ABSTRACT
The p75 neurotrophin receptor (p75NTR; also known as NGFR) can mediate neuronal apoptosis in disease or following trauma, and facilitate survival through interactions with Trk receptors. Here we tested the ability of a p75NTR-derived trophic cell-permeable peptide, c29, to inhibit p75NTR-mediated motor neuron death. Acute c29 application to axotomized motor neuron axons decreased cell death, and systemic c29 treatment of SOD1G93A mice, a common model of amyotrophic lateral sclerosis, resulted in increased spinal motor neuron survival mid-disease as well as delayed disease onset. Coincident with this, c29 treatment of these mice reduced the production of p75NTR cleavage products. Although c29 treatment inhibited mature- and pro-nerve-growth-factor-induced death of cultured motor neurons, and these ligands induced the cleavage of p75NTR in motor-neuron-like NSC-34 cells, there was no direct effect of c29 on p75NTR cleavage. Rather, c29 promoted motor neuron survival in vitro by enhancing the activation of TrkB-dependent signaling pathways, provided that low levels of brain-derived neurotrophic factor (BDNF) were present, an effect that was replicated in vivo in SOD1G93A mice. We conclude that the c29 peptide facilitates BDNF-dependent survival of motor neurons in vitro and in vivo.
INTRODUCTION
The selective death of motor neurons is the central pathology of neurodegenerative diseases such as amyotrophic lateral sclerosis and spinal muscular atrophy (Murray et al., 2010). Motor neurons are also vulnerable following axonal injury, with motor neuron dysfunction resulting in paralysis (Branco et al., 2007). During development, motor neurons are dependent on growth factors such as the neurotrophins for their survival and function. In developing motor neurons, brain-derived neurotrophic factor (BDNF) and neurotrophins (NT) 3 and 4 mediate cell survival through their cognate Trk receptors, TrkB or C (Kaplan and Miller, 2000). Similarly, after nerve lesion and in animal models of motor neuron disease, BDNF and NT3 can protect and repair motor neurons (Yuen and Mobley, 1995; Wiese et al., 2004; Henriques et al., 2010). During development and following injury, the p75 neurotrophin receptor (p75NTR; also known as NGFR) has been shown to mediate motor neuron death (Seeburger et al., 1993; Lowry et al., 2001; Dechant and Barde, 2002; Ibanez and Simi, 2012) in response to nerve growth factor (NGF) (Sedel et al., 1999; Ernfors, 2001) or its immature precursor (proNGF), released by activated astrocytes (Pehar et al., 2004; Domeniconi et al., 2006). However, p75NTR is a multifunctional receptor for both mature and pro-neurotrophins, capable of regulating different biological effects in response to its ligands (Roux and Barker, 2002; Blochl and Blochl, 2007; Underwood and Coulson, 2008). In addition to its death-signaling capability, p75NTR also acts as a co-receptor that can assist Trk receptors to bind their ligands with higher affinity and specificity, thereby enhancing their trophic functions (Hempstead et al., 1991; Barker, 2007). An important mechanism in this process appears to be the generation of biologically active receptor fragments following enzymatic cleavage of p75NTR (Ceni et al., 2010; Kommaddi et al., 2011; Matusica et al., 2013).
In a two-step process, p75NTR is enzymatically initially cleaved by an α-secretase (ADAM17) that removes the extracellular domain of the receptor, leaving a membrane-bound C-terminal fragment. This fragment is subsequently cleaved by γ-secretase, releasing the intracellular domain into the cytoplasm (Jung et al., 2003; Kanning et al., 2003). The intracellular domain contains two regions that mediate intracellular signaling through protein–protein interactions: the juxtamembrane domain, and a tumor-necrosis-factor-receptor-like death domain (Roux and Barker, 2002; Skeldal et al., 2011). Increased metalloprotease cleavage of p75NTR has been associated with the death of sciatic neurons following injury (DiStefano et al., 1993; Murray et al., 2003), and evidence of increased levels of p75NTR proteolytic cleavage has been reported using urine analysis in both SOD1 transgenic mice and people living with amyotrophic lateral sclerosis (Shepheard et al., 2014). Elevated levels of the C-terminal and intracellular domain fragments have also been reported in the degenerating motor axons of SOD1G93A transgenic mice (Perlson et al., 2009), and both p75NTR cleavage fragments have been linked to death signaling (Skeldal et al., 2011). Although the nature of p75NTR fragment signaling appears to be cell-type specific (Vicario et al., 2015), these studies indicate that the proteolytic cleavage of p75NTR might be a key mechanism by which p75NTR mediates death signaling after motor neuron injury and disease.
We have previously shown that the juxtamembrane region of p75NTR has the ability to induce cell death when bound to the cell membrane (Coulson et al., 2000; Underwood et al., 2008). We also demonstrated that a soluble peptide mimic (c29) of the juxtamembrane intracellular domain fragment of p75NTR can act as a dominant-negative inhibitor of p75NTR death signaling (Coulson et al., 2000). Recently, we have shown that the intracellular domain fragment of p75NTR and the c29 peptide are able to increase NGF survival signaling through interaction with TrkA receptors (Matusica et al., 2013). This in turn leads to increased neurite outgrowth at low NGF concentrations, and enhanced survival of sympathetic neurons undergoing growth factor withdrawal (Matusica et al., 2013). Here we investigated whether c29 plays a similar role in TrkB-dependent cells, determining the role of p75NTR cleavage in neurotrophin-induced motor neuron death and exploring whether c29 can inhibit this process in vitro and in vivo.
RESULTS
c29 prevents the death of axotomized ulnar motor neurons and delays disease progression of SOD1G93A-induced motor neuron disease
It is clear that a balance between Trk-receptor-mediated survival and p75NTR-mediated cell death signaling pathways regulates motor neuron survival both during development and following injury or disease. In particular, the role of these competing neurotrophin signaling pathways has been well studied in motor neurons following axotomy and in degenerative conditions including amyotrophic lateral sclerosis. Therefore, to determine if c29 is capable of inhibiting motor neuron cell death in vivo, we first used the injury paradigm of ulnar nerve axotomy model in which 50% of the ipsilateral motor neurons typically die, and where both downregulation of p75NTR by antisense oligonucleotides (Lowry et al., 2001) and BDNF application (Yuan et al., 2000) have been shown to promote motor neuron survival. Following ulnar nerve axotomy of newborn rat pups, c29 or protein transduction domain (PTD) vehicle-fused control peptide was applied to the cut nerve stump in gel foam. After five days, the number of live motor neurons was determined by stereological counts following histological processing of the spinal cords. In the axotomized animals treated with control peptide, axotomy resulted in a loss of ∼45% of choline acetyl transferase (ChAT)-positive motor neurons in the ipsilateral cervical enlargement (C7–C8) compared with the contralateral side (Fig. 1). By contrast, the number of surviving motor neurons in the spinal cord of animals treated with c29 was significantly higher, with the average neuronal loss being less than 15% that of the contralateral side (Fig. 1).
