility. Namely that after injury, Schwann cells dedifferentiate past the immature Schwann cell stage to a cell that shares some properties in common with the Schwann cell precursor.

c-Jun is not significantly expressed in Schwann cell precursors (D.K.W., unpublished). It is therefore possible that the unique identity of the Bungner repair cell in adult nerves consists of a c-Jun-activated repair program in a cell that in significant other aspects has dedifferentiated more completely than hitherto envisaged.

It is clear that the transdifferentiation of myelinating cells to Bungner cells is central to nerve repair. But much remains to be learned about the twin components of this process, the dedifferentiation and repair programs, and about the molecular links that integrate them. This includes issues of practical importance such as the identification of methods to sustain expression of the repair program over the long periods required for nerve repair in humans, and the question of whether the repair program can be activated in other glial cells.

Sciatic nerves of adult mice were cut or crushed at the sciatic notch.

RNA was extracted, cDNA generated and applied to Mouse 430 2.0 array (Affymetrix, MA). Significantly different genes were selected with Bayes’ t test. After control for false discovery rate, genes with a p value of less than 0.05 were filtered out. The microarray data are MIAME compliant.

This was performed as described (Lee et al., 1997).

QPCR was performed with Sybrgreen SYBR Green JumpStart (Sigma) and carried out using Chromo4 Real Time Detector (Bio-Rad). For primers see Table S5. Data was analyzed using Opticon monitor 3 software and fold-changes determined with the Livak method (see Supplemental Information).

Adenovirus expressing c-Jun was generated by cloning mouse c-jun sequence into pAdTrack-CMV. This was recombined into adenoviral backbone plasmid pAdEasy-1 in bacteria. Schwann cell cultures (Dong et al., 1999) were infected with purified adenoviral supernatants (Parkinson et al., 2001, 2008).

Nerve segments, spinal cords or Schwann cell cultures were fixed in paraformaldehyde (PF)/PBS for 10 min–2 hr. Sections were fixed in 2% or 4% PF/PBS for 10 min or methanol for 30 min prior to immunolabeling. Alternatively, nerves were fixed in PF/PBS for 24 hr and wax embedded. Four micrometer sections were deparaffinized and antigen retrieved prior to immunolabeling. Blocking solution was used before incubation with primary antibodies overnight followed by secondary antibodies for 30 min to 1 hr. The first layer was omitted as a control.

The nerve pinch test was used to assess axonal regeneration distance in vivo. Sensory motor coordination was assessed using mouse footprints to calculate the sciatic functional index. Sensory function was assessed by Von Frey Hair analysis, the Hargreaves test and response to toe pinching. Motor function was analyzed by observing toe spread (see Supplemental Information).

True Blue (2 μl) was injected into the tibialis anterior muscle at three sites to label motor neurons in spinal cord segments L2 to L6. Seven days later, mice were perfused. Serial 30 μm sections were collected and the number of labeled neurons was counted (Supplemental Information).

The L4 DRG was cryosectioned. DRG neurons (nuclei) were counted as described (Puigdellívol-Sánchez et al., 2000). Ten micrometer serial sections were labeled with Neurotrace fluorescent Nissl green stain. Every sixth section was analyzed and systematic random sampling (SRS; see Supplemental Information) applied to ensure unbiased estimation of neuron numbers. A and B cells were differentiated on size and morphological criteria as described (Tandrup et al., 2000). For further confirmation, A cells in 10 week cut WT and mutant DRG were quantified by nucleolar counts (Jiang and Jakobsen, 2010). Both nuclear and nucleolar counts were corrected as described in Abercrombie (1946). Schwann cells and macrophages in injured tibial nerves were counted in whole transverse sections in the electron microscope using SRS (see Supplemental Information).

Following PF fixation, 10 μm sections were treated with 2% OsO₄-PBS solution overnight. Percentage stained nerve area relative to that in uninjured nerves was quantified using NIH ImageJ.

Frozen nerve samples or cell lysates were blotted as described (Parkinson et al., 2004).

Using a three-compartment microfuidic chamber (Taylor et al., 2005), 5,000 adult DRG neurons were plated in the central compartment in defined medium with 50 mM glucose (Dong et al., 1999). 2 × 10⁵ WT Schwann cells, c-Jun null cells or c-Jun null cells infected with c-Jun adenovirus were plated in the side chambers. The number of axons longer than 50 μm growing into the side compartment was counted. Alternatively the area of axonal elongation into the side chambers from the microgrooves up to 300 μm beyond them was measured using ImageJ.

Macrophage migration was assessed using 6.5 mm Transwells with 5 μm pores (Corning Costar; see Supplemental Information).

Data are presented as arithmetic mean ± standard error of the mean (SEM) unless otherwise stated. Statistical significance was estimated by Student’s t test, two-way ANOVA, or Mann-Whitney U-test.