InteractiveFly: GeneBrief

axundead: Biological Overview | References


Gene name - axundead

Synonyms -

Cytological map position - 65C1-65C1

Function - signaling - adaptor protein

Keywords - a mediator of axon death - axed mutants suppress axon death - acts in glia downstream of sarm, mutants - pro-degenerative pathways activated by Sarm signaling or Nmnat elimination ultimately converge on Axed - possibly involved in recruitment of substrates to Cullin Ring Ubiquitin Ligase complexes

Symbol - axed

FlyBase ID: FBgn0035708

Genetic map position - chr3L:6,609,325-6,619,113

NCBI classification - BACK (BTB and C-terminal Kelch) domain

Cellular location - cytoplasmic



NCBI links: EntrezGene, Nucleotide, Protein

Axundead orthologs: Biolitmine
BIOLOGICAL OVERVIEW

Axon degeneration is a hallmark of neurodegenerative disease and neural injury. Axotomy activates an intrinsic pro-degenerative axon death signaling cascade involving loss of the NAD+ biosynthetic enzyme Nmnat/Nmnat2 in axons, activation of dSarm/Sarm1, and subsequent Sarm-dependent depletion of NAD+. This study has identified Axundead (Axed) as a mediator of axon death. axed mutants suppress axon death in several types of axons for the lifespan of the fly and block the pro-degenerative effects of activated dSarm in vivo. Neurodegeneration induced by loss of the sole fly Nmnat ortholog is also fully blocked by axed, but not dsarm, mutants. Thus, pro-degenerative pathways activated by dSarm signaling or Nmnat elimination ultimately converge on Axed. Remarkably, severed axons morphologically preserved by axon death pathway mutations remain integrated in circuits and able to elicit complex behaviors after stimulation, indicating that blockade of axon death signaling results in long-term functional preservation of axons (Neukomm, 2017).

Maintenance of the morphological integrity of neurons is essential for sustained nervous system function throughout an animal's lifespan. Nervous system injury or neurological disease leads to axonal and synaptic degeneration and, in turn, loss of neural circuit connectivity and function. Molecular pathways driving axonal degeneration remain poorly defined in any context; however, recent work on Wallerian degeneration (WD) has revealed that axon injury activates an intrinsic, conserved, pro-degenerative (axon death) signaling pathway. Previously identified dSarm/Sarm1 (sterile α/Armadillo/Toll-Interleukin receptor homology domain protein) as a key mediator of axon death signaling. Loss of dSarm in Drosophila, or Sarm1 in mouse, resulted in severed distal axons remaining morphologically preserved for weeks after injury (Gerdts, 2013, Osterloh, 2012), indicating that dSarm/Sarm1 pro-degenerative signaling is an ancient mechanism used by axons to drive self-destruction. How dSarm/Sarm1 signals to execute axon death remains unclear, but dSarm/Sarm1 has recently been linked to the NAD+ metabolic pathway, which based on extensive evidence appears to be a central mediator of axonal integrity (Neukomm, 2017).

The first evidence supporting a role for NAD+ in axon maintenance came from the identification and characterization of the slow Wallerian degeneration (WldS) mouse, where severed axons fibers survived for weeks when detached from their cell bodies. This remarkable neuroprotective phenotype effect was due to a chromosomal rearrangement that led to the generation of the novel WldS molecule, a fusion protein consisting of the NAD+ biosynthetic enzyme nicotinamide mononucleotide adenylyltransferase 1 (Nmnat1) and a short fragment of Ube4b. Consistent with a positive role for NAD+ in sustaining axons, NAD+ levels were found to plummet in axons immediately prior to granular fragmentation. Axon degeneration could be rescued by exogenous NAD+ or its precursors, and axotomy-induced NAD+ depletion was blocked by WldS. Numerous studies have demonstrated neuroprotective roles for NAD+-related metabolites. The current proposed mechanism for activation of axon degeneration after injury is the depletion of Nmnat2, a labile form of Nmnat found in mammalian axons that is normally transported down axons from the soma. Nmnat2 is seen as a critical regulator of axon survival: its half-life approximates the latent phase prior to explosive axon fragmentation, depletion of Nmnat2 from axons induces spontaneous degeneration, and stabilization of Nmnat2 can phenocopy the effects of WldS (Neukomm, 2017).

The second line of evidence supporting a role for NAD+ in axonal protection came from drug screens for molecules that promoted neurogenesis (or neuroprotection) in vivo. P7C3 was identified as an activity-enhancing compound of nicotinamide phosphoribosyltransferase (Nampt), a rate-limiting enzyme in the NAD+ salvage pathway, which likely leads to increased levels of NAD+ in injured axons to help sustain integrity. Derivatives of the P7C3 series have since been shown to be neuroprotective in many models of neurodegenerative disease, and in some models of neural injury . However, whether P7C3 harbors the ability to attenuate axon death remains to be determined (Neukomm, 2017).

Finally, recent work has demonstrated that NAD+ depletion after axotomy is blocked in Sarm1-/- mutants, and dimerization of the Sarm1 TIR domain can drive rapid depletion of NAD+ from cells and axons (Gerdts, 2015). Surprisingly, it appears that the Sarm1 TIR domain harbors endogenous NAD+ hydrolase activity (Essuman, 2017). Thus an emerging stepwise model of WD is that axotomy leads to depletion of labile pools of Nmnat2 in axons detached from their cell bodies; this in turn results in an initial depletion of NAD+ levels due to lack of new NAD+ synthesis and/or salvage, dSarm/Sarm1 signaling is then activated followed by rapid depletion of NAD+ levels below a threshold needed to maintain axonal integrity, and finally explosive degeneration of the axon ensues. NAD+ depletion from axons is a compelling model for WD, although alternative mechanisms cannot be excluded, due to limited understanding of the genetics of axon death signaling. Besides dSarm/Sarm1, only one other molecule, the E3 ubiquitin ligase Highwire/Phr1, has been shown to be required for axon death in vivo: loss-of-function mutations in highwire/Phr1 potently suppress axon death in fly nerve injury models and mammalian sciatic nerve lesion experiments. Highwire/Phr1 likely acts upstream of dSarm/Sarm1, and is required for the normal turnover of Nmnat/Nmnat2. Loss of Highwire/Phr1 is therefore proposed to stabilize Nmnat/Nmnat2 and maintain axonal pools of NAD+. Are there other signaling molecules acting downstream of dSarm/Sarm1 to execute axonal death, or is NAD+ depletion the final step? Sarm1 has also been reported to drive axon death through a downstream MAPK signaling cascade (Yang, 2015), similar to the signaling mechanism used by C. elegans TIR-1 during regulation of odorant receptor expression; however, this remains controversial. A recent contradictory study argues that MAPKs exert their effect by fine-tuning levels of Nmnat2 upstream of Sarm1 signaling (Neukomm, 2017).

This paper reports the identification and characterization of Axundead (Axed), an axonal BTB and BACK domain protein that signals downstream of dSarm. Axed is essential for injury-induced axon death signaling, and axed mutations can fully suppress degeneration induced by activated dSarm, or complete elimination of Nmnat activity from axons. Thus, Axed is a novel pro-degenerative signaling molecule, and the neuroprotective effects of loss of Axed exceed those of dsarm null mutants (Neukomm, 2017).

This study presents the identification and initial characterization of the BTB and BACK domain molecule Axed. Loss of Axed function was sufficient to block axon death for the lifespan of the fly, and axed mutants were neuroprotective in all neurons tested. Axed, like dSarm, appears to function selectively in axon death during WD, as axed mutants blocked neither cell death nor developmental pruning of axons or dendrites, supporting the notion that the Nmnat-depletion/dSarm/Axed signaling pathway is engaged specifically in response to axonal injury. However, it remains possible that while Axed or dSarm elimination is not sufficient to block cell death, redundant genetic pathways might work together with dSarm and Axed in the context of apoptotic cell death (Neukomm, 2017).

Axed encodes a previously uncharacterized BTB and BACK domain protein. The best-characterized role for BTB-containing proteins is the recruitment of substrates to Cullin Ring Ubiquitin Ligase (CRL) complexes for ubiquitin tagging and proteasome degradation, where BTB domains function in homo- or hetero-dimerization while the BACK domain (often with the BTB domain) interacts directly with Cullins. Typically, the C-terminal regions in BTB domains molecules function within CRL complexes to bind target substrates designated either for ubiquitination by the CRL and subsequent degradation by the proteasome or for signaling. The BTB and BACK domains of Axed are required for optimal Axed function in vivo based on their partial abilities to rescue axon death phenotypes in axed mutant backgrounds. In contrast, the C terminus appears to be absolutely essential for Axed function as AxedδCterm completely fails to rescue axon death. This is not likely due to destabilization of the molecule lacking the C terminus, since AxedδCterm can be stably expressed in Drosophila S2 cells. Despite extensive attempts to implicate the wide array of CRLs in Drosophila in axon death and Axed function, no evidence was found supporting a role for CRLs in axon degeneration. Given that there are five Cullin-like molecules in the Drosophila genome, it seems possible that genetic redundancy between Cullins might explain the lack of phenotype. Alternatively, Axed could be functioning to promote axon death in a completely novel Cullin-independent manner. Future studies aimed at identifying direct binding partners for Axed will be essential to resolve these issues. If Axed functions as a bridging molecule for CRL ubiquitination activity, of particular interest will be identifying molecules that bind to the Axed C-terminal putative substrate binding domain (Neukomm, 2017).