As PTD-fused peptides and proteins given to mice systemically have previously been shown to cross the blood–brain barrier (Schwarze et al., 2000), we confirmed that biotinylated-c29 peptide reached the spinal neurons after intraperitoneal injection of wild-type mice. In mice sacrificed one hour after 5 mg/kg peptide injection, immunostaining for biotin was observed in grey matter of the spinal cord in c29-treated animals but not in controls (non-biotinylated-peptide-treated animals) (Fig. 2A,B). Next, we tested the effect of systemic c29 treatment in the SOD1G93A mouse model of amyotrophic lateral sclerosis. c29 was administered chronically (5 mg/kg/day) to the mice from 9 weeks of age by means of a subcutaneous osmotic minipump, and the course of disease was monitored. No significant difference between c29-treated mice and vehicle-treated animals was observed in terms of lifespan (Fig. 2C). However, c29-treated mice had a delayed disease onset, taking significantly longer to lose 10% of their body weight (P<0.01; median SOD1+vehicle: 31 days; SOD1+c29: 42 days; Fig. 2D) and c29-treated animals maintained their performance on the rotarod between 100 and 125 days, the time frame during which the performance of the vehicle-treated animals declined dramatically (Fig. 2E). Furthermore, at this time, the number of motor neurons in the spinal cord of c29-treated SOD1G93A animals was significantly higher than in vehicle-treated animals (Fig. 2F,G), and fewer cells appeared atrophic (Fig. 2G).
p75NTR cleavage associates with neurotrophin-induced cell death in vivo and in vitro
As upregulation of p75NTR expression, proteolysis and death signaling are associated with motor neuron degeneration, we asked whether the reduced neuronal loss was associated with altered p75NTR cleavage and/or expression by measuring the levels of p75NTR extracellular domain (ECD) present in the urine of SOD1G93A animals over the course of the disease. In SOD1G93A animals, the level of p75NTR ECD present in urine was significantly higher than in C57BL/6J (wild-type) healthy controls presymptomatically from 60 days of age and throughout disease progression (Fig. 3A). However, in SOD1G93A animals treated with c29, the level of p75NTR ECD present in the urine between 80 and 120 days of age was significantly lower than that found in the urine of vehicle-treated animals (Fig. 3B). Together, these results indicate that c29 can potentiate the survival of injured or dying motor neurons in vivo, which corresponds to lowered p75NTR expression and/or cleavage.
To determine whether cleavage of p75NTR is important for neurotrophin-mediated motor neuron death, we asked whether neurotrophins stimulate the cleavage of p75NTR using the motor-neuron-like NSC-34 cell line. As previously reported for other cell types, treatment of cultures with PMA, which activates metalloproteases (Ni et al., 2001), stimulated the cleavage of p75NTR, resulting predominantly in the generation of its intracellular domain fragment. This cleavage could be prevented with the metalloprotease inhibitor TAPI-2, whereas treatment of cells with the γ-secretase inhibitor Compound E together with PMA resulted in accumulation of the C-terminal p75NTR fragment (Fig. 3C) (Jung et al., 2003; Kanning et al., 2003; Zampieri et al., 2005). Treatment of NSC-34 cells with BDNF or proBDNF produced predominantly intracellular domain fragments of p75NTR whereas either NGF or proNGF treatment predominantly resulted in generation of the C-terminal fragment (Fig. 3C).
The ability of NGF and the proneurotrophins to induce p75NTR-mediated cell death in the presence of low concentrations of the survival factor BDNF was then tested. Treatment of E13 TrkB-dependent motor neuron cultures with NGF (Fig. 4A,D) or proNGF (Fig. 4C), resulted in significant neuronal death over 48 h, whereas proBDNF treatment had no effect (Fig. 4C). NGF-induced death was prevented by treatment with the p75NTR cleavage inhibitor TAPI-2 (Fig. 4A). Taken together, these results suggest that cleavage of p75NTR resulting in accumulation of the C-terminal fragment is associated with death signaling in motor neurons.
c29 promotes survival of BDNF-maintained neurons
We next asked whether c29 inhibits the (pro)NGF-induced death of motor neurons. Co-treatment of motor neuron cultures with c29 at the time of proNGF or NGF addition resulted in significantly reduced cell death of cultures grown in BDNF (Fig. 4B–D). The scrambled control peptide had no significant effect on either cell death or survival, and c29 had no effect on proBDNF-treated motor neurons (Fig. 4C).
To determine whether c29 was acting by affecting p75NTR cleavage, NSC-34 cells were treated with NGF or proNGF together with c29 or scrambled control peptide. No difference in the level of p75NTR cleavage was observed in cells treated with the peptides (Fig. 4E), indicating that c29 was more likely to be affecting the balance of survival signaling downstream of p75NTR activation.
To determine whether c29 was acting to prevent p75NTR-mediated death signaling or promote TrkB-mediated signaling, neurons cultured in ciliary neurotrophic factor (CNTF) or glial cell line-derived neurotrophic factor (GDNF) instead of BDNF were treated with NGF, which also resulted in a 40–50% reduction in survival, similar to that induced in BDNF conditions. The cleavage inhibitor TAPI-2 significantly inhibited NGF-induced cell death in CNTF- and GDNF-supported cultures (Fig. 4F,G). However, c29 did not inhibit the NGF-induced death of CNTF- or GDNF-dependent neurons (Fig. 4F,G). These results demonstrate that NGF-induced death signaling relies on a cleaved form of p75NTR, that its death signaling is independent of TrkB signaling, and, finally, that c29 does not appear to act in a dominant-negative manner in vitro to inhibit p75NTR signaling, but instead acts by promoting neuronal survival only in the presence of BDNF.
