SEARCH

SEARCH BY CITATION

Keywords:

  • ciliary neurotrophic factor;
  • motoneuron disease;
  • spinal muscular atrophy;
  • stathmin;
  • STAT3

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms used by CNTF signaling for axon maintenance
  5. Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin
  6. Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics
  7. Local protein synthesis in motor axons: function in axon dynamics and plasticity
  8. Defective mRNA processing and axonal defects in SMA
  9. Concluding remarks
  10. Acknowledgement
  11. References

In motoneuron disease and other neurodegenerative disorders, the loss of synapses and axon branches occurs early but is compensated by sprouting of neighboring axon terminals. Defective local axonal signaling for maintenance and dynamics of the axonal microtubule and actin cytoskeleton plays a central role in this context. The molecular mechanisms that lead to defective cytoskeleton architecture in two mouse models of motoneuron disease are summarized and discussed in this manuscript. In the progressive motor neuropathy (pmn) mouse model of motoneuron disease that is caused by a mutation in the tubulin-specific chaperone E gene, death of motoneuron cell bodies appears as a consequence of axonal degeneration. Treatment with bcl-2 overexpression or with glial-derived neurotrophic factor prevents loss of motoneuron cell bodies but does not influence the course of disease. In contrast, treatment with ciliary neurotrophic factor (CNTF) significantly delays disease onset and prolongs survival of pmn mice. This difference is due to the activation of Stat-3 via the CNTF receptor complex in axons of pmn mutant motoneurons. Most of the activated Stat-3 protein is not transported to the nucleus to activate transcription, but interacts locally in axons with stathmin, a protein that destabilizes microtubules. This interaction plays a major role in CNTF signaling for microtubule dynamics in axons. In Smn-deficient mice, a model of spinal muscular atrophy, defects in axonal translocation of β-actin mRNA and possibly other mRNA species have been observed. Moreover, the regulation of local protein synthesis in response to signals from neurotrophic factors and extracellular matrix proteins is altered in motoneurons from this model of motoneuron disease. These findings indicate that local signals are important for maintenance and plasticity of axonal branches and neuromuscular endplates, and that disturbances in these signaling mechanisms could contribute to the pathophysiology of motoneuron diseases.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms used by CNTF signaling for axon maintenance
  5. Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin
  6. Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics
  7. Local protein synthesis in motor axons: function in axon dynamics and plasticity
  8. Defective mRNA processing and axonal defects in SMA
  9. Concluding remarks
  10. Acknowledgement
  11. References

When neurons signal to other neurons or to target cells such as muscle, they normally do this through their axons by release of neurotransmitters from axon terminals. Axons of motoneurons can reach a length of more than 1000 mm. This is more than 20 000 times larger than the diameter of their cell body. Motor axons project to neuromuscular endplates that are among the largest synapses, and the number of these motor endplates that are served by one single motoneuron cell body can exceed several thousand in mammals (McPhedran et al. 1965). Thus, the axonal side of these neurons represents the major part of the cell, with high numbers of nerve terminals residing long distances apart from the cell body. This raises the question how these axons and axonal processes are maintained throughout life. In mammals, motoneurons are generated early during development. They grow out their axons towards developing skeletal muscle, and developing myofibers are usually innervated by more than one axon branch. The excess of axon terminals that results in double and even triple innervation of single muscle fibers is corrected later in development (Sanes & Lichtman, 1999), in rodents in the second and third postnatal week. During this period, extensive pruning occurs indicating that the potential of axons to remodel is very high and that efficient mechanisms exist which maintain some axonal branches but not others during this period.

