Local axonal protection by WldS as revealed by conditional regulation of protein stability

Significance

We investigated the molecular basis of Wallerian degeneration slow (WldS) axon protection by defining a spatial and temporal requirement of WldS activity to promote axonal survival. The findings carry clinical significance in showing that the cellular mechanisms governing degeneration of mammalian axons exhibit a latency period before the axon is irreversibly committed to degradation. Moreover, our findings indicate that the progression of Wallerian axon degeneration can be attenuated or halted altogether even long after an injury has occurred. We believe these findings narrow the molecular mechanisms responsible for orchestrating axonal degeneration to events within the axonal compartment and help identify therapeutic targets to halt or reverse the course of axon degeneration in neurological injuries and diseases.

Abstract

The expression of the mutant Wallerian degeneration slow (WldS) protein significantly delays axonal degeneration from various nerve injuries and in multiple species; however, the mechanism for its axonal protective property remains unclear. Although WldS is localized predominantly in the nucleus, it also is present in a smaller axonal pool, leading to conflicting models to account for the WldS fraction necessary for axonal protection. To identify where WldS activity is required to delay axonal degeneration, we adopted a method to alter the temporal expression of WldS protein in neurons by chemically regulating its protein stability. We demonstrate that continuous WldS activity in the axonal compartment is both necessary and sufficient to delay axonal degeneration. Furthermore, by specifically increasing axonal WldS expression postaxotomy, we reveal a critical period of 4–5 h postinjury during which the course of Wallerian axonal degeneration can be halted. Finally, we show that NAD⁺, the metabolite of WldS/nicotinamide mononucleotide adenylyltransferase enzymatic activity, is sufficient and specific to confer WldS-like axon protection and is a likely molecular mediator of WldS axon protection. The results delineate a therapeutic window in which the course of Wallerian degeneration can be delayed even after injures have occurred and help narrow the molecular targets of WldS activity to events within the axonal compartment.

Axon degeneration is a characteristic event in many neurodegenerative conditions including stroke, glaucoma, and motor neuropathies. Remarkably, expression of the Wallerian degeneration slow (WldS) transgene delays nerve degeneration in these events, and the protection is conserved across many species, including rats (1), Drosophila (2, 3), and even in human neurons (4). Thus, identifying the molecular components of the degeneration pathway with which WldS interferes provides a window of opportunity to understand how axons are normally lost after injury.

The WldS mutation results in the formation of a chimeric gene product consisting of the N-terminal 70 amino acids of ubiquitination factor 4B (Ube4B), which contains no enzymatic activity, and the full functional sequence of a NAD⁺ synthetic enzyme, nicotinamide mononucleotide adenylyltransferase (Nmnat1) (5). The Ube4B portion in WldS contains a binding site for valosin-containing protein (VCP) (6), a cytoplasmic protein with diverse cellular functions (7). Both this VCP-binding domain and the enzymatic activity of Nmnat1 are required for WldS-mediated axon protection (8, 9). Although the WldS protein is localized predominantly in the nucleus because of the endogenous nuclear localization of Nmnat1, trace amounts of WldS protein also have been identified in extranuclear compartments in the axoplasm and in axonal organelles including the mitochondria and phagosomes (10, 11), suggesting that the N-terminal Ube4B region of WldS partially redistributes the nuclear Nmnat1 to the axon.

Despite the remarkable phenotype, little is known regarding the mechanism by which WldS protein delays axon degeneration. The different regions of WldS expression in the neuron have given rise to competing theories regarding the location within the neuron at which WldS exerts axonal protection. One hypothesis proposes that WldS mediates protection by increasing nuclear NAD⁺ levels and thus regulates global gene expression to confer resistance to axonal degeneration. Consistent with this hypothesis, an earlier study showed that, compared with WT neurons, WldS neurons express elevated levels of the NAD⁺-dependent deacetylase Sirt1 and that overexpression of Sirt1 in WT neurons robustly delayed axonal degeneration (12). An alternate hypothesis proposes that the extranuclear WldS expression provides ectopic Nmnat activity in the axon to mediate the axon protection. Consistent with this alternate hypothesis, several studies have shown that misexpression of Nmnat1 alone outside of the nucleus by deleting its nuclear localizing sequence (13, 14), virally transducing Nmnat1 in injured axons (15), or fusing it to the N-terminal sequence of APP protein to increase expression in axonal compartments (16) leads to robust axon protection comparable to that of WldS neurons.

To assess the relative contribution of WldS-mediated gene-expression changes (17) and local axonal WldS enzymatic activity in axonal protection, we adapted a ligand-based genetic and chemical method for temporally regulating posttranslational protein stability (18) to tune the expression of WldS protein. This ligand-based tool allows us to control the timing in which the axonal pool of WldS is expressed before or after axotomy. Using this approach, we show that WldS axonal protection requires local, continuous Nmnat enzymatic activity in the axonal compartment and is independent of nuclear gene transcription. Moreover, we reveal a critical window of 4–5 h after axonal injury during which the course of degeneration can be halted in injured axons. Finally, we demonstrate that increased production of NAD⁺ is sufficient to rescue axon degeneration, supporting its role as a molecular mediator of WldS axon protection.

Results

Differential mRNA Expression by WT and WldS Neurons.

Given the enrichment of WldS protein in the neuronal nucleus (Fig. S1), we suspect that WldS may mediate axonal protection by regulating new gene transcriptions. Because WldS delays degeneration even in axotomized nerves where there is no communication between the axon and cell body, we reason that gene changes that promote axonal survival, if any, must occur before the onset of injury. We therefore profiled the mRNA expressions of acutely isolated retinal ganglion cells (RGCs) in WT and WldS mice and in WT and WldS rats, in which WldS also has been shown to confer robust axonal protection (1). Although we identified a restricted subset of genes that are differentially expressed in both mouse and rat WT and WldS neurons, we found no easy correlation or overlap in the candidate genes between the mouse and rat datasets (Fig. S1). Thus, despite the predominantly nuclear localization of WldS protein and restricted gene changes in WldS neurons, WldS axon protection is unlikely to require new gene transcription.

Fig. S1.

Expression profiling of WT and WldS neurons. (A and B) Live images of RGCs transfected with the control PN1-EGFP construct and merged with brightfield show diffuse fluorescence in the cytoplasm and axons. (C and D) Live images of RGCs transfected with PN1-WldS-EGFP and merged with brightfield show predominantly nuclear localization of WldS protein. (E) Genomic analysis of postnatal day 5 mouse WldS and WT RGCs. The scatter plot of the microarray analysis shows a restricted subset of genes that change in either direction between WldS and WT mouse RGCs. (F) List of top differentially expressed gene candidates in WT and WldS RGCs in mouse and rat. No easy overlap is found in the top up-regulated or down-regulated gene candidates between expression-profiling datasets in mice and rats.

Temporal Control of WldS Protein Expression by Regulating Protein Stability.

An alternate account of WldS axon protection is that the smaller pool of WldS protein in the axon locally mediates axonal protection independent of the soma. Consistent with this alternate account, previous reports identified small but detectable levels of WldS in the axonal compartment, and mistargeting Nmnat1 outside the nucleus significantly delayed Wallerian degeneration (13, 14, 16). However, these experiments still involve continuous communications with the neuronal cell body. To test whether the axonal fraction of WldS is required for axonal protection, we adapted a method to regulate the temporal expression of WldS protein in the axon. Because no cargo transport or communication can occur between the soma and axonal compartments after axotomy, depleting or increasing WldS levels in the axon after axotomy and monitoring the subsequent effect on axon protection should allow us to assess definitively the spatial requirement of axonal WldS pool in delaying axon degeneration.

We therefore took advantage of a previously developed ligand-based system that allows temporal regulation of protein activity by altering the stability of the protein of interest (Fig. 1 A and B). In this system, fusion of a destabilizing domain (DD), a short peptide sequence from FKBP12 protein, confers protein-folding instability to the fused protein of interest and leads to rapid degradation of the attached protein by the endogenous proteasome system in the cell. The process is reversed by the addition of a cell-permeable ligand (Shield-1) that binds to and inhibits the activity of the DD (18).

Fig. 1.

Controlling the temporal expression of WldS by regulating protein stability. (A) Schematic of the FKBP12/Shield-1 system. Expression of a DD fused to a protein of interest confers structural instability to the fused protein and leads to rapid degradation of the protein by the ubiquitin–proteasome system. However, this process can be inhibited by adding a cell-permeable stabilizing ligand (Shield-1). (B) A fusion construct consisting of the DD and WldS flanked by a reporter GFP construct (DD-WldS-IRES-EGFP) is generated and packaged into lentivirus for efficient (>92% transduction) expression in rat DRG neurons. (C) Whole rat DRG neurons expressing DD-WldS are collected 45 min to 1 h after the addition or removal of Shield-1 in culture medium. WldS protein levels subsequently are probed with WldS-18 antibody. The expression of destabilized WldS protein is attenuated in the absence of Shield-1 ligand (−Shield1), but the presence of Shield-1 (+Shield1) rapidly stabilizes the WldS protein and increases WldS protein expression 30–45 min after the addition of Shield-1. This effect is abolished again by withdrawing Shield-1 from medium 30–45 min after medium replacement (washout). β-Actin control indicates equal loading of protein lysates. (D) Representative images of severed distal axons in various conditions at selective time points after axotomy. Shield-1 ligand is continuously present, continuously absent, applied at time of axotomy, or withdrawn at time of axotomy. The effect of regulating WldS protein stability, and therefore WldS protein expression, was measured at 72 h postinjury. Corresponding images at 72 h postaxotomy and binarized by computer software to white (background) and black (axonal material) pixels before analysis of the degeneration index (“72hr Binarized”) are shown for easier visualization of axonal integrity. The presence of Shield-1, and thus the stabilization of WldS protein activity, at time of axotomy results in robust axonal protection at 72 h after axotomy, whereas the withdrawal of Shield-1, and thus the destabilization of WldS protein, at time of axotomy results in axon degeneration at 72 h after axotomy. The lack of axon protection following destabilization of WldS protein after axotomy demonstrates that the axonal pool of WldS protein is required for axonal protection. (Scale bar, 100 um.) (E) Quantification of axonal survival in various conditions at selective time points after axotomy. As a control, Shield-1 alone in the absence of DD-WldS expression exhibits no independent axonal-protective effect. Results are quantified and graphed ± SEM. *P < 0.005 by Student t-test.

To assess whether expression of WldS protein can be reversibly regulated by the DD-stabilizing ligand Shield-1, we infected cultured embryonic day 15 (E15) rat dorsal root ganglion (DRG) neurons with a lentivirus carrying the DD-WldS-EGFP expression construct. In the absence of Shield-1 ligand, WldS protein levels are low to undetectable, as verified by both EGFP fluorescence and Western blot. In contrast, both EGFP fluorescence and WldS protein expression are elevated significantly when Shield-1 is added to the medium. The WldS protein levels returned to basal levels as soon as 30 min after the medium was replaced with fresh medium containing no Shield-1 ligand (Fig. 1C). Thus, this method rapidly and reversibly alters the posttranslational levels of WldS protein and provides a potent means to control WldS protein expression temporally in the axon.

Local WldS Activity in the Axon Is Required to Delay Axonal Degeneration.

To determine whether the trace WldS pool in the axon is required for axonal protection, we incubated soma-free axonal segments (from DRGs expressing DD-WldS-IRES-EGFP) with or without Shield-1 ligand at selective time points after axotomy and quantified subsequent axonal survival. We reason that increasing axonal WldS expression after axotomy should fail to exert protection of the distal axon if the nuclear/soma fraction of WldS is responsible for conferring protection. On the other hand, continued protection of the distal axon with increased WldS expression after axotomy indicates that the axonal fraction of WldS is critical for axon protection.

We observe that in the absence of Shield-1 ligand, the distal axons degenerate rapidly at a rate and in a manner similar to injured WT axons. Notably, a small amount of “leaky” WldS expression is still detected in the absence of Shield-1, although the level of protein expression is insufficient to exert axonal protection (Fig. 1 C and D). Interestingly, the addition of Shield-1 at the time of axotomy was sufficient to confer protection to the distal axon for up to 72 h in vitro, comparable to the degree of protection seen in WldS mutants (Fig. 1 D and E). Furthermore, initially incubating DRGs with Shield-1 but later withdrawing Shield-1 at the time of axotomy completely abolished subsequent axonal protection (Fig. 1 D and E). Western blots for WldS protein harvested from axonal lysates after the addition or withdrawal of Shield-1 show corresponding changes in WldS protein level, suggesting that the phenotypes are consistent with changes in WldS expression levels in the axon (Figs. 2B and 3B). Together, the results demonstrate that increasing axonal WldS expression is sufficient to rescue axons otherwise destined to undergo degeneration even after injury has occurred and that axonal fraction of WldS is required continuously to delay axonal degeneration.

Fig. 2.

Wallerian axon degeneration is reversible within 4–5 h after injury. Rapid stabilization of WldS protein postinjury delineates a critical window for halting the course of axonal degeneration. (A) Representative images of severed distal axons at selective time points after axotomy. The Shield-1 ligand is applied to DRGs expressing DD-WldS at 0, 2, 4, 6, and 8 h after axotomy. The effects of Shield-1 application on distal axotomized axons were monitored every 12–24 h. (Scale bar, 100 um.) (B) Western blots probing for the axonal pool of WldS protein from distal DRG axonal lysates collected 45 min to1 h after the addition of Shield-1 at selective time points after axotomy. β-Actin bands indicate similar protein loading. (C) Quantification of axonal survival following the addition of Shield-1 at selective time points after axotomy. Stabilizing WldS protein between 0–4 h after axotomy resulted in robust axonal protection up to 108 h postinjury. However, stabilizing WldS protein at any time point beyond 6 h after axotomy failed to confer axonal protection. Stabilizing WldS protein 5 h after axotomy resulted in intermediate protection of distal axons in comparison with WldS protein stabilization at 0–4 h postinjury. The data indicate that local WldS activity in the axon can reverse the course of axon degeneration if present within 4–5 h postinjury. Results are quantified and graphed ± SEM.

Fig. 3.

Continuous WldS activity is required to sustain axonal protection. Rapid destabilization of WldS protein postinjury shows that continuous WldS activity is required to delay axonal protection. (A) Representative images of severed distal axons at selective time points after axotomy. Previously applied Shield-1 ligand was withdrawn from DRGs expressing DD-WldS at 0, 2, 4, 6, and 8 h after axotomy. The effects of Shield-1 withdrawal on distal axotomized axons were monitored every 12–24 h. (Scale bar, 100 um.) (B) Western blots probing for the axonal pool of WldS protein from distal DRG axonal lysates collected 45 min to 1 h following the withdrawal of Shield-1 at selective time points after axotomy. β-Actin bands indicate equal protein loading. (C) Quantification of axonal survival following the withdrawal of Shield-1 at selective time points after axotomy. Destabilizing WldS protein at all time points tested resulted in axonal degeneration postinjury, indicating that continuous WldS activity in the axon is required to sustain axonal protection. Results are quantified and graphed ± SEM.

Rapid Stabilization of WldS Protein Reveals a Critical Window to Rescue Axonal Degeneration.

We next asked whether we can further extend the window to halt the course of axon degeneration by applying Shield-1 ligand and thus increasing WldS protein levels at selective time points after axotomy. Surprisingly, we found that the addition of Shield-1 at any time within 4 h postinjury is sufficient to rescue axonal degeneration. However, no axonal protection occurs if WldS is stabilized at any point more than 4 h after the time of axotomy, regardless of the concentration of Shield-1 applied (Fig. 2 A and C). Thus, the axons remain protected so long as Shield-1 and corresponding WldS levels (Fig. 2B) are increased within the 4 h postinjury period, indicating a critical window after nerve injury during which the course of axon degeneration can be halted.

Conversely, we find that no protection occurs when Shield-1 is withdrawn from the medium at any time after axotomy, regardless of the duration of the neurons’ previous incubation with Shield-1 (Fig. 3 A and C). Notably, the time from the withdrawal of Shield-1 to the initial morphological sign of fragmentation remains fairly consistent (∼12 h). Both results above indicate that continuous WldS activity delays the triggering of the cellular commitment to degeneration rather than merely prolonging the duration of morphological fragmentation. Thus, our data show that continuous WldS activity in the axon is required to delay or prevent cellular commitment to degeneration, but only if WldS activity is present within a 4- to 5-h critical window after the onset of injury.

NAD⁺ Is Sufficient for WldS Axonal Protection.

We next sought to determine the critical downstream signals or molecular mediators that translate WldS enzymatic activity into axonal protection. Previous studies have shown that preincubation with excess extracellular NAD⁺, the enzymatic product of Nmnat, is sufficient to delay axon degeneration (12, 19). Having demonstrated a local requirement for WldS activity in the axon to delay degeneration, we asked whether application of NAD⁺ to soma-free axonal segments after axotomy is also sufficient to delay or halt axon degradation from traumatic injury.

We found that application of 1–5 mM NAD⁺ to WT DRG cultures immediately after axotomy significantly delayed axonal degeneration (Fig. 4 B and C). The ability of extracellular NAD⁺ to delay axon degeneration efficiently when added after axotomy suggests that locally supplied NAD⁺ in the axon is sufficient to prevent axon degeneration. However, a very high concentration of NAD⁺ (>1 mM) is required to exert axon protection, and the overall effects remain weaker than in WldS neurons.

Fig. 4.

Exogenous addition of NAD⁺ or its precursor NMN is sufficient to protect axons in vitro and is sufficient to rescue axon degeneration from depletion of WldS protein. (A) Schematic of the NAD⁺ synthesis and salvage pathway in mammals. Nicotinamide (NAM) is taken up and converted into NMN by the NAMPT enzyme. NMN then is converted into NAD⁺ by the NMNAT1 enzyme. NAD⁺ can be metabolized further into cADP ribose (cADP) by the CD38 enzyme. Finally, NAD⁺, in its reduced form NADH, can be phosphorylated (NADP) by NADH phosphorylase. (B) Representative images of severed distal axons at 60 h after axotomy following the addition of 1 mM NAD⁺ precursors or metabolites. (Scale bar, 100 um.) (C) Quantification of axonal survival following the addition of 1 mM of known NAD⁺ precursors or metabolites to DRG cultures at time of axotomy. Survival of distal axons was assessed 60 h after injury. Only the addition of NAD⁺ or its immediate precursor NMN resulted in significant axon protection 60 h after axotomy. Results are graphed ± SEM. *P < 0.01 by Student t-test. (D) Representative images of severed distal axons at selective time points after axotomy. Previously applied Shield-1 ligand was withdrawn from DRGs expressing DD-WldS at the time of axotomy. Then 1 mM NAD⁺ was added to the medium at 0, 2, 4, 6, or 8 h after axotomy. The effects of NAD⁺ on distal axotomized axons depleted of WldS protein were monitored after 60 h. (Scale bar, 100 um.) (E) Quantification of axonal survival following the addition of 1 mM NAD⁺ after depleting WldS in axotomized neurites. The addition of NAD⁺ within 0–4 h after axotomy was sufficient to rescue axonal degeneration despite depletion of WldS protein, but no axonal protection was observed if NAD⁺ was applied more than 6 h after injury. The results indicate that NAD⁺ rescue of axon degeneration is limited to the same critical window as WldS protein. Results are quantified and graphed ± SEM. **P < 0.005 by Student t-test.

Furthermore, to examine if other precursors and metabolites in the NAD synthetic pathway also are capable of delaying axonal degeneration, we added excess amounts of known NAD⁺ precursors and metabolites to cultured DRG neurons at the time of axotomy. We found that addition of at least 1 mM nicotinamide mononucleotide (NMN), the immediate precursor of NAD and a substrate of Nmnat enzyme, is sufficient to exert axonal protection, but the addition of other immediate NAD⁺ metabolites, including cyclic ADP, NADP, or an earlier NAD⁺ precursor, nicotinamide, is not (Fig. 4 B and C). That the addition of only NAD⁺ or its immediate precursor NMN, but not other NAD metabolites, resulted in axonal protection demonstrates that NAD⁺ is specifically capable of delaying the course of axonal degeneration after injury.

To assess if NAD⁺-mediated axonal protection, like WldS-mediated axon protection, also requires activity within the critical window of 4–5 h after injury, we applied NAD⁺ to neurons expressing DD-WldS after WldS had been depleted by Shield-1 withdrawal at the time of axotomy. Interestingly, NAD⁺ was sufficient to delay axonal degeneration, but only if added within 4 h after axotomy (Fig. 4 D and E), consistent with the critical window previously identified for WldS-mediated rescue of axon degeneration. Together, the results indicate that NAD⁺ is specific and sufficient to confer local axonal protection and that the mechanism requires the same 4- to 5-h critical period from initial injury as WldS. Our data thus provide supporting evidence that NAD⁺ and WldS use similar pathways in delaying axonal degeneration after injury.

Discussion

Regulating WldS Protein Stability to Reveal That Local Axonal WldS Activity Is Required to Delay Axonal Degeneration.

Despite the potent phenotype of WldS-mediated axon protection, the subcellular site of WldS activity that confers axonal protection remains controversial. Although prior studies have shown that targeting WldS or its enzymatic domain Nmnat1 to axons was sufficient to protect against axonal degeneration (13, 15, 16), direct evidence for the requirement of the axonal pool of WldS remained scant (20) because a method to deplete WldS from axons selectively and directly without affecting cell body expression was lacking.

Here, we adapted a previously described chemical–genetic approach to control the temporal expression of WldS in the distal injured axon (18, 20). Because no nucleus–axon communication can occur after axotomy, we also confine the changes in WldS spatial expression to within the distal axon. We show that, despite the abundant expression of WldS in the nucleus, it is the smaller axonal pool of WldS that is critical to mediate its axon-protective effect. We further demonstrate that continuous expression of WldS is required for axonal protection, because depleting WldS levels at any time point after axotomy results in rapid axonal degeneration. Thus, the approach described here provides definitive evidence that axonally localized WldS confers protection from Wallerian degeneration and that downstream effectors of WldS activity are not in the nucleus, as previously suggested (12), but are localized in the axon. The technique allows the WldS protein in neurons to be regulated in a temporally and spatially restricted manner and provides a method to establish the subcellular role of the axonal pool of WldS protein.

Identification of a Critical Window for Reversing Axonal Degeneration.

Using the same method, we unexpectedly reveal a critical period of ∼4 h after nerve injury in which the course of Wallerian axon degeneration is reversible. This course of axonal degeneration can be halted so long as WldS activity is continuously present in the axon within 4 h postinjury. At the same time, the ability of NAD⁺ also to rescue axon degeneration within the same protective window suggests the likely existence of a parallel, intrinsic “degeneration clock” that irreversibly commits the axon to degeneration. Increasing WldS activity or the NAD⁺ level within this window thus likely delays the commitment phase so long as WldS enzymatic activity or NAD⁺ is continuously present in the axon.

The ability to increase WldS activity even in severed axons raises a question regarding the source of WldS protein that the Shield-1 ligand stabilizes. Notably, in our system WldS expression in the absence of Shield-1 remains barely detectable, but the addition of Shield-1 significantly and rapidly raises WldS levels even in axotomized neurites (Fig. 2B). The increased WldS expression with Shield-1 addition at further time points (i.e., at 4 h) despite the lack of new protein supply from the soma within transected axons suggests that existing pool of axonal WldS alone is unlikely to account for the source of increased WldS protein and indicates that new WldS protein may be locally translated within the axon. A previous report using a similar Shield-1 system to regulate protein stability showed no evidence of additional WldS protein when stabilized after axotomy (20), although the study was performed using harvested protein samples within a microfluidics chamber and thus limited the protein quantity. Greater protein samples were harvested using the current explant method, and thus this method may be more sensitive to changes in WldS protein levels within the collected axonal fraction.

Indeed, recent studies have identified numerous axonal proteins that are locally translated (21, 22), although currently no report has shown that the Nmnat isomers or WldS is locally translated in the axon. Another study reports the transfer of ribosomes from Schwann cells to peripheral nervous system axons (23); however, our culture is devoid of mitotic glial cells, and our results are unlikely to be explained by the non–cell-autonomous transfer of ribosomal material. Our study thus suggests that even though the bulk of the preexisting pool of WldS is degraded, a small amount of WldS protein may be newly translated in the axon and stabilized in sufficient in quantity to confer axonal protection. Future experiments looking at DD-WldS expression while simultaneously applying translational inhibitors after axotomy would help address the relative contribution of local protein translation in axonal protection.

NAD Is a Mediator of WldS Protection.

Because all three Nmnat isoforms contain the highly conserved catalytic domain for the synthesis of NAD⁺ (24), the common metabolite of Nmnat enzyme activities, one potential mechanism for Nmnat-dependent axon protection is through its known role in NAD⁺ biosynthesis. Consistent with the requirement for functional Nmnat activity, which metabolizes NMN to NAD⁺, for axon protection, we showed that NAD⁺ is specifically capable of providing local axonal protection to distally severed axons. This finding is surprising, given recent reports of NMN acting as a trigger of axonal degeneration. In the report it was shown that FK866, a presumed chemical inhibitor of the nicotinamide phosphoribosyltransferase (NAMPT) enzyme, which catalyzes the synthesis of NMN, protected against axonal degeneration and that this axonal protection was reversed by the addition of NMN (25). Moreover, NMN was shown to accumulate before the onset of axonal degeneration. It is unclear, however, if FK866 has additional targets in addition to NAMPT enzyme in exerting axonal protection. It is possible that, depending on the condition, NMN is able to sustain axonal survival by promoting NAD⁺ levels and raising bioenergetics but also acts independently to trigger separate mechanisms leading to axonal destruction.

Nonetheless, our results show that, like WldS protein, NAD⁺ is able to delay axonal degeneration as long as it is present in sufficient amounts within the same critical period and indicate that WldS acts, in part, through NAD⁺ to delay the course of axonal degeneration locally.

Mechanism of WldS Axon Protection.

How does local axonal WldS exert protection against axon degeneration? Our data support a “degeneration clock” that triggers an irreversible commitment to degeneration 4–5 h after neuronal injury. On a molecular level, the timing to degeneration may be explained by a certain rate of decline or turnover of an axon-survival factor, which triggers the irreversible commitment to degeneration when its level falls below a critical threshold. Consistent with this hypothesis, a prior report identified the cytoplasmic Nmnat isoform Nmnat2 as a labile axonal survival factor necessary and sufficient for axonal maintenance (26). This Nmnat isomer is highly expressed in the neuron, and its expression decreases dramatically after axotomy. Interestingly, our identification of a critical window to rescue axon degeneration corroborates with the half-life of Nmnat2 (26), lending support to the notion that WldS functions to substitute for Nmnat2 depletion. We surmise that the inability of WldS activity to delay axon degeneration after 4–5 h may be explained by the parallel depletion of Nmnat2 activity after 5 h, which triggers a cellular commitment to degeneration. WldS therefore may function to substitute for Nmnat2 before its enzymatic activity drops below the threshold.

Together, our findings show that protection from Wallerian degeneration requires continuous enzymatic activity from axonal pools of WldS protein. We further show that both a postinjury rise in WldS activity and exogenous addition of NAD⁺ are sufficient to rescue axonal degeneration. The ability of NAD⁺ to rescue degeneration from the depletion of WldS suggests that NAD⁺ acts downstream of WldS activity and may be a specific molecular mediator of WldS axon protection. Finally, we define a critical window of 4–5 h to reverse degeneration following nerve injury in mammalian axons; this window is consistent with the known rate of depletion of a previously identified axonal survival factor. WldS thus may act to delay the trigger of such a degeneration clock, putatively by substituting for the activity of a soma-derived survival factor such as Nmnat2. Moreover, our results confine the molecular targets of WldS and/or Nmnat2 to within the axon and thus will help identify potential therapeutic targets to slow or prevent axon degeneration in neurological injuries and diseases.

Materials and Methods

Animal Work.

All procedures were approved by the Stanford University Administrative Panel on Laboratory Animal Care (APLAC) under protocol 10726. Rats were euthanized with CO₂ or rapid decapitation (embryos).

Construction of Lentiviral Expression Plasmids.

The lentiviral vector containing IRES-GFP reporter (FUIGW) was a generous gift from Ricardo Dolmetch (Stanford University, Stanford, CA). WldS plasmid in the PN-1 expression vector was a generous gift from Michael Coleman (Cambridge University, Cambridge, UK). The DD plasmid and Shield-1 ligand were generous gifts from Tom Wandless (Stanford University, Stanford, CA). The DD domain was first subcloned into the FUIGW vector by adding flanking restriction sites via PCR. WldS then was subcloned immediately 3′ of the DD plasmid by similar addition of complementary restriction sites via PCR to generate DD-WldS. The following primers were used to clone DD and WldS plasmids sequentially into the FUIGW vector: DD: forward, 5′-TAATATGGATCCATGGGAGTGCAGGTGGAAACCA-3′, reverse, 5′-TCTTCGATGTGGAGCTTCTAAAACCGGAAGCTAGCGATGATGAATTCTAATAT-3′; WldS: forward, 5′-TAATATTCTAGAGCCACCATGGGAGTGCAGGTGGAAACCA-3′, reverse, 5′-CCAAGCACAACCATTCCACTCTGCCGGATCCTAATAT-3′.

Lentiviral Infection of DRG Neurons.

For lentivirus generation, the DD-WldS plasmid was transfected into HEK 293T cells using Lipofectamine 2000 per the manufacturer’s instructions (Life Technologies). After 48 h, supernatant containing virus particles was spun down and extracted. For lentiviral infection of DRG neurons, lentivirus (∼100 cfu) was added to an individual well of a six-well plate containing 200,000–250,000 DRG neurons per well. Transgene expression from the lentivirus was allowed to proceed for 4–10 d before the infected neurons were used for experiments. The efficiency of viral transduction and transgene expression were monitored, where applicable, by GFP reporter fluorescence via fluorescent microscopy.

Cell Culture and Transfection.

DRG neurons were dissociated from E15 Sprague–Dawley rats as previously described (27) and were maintained for 4–7 d in culture in NeuroBasal Medium Eagle with B27 and NGF.

HEK 293T cells and Neuro2A cells were cultured in DMEM (Life Technologies) containing 10% (vol/vol) FBS and l-glutamine. All cells conditions were visualized by conventional fluorescence microscopy coupled to an inverted Eclipse microscope (Nikon Corp.).

For microarray analysis, mouse RGCs from the postnatal day 7 WT C57/BL6 mouse strain and age-matched WldS mice bred for >10 generations in the C57/BL6 background were acutely purified by sequential immunopanning to 99.5% purity as described previously (28). Rat RGCs from WT Sprague–Dawley rats and from WldS transgenic rats (a generous gift from Michael Coleman) maintained in the Sprague–Dawley background were acutely purified using same method.

RNA Preparation, Microarray Hybridization, and Data Analysis.

Total RNA from WT and WldS mouse or rat RGCs was extracted using the RNeasy kit (Qiagen). RNA quality was assessed by spectrophotometry and gel electrophoresis. Total RNA (1 μg) was reverse transcribed into cDNA, amplified once by in vitro transcription, reverse transcribed into cDNA, and then used to generate biotin-labeled cRNA, according to Affymetrix protocols. Fragmented cRNA was hybridized onto mouse Genome Array 430 2.0 or rat Genome Array 230 2.0 gene chips (Affymetrix). Scanned output files were analyzed with Microarray Suite 5.0 (Affymetrix) and by ANOVA. Further analyses were performed using Microarray Suite 5.0 and Excel (Microsoft) for ANOVA confirmation of statistical reproducibility.

Shield-1 Regulation of WldS Protein Stability.

Dissociated DRGs from E15 rats (∼250,000 cells per well) were plated onto cell-culture plates. Five days after plating, DRGs were infected with lentivirus expressing DD-WldS-EGFP or control vehicle. At 7 d after lentivirus infection, Shield-1 (20 nM) was added to fresh culture medium. The axons were transected in the cell-culture plates using a sterile surgical knife, and the cell bodies subsequently were suctioned off after the axons around the soma were circumferentially excised. Unless otherwise indicated, medium in all wells was replaced after 48 h with fresh culture medium containing Shield-1 or vehicle. Following axotomy, phase-contrast images of DRG axons were acquired from an inverted Nikon microscope (Nikon Corp.) with a 20× objective.

Quantification of Axon Survival.

Dissociated DRG neurons were treated with the indicated Shield-1/chemical agents and imaged using phase-contrast microscopy with a 20× lens at the indicated time points. A grid was created over each image using ImageJ software with publically available grid plugin software (NIH), and images were binarized into black (axonal segments) and white (background) pixels. Axonal segments were considered degenerated if they showed evidence of swellings and/or blebbing representing discontinuous particles on the image. The cell-counting plugin was used to score particles representing degenerated axonal fragments, and the resultant values then were divided over total axonal surface area to generate a degeneration index. The inverse of the degeneration index was taken to generate an axonal survival index. Degenerating and healthy axons were counted in at least three high-power fields per image (∼20 neurites) for each well. Data analysis was performed blind to experimental conditions.

NAD⁺ Metabolite Assay.

Dissociated DRG cultures were plated as described above. When used, all nucleotides (Sigma-Aldrich) and metabolites of NAD⁺ (Sigma-Aldrich) were added at the time of axotomy at 1 mM unless otherwise indicated in the text.

Western Blot Detection of WldS Protein.

WldS protein was detected using conventional Western blotting techniques. Axonal lysates were lysed, and the proteins were eluted in SDS sample buffer, separated on a polyacrylamide gel at a minimum of 10 μg protein per condition, and subsequently transferred to a PVDF membrane. The membranes were blocked and probed with a 1:200 dilution of rabbit WldS-18 polyclonal antibody (a generous gift from Michael Coleman), and the presence of protein was detected by standard chemiluminescence techniques.

Statistical Analysis.

Data were analyzed with Prism5 (GraphPad) using ANOVA and Student t-tests. P values less than 0.05 were considered significant for any set of data.