c29 protects motor neurons from BDNF, but not CNTF or GDNF withdrawal
We next tested whether c29 could potentiate BDNF survival signaling in conditions of growth factor withdrawal. Neurons plated and maintained in reduced BDNF concentrations (0.1 or 0.01 ng/ml) and treated with c29 had increased levels of survival compared with control-peptide-treated cultures (Fig. 5A). Similarly, in motor neuron cultures undergoing BDNF withdrawal (cultures changed to 0.1 ng/ml BDNF-containing medium following initial plating in 1 ng/ml BDNF), treatment with c29 resulted in significantly increased neuronal survival compared with that of cultures treated with control peptide, with survival at levels greater than or equal to that of cultures grown in 10-fold higher BDNF concentrations (Fig. 5B). c29 treatment had no survival effect in the absence of any exogenous growth factor (Fig. 5A), nor any effect on the survival of motor neurons undergoing CNTF (Fig. 5C) or GDNF withdrawal (Fig. 5D). The scrambled control peptide again had no significant effect on either cell death or survival. These results indicate that c29 has no intrinsic cell-survival-inducing properties, but rather that coincident BDNF–TrkB receptor-mediated signaling is required for c29 to promote neuronal survival in these conditions.
c29 does not regulate p75NTR cleavage but interacts with and promotes TrkB survival signals
To determine whether the c29 sequence was able to interact with TrkB (in addition to TrkA; Matusica et al., 2013) and thereby potentially affect activation of TrkB signals, we performed pull-down experiments. Biotinylated c29 peptide, but not a control biotinylated peptide (LC1), was able to pull down overexpressed TrkB from lysates of transfected HEK293 cells (Fig. 6A), as well as endogenously expressed TrkB from lysates of NSC-34 cells (Fig. 6B). These results indicate that the 29-amino acid intracellular juxtamembrane sequence is sufficient for p75NTR to interact with TrkB.
We next investigated whether c29 was potentiating survival signals by examining the activation of MAPK3 and MAPK1 (Erk1/2), and Akt in NSC-34 cells treated with BDNF for various periods of time (0–90 min). We found that c29 treatment resulted in a higher level of peak activation at all times measured for both Erk1/2 (Fig. 6C) and Akt (Fig. 6D). As Erk1/2 is regulated not only by BDNF (McCarty and Feinstein, 1999), but also by CNTF (Frebel and Wiese, 2006) and GDNF (Chiariello et al., 1998; Trupp et al., 1999) following activation of their specific receptors, we asked whether this response was specific for TrkB-mediated signaling. Analysis of activated Erk1/2 in NSC-34 cells treated with these growth factors for 15 min demonstrated that although BDNF-induced phosphorylation of Erk1/2 was significantly enhanced in the presence of c29, c29 treatment of CNTF- or GDNF-supported cultures had no effect on the level of phosphorylated Erk1/2, which was equivalent to that of cultures treated with scrambled control peptide (Fig. 6E). These results indicate that, in conditions where growth factors are limited, c29 enhances survival signaling and thus neuronal survival in a BDNF–TrkB-dependent manner.
Finally, to establish whether c29 could activate BDNF-mediated survival signaling within the spinal cords of SOD1G93A mice, 16-week-old SOD1G93A animals were treated acutely with c29 or control peptide. SOD1G93A mice treated with the control peptide had very low levels of phosphorylated TrkB (Fig. 7A), Erk1/2 and Akt (Fig. 7B), and CREB (Fig. 7C), in comparison to the levels in lysates from wild-type animals. However, the activity of these survival-signaling proteins was significantly increased by acute c29 treatment in comparison to that in control peptide-treated animals (Fig. 7A–C). Conversely, levels of cleaved (activated) caspase 3 in SOD1G93A spinal cord lysates were reduced by c29 treatment compared with the control treatment (Fig. 7D). Together, these findings indicate that c29 can enhance neuronal survival in a BDNF–TrkB-dependent manner in vitro and in diseased motor neurons in vivo.
DISCUSSION
In this study we demonstrated that the c29 peptide, derived from the intracellular juxtamembrane domain of p75NTR, can promote neuronal survival in motor neurons in vitro, as well as following axotomy and in disease conditions in vivo. Although c29 treatment reduced the amount of the ECD fragment of p75NTR in vivo, and motor neuron cell death induced by neurotrophins in vitro was accompanied by increased cleavage of p75NTR, c29 did not directly affect p75NTR cleavage in vitro. Rather, it interacted with TrkB and acted in a BDNF-dependent manner to promote survival signaling, without directly inhibiting NGF-induced p75NTR death signaling.
NGF triggers an active motor neuron death pathway mediated by the p75NTR C-terminal fragment
p75NTR is well known to mediate neuronal death in conditions of injury and disease, typically triggered by p75NTR ligands that do not bind to co-expressed Trk receptors (Ibanez and Simi, 2012), i.e. in motor neurons, NGF or proNGF elicit p75NTR-mediated death signals (Henderson et al., 1998; Sedel et al., 1999; Wiese et al., 1999; Sendtner et al., 2000). There is also emerging evidence that regulated intramembrane proteolysis might be important for p75NTR functions, including induction of cell death signaling pathways and promotion of cell survival in vitro, although reports of ligand activation of p75NTR cleavage are inconsistent (Kenchappa et al., 2006; Skeldal et al., 2011; Vicario et al., 2015). Our in vitro experiments measuring cleavage of endogenous p75NTR in response to exogenous ligand application established that cell death ligands increase the production of the C-terminal fragment, which is associated with increased death signaling (Coulson et al., 1999; Underwood et al., 2008; Vicario et al., 2015), whereas cell survival ligands result in greater production of the intracellular domain fragment, consistent with Trk-mediated cleavage of p75NTR facilitating survival through this fragment (Ceni et al., 2010; Kommaddi et al., 2011; Matusica et al., 2013).
In vitro, motor neuron death was induced through activation of p75NTR by either NGF or proNGF in the presence of BDNF survival signaling, consistent with previous reports that survival or death is regulated by the strength of competing signaling pathways (Song et al., 2010). However, NGF-induced cell death ensued even in the absence of Trk activity, i.e. in cultures maintained in GDNF or CNTF, indicating that NGF triggers an active cell death pathway which is not simply a result of passive loss of growth factor support resulting from sequestration of p75NTR away from a BDNF–TrkB high affinity complex by NGF, as previously suggested (Davies et al., 1993; Sendtner et al., 2000; Vesa et al., 2000). In addition, ligand-dependent death mediated by p75NTR in any of the growth factor conditions tested was blocked by inhibiting p75NTR cleavage, indicating that cleavage and ligand were equally important in inducing death signaling.
c29 promotes motor neuron survival in vivo
Our experiments aimed to test the effect of the c29 peptide in vivo in conditions in which p75NTR death signaling was active, namely, following axotomy and in the SOD1G93A model of amyotrophic lateral sclerosis. Notably, both of these conditions trigger increased cleavage of p75NTR (DiStefano et al., 1993; Perlson et al., 2009), and we observed increased p75NTR ECD levels with SOD1G93A disease progression, as previously observed in human amyotrophic lateral sclerosis (Shepheard et al., 2014). As a result of c29 treatment, we observed a significant improvement in motor neuron survival after axotomy and in SOD1G93A animals, and in the latter model, reduced levels of p75NTR ECD, and a coincident improvement in rotarod performance and a delay in disease-induced weight loss, suggesting that the functional outcomes could be the result of reduced p75NTR activation and cell death signaling. However, the observed reduction in the level of p75NTR ECD might simply reflect a slower disease progression. Although p75NTR expression and the generation of the cleavage fragment are upregulated in motor neurons coincident with axonal degeneration, p75NTR is also upregulated in oligodendrocytes, Schwann cells and microglia following nerve injury (reviewed in Meeker and Williams, 2014). Therefore, it is not clear from our experiments whether the p75NTR ECD is derived from motor neurons and/or other cell types and whether it reflects basal levels of p75NTR cleavage accompanying broader expression, or increased cleavage specifically reflecting activation of p75NTR degenerative processes. The reduction in the level of p75NTR ECD afforded by c29 treatment is therefore difficult to interpret in isolation.
Although we cannot rule out the possibility that c29 acted to prevent p75NTR death signaling in vivo, our in vitro experiments strongly indicate that it did not directly affect the level of p75NTR cleavage or block p75NTR death signaling. Firstly, no effect on the level of NGF- or proNGF-induced cleavage of p75NTR was observed following treatment of cultures with c29, nor did c29 treatment affect the level of NGF-induced death of neurons cultured in the absence of BDNF, even though p75NTR cleavage inhibitors promoted neuronal survival in the same conditions. Rather, our data suggest that c29 treatment, by keeping motor neurons alive or healthier for longer (as discussed below), indirectly affected other disease mechanisms, including reducing the upregulation and/or cleavage of p75NTR. Therefore, the specific process being measured by quantifying p75NTR ECD levels remains unclear. Nonetheless, as p75NTR ECD levels increased with disease progression and c29 treatment resulted in reduced steady-state levels of p75NTR ECD in mid disease, we suggest that further validation of this fragment as a biomarker of motor neuron disease severity is warranted.
c29 promotes motor neuron survival in vitro but requires coincident TrkB activation
Although we hypothesized that c29 might act to block p75NTR death signaling, our results consistently demonstrated that c29 only promotes motor neuron survival when coincident with TrkB activation in vitro. Specifically: (i) c29 overcame NGF-induced death signaling in neurons cultured in BDNF but not in neurons cultured in CNTF; (ii) in low concentrations of BDNF that would not otherwise support motor neuron survival, co-treatment of cultures with c29 promoted levels of survival at or above that of cultures in 10-fold higher BDNF concentrations, but had no effect on cultures grown in low concentrations of CNTF or GDNF; and (iii) c29 treatment facilitated both increased activation of and longer-lived survival signaling in motor-neuron-like NSC-34 cells, but only in cultures grown in the presence of BDNF. This conclusion is consistent with our in vivo findings.
In the axotomy paradigm, an exogenous supply of BDNF is reported to promote neuronal survival (Sendtner et al., 1992; Boyd and Gordon, 2001; Murray et al., 2010). Thus, given that at least low endogenous levels of neurotrophins are present at the axotomized cell bodies, c29 could have acted by enhancing the survival signaling of locally produced BDNF. Similarly, previous work has indicated that BDNF is upregulated in neurons, glia and immune cells within the spinal cord following injury (Li et al., 2006,, 2009; Xin et al., 2012), which might allow c29 to act in motor neurons. Indeed, acute treatment of SOD1G93A animals with c29 resulted in enhanced phosphorylation of TrkB, Akt, Erk1/2 and CREB within the spinal cord coincident with reduced levels of cleaved caspase 3, which was not induced by control peptide treatment. Our current results therefore provide strong evidence to support the idea that c29 enhances BDNF-induced TrkB survival signaling both in vitro and in vivo, which in turn blocks death signals.
Role of the p75NTR juxtamembrane (c29) region in Trk signaling
One characterized role of p75NTR is its ability to increase ligand-binding affinity for the TrkA receptor and increase neuronal survival (Hempstead et al., 1991; Lee et al., 1994; Horton et al., 1997; Esposito et al., 2001), and it is widely accepted that p75NTR also interacts with other Trk receptors with equivalent effect (Hempstead et al., 1991; Davies et al., 1993; Huber and Chao, 1995; Bibel et al., 1999; Vesa et al., 2000; Ceni et al., 2010; Kommaddi et al., 2011). Although the precise mechanism responsible for this outcome remains elusive (Barker, 2007), the transmembrane and juxtamembrane regions of p75NTR have been implicated in this interaction (Hempstead et al., 1991; Bibel et al., 1999; Esposito et al., 2001; Iacaruso et al., 2011), and more recent studies also suggest that cleavage of p75NTR plays an important role in this process (Ceni et al., 2010; Kommaddi et al., 2011; Matusica et al., 2013). Our observation that a soluble peptide mimicking the juxtamembrane region of p75NTR is able to enhance the function of TrkA (Matusica et al., 2013) and TrkB (herein) accords with the previous studies. In particular, the high conservation of the juxtamembrane region of p75NTR supports the idea that this domain is essential for p75NTR function (Hutson and Bothwell, 2001). We have previously shown its importance in p75NTR-mediated death signaling when attached to the membrane as well as demonstrating that the function of this domain differs following intramembranous cleavage (Coulson et al., 2000; Matusica et al., 2013), i.e. cleavage of p75NTR to its intracellular domain fragment is a requirement for enhancing NGF affinity for TrkA in vitro, and this can be reproduced using the c29 peptide (Matusica et al., 2013).
Proposed mechanism of c29 action
As the current findings using c29 in vitro mirror our recent work in TrkA-dependent neurons (Matusica et al., 2013), we propose that c29 interacts with all Trk receptors to facilitate ligand binding. Increased numbers of ligand-bound TrkB receptors per cell allow activation of significantly more receptors than are normally activated in equivalent concentrations of neurotrophins (unless the concentrations were already saturating). Thus, the threshold for BDNF survival and trophic signaling per neuron would be reached sooner and/or be higher in the presence of c29. In the context of our experiments, the resulting increased TrkB survival signaling was sufficient to promote survival, including overriding coincident p75NTR-mediated death signals. Although recent work has suggested that TrkB expression causes a more severe SOD1 disease phenotype (Zhai et al., 2011), it appears that this deleterious effect is mediated by the truncated TrkB isoform T1 (Yanpallewar et al., 2012). TrkB.T1 lacks the intracellular kinase domain and, unlike full-length TrkB, is upregulated in motor neurons in SOD1 animals (Zhang and Huang, 2006). Thus, the effect of promoting motor neuron survival in SOD1 animals by c29 can be explained by selective enhancement of full-length TrkB trophic signaling, which can counteract p75NTR-induced death (Casaccia-Bonnefil et al., 1996; Yoon et al., 1998; Friedman, 2000).
Alternatively or additionally, c29 could also have acted in vivo to promote motor neuron survival through TrkC (or TrkA), which is upregulated in the longer-lived motor neurons in motor neuron disease patients (Duberley et al., 1997; Nishio et al., 1998). Unlike BDNF, exogenous delivery of the TrkC ligand NT3 has shown promise in slowing disease progression and maintaining motor neuron–muscle connectivity in a motor neuron disease mouse model (specifically the pmn mouse; Haase et al., 1998). Given that c29 acts in a similar fashion in vitro to promote both TrkA and TrkB signals, it would be surprising if c29 did not affect TrkC in the same manner. Thus, the ability of c29 to promote motor neuron survival in vivo can be explained by our in vitro findings that Trk receptor function (TrkB and/or TrkC) is enhanced in neurons in conditions of low-level neurotrophin exposure or activation of p75NTR death signaling. Even though c29 treatment did not alter the lifespan of the aggressive SOD1G93A model mice, it remains to be determined whether earlier treatment (e.g. prior to when p75NTR cleavage is first evident), a different dose, a more biostable mimetic, or a targeted treatment route of c29, either alone or in combination with other candidate therapeutics, would provide any additional benefit to animals undergoing motor neuron degeneration.
In summary, our results show that regulated intramembrane proteolysis of p75NTR plays an important role in (pro)NGF-mediated motor neuron death, which can be induced even in the presence of coincident survival signals. We also demonstrate that a peptide mimetic of the p75NTR intracellular domain (c29), that interacts with TrkB receptors, potentiates the survival of motor neurons in a BDNF- and TrkB-dependent manner, under conditions of growth factor withdrawal as well as active ligand killing, by redressing the neurotrophin signaling imbalance in favor of trophic support.
MATERIALS AND METHODS
c29 and control peptide synthesis
A 29 amino acid peptide of the juxtamembrane ‘Chopper’ domain (Coulson et al., 2000) (c29: KRWNSCKQNKQGANSRPVNQTPPPEGEKL) and a randomly scrambled version (SC: SKGQVCRNQPGQNKPEPANKSWKETPLRN) were synthesized as N-terminal fusions to a non-naturally occurring TAT-like protein transduction domain (PTD) peptide (YARAAARNARA) (Ho et al., 2001) using t-boc chemistry then purified using reverse phase HPLC by Dr James I. Elliott (Yale University). For axotomy experiments, PTD was conjugated to c29 by means of a disulfide link (Coulson et al., 2000) by Auspep. No difference in functionality was seen between the two conjugation methods in cell culture assays, and no effects were seen in cells treated with PTD peptide alone, peptides without carrier or disulfide-linked peptides that had been reduced with DTT. For immunoprecipitation experiments, the unconjugated c29 peptide was labeled on the amino terminus with biotin through a six-carbon spacer. Except for these immunoprecipitation experiments, ‘c29’ refers to the PTD-conjugated form of the c29 peptide. The biotinylated control peptide mimicking the p75NTR extracellular juxtamembrane domain LC1 (H-RGTTDNLIGGSC-NH2) was manufactured by Auspep.
Motor neuron axotomy
The Institutional Animal Ethics Committee approved all experiments involving the use of animals, in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. Ulnar nerve axotomy was performed on newborn Wistar rat pups as previously described (Lowry et al., 2001). Peptides (PTD–c29 and PTD only; 100 nM) were applied by means of a 1 mm3 cube of pluronic gel foam that was sutured to the proximal axonal stump (Cheema et al., 1996). After 5 days the animals were perfused with a 4% solution of paraformaldehyde in phosphate buffer. Cervical spinal cords were removed, after which serial sections were cut at 40 µm, mounted onto gelatin-coated slides and stained in 0.1% Cresyl Violet. Details of the counting procedure have been described previously (Lowry et al., 2001). Briefly, motor neurons in the spinal cord displaying a prominent nucleolus were counted in every fifth section. The percentage neuronal survival after axotomy was calculated based on the nucleolar number on the axotomized side as a percentage of the nucleolar number in the contralateral side. Data were analyzed by two-tailed t-test.
SOD1 animal experiments
Treatments
Male and female transgenic mice with the G93A human SOD1 mutation (SOD1G93A; line B6SJL-TgN [SOD1-G93A] 1Gur; Jackson Laboratories), and in some cases age-matched C57Bl6 mice, were used. Some cohorts of mice were administered peptides by a single intraperitoneal injection (5 mg/kg) and sacrificed 60–90 min later by perfusion (for histology) or cervical dislocation (biochemical analyses). Alternatively, mice were treated systemically with peptides starting from 9 weeks of age (the time at which p75NTR expression is upregulated in spinal motor neurons; Lowry et al., 2001). The animals were randomly assigned a treatment group and implanted with a 28-day subcutaneous osmotic minipump (Alzet model 1004) delivering ∼5 mg/kg/day PTD–c29 (flow rate: 0.11 µl/h) or vehicle and the pumps were replaced monthly until death or euthanasia.
Behavioral analyses
Mice were weighed weekly until close to end stage when they were monitored daily. Urine discharged during weighing was collected and frozen. Testing of motor function using a rotarod device began at 10 weeks of age. Each weekly session consisted of three trials on the elevated accelerating rotarod (Ugo Basile) beginning at 5 rpm and accelerating to 20 rpm over 20 s for a total duration of 3 min. To determine mortality in a reliable and humane fashion, we used an artificial end point, defined as 80% minimum original weight threshold. The moribund mice were scored as ‘dead’ and were euthanized. Survival analysis of the SOD1G93A transgenic mice was performed using a log-rank (Mantel–Cox) test and analysis of motor neuron cell numbers in spinal cord sections was performed using a Student's t-test. Between-group analyses of weight were performed by repeated measure ANOVA followed by a Sidak's test of analysis of the differences between groups. A two-tailed Mann–Whitney test was used to compare the time it took mice to lose 10% of their maximal weight. Rotarod data were analyzed by ANOVA with a Sidak's multiple comparison test.
Biochemical analysis of p75NTR extracellular domain fragment
Urine was collected from SOD1G93A and wild-type mice during weighing, using metabolic cages or from the bladder after euthanasia. However, insufficient urine for analysis was obtained from some mice at each time point. For peptide-treated mice, urine was collected from six c29-treated and six saline-treated SOD1 animals on an at least weekly basis. For analysis, we pooled data obtained from consecutive samples such that many data points contain data derived from two samples from the same animals but collected a week or less apart. A sandwich ELISA was used to measure the extracellular domain fragment of p75NTR (p75NTR ECD) in urine as previously described (Shepheard et al., 2014). Briefly, ELISA plates (96-well plates, Costar Corning) were coated for 18 h with anti-p75NTR ECD MLR1 (4 µg/ml in 25 mM sodium carbonate 25 mM, sodium hydrogen carbonate, 0.01% thimerosol, pH 9.6) at 4°C. Wells were blocked with sample buffer comprising phosphate buffered saline (PBS), 2% bovine serum albumin (BSA; Sigma) and 0.01% thimerosol, pH 7.4, for 1 h at 37°C. Recombinant human mouse p75NTR ECD (control antigen, aa:20–243; R&D Systems) and urine samples were diluted in sample buffer and incubated for 20 h at room temperature in the antibody coated wells. For detection, wells were incubated in goat anti-p75NTR ECD (R&D Systems) for 1 h followed by a bovine anti goat IgG-HRP antibody (Jackson ImmunoResearch) for 1 h at room temperature. Visualization was by means of a peroxidase reaction developed using the color reagent 3,3′,5,5′-tetramethylbenzidine (TMB; Bio-Rad) and stopped with 2 M sulfuric acid. Between steps, plates were washed 4 times with wash buffer (PBS, 0.05% Tween 20, 0.01% thimerosol, pH 7.4). Plates were read at 450 nm with a PerkinElmer Victor-x4 Plate Reader. Mouse creatinine was used as a standard for urinary protein measurements and was assessed using a creatinine detection kit (Enzo Life Sciences) as previously described (Shepheard et al., 2014). p75NTR-ECD-specific ELISA results were analyzed by one-way ANOVA, with a Bonferroni's multiple comparison test.
Peptide distribution and motor neuron survival
To assess the biodistribution of c29 and motor neuron survival in treated postnatal day 115 SOD1G93A mice, animals were first anesthetized and then intracardially perfused with PBS, followed by 4% paraformaldehyde in PBS. Spinal cords were dissected and post-fixed in 4% paraformaldehyde at room temperature for 2 h.
Fixed spinal cord samples were cryoprotected overnight by immersion in PBS with 20% sucrose, embedded in optimal cutting temperature compound (OCT; Tissue Tek, Raymond Lamb Ltd Medical Supplies) and frozen on dry ice. All spinal cord samples were stored at −80°C until needed, and then cryosectioned.
For biodistribution, 20 µm-thick lumbar spinal cord sections were stained using streptavidin-conjugated horseradish peroxidase. Motor neuron numbers were determined on 20 µm-thick lumbar spinal cord sections stained with ChAT. Sections were washed in 0.1 M phosphate buffer, pH 7.4 and then incubated in 3% H2O2, 50% ethanol to inactivate endogenous peroxidase. Sections were subsequently washed in 0.1 M phosphate buffer, and incubated for 1 h at room temperature in blocking solution (phosphate buffer, 10% horse serum, 0.1% Triton X-100). Primary antibody (goat anti-ChAT, 1:1000; Millipore, AB144P) was diluted in the above blocking buffer and applied onto sections for overnight incubation at room temperature. Sections were then washed at room temperature in phosphate buffer containing 0.1% Triton X-100 before incubation with biotinylated secondary antibody (1:1000; Jackson Immunoresearch Laboratories, #705-067-003) and avidin-biotin complex (ABC) reagents (Vector Elite Kit, Vector Laboratories; 6 µl/ml avidin and 6 µl/ml biotin). Black immunoreactivity was revealed by a nickel-intensified diaminobenzidine reaction. Sections were finally washed in phosphate buffer and gradually dehydrated in ethanol and xylene before being mounted in DePex mounting medium (VWR International). A total of 10 sections per mouse were analyzed. Motor neuron numbers were compared using a two-tailed t-test.
Spinal cord lysate preparation
Following a 60 min treatment of wild-type C57Bl6 and SOD1G93A mice with c29 or control peptide delivered intraperitoneally, mice were killed by cervical dislocation and spinal cords were dissected immediately after sacrifice and snap-frozen. The lumbar spinal cords were thawed and dissected in RIPA buffer containing 50 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 1 mM EDTA, 1 mM sodium orthovanadate, 1 mM PMSF, 1 µM BB94 (Betimastat), 15 mM sodium fluoride, 40 mM β-glycerophosphate, 0.5 mM 4-deoxypyridoxine, 1% (v/v) Roche Complete-Mini inhibitor cocktail stock and 1 mM DTT at pH 7.4. They were then lysed in 300 µl of the same buffer using a Retch Tissue Lyser (Qiagen) for 3 min at 4°C. Lysates were centrifuged at 16,000 g and pellets discarded. Protein estimation of all lysates was done using the Pierce BCA protein assay kit (Thermo Fisher). Western blots were performed as described below.
Motor neuron isolation and culture
Cultures of spinal motor neurons were prepared from embryonic day (E)13 wild-type C57BL/6J mice and isolated by immunopanning as previously described (Wiese et al., 2010). Briefly, pregnant dams were sacrificed by cervical dislocation and the uterus removed by Cesarean section and placed in neurobasal medium (Gibco). Embryos were removed from the uterine sac and the lumbar spinal cords were dissected on ice. Batches of two or three spinal cords were digested in 1 ml of 0.05% trypsin–ETDA (Gibco) for 10 min at 37°C in 15 ml tubes. To prevent further digestion, 0.14 mg/ml trypsin inhibitor (Sigma) solution was added, and the tissue was centrifuged at 100 g for 7 min. Spinal cords were trituated manually for 2 min with a polished glass pipette, after which the suspension was filtered through a 100 μm sieve (Falcon). The remaining tissue was further triturated ∼8 times and filtered through another 100 μm sieve. The combined cell suspensions were then immunopanned by being plated onto a 10 cm culture dish (Nunc) pre-coated with MLR-2 mouse anti-p75NTR antibody (Rogers et al., 2006). The dish was pre-coated with 8 ml of immunopanning solution containing 1.5 μg/ml MLR-2 in 10 mM Tris-HCl, pH 9.5, for 2 h and washed with neurobasal medium before the spinal cord suspension was applied. After 45 min incubation, the dish was washed three times with neurobasal medium and the remaining attached cells were isolated from the plate by incubation for 1 min in 0.85 M KCl. The collected cell suspension was centrifuged at 400 g for 5 min at room temperature. The supernatant was removed and the cell pellet was resuspended in 1 ml of neurobasal medium. Neurons were plated at a density of 3000 live cells per well in 4-well culture dishes (Greiner) pre-coated with 0.0015% poly-ornithine (overnight at 4°C) and 3.75 μg/ml laminin (for 30 min at 37°C). Cells were allowed to adhere for 1 h at 37°C, after which 2 ml of motor neuron culture medium was added [neurobasal medium supplemented with B27 (Life Sciences), 500 μM glutamax (Life Sciences), 10% horse serum (Linaris), 25 μM β-mercaptoethanol (Merck) and 1 ng/ml recombinant human BDNF (Peprotech)]. Cells were maintained at 37°C in a humidified atmosphere with 5% CO2 for 24 h, after which the culture medium was replaced, live cells counted (day 1), and the experimental treatments administered. Phase contrast images of motor neurons were taken using an Olympus (6IX81) microscope.
Neuronal survival assays
In addition to BDNF, the following growth factors were used in primary neuronal culture assays at concentrations indicated in individual experiments: purified mouse 2.5S NGF (Biosensis), proNGF (SCIL proteins GmbH), proBDNF (Biosensis), recombinant rat cilliary neurotrophic factor (CNTF; R&D Systems) and recombinant human glial-derived neurotrophic factor (GDNF; Peprotech). For the neuronal survival assays, BDNF, CNTF or GDNF was added at the time of plating as indicated. A concentration of 1 ng/ml for each trophic factor has previously been determined to result in optimal rates of survival (Wiese et al., 1999). In order to count the same field at each time point, grids were etched onto the plates using an 18-gauge needle. Motor neurons were counted on day 1 (four separate fields for each well) and the same fields were recounted on day 3. Cell survival was expressed as a percentage compared with survival in BDNF alone on day 1.
For assays investigating the effects of the induction of p75NTR cleavage by the metalloproteinase enhancer phorbol 12-myristate 13-acetate (PMA; 200 nM), and blocking p75NTR cleavage with the selective non-competitive metalloproteinase inhibitors TAPI-2 (200 nM, extracellular cleavage) and Compound E (20 µM, intracellular cleavage; Tian et al., 2002), the compounds were added to established cultures on day 1 after the initial cell counts and cells were then counted again on day 3.
To assess the effects of c29 peptide on motor neuronal survival in limiting growth factor concentrations, isolated spinal motor neurons were plated and treated with BDNF, as indicated, in the presence of 1 µM PTD-linked c29 or scrambled peptide, then cultured for three days with the addition of fresh BDNF and peptide treatments on day 2. For growth-factor-withdrawal assays, following initial plating in 1 ng/ml of BDNF, CNTF or GDNF, cultures were washed four times with serum-free neurobasal medium, after which the medium was replaced with medium containing 0.1 ng/ml BDNF, CNTF or GDNF and 1 µM PTD-linked c29 or scrambled peptide, as indicated.
For assays investigating the effect of c29 on NGF-mediated motor neuron death, cultures were grown in 1 ng/ml BDNF, CNTF or GDNF for one day, before being treated with 100 ng/ml 2.5S NGF and 1 µM PTD-linked c29 or scrambled peptide, as indicated. Motor neurons were counted on day 1 and, to assess the effects of various treatments, again on day 3.
Rates of survival were compared between conditions by ANOVA with Bonferonni post-hoc testing.
Immortalized cell culture, pull-down experiments and western blotting
Mycoplasma-free neuroblastoma×spinal cord (NSC-34) cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Gibco) supplemented with 10% FBS and 1% penicillin/streptomycin/glutamine solution (PSG; Cashman et al., 1992). For experiments, cells were grown to 60% confluency in medium comprising 1:1 DMEM:Ham's F12, plus 1% N-2 supplement, 1% PSG, 1% modified Eagle's medium non-essential amino acids, and 1 µM retinoic acid (Sigma). The cells were allowed to differentiate for up to three days, with the medium being replaced after 48 h.
For p75NTR-shedding experiments, NSC-34 cells were cultured in poly-L-ornithine-coated 6-well plates, starved for 1 h in serum-free DMEM, and treated with 200 nM PMA, 200 nM TAPI-2, or 20 µM Compound E; 50 ng/ml NGF, 5 ng/ml proNGF, 50 ng/ml BDNF or 5 ng/ml proBDNF for 6 h. To inhibit degradation of the p75NTR intracellular domain fragment, cells were treated with 1 µM epoxomicin (Sigma) 1 h prior to the addition of other compounds.
For experiments measuring phosphorylated TrkB, Akt, and Erk1/2, NSC-34 cells were differentiated for five days before being serum starved for 4 h in serum-free DMEM. Cells were then pretreated with 1 µM c29 or scrambled control peptide for 1 h prior to addition of 10 ng/ml of BDNF followed by treatments with growth factors as indicated for time intervals ranging from 1–90 min. Following treatments, cells were immediately washed with ice cold PBS, lyzed and prepared for SDS-PAGE analysis as described below.
Cells were lyzed using chilled lysis buffer containing 10 mM Tris-HCl, pH 8.0; 150 mM NaCl; 2 mM EDTA; 1% NP-40; 1% Triton X-100; 10% glycerol; 1 mM phenylmethanesulfonyl fluoride; 1 mM sodium orthovanadate; 1 µM batimastat (BB-94) and 1% Roche protease inhibitor cocktail (Ceni et al., 2010).
HEK293 cells used for pull-down assays were routinely cultured in DMEM supplemented with 10% FBS, in 5% CO2 at 37°C. For experiments, cells were transfected with TrkB plasmid DNA using FuGENE 6 transfection reagent (Roche Applied Science), lyzed 48 h after transfection and the lysates used for c29 pull-down experiments as described below.
To perform pull-down assays, biotinylated c29 or LC1 control peptide was bound to Dynabeads MyOne Streptavidin T1 (Invitrogen) before incubation with transfected HEK293 or NSC-34 cell lysates in the presence of 2% BSA. Nonspecific binding was removed by washing the beads five times in Tris-buffered saline (TBS) supplemented with 0.05% Tween 20, and the proteins were eluted by boiling samples in reducing sample buffer (20 mM dithioerythritol, 2.5% SDS).
Cell lysates, lumbar spinal cord lysates or pulled-down proteins were electrophoresed through 4–12% Bis-Tris buffered SDS gels (Life Sciences), transferred onto a PVDF or nitrocellulose membrane and probed using the following antibodies: rabbit anti-p75NTR intracellular domain #9992 (1:5000; a gift from Moses Chao), rabbit anti-TrkB (1:500; Biosensis, #R121-100), phospho-TrkB (Ser478) (1:250; a gift from Biosensis), mouse pan anti-Akt (1:2000; Cell Signaling Technologies, #2920S), XP™ rabbit mAb rabbit anti-phosphorylated serine 473 (Ser 473) Akt (1:2000; Cell Signaling Technologies, #4060), XP™ rabbit mAb phospho-p44/42 MAP kinase (Erk1/2; Thr202 and Tyr204) (1:2000; Cell Signaling Technologies, #4370), rabbit phospho-p44/42 MAP kinase (Erk1/2; Thr202 and Tyr204) (1:1000; Millipore, #05-797R), rabbit anti-pan p44/42 Erk1/2 (1:1000; Cell Signaling Technologies, #9101), rabbit cleaved caspase-3 (Asp175) (1:1000; Cell Signaling Technologies, #9661) and rabbit caspase-3 (8G10) (1:1000; Cell Signaling Technologies, #9665), CREB (48H2, #9197) and phospho-(Ser133) CREB (87G3, #9198) (1:1000; Cell Signaling Technologies), and mouse anti-beta-III tubulin (1:2000; Promega, #G7121). The membranes were blocked for 1 h at room temperature in 0.1% Tween-20, and 0.02% NaN3 in TBS, pH 8.0, and either 4% skim milk powder for p75NTR detection, 3% BSA for phosphorylated Erk1/2 and Akt, or 1% BSA for phosphorylated TrkB, before being incubated overnight with primary antibodies at 4°C. Membranes were then washed in TBS-Tween 20 (TBST), and incubated for 1 h with LICOR donkey anti-rabbit 680 or donkey anti-mouse 800 secondary antibody (1:50,000; Invitrogen, #925-68073, #926-32212) in TBS at room temperature after which they were washed and imaged using the Gel Doc XR imaging system (Bio-Rad). Alternatively, blots were probed with appropriate species-specific HRP-conjugated secondary antibodies (1:10,000; Jackson ImmunoResearch, #115-035-003/ #111-035-003) and visualized with either Supersignal West Femto Sensitivity Substrate (Pierce) or Clarity™ Western ECL substrate (Bio-Rad) using a Fujifilm ImageQuant™ LAS-4000 (Fuji Corporation). For stripping and re-probing, western blot membranes were treated with 10 ml of Restore™ PLUS Western Blot Stripping Buffer (Thermo Fisher) for 30 min at room temperature on a orbital shaker, washed three times with TBST and blocked with either 5% skim milk powder or 3% BSA in TBST for 1 h. The membranes were then incubated with appropriate primary and secondary antibodies as described above.
Statistical analyses
N represents the number of individual animals; n represents the number of individual cell culture experiments, each with four replicates. Stereology, rotarod and weight measurements, p75NTR ECD analysis and, for key experiments, the motor neuron culture counts were conducted by researchers blinded to the treatment conditions. Densitometry analysis of relative protein levels on western blots was performed using Image Studio Digits (Licor), Image Lab (Bio-Rad) and Fiji (ImageJ, NIH). All representative graphs measuring relative densitometry of proteins on western blots were calculated from three independent experiments, with bands normalized to loading controls (tubulin or actin) probed and exposed at the same time. Statistical analysis was performed using GraphPad Prism (GraphPad Software) as described in each method. P<0.05 was considered statistically significant.
Acknowledgements
Robert Rush (Flinders University, Adelaide, Australia) kindly provided the TrkB antibody used for western blotting and immunoprecipitation experiments. Moses Chao (Skirball Institute, New York, USA) kindly provided the 9992 p75NTR antibody for western blotting. We thank Stephan Weise and Michael Sendtner (Würzburg University, Germany) for assistance in establishing the NGF-killing motor neuron assay. NSC-34 cells were kindly provided by Dr Neil Cashman (University of British Columbia). We thank other members of the Coulson laboratory past and present for helpful discussions and Rowan Tweedale for critical reading of the manuscript and editorial assistance.
Footnotes
Author contributions
All authors executed and interpreted experiments and contributed to the drafting the article. D.M., F.A., C.U. and E.J.C. conceived and designed the experiments and D.M. and E.J.C. revised the article.
Funding
This work was supported by the National Health and Medical Research Council of Australia [Career Development Fellowship 569601 and project grant 10012610 to E.J.C.]; the Australian Research Council; and NuNerve Pty Ltd [Linkage grant LP10012610 to E.J.C.]; Goodenough and Wantoks bequest (to E.J.C and M.M.); The Australian Government (Postgraduate Award to C.K.U); a Ross Maclean Fellowship (to M.M.); the Centre for Neuroscience, Flinders University and the Flinders Medical Centre Research Foundation (grants to D.M.); Australian Rotary Health (Neville & Jeanne York Motor Neuron Disease PhD Scholarship to S.R.S); and the Motor Neuron Disease Reseach Instititue of Australia (Rosalind Nicholson Grant to M-L.R.).
References
Competing interests
Patents covering the use of c29 as a therapeutic in a range of neurological conditions are held by the University of Queensland.