Also in the adult, axon branches appear as dynamic structures. Normal muscle use and exercise modulate existing motor units not only on the level of skeletal muscle where new muscle fibers are generated but also on the level of motoneurons themselves (Gardiner et al. 2006). Exercise-induced nerve terminal branching at the neuromuscular junctions (Deschenes et al. 1993) and enhanced neurotransmission (Dorlochter et al. 1991) modulate the axonal compartment of motoneurons under such conditions. Under pathophysiological circumstances, when axons are lesioned or motoneurons are lost, sprouting can be excessive, which leads to extensive remodeling of motor units. However, such axons with large arborization have been found to be more vulnerable, and clinical conditions such as post-polio syndrome (Birk, 1993) reflect a situation in which such large motor units are destabilized to a stage where motoneurons and axons cannot be maintained. Similar situations have also been observed at individual motor endplates that have been reinnervated, i.e. after nerve lesion (Thompson & Jansen, 1977). The sprouting axons in the adult have been found to sometimes incompletely occupy synaptic sites (Schaefer et al. 2005), and the question arises how long such synapses can be maintained, whether they remain functional for prolonged periods, and what the molecular mechanisms are that determine whether reinnervation leads to a complete restoration of synaptic morphology and function or only to incomplete occupancy of synaptic sites.

These observations indicate that local signaling pathways exist that protect and modulate axonal structure by inducing sprouting of axonal branches and remodeling of individual neuromuscular endplates in the periphery. This article should give a short overview of the signaling pathways initiated by ciliary neurotrophic factor (CNTF) and the role of local protein synthesis, which seems to be altered in spinal muscular atrophy (SMA) and possibly also other forms of motoneuron disease.

Mechanisms used by CNTF signaling for axon maintenance

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms used by CNTF signaling for axon maintenance
  5. Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin
  6. Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics
  7. Local protein synthesis in motor axons: function in axon dynamics and plasticity
  8. Defective mRNA processing and axonal defects in SMA
  9. Concluding remarks
  10. Acknowledgement
  11. References

Survival of motoneurons in cell culture can be supported by a large variety of neurotrophic molecules (Hughes et al. 1993). How these molecules interact and cooperate in promoting developmental axon growth and control axonal maintenance and remodeling of motor units in the adult is not fully understood. Members of the CNTF and leukemia inhibitory factor (LIF) gene family are potent survival factors for motoneurons in vitro (Arakawa et al. 1990) and in vivo (Sendtner et al. 1990). Gene targeting in mice has revealed that CNTF is necessary for postnatal maintenance of motoneurons (Masu et al. 1993), and that about 20% of motoneurons are lost when this neurotrophic factor is deficient. Nevertheless, loss of muscle strength amounts to only 10%, a range that normally would not reach pathological significance. This also explains why a polymorphism in the CNTF gene (Takahashi et al. 1994) that leads to an altered splice acceptor site in exon 2 and abolishes biological activity of the CNTF protein is not associated with neurological disease such as amyotrophic lateral sclerosis (ALS) or SMA. This genetic alteration is quite common and affects about 1–2% of the population worldwide. A possible reason for the lack of significant deficits in bearers of this CNTF polymorphism is the fact that other factors of the same family, such as LIF and cardiotrophin-1 (CT-1), can bind to the same receptor complex on motoneurons that mediates effects on survival and axon maintenance. This receptor complex involves the LIF receptor (LIFR)β and gp130 (Davis et al. 1993). LIF is expressed in Schwann cells (Murphy et al. 1993), and CT-1 is synthesized in skeletal muscle (Oppenheim et al. 2001). By binding to the LIFRß/gp130 complex these factors can compensate for the deficiency of CNTF. This becomes apparent in mice that are double-deficient for Cntf and Lif (Sendtner et al. 1996) and in Cntf/Lif/Ct-1 triple knockout mice (Holtmann et al. 2005), which exhibit more than 30% grip strength loss and reduced voluntary motor activity when they become older than 5 or 6 months.

Null mutations in the CNTF gene lead to an early onset of disease both in patients with sporadic and familial ALS and in the human superoxide dismutase 1 G93A mouse model of familial ALS (Giess et al. 2002). Moreover, in a mouse model of multiple sclerosis, in which experimental autoimmune encephalomyelitis (EAE) is induced by myelin oligodendrocyte glycoprotein peptides, CNTF-deficient mice show higher grades of disease severity that are associated not only with enhanced loss of oligodendrocytes, but also with massive axon loss in spinal cord and other regions of the central nervous system (Linker et al. 2002). Also, axon loss seems to correlate with disease severity in this mouse model of multiple sclerosis. Similar experiments have been performed with Lif knockout mice, and these experiments showed that CNTF and LIF protect axons in EAE via an oligodendrocyte-independent mechanism (Gresle et al. 2012).