Sarm1 functions in axons after injury (Gerdts, 2015). Given that Axed functions genetically downstream of Sarm1, it is proposed that Axed also functions in axons. Using a functional, endogenously enhanced GFP-tagged version of Axed (AxedeGFP), it was found that Axed protein is enriched in the neuropil in both the larva and the adult nervous system. The Drosophila larval neuropil is highly enriched for dendrites and axons and is the site of all CNS synapse formation. AxedeGFP signals overlapped extensively with axons, dendrites, and synapses, further supporting a cell-autonomous role for Axed in neurons during axon death. AxedeGFP is also present in the adult neuropil, and AxedeGFP signals transiently increased at 4 and 6 hr in the antennal lobe after injury but returned to baseline levels by 24 hr, a time point at which most severed axons have fragmented. The return to baseline levels likely represents AxedeGFP staining in local and projection interneurons in the antennal lobe that were not severed by antennal ablation. Why AxedeGFP signals might transiently increase remains unclear. Perhaps AxedeGFP is relocalized in a way that increases GFP exposure, or new AxedeGFP protein may be locally synthesized. Regardless of the precise mechanisms, the observations indicate that AxedeGFP dynamics in vivo are sensitive to axonal injury (Neukomm, 2017).

axed mutant phenotypes are indistinguishable from dsarm mutants with respect to axon preservation after injury. In addition, axed mutants completely blocked the pro-degenerative activity of the gain-of-function dSarm molecule dSarmΔdsarm in vivo. These genetic data strongly support a model whereby Axed functions genetically downstream of dSarm, and argue that dSarm and Axed drive axon death through the same genetic pathway (Neukomm, 2017).

dSarmΔdsarm appears to induce a Wallerian-like program in vivo based on several observations. First, the explosive fragmentation observed is morphologically similar to WD. Second, dSarmΔdsarm-induced degeneration can be suppressed partially by WldS, whose activity is highly selective to WD. Third, dSarmΔdsarm pro-degenerative activity can be fully suppressed in axed mutants. Finally, all structural and biochemical studies to date support the notion that elimination of the N-terminal ARM domain of dSarm (Drosophila), TIR-1(C. elegans), or Sarm1 (mammals) leads to the production of a gain-of-function molecule that activates signaling. That dSarmΔdsarm signaling in the absence of injury is potently suppressed by axed mutants provides a compelling argument for Axed to act downstream of dSarm. The nature of the genetic program that drives death of neuronal cell bodies after dSarm/Sarm1 activation remains mysterious. It could not be blocked with the broad caspase inhibitor P35. Related studies in mouse found a lack of evidence for apoptotic signaling molecules, necroptosis, and parthanatos in death induced by activated Sarm1. The observation that it was possible to induce robust degeneration of axons, dendrites, and cell bodies by adult-specific induction of dSarmΔdsarm in adult PDF+ neurons, even in the absence of injury, suggests that dSarmΔdsarm could provide a useful tool for conditional removal of selected neurons from adult circuits for functional studies (Neukomm, 2017).

In the larger context of axon death signaling, the finding that mutations in hiw do not suppress dSarmΔdsarm in vivo agrees with the proposed model that Hiw/Phr1 acts genetically upstream of dSarm/Sarm1 and Nmnat degradation. The position of the MAPK signaling cascade relative to dSarm/Sarm1 is still debated in the field and remains to be determined. The fact that the Drosophila genome houses only a single JNK family member, and the observation that null alleles had no axon death phenotype, argues strongly against a central role for JNK signaling in axon death (Neukomm, 2017).

With respect to their roles in the functional disassembly of axons, this study has demonstrated that blocking axon death with mutations in dsarm, axed, or highwire is sufficient to preserve axons in a functional state in neural circuits where they can elicit complex grooming behaviors for weeks after axotomy. Therefore, axon death mutants, like the WldS molecule, exert their neuroprotective effects very early in the axon death genetic program, and their therapeutic blockade is likely to lead to the preservation of functional axons in the context of neurological disease (Neukomm, 2017).

A number of recent studies have led to the following model for axon death signaling: (1) axotomy results in the degradation of pools of labile Nmnat/Nmnat2 in distal severed axons; (2) depletion of Nmnat/Nmnat2 in turn leads to decreases in axonal NAD+; (3) loss of Nmnat/Nmnat2 or decreased NAD+ somehow activates dSarm/Sarm1 signaling, and (4) Sarm1 NAD+ hydrolase activity then drives pathological depletion of axonal NAD+ pools, causing catastrophic energy failure that ultimately drives axon degeneration. A number of observations were made that are difficult to rectify with a simple Nmnat2/Sarm1-dependent NAD+ depletion model. The first is that expression of dSarmΔdsarm, which based on previous reports should rapidly degrade axonal NAD+ and normally leads to rapid axonal degeneration (Essuman, 2017), can be completely suppressed by loss of Axed. If the terminal step in axonal death were Nmnat2/Sarm1-dependent depletion of NAD+ pools, it is hard to imagine how the NAD+ hydrolase activity of the Sarm1 TIR domain, once unleashed, could be suppressed by the loss of a BTB and BACK domain protein (i.e., Axed). The alternative hypothesis is favored that Sarm1 signaling, perhaps through its TIR domain NAD+ hydrolase activity, activates an Axed-dependent downstream signaling pathway essential for axon death. ADP-ribosylation of targets is one mechanism that allows certain E3 ubiquitin ligases to bind substrates. There is certainly a tight association between NAD+ metabolism and axon death signaling, and the NAD+ hydrolase activity associated with the Sarm1 TIR domain can lead to the production of ADPr. Perhaps BTB and BACK domain proteins like Axed also require ADP-ribosylation of selected targets during axon death. In this model, the degradation of NAD+ would be a byproduct of the reaction, not a driving force (Neukomm, 2017).

Nmnat2-/- mutant mice die perinatally with neurons containing short axons, but these embryos can be rescued to adulthood by Sarm1-/- null mutations. It is likely that Nmnat2-/- animals and axons can survive in Sarm1-/- backgrounds because Nmnat1 and Nmnat3, the additional isoforms of Nmnat found in mice, could partially compensate. An nmnatRNAi experiment may therefore more closely resemble conditions in the Nmnat2-/-, Sarm1-/- mouse, with respect to the partial, but not complete elimination of Nmnat function. dsarm null alleles could only partially protect axons from Nmnat depletion with nmnatRNAi, as far fewer axons and cell bodies were observed compared to controls. Consistent with a requirement for Nmnat activity in cell survival, this study found null clones of the sole fly Nmnat molecule die within 1 day after eclosion. Remarkably, this phenotype can be completely overcome by loss of Axed, since axed, nmnat double mutant clones survive for weeks. However, it was surprising to find that loss of dSarm fails to block this degeneration. Whether Sarm1-/- animals could suppress the degeneration of mammalian neurons completely lacking Nmnat function would require the simultaneous elimination of all Nmnat isoforms (i.e., Nmnat1,2,3 triple mutants), which has not been explored. In the absence of Nmnat activity, invertebrate cells should not be able to autonomously synthesize NAD+. How these cells survive remains a mystery, but a number of possibilities might explain this phenotype. Perhaps yet-to-be-identified NAD+ biosynthetic pathways exist that can function without Nmnat. Alternatively, axons in vivo could be supplemented with NAD+ or key intermediates by surrounding glia and thereby sustain their integrity in axed, nmnat double mutant axons. Finally, it is possible that in the absence of Axed, Nmnat protein may not be turned over appropriately in axons and that perdurance of small amounts of Nmnat may generate sufficient levels of NAD+ for axon survival. Direct, in vivo measurement of NAD+ levels in axons in Drosophila will be essential to answer this question, but current NAD+ sensors have not yet provided sufficient sensitivity for this analysis (Neukomm, 2017).

Axed appears to sit genetically at a convergence point in axon death signaling. There are three ways to activate axon death: axotomy, Nmnat/Nmnat2 depletion, or expression of gain-of-function dSarm/Sarm1. Axed mutants, but not dsarm, can suppress each of these treatments, and, amazingly, axed mutants can even survive the combination of all three insults at once. The neuroprotective effects of axed therefore exceed those of dsarm mutants. It remains to be determined whether any of the four putative mammalian paralogs (BTBD1, BTBD2, BTBD3, and BTBD6) play a role in axon death, but this seems likely based on the strong conservation of dSarm/Sarm1 function in axon degeneration. If so, these would represent important new therapeutic targets for blocking axon death in neurological diseases such as traumatic brain injury, peripheral neuropathy, or nerve injury, where dSarm/Sarm1 signaling is known to drive axon loss (Neukomm, 2017).

Injury-induced inhibition of bystander neurons requires dSarm and signaling from glia

Nervous system injury and disease have broad effects on the functional connectivity of the nervous system, but how injury signals are spread across neural circuits remains unclear. This study explored how axotomy changes the physiology of severed axons and adjacent uninjured "bystander" neurons in a simple in┬ávivo nerve preparation. Within hours after injury, suppression of axon transport was observed in all axons, whether injured or not, and decreased mechano- and chemosensory signal transduction was observed in uninjured bystander neurons. Unexpectedly, it was found the axon death molecule Sterile alpha and Armadillo motif (dSarm), but not its NAD(+) hydrolase activity, was required cell autonomously for these early changes in neuronal cell biology in bystander neurons, as were the voltage-gated calcium channel Cacophony (Cac) and the mitogen-activated protein kinase (MAPK) signaling cascade. Bystander neurons functionally recovered at later time points, while severed axons degenerated via α/Armadillo/Toll-interleukin receptor homology domain (dSarm)/Axundead signaling, and independently of Cac/MAPK. Interestingly, suppression of bystander neuron function required Draper/MEGF10 signaling in glia, indicating glial cells spread injury signals and actively suppress bystander neuron function. This work identifies a new role for dSarm and glia in suppression of bystander neuron function after injury and defines two genetically and temporally separable phases of dSarm signaling in the injured nervous system (Hsu, 2020).

Nervous system injury or neurodegenerative disease can lead to profound alterations in neural circuit function. The precise cellular basis is poorly defined in any context, but disruption of circuit signaling is generally thought to occur as a result of a loss of physical connectivity between damaged neurons. Indeed, axon and synapse degeneration are among the best correlates of functional loss in patients with a variety of brain injuries or neurological diseases. But whether, and the extent to which, an injured or diseased neuron might also alter the functional properties of neighboring healthy 'bystander' neurons (i.e., those not damaged or expressing disease-associated molecules) is an important and open question. If the physiology of bystander neurons is radically altered by their damaged neighbors, this would force reconsideration of the simple loss-of-physical-connectivity model as the appropriate explanation for functional loss in neural circuits after trauma (Hsu, 2020).

It is well documented that bystander neurons can change their physiology in response to their neighbors being injured. For instance, mouse L5 spinal nerve transection results in the degeneration of distal L5 afferents in sciatic nerve alongside intact L4 C fiber afferents. Within 1 day after L5 lesion, L4 C fibers develop spontaneous activity that lasts for at least 1 week and appears to mediate injury-induced pain and hyperalgesia behaviors. Bystander effects have also been observed in the central nervous system (CNS). In a mouse model of mild traumatic brain injury (TBI), 1 day after injury, pyramidal neurons with severed axons and intact bystander neurons both exhibited injury-induced changes in action potential firing and afterhyperpolarization. Injured neurons failed to recover, while bystander neurons ultimately exhibited a return to normal firing properties. How injured neurons or surrounding glia signal to bystander neurons, or how bystander neurons receive this signal, is not known, but the similar electrophysiological changes observed in axotomized and intact dorsal root ganglion neurons have been proposed to be associated with Wallerian degeneration (Hsu, 2020).

Recent work has begun to illuminate the mechanisms by which damaged axons autonomously drive their own degeneration during Wallerian degeneration. A forward genetic screen in Drosophila identified the sterile α/Armadillo/Toll-interleukin receptor homology domain (dSarm) molecule as essential for axon auto-destruction, as loss of dSarm completely blocked Wallerian degeneration (Osterloh, 2012). All known dSarm pro-degenerative function requires the BTB and BACK domain molecule Axundead (Axed), another powerful regulator of axon degeneration (Neukomm, 2017). dSarm function in axon degeneration after injury is conserved in mouse: Sarm1-/- mutants block Wallerian degeneration, and loss of Sarm1 also suppresses axon degeneration in mouse models of TBI and peripheral neuropathy. Sarm1 inhibition is thus an exciting potential approach for blocking axon loss and neuroinflammation in human disease (Hsu, 2020).

dSarm/Sarm1 has been studied primarily in the nervous system as a positive regulator of axonal degeneration. In mammals, axotomy leads to the depletion of the labile NAD+ biosynthetic enzyme Nmnat2 and a decrease in NAD+ in severed axons. Nmnat2 loss somehow activates Sarm1 (Gilley, 2015), which is proposed to lead to further NAD+ depletion and metabolic catastrophe in the severed axon (Gerdts, 2015) through a Sarm1-intrinsic NAD+ hydrolase activity (Essuman, 2017). The Sarm1 NAD+ hydrolase activity appears to be activated directly by the NAD+ precursor, NMN, presumably through allosteric conformational changes in Sarm1 upon NMN binding. This NAD+ depletion model has been proposed as the primary mechanism by which Sarm1 drives axon loss, and to explain the mechanistic basis of protection by several other neuroprotective molecules (Gerdts, 2016). For instance, the slow Wallerian degeneration molecule (WldS), which includes the highly stable NAD+ biosynthetic enzyme Nmnat1, is thought to protect axons by substituting for the labile Nmnat2 molecule, thereby reducing NMN levels and avoiding NAD+ depletion. Similarly, the protective effects of loss of the E3 ubiquitin ligase Highwire/Phr1 is thought to result from blockade of its direct role in degrading Nmnat2, such that in hiw/phr1 mutants Nmnat2 is stabilized and continues to maintain NAD+ levels (Hsu, 2020).

Elegant genetic studies in C. elegans demonstrated that TIR-1 (the worm homolog of dSarm/Sarm1) is part of a signaling cascade downstream of the voltage-gated calcium channel UNC-36 and CamK-II and signals via the mitogen-activated protein kinase (MAPK) signaling cascade. Based on this work, MAPK signaling was examined for roles in Wallerian degeneration but met with mixed results. Changes in MAPK signaling (i.e., phosphorylation of MAPK pathway members) were found in axons within 15-30 min after axotomy, were Sarm1 dependent and suppressed by Nmnat overexpression, and partial suppression of axon degeneration was observed after simultaneous blockade of multiple MAPK components. But how MAPK signaling modulates axon degeneration, particularly in the context of Sarm1 signaling, remains controversial, as one study proposed MAPK signals downstream of Sarm1, while another argued Sarm1 was upstream of MAPK signaling, and the neuroprotective phenotypes resulting from MAPK blockade do not approach levels afforded by loss of Sarm1 in vivo (Hsu, 2020).

This study used a partial nerve injury model to examine early changes in the physiology of severed axons and neighboring uninjured bystander neurons. Axotomy of even a small subset of neurons was shown to leads to inhibition of cargo transport in all axons within the nerve and suppression of sensory signal transduction in bystander neurons. Surprisingly, this early blockade of axon transport and sensory signal transduction required dSarm in both severed and uninjured bystander neurons, where it signaled via the conserved UNC-36/MAPK signaling pathway. Early suppression of axon transport and bystander neuron function did not require dSarm NAD+ hydrolase function or Axed, was not modulated by NMN, and was not induced by depletion of dNmnat. This suggests it is mechanistically different from later events in axon death, where dSarm drives axon degeneration with Axed. Intriguingly, this study found that this early spreading of injury signals to bystander neurons required the Draper receptor in surrounding glia, indicating that glial cells actively signal to inhibit the function of bystander neurons in vivo. This work identifies new roles for dSarm and glia in modifying neurophysiology early after injury, assigns the NAD+ hydrolase function exclusively to later axon degenerative events, and reveals a new role for UNC-36/MAPK signaling in promoting these dSarm-dependent changes early after an injury has occurred in the nervous system. It is proposed that two temporally and genetically separable phases of dSarm signaling exist that mediate these distinct injury-induced changes in neurophysiology and axon degeneration (Hsu, 2020).

This study shows that relatively small injuries can lead to the rapid and efficient spreading of injury signals across nerves that potently suppress axon transport throughout the nerve, and broadly inhibit neurophysiology in uninjured bystander neurons. Surprisingly, the same molecule was found to be required to drive explosive axon degeneration in severed axons at later stages, dSarm/Sarm1, is required for this early suppression of neuronal function, although the signaling mechanisms at each stage appear to be different. The data support a model whereby early (i.e., 1-3 h after injury) dSarm signals with Cac and MAPK components, but independent of its NAD+ hydrolase activity, to suppress axon transport and neurophysiology, while at later stages (8-12 h), dSarm signals with Axed to promote explosive axon degeneration. This significantly expands the role for dSarm/Sarm1 in regulating nervous system responses to injury to include even uninjured bystander neurons. Furthermore, a critical role was discoverd for glial cells, through Draper, in signaling to bystander neurons to inhibit their axon transport and neurophysiology. Together, this work suggests that a significant amount of functional loss after neural trauma is a result of not only frank degeneration but also more widespread changes in neuronal function, and it occurs in uninjured neurons through glial spreading of injury signals (Hsu, 2020).

The data support the notion that widespread signaling occurs between cells in injured neural tissues immediately after injury and that injury signals can radically alter neuronal function. Severing even a small number of axons led to a suppression of axon transport within hours in all axons in the adult wing nerve, even in uninjured bystander neurons. Beyond axon transport, local uninjured bystander sensory neurons also exhibited a disruption of mechano- and chemosensory signal transduction, which was partially reversible within a few hours. These observations suggest that beyond simple breakage of connectivity, a significant part of functional loss after brain injury or in neurodegenerative disease may also be occurring in healthy, intact neurons that have received function-suppressing signals from nearby damaged neurons (Hsu, 2020).

Surprisingly, dSarm was found to be required cell autonomously in bystander neurons to alter axon transport and nerve function in response to injury, and this role did not require its NAD+ hydrolase activity. Reception of this injury signal in the bystander neuron (and severed axons) requires the VGCC Cac and the MAPK signaling cascade, similar to Tir-1 signaling in C. elegans, but not Axed. Reciprocally, Cac and MAPK components are not required for Wallerian degeneration at later stages. Explosive axon degeneration requires dSarm, its NAD+ hydrolase activity, and Axed. Based on the timing of these different events (i.e., changes in neuronal function versus frank degeneration) with the genetic studies indicating they are separable, a two-phase model is proposed for dSarm signaling in injured neural tissues: early dSarm-dependent changes in axon biology and neurophysiology that occur within hours after injury are mediated by the Cac/dSarm/MAPK signaling cascade (phase I), while late-stage axon degeneration is driven by dSarm signaling through Axed (phase II). The existence of these temporally distinct phases of dSarm signaling likely explain previous results that seemed in conflict, where MAPK signaling was proposed to act both upstream (Yang, 2015) and downstream (Walker, 2017) of Sarm1 after axotomy. According to the current model, both of these assertions would be correct, with dSarm/Sarm1 acting upstream of MAPK early (phase I) and independent but ultimately downstream of MAPK later to drive dSarm/Axed-dependent axon degeneration (phase II)(Hsu, 2020).

To date, dSarm/Sarm1 has been thought of primarily as a cell-autonomous regulator of explosive axon degeneration, but the current work shows that dSarm can also drive important changes in circuit function through altering neuronal cell biology and neurophysiology. That bystander neurons recover and remain viable also demonstrates that activation of dSarm after injury does not necessarily lead to axon death. It is suspected that recovery occurs in large part because bystander neurons have not been severed, which is an extreme injury, and depends on their connection to the cell body, which is a source of axon survival factors like Nmnat2. Connection to the cell body may also explain why axon transport was less severely suppressed in the bystander neurons; additional transport factors can still be continuously supplied to the distal axon from the soma. Defining how dSarm activity is regulated in each of these contexts to interact with Cac/MAPKs versus Axed, and why the first phase does not require NAD+ hydrolase function, are key questions for the future (Hsu, 2020).

A compelling case exists for the NAD+ depletion hypothesis for dSarm/Sarm1 function in axon degeneration (Essuman, 2017; Gerdts, 2015, 2016), although arguments have been made this dSarm/Sarm1 signaling is likely more complex (Neukomm, 2017). In this model, depletion of Nmnat2 via Hiw/Phr1 results in the accumulation of NMN, which functions as an activator of Sarm1, with Sarm1 NAD+ hydrolase activity driving metabolic catastrophe. This study provides several lines of evidence that the above, newly described early dSarm signaling events (i.e., suppression of axon transport and neurophysiology) are mechanistically distinct but are nevertheless also regulated by some axon-death-associated molecules. First, while NMNd can suppress axon degeneration in flies and other species, it cannot block early suppression of axon transport or changes in bystander neuron function. This argues that NMN is not a driving force for dSarm activation in the early phase. Second, although limited to tagged versions of dNmnat for this analysis, no depletion of dNmnat was observed within the time frame of 6 h after injury. Previous studies in SCG or DRG cultures in vitro suggest Nmnat2 depletion takes 4-6 h and NAD+ depletion begins ~2-3 h after axotomy, which is slightly later than the bystander effect was observed in vivo. Because full axon degeneration is prolonged in vivo compared to in vitro studies, the timing of Nmnat2 loss and NAD+ depletion is likely also prolonged in vivo, further suggesting this likely happens after cessation of axon transport. Third, Axed, which is genetically downstream of dSarm during axon degeneration (Neukomm, 2017), is not required for early suppression of nerve responses to injury in either severed or intact neurons, only later axon degenerative events in the severed axons. Finally, this study shows that while the NAD+ hydrolase function of dSarm is required in vivo for efficient axon degeneration, it is dispensable for early suppression of axon transport (Hsu, 2020).

Despite these clear molecular and genetic differences between early- and late-phase signaling events, WldS or dNmant expression or hiw mutants are capable of suppressing early changes in axon transport and neurophysiology, even in bystander neurons. This could be interpreted as evidence for similarity in signaling mechanisms at early and late stages of dSarm signaling (i.e., that they act by maintaining NAD+). However, the alternative possibility is favored that these data point to an important role for dNmnat in mediating early dSarm signaling events during suppression of bystander neuron function. Loss of Axed does not affect the bystander effect, and axon transport is suppressed. However, this study found that loss of dNmnat in axed null backgrounds (which allows for preservation of neuronal integrity despite loss of dNmnat) blocked the ability of injury to induce the bystander effect. This result reveals a paradoxical, positive role for dNmnat in promoting the bystander effect early. It is suspected that dNmnat exerts this effect through modulating MAPK signaling, whose interactions are complex: loss of Nmant has been shown to suppress MAPK signaling, while increased Nmnat activity can also potently block the activation of MAPK signaling within the first few hours after axotomy. It is proposed that dNmnat activity is required early for the bystander effect and that dNmnat levels need to be precisely tuned for proper signaling at each phase (Hsu, 2020).

Glial cells are well positioned to rapidly spread signals to all axons in the wing nerve. Much like Remak bundles in mammals, the Drosophila L1 wing nerve has glial cells that appear to wrap axons individually, which would imply that axon-to-axon signals must pass through glia. The observation that selective elimination of Draper signaling in glia is sufficient to inhibit the spreading of injury signals to bystander neurons is consistent with an axon->glia->bystander neuron signaling event, although it is also possible that glia are directly injured by the axotomy and signal to bystander neurons without input from the severed axons. Given the similarities in the response of severed axons and those of bystanders (i.e., both block axon transport on the same timescale), and the selective effects of Draper on the bystander neuron axons, the former model is favored rather than the latter (Hsu, 2020).

Draper signals to bystander neurons through a transcriptional JNK/dAP-1 cascade, likely through activating MMP-1. Nerve injury also rapidly activates JNK/c-Jun signaling in mammalian Schwann cells, where JNK/c-Jun mediate most aspects of Schwann cell injury responses and reprogramming events. This conserved glial response likely occurs in Schwann-cell-like wrapping glia present in the Drosophila L1 wing nerve, although it may be activated in the subperineurial glia, which can act in a partially redundant fashion with wrapping glia. The involvement of Mmp-1 is intriguing given its well-known role in neuroinflammatory responses to brain injury in mammals, where it functions to break down the extracellular matrix and has been proposed to promote diffuse axon injury. Other key components of the Draper signaling pathway (dCed-6 in particular, which is required for Draper signaling in all other known contexts) were not required for suppression of bystander neuron neurophysiology (Hsu, 2020).

How bystander neurons receive injury signals and respond has remained unclear, although injury- or disease-induced effects on bystander neurons is well documented. In most cases, this has been explored in the context of bystander neuron cell death driven by neuroinflammatory cells. For instance, release of C1q, interleukin-1α (IL-1α), tumor necrosis factor (TNF) from microglia following brain injury drives the formation of neurotoxic astrocytes, which can promote the death of neurons through release of yet-to-be-identified toxins. Bystander neuronal cell death is also driven by brain-infiltrating inflammatory monocytes in viral encephalitis, in a way that is mediated by calpains, which are also important regulators of axon degeneration. Secondary axon degeneration (i.e., that occurring in neurons not damaged by the initial injury) can be driven in a way that requires intracellular Ca2+ release through IP3Rs and ryanodine receptors. These represent extreme cases of bystander effects, where cells undergo apoptosis or their axons degeneration. Whether dSarm/Sarm1 is involved in these effects is an open question. The model employed by this study is likely most relevant to partial nerve injury, where non-autonomous changes in bystander neurons have been well documented. Uninjured bystander neurons in mild TBI models are certainly altered physiologically in a reversible way. The molecular basis of any of these signaling events remains unknown, but this study points to dSarm/Sarm1 as a candidate mediator. It is interesting to note that in contrast to control mice, which show significant behavioral defects for hours after mild TBI, Sarm1-/- animals exhibited almost immediate recovery, and this was at a time point long before diffuse axon injury is observed in TBI models. It is plausible that this early loss of function is mediated in part by the bystander effect (Hsu, 2020).

In summary, this study defines two genetically separable phases of dSarm signaling, places dSarm/Sarm1 at the heart of neuronal injury signaling throughout neural tissues, identifies new signaling partners for dSarm, and expands its role to regulating the responses of uninjured neurons to local tissue injury (Hsu, 2020).


REFERENCES

Search PubMed for articles about Drosophila Axundead

Essuman, K., Summers, D. W., Sasaki, Y., Mao, X., DiAntonio, A. and Milbrandt, J. (2017). The SARM1 Toll/Interleukin-1 receptor domain possesses intrinsic NAD(+) cleavage activity that promotes pathological axonal degeneration. Neuron 93(6): 1334-1343. PubMed ID: 28334607

Gerdts, J., Summers, D. W., Sasaki, Y., DiAntonio, A. and Milbrandt, J. (2013). Sarm1-mediated axon degeneration requires both SAM and TIR interactions. J Neurosci 33(33): 13569-13580. PubMed ID: 23946415

Gerdts, J., Brace, E. J., Sasaki, Y., DiAntonio, A. and Milbrandt, J. (2015). SARM1 activation triggers axon degeneration locally via NAD(+) destruction. Science 348(6233): 453-457. PubMed ID: 25908823

Gerdts, J., Summers, D. W., Milbrandt, J. and DiAntonio, A. (2016). Axon self-destruction: new links among SARM1, MAPKs, and NAD+ metabolism. Neuron 89(3): 449-460. PubMed ID: 26844829

Gilley, J., Orsomando, G., Nascimento-Ferreira, I. and Coleman, M. P. (2015). Absence of SARM1 rescues development and survival of NMNAT2-deficient axons. Cell Rep 10(12):1974-81. PubMed ID: 25818290

Hsu, J. M., Kang, Y., Corty, M. M., Mathieson, D., Peters, O. M. and Freeman, M. R. (2020). Injury-induced inhibition of bystander neurons requires dSarm and signaling from glia. Neuron. PubMed ID: 33296670

Neukomm, L. J., Burdett, T. C., Seeds, A. M., Hampel, S., Coutinho-Budd, J. C., Farley, J. E., Wong, J., Karadeniz, Y. B., Osterloh, J. M., Sheehan, A. E. and Freeman, M. R. (2017). Axon death pathways converge on Axundead to promote functional and structural axon disassembly. Neuron 95(1): 78-91.e75. PubMed ID: 28683272

Osterloh, J. M., Yang, J., Rooney, T. M., Fox, A. N., Adalbert, R., Powell, E. H., Sheehan, A. E., Avery, M. A., Hackett, R., Logan, M. A., MacDonald, J. M., Ziegenfuss, J. S., Milde, S., Hou, Y. J., Nathan, C., Ding, A., Brown, R. H., Jr., Conforti, L., Coleman, M., Tessier-Lavigne, M., Zuchner, S. and Freeman, M. R. (2012). dSarm/Sarm1 is required for activation of an injury-induced axon death pathway. Science 337(6093): 481-484. PubMed ID: 22678360


Biological Overview

date revised: 28 February 2021

Home page: The Interactive Fly © 2011 Thomas Brody, Ph.D.