Earlier studies have shown that CNTF plays an essential role for long-term maintenance of axons and paranodal networks, the specific structures that are formed by Schwann cells and axons at the nodes of Ranvier (Gatzinsky et al. 2003). When progressive motoneuropathy (pmn) mice, which develop motoneuron disease on the basis of a mutation in the tubulin-specific chaperone E (TBCE) gene (Bommel et al. 2002; Martin et al. 2002), are treated with CNTF (Sendtner et al. 1992), delayed disease onset and improved survival are detected. However, when untreated pmn mutant mice die, loss of motoneuron cell bodies is only in a range of 30–40%, whereas motor axon loss, e.g. in the phrenic nerve, affects more than 60% of the fibers. CNTF treatment has a strong effect on the preservation of axons (Sendtner et al. 1992). Whereas untreated pmn mice exhibit only about 85 axons at 5 weeks after birth in the phrenic nerve, compared with about 300 in healthy controls, CNTF treatment from day 21 to day 35 significantly enhances the number of myelinated nerve fibers to more than 140 after this period (Sendtner et al. 1992). When these mice either are treated with other neurotrophic factors, e.g. glial-derived neurotrophic factor (GDNF) (Sagot et al. 1996) or when bcl-2 is overexpressed in motoneurons (Sagot et al. 1995), improved survival of cell bodies can be observed. However, there is no beneficial effect on the maintenance of neuromuscular endplates and axons, indicating that CNTF and these neurotrophic factors although having similar potency in promoting survival of isolated embryonic motoneurons differ significantly in how they physiologically act on the maintenance and regeneration of motor axons.

Isolated motoneurons from pmn mice do not show any abnormalities in their survival response to various neurotrophic factors such as CNTF, IGF, GDNF or brain-derived neurotrophic factor (BDNF), respectively (Selvaraj et al. 2012). When these motoneurons are cultured in the presence of BDNF they survive normally but exhibit severe alterations in axon growth. Axons are shorter and they contain swellings that include disorganized cytoskeletal elements and organelles, in particular mitochondria (Fig. 1). When pmn mutant motoneurons are cultured with CNTF, significant changes are observed in axons. Total axon length after 7 days is restored to normal levels and the number of axonal swellings is also reduced (Selvaraj et al. 2012). This correlates with normalization of defective axonal transport of mitochondria. The number of stationary mitochondria that are elevated in pmn mutant motoneurons decreases to normal levels and the number of mitochondria that are transported in the anterograde and retrograde direction is also normalized. This effect correlates with the activation of Stat-3 in CNTF-treated pmn mutant motoneurons. When Stat-3 is knocked out in pmn mutant motoneurons, CNTF cannot rescue these axonal defects anymore, despite the fact that survival effects of CNTF in motoneurons are not abolished in Stat-3-deficient embryonic motoneurons (Schweizer et al. 2002). This indicates that the signaling pathways for survival and axon growth and maintenance are distinct, and that CNTF uses Stat-3 for the latter effect on axons.

image

Figure 1. Ultrastructural morphology of pmn mutant motor axons. Motoneurons were isolated from E13.5 mouse embryos and cultured for 7 days on laminin-1 (111) with BDNF. Under these circumstances, axons are shorter and exhibit many axonal swellings (Bommel et al. 2002; Selvaraj et al. 2012). Ultrastructural analysis of these swellings is shown in examples of three motor axons (A–C), exhibiting the morphology of mitochondria that accumulate in these axonal segments. Scale bars: 500 nm.

Download figure to PowerPoint

Stat-3 is known as a classical signaling mediator that is activated at the cell surface receptor and then translocates to the nucleus to activate transcription of specific genes (Taga, 1996). In axons of isolated motoneurons most of the activated Stat-3 does not translocate to the nucleus but instead acts locally. Interestingly, when the transactivation domains of Stat-3 that mediate these transcriptional effects are depleted, Stat-3 is still capable of rescuing axons. The axonal effects are only abolished when tyrosine 705, which is phosphorylated after receptor activation, is mutated to phenylalanine. Taken together, this indicates that the axonal effects by which CNTF acts on axon growth and on restoration of axonal transport and axonal swellings is dependent on Stat-3, but not on the transcriptional effects of Stat-3.

Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms used by CNTF signaling for axon maintenance
  5. Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin
  6. Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics
  7. Local protein synthesis in motor axons: function in axon dynamics and plasticity
  8. Defective mRNA processing and axonal defects in SMA
  9. Concluding remarks
  10. Acknowledgement
  11. References

The finding that axonal transport is severely disturbed in pmn motoneurons points to defects in microtubules in this model of motoneuron disease. This is plausible, given that the underlying gene defect in the TBCE gene (Bommel et al. 2002; Martin et al. 2002) also emphasizes defects in microtubule assembly in this mouse model (Schaefer et al. 2007).

Axons normally contain stable, long-living microtubules that are acetylated (Witte et al. 2008). These acetylated microtubules are usually found in all parts of the axons, but are relatively excluded from axonal growth cones and dendrites. Highly dynamic microtubules are tyrosinated (Witte et al. 2008) and found in all regions of motoneurons, including axonal growth cones and dendrites. The levels of tyrosinated tubulin are increased in isolated axons of pmn mutant motoneurons when cultured with BDNF. Interestingly, upon CNTF treatment the levels of tyrosinated tubulin in axons are reduced, indicating that downstream signaling from CNTF receptor complexes influences the dynamics of microtubules by stabilizing them.

Phosphorylated Stat-3 can interact with stathmin (Ng et al. 2006), a microtubule destabilizing protein that binds to α/β tubulin heterodimers and thereby reduces the pool of available tubulin subunits for microtubule elongation (Amayed et al. 2002). Stathmin, which is also named Op18 (oncoprotein-18), is a member of a family of microtubule-interacting proteins that also includes SCG10 (superior cervical ganglion protein-10), SCLIP (SCG10-like protein) and RB3 (rice XA21-binding protein 3) (Charbaut et al. 2001). All of these proteins carry a conserved tubulin-binding domain in their C-terminus, and they directly induce catastrophe-promoting microtubule depolymerization by sequestering tubulin heterodimers. Activated Stat-3 can interact with the C-terminus of stathmin, the same region that also interacts with tubulin heterodimers (Ng et al. 2006). When cultured motoneurons overexpress green fluorescent protein (GFP)-coupled Stat-3, allowing analysis of movement of activated Stat-3 after CNTF addition, most of the activated Stat-3 stays locally and does not move towards the nucleus. Biochemical pull-down experiments from cytoplasmic extracts of cultured motoneurons show increased Stat-3/stathmin interaction and reduced interaction of stathmin with tubulin (Selvaraj et al. 2012). In order to prove that this interaction is responsible for the observed effect of CNTF signaling on microtubule stability and axon elongation, stathmin was depleted by lentiviral shRNA suppression. Under these circumstances, axon length normalized and the relative levels of tyrosinated microtubules decreased, indicating that indeed the local interaction of Stat-3 with stathmin in axons is responsible for the effect of this neurotrophic factor on microtubule dynamics in axons.

Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms used by CNTF signaling for axon maintenance
  5. Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin
  6. Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics
  7. Local protein synthesis in motor axons: function in axon dynamics and plasticity
  8. Defective mRNA processing and axonal defects in SMA
  9. Concluding remarks
  10. Acknowledgement
  11. References

The mRNA for Stat-3 is highly expressed in motoneurons, and it exists in several isoforms. A Stat-3 mRNA with long 3′UTR (untranslated region) has been shown to specifically translocate to axons and to be locally translated in distal axonal components (Ben-Yaakov et al. 2012). Thus, Stat-3 levels can be modulated locally in axons and axon branches, and differences in Stat-3 levels could regulate the strength of CNTF signaling and downstream effects on microtubule stability in individual axon segments. Among various forms of neurodegenerative and inflammatory diseases, for example EAE, a model of multiple sclerosis, axon damage occurs and appears responsible for permanent neurological deficits. Under conditions of inflammatory axon damage axons are locally destabilized (Nikic et al. 2011). They show focal swellings that appear reversible but also can proceed to axonal fragmentation. This raises the question whether disturbed local Stat-3–stathmin signaling is responsible for focal destabilization of axons under the specific pathological conditions in multiple sclerosis. Stat-3 becomes highly phosphorylated within a short time after axon injury (Lee et al. 2004). When Stat-3 is depleted by conditional gene ablation in spinal motoneurons and dorsal root sensory neurons, developmental cell death of these neural populations is not affected (Schweizer et al. 2002). However, after axotomy, cell death is enhanced and regeneration is impaired. Recently, it has been shown that overexpression of Stat-3 by viral gene transfer increases both axonal growth and collateral sprouting of sensory neurons, that project to the spinal cord, at least fourfold (Bareyre et al. 2011). Even more interesting, these in vivo imaging studies revealed that the initiation of axon growth is highly affected, which is compatible with our data that Stat-3–stathmin interaction promotes axon growth by increasing the pool of available α/β-tubulin heterodimers for microtubule elongation. When microtubules are destabilized in cultured motoneurons by nocodazole treatment (Selvaraj et al. 2012), re-growth after nocodazole washout is greatly enhanced by CNTF addition, which correlates with these in vivo findings. Similar observations have also been made in EAE mouse models of multiple sclerosis. CNTF depletion massively enhances axonal degeneration (Linker et al. 2002), and LIF, which also uses LIFRβ and gp130 for its signal transduction in responsive cells, protects axons under the conditions of EAE (Gresle et al. 2012). This suggests that Stat-3–stathmin signaling is not only relevant for the specific conditions of motoneuron disease caused by mutations in the TBCE gene, but for a much broader spectrum of pathophysiological conditions, in particular neuroinflammation.

Local protein synthesis in motor axons: function in axon dynamics and plasticity

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms used by CNTF signaling for axon maintenance
  5. Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin
  6. Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics
  7. Local protein synthesis in motor axons: function in axon dynamics and plasticity
  8. Defective mRNA processing and axonal defects in SMA
  9. Concluding remarks
  10. Acknowledgement
  11. References

So far, the role of local protein synthesis and translational regulation in synaptic plasticity has mainly been assessed in dendrites (Sossin & Lacaille, 2010). Local protein translation is important for several aspects of synaptic plasticity. Inhibition of protein synthesis by injection of cycloheximide inhibits the potentiation of synaptic strength that is induced by repeated neural activity at individual neuromuscular synapses from Xenopus nerve–muscle cocultures (Zhang & Poo, 2002). This effect occurs within a very short time. Therefore, it is unlikely that retrograde signaling to the nucleus and subsequent movement back of the response signal from the cell body is involved which then executes the effect on excitability and neural transmitter release at axon terminals. The PI3-K/Akt-mTOR pathway is a well-characterized mediator of local signaling on protein translation at individual synapses (Tsokas et al. 2007). This raises the question whether mRNAs both for proteins that mediate plasticity in axonal subcompartments and for mediators of extracellular signals are locally translated and thus influence axon growth, maintenance and plasticity. Indeed, the Stat-3 mRNA isoform with long 3′UTR is translocated into axons and synthesized in a Ca2+-dependent manner under conditions when axons are lesioned (Ben-Yaakov et al. 2012). This points to neural activity-related mechanisms that control the local synthesis of this transcription factor in axons. The increased local availability of Stat-3 could in turn enhance the capacity both for local actions of Stat-3 and for retrograde actions that lead to transcriptional responses.

Local protein synthesis in axons and axon terminals depends on the presence of ribosomes and other components of the mRNA translational machinery. There is still an ongoing debate whether ribosomes are present in axons of neurons and whether local translation of mRNAs can occur in this subcellular compartment (reviewed in Twiss & Fainzilber, 2009). The axonal localization of mRNAs could reflect incomplete axonal-dendritic polarity and/or diffusion of highly expressed mRNAs into early axons, in particular in cell culture. We have studied the presence of polyribosomes in isolated motoneurons (Fig. 2). Such polyribosome-like structures are present in axon terminals within regions close to the active zones that are forming at axon tips when these cells are cultured for more than 5 days. At that stage, axons have reached a length of more than 400 μm and can be clearly distinguished from dendrites, arguing against the possibility that the presence of polyribosomes reflects diffusion from the cell body or incomplete axonal-dendritic polarity.

image

Figure 2. Electron micrographic overview of an axonal growth cone from an embryonic motoneuron grown for 7 days in culture. Inset 1 shows higher magnifications of the boxed area (1), and inset 2 and 3 the same area in a serial section of another growth cone at medium (2) and high magnification (3). Arrowheads point to polyribosome-like structures with a size of ~25 nm. Scale bars: 1 μm in the overview and 200 nm in the insets.

Download figure to PowerPoint

As a second requirement, efficient mechanisms for sorting and axonal transport of specific mRNAs are necessary. This includes specific proteins that guide the traffic of these mRNAs into axons (Kiebler & Bassell, 2006). The formation of these RNA transport granules depends on sequence structures that are normally found in the UTRs of these mRNAs. These cis-elements are recognized by RNA-binding proteins (RBPs) or other trans-acting factors that guide their transport to specific cytoplasmic regions. This mechanism has been studied in detail for the 3′UTR of β-actin mRNA. The 3′UTR region of the β-actin mRNA contains elements that are recognized by RBPs, such as ZIP-code-binding protein-1 (ZBP-1) (Bassell et al. 1998), also known as the insulin-like growth factor II mRNA-binding protein-1 (IMP-1), and the hnRNP R protein, respectively (Rossoll et al. 2003). ZBP-1 and hnRNP R both interact with the β-actin mRNA, but within distinct regions in its 3′UTR of. They form a complex that stabilizes and protects the transcript from premature translation. This complex is translocated into axons, and local signaling through Src family kinases can then phosphorylate ZBP-1 (Huttelmaier et al. 2005), which subsequently leads to the release of β-actin mRNA and initiation of translation. It is currently unknown whether Stat-3 mRNA also binds to ZBP-1/IMP-1 or hnRNP R, and how many distinct RBPs exist that guide mRNAs into the axonal compartment.

Another mRNA that is locally translated in axons encodes for the low molecular weight chain of neurofilament (NFL) (Strong et al. 2007). This specific mRNA is interesting for several aspects. First, alterations in axonal NFL content and composition have been observed in several forms of motoneuron disease (Lin & Schlaepfer, 2006). Increased axonal intermediate filament content has also been detected in axons of motoneurons from pmn mutant mice that primarily suffer from a defect in microtubule assembly and stability (Selvaraj et al. 2012). Second, the axonal translocation of the NFL-encoding mRNA involves the TDP43 protein that is found in aggregates and axonal swellings being characteristic for many forms of motoneuron disease and other neurodegenerative disorders. Third, massive accumulation of phosphorylated NFL in terminal axons and remaining neuromuscular junctions appears as a pathological hallmark also in SMA (Cifuentes-Diaz et al. 2002). Similarly, alterations in microtubule and actin organization are found in cultured Smn-deficient motoneurons (Fig. 3 A–B, E–F), which corresponds to reduced β-actin, in particular F-actin, levels in axon terminals (Fig. 3 C–D, G–I, K–M, O–P). These alterations indicate that disorganization of the cytoskeleton also plays a major role in SMA, and that loss of motoneuron cell bodies in this form of motoneuron disease is likewise a consequence of an axonal ‘dying-back’ process (Cifuentes-Diaz et al. 2002).

image

Figure 3. Distribution of cytoskeletal proteins in growth cones from isolated motoneurons. (A–H) Tyrosinated and glutamylated tubulin and F-actin were detected in axonal growth cones of isolated motoneurons from E13.5 control (A–D) and Smn-deficient mouse embryos (E–H) grown for 5 days in culture on laminin-1 (111). (I–P) I–L shows the arrangement of β-actin protein, tau and F-actin in growth cones of DIV5 control motoneurons and M–P in Smn-deficient embryos. Immunocytochemistry for (A–H) was performed with monoclonal antibodies against glutamylated α tubulin and tyrosinated α tubulin (against the glutamylated/tyrosinated motif at amino acids 445–457 of α tubulin, Synaptic Systems 1 : 1000, Abcam 1 : 2000, respectively), and the F-actin dye Phalloidin (Invitrogen, 1 : 30). (I–P) Immunocytochemistry for β-actin with a monoclonal antibody (GeneTex, 1 : 500) which also detects unpolymerized actin. A polyclonal antibody against tau (Sigma-Aldrich, 1:500) was used as an axonal marker, and phalloidin (Invitrogen, 1:30) as a marker of F-actin.

Download figure to PowerPoint

Defective mRNA processing and axonal defects in SMA

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms used by CNTF signaling for axon maintenance
  5. Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin
  6. Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics
  7. Local protein synthesis in motor axons: function in axon dynamics and plasticity
  8. Defective mRNA processing and axonal defects in SMA
  9. Concluding remarks
  10. Acknowledgement
  11. References

Defects in axonal mRNA translocation appear as a central theme in motoneuron disease, in particular in SMA. SMA is caused by deficiency of the SMN protein due to homozygous deletion or mutation of the SMN1 gene on human chromosome 5 (Burglen et al. 1995; Lefebvre et al. 1995). On human chromosome 5, a second copy of the SMN gene termed SMN2 is found which is usually preserved in SMA patients. However, based on a single nucleotide exchange in the domain coding for exon 7, the majority of transcripts from this second gene copy lacks the corresponding exon 7-encoded domains, with only a minority of transcripts giving rise to a full-length and functional protein. In contrast to humans, the mouse genome contains one single allele for Smn, and the Smn gene knockout is lethal at early developmental stages in this species (Schrank et al. 1997). Only when two copies of human SMN2 are introduced into the mouse genome, these null mice develop to term showing a severe motoneuron disease phenotype that leads to death within a few days after birth (Monani et al. 2000).

The Smn protein forms the core of a complex for the assembly of small spliceosomal U snRNP complexes (Fischer et al. 1997). However, the protein is also found in axons and axon terminals, indicating that it might serve additional functions in this cellular region (Jablonka & Sendtner, 2003). In the cytosol of motoneurons, the Smn protein has been found to interact with hnRNP R (Rossoll et al. 2002). Overexpression of hnRNP R in PC 12 cells enhances NGF-induced neurite growth (Rossoll et al. 2003). When a hnRNP R isoform is overexpressed lacking amino acids 522–556, which have been shown to mediate interaction with Smn, this mutated hnRNP R protein does not enhance neurite growth; neither is this achieved by a hnRNP R protein lacking the RNA recognition motifs 1 and 2. This effect correlates with the binding capacity of hnRNP R with the 3′UTR of β-actin mRNA (Rossoll et al. 2003). The binding activity for this mRNA is enhanced when hnRNP R and Smn are co-expressed. This correlates with the observation that the β-actin mRNA translocation into axons is severely impaired both in motoneurons derived from a mouse model of type 1 SMA and in motoneurons in which hnRNP R has been depleted. Similar observations have also been made in vivo. When Smn (McWhorter et al. 2003) or hnRNP R (Glinka et al. 2010) is depleted from zebrafish embryos, axon growth is severely disturbed in a virtually indistinguishable manner.

The lack of β-actin mRNA in axons of Smn-deficient motoneurons (Rossoll et al. 2003) indicates that defects in axonal mRNA translocation could contribute to the disease pathomechanism in SMA and possibly also other forms of motoneuron disease. However, it is neither clear whether and how much β-actin mRNA is translated in axons nor how much this translation contributes to axonal function and whether lack of axonal translocation and local translation of the β-actin mRNA can explain the specific disease phenotype in SMA. In sensory neurons, ß-actin mRNA and GAP-43 mRNA compete for transport into neurites. Depleting ß-actin mRNA in neurites by increased expression of GAP-43 mRNA levels results in extended neurite growth, whereas increased ß-actin production in neurites enhances branching (Donnelly et al. 2013). Thus, local synthesis of ß-actin protein in neurites of sensory neurons does not only regulate neurite extension but also branching of neuronal processes.

In order to investigate local translation of β-actin mRNA in motoneurons and how this is altered when Smn is lacking, a lentiviral reporter construct was studied in which myristoylated eGFP was axonally translocated via the 3′UTR of the β-actin mRNA (Rathod et al. 2012). This reporter was strongly activated in axons of cultured motoneurons and, after photobleaching, the fluorescent signal returned within minutes in axon terminals, long before it was detected in proximal axons (Fig. 4). This fluorescence recovery could be blocked by protein synthesis inhibitors and by rapamycin (Rathod et al. 2012). Interestingly, fluorescence recovery was not so much different between control and Smn-deficient motoneurons cultured under standard conditions on laminin-1 (111) (Aumailley et al. 2005) (Fig. 4). When motoneurons were cultured on synapse-specific laminin-4/merosin (laminin-2/4, laminin-211/221), recovery was greatly enhanced in Smn-deficient motoneurons, but lower in control motoneurons, indicating that laminin-2/4 (211/221) influences the protein synthesis rate of β-actin in axon terminals of motoneurons, and that this specific type of regulation of β-actin protein synthesis by laminin-2/4 (211/221) is altered in Smn-deficient motoneurons. The mechanisms underlying this defect are unclear. However, these data also suggest that reduced availability of β-actin mRNA in axon growth cones could, under circumstances, be compensated, at least in part, by enhanced protein translation. Experiments with migrating chick fibroblasts have shown that only about 7% of the β-actin protein necessary for cell migration is locally translated (Condeelis & Singer, 2005). This finding suggests that local translation does not fully provide the actin monomers that are necessary for building the actin cytoskeleton, but that local synthesis could increase the local concentration to a level that forms nuclei for actin polymers. This could then alter the cytoskeleton in axonal branch points and axon terminals, thus providing a major tool for plasticity under physiological and pathophysiological conditions.

image

Figure 4. Laminin-1 (111)- and laminin-2/4 (211/221)-mediated effects on local β-actin production in axon terminals of control and Smn-deficient motoneurons. (A) Fluorescence recovery in the axonal growth cone of a bleached Myr-eGFP 3′UTR (β-actin) transfected Smn+/+;SMN2 motoneuron cultured on laminin-1 (111) (left) and laminin-2/4 (211/221) (right) (scale bar: 10 μm). (B) Quantification of FRAP during 60 min in growth cones of Smn+/+;SMN2 motoneurons cultured on laminin-1 (111) (black, n = 8) or laminin-2/4 (211/221) (green = 8). (C) 1/slope for FRAP in Smn+/+;SMN2 MN cultured on laminin-1 (111) vs. laminin-2/4 (211/221), ***P = 0.0003, tested by Student′s t-test with Mann–Whitney test, respectively. (D) Fluorescence recovery in axonal growth cones of bleached β-actin reporter transfected Smn−/−;SMN2 motoneurons cultured on laminin-1 (111) (left) and laminin-2/4 (211/221) (middle) showing increased recovery and, on laminin-2/4 (211/221) (right), showing rapid extension of axon length (scale bar: 10 μm). (E) Quantification of FRAP during 60 min in growth cones of Smn−/−;SMN2 motoneurons cultured on laminin-1 (111) (black, n = 12) and laminin-2/4 (211/221) (green, n = 4). (B, E) Mean intensity ± SEM are shown. (F) 1/slope for FRAP in Smn−/−;SMN2 MN cultured on laminin-1 (111) vs. laminin-2/4 (211/221), **P = 0.0081, tested by Student′s t-test with Mann–Whitney test, respectively. Reproduced in modified form with permission from Rathod et al. (2012).

Download figure to PowerPoint

When β-actin mRNA is depleted by conditional gene ablation in postnatal mice, no major effect on motor performance occurs and no disease resembling SMA develops (Cheever et al. 2011). This is not unexpected, given that actin monomers can also be produced from the α- and γ-actin genes and mRNAs. Indeed, when β-actin is depleted in the nervous system, increased levels of α- and γ-actin can be observed (Cheever et al. 2012). On the basis of the conclusions drawn from the study by Condeelis & Singer (2005), less than 10% of actin filaments are generated from actin monomers produced by local translation in the distal region of the cell. It would be interesting to investigate in this mouse model whether plasticity of motor units is disturbed under physiological conditions such as enhanced exercise or pathological conditions, for example after nerve lesion or when these mice are crossbred with other mouse models of motoneuron disease. Additionally, it needs to be investigated whether α- and γ-actin mRNAs are also translocated to axons via mechanisms that possibly do not involve ZBP-1/IMP-1, and whether actin monomers could be synthesized from these distinct actin isoforms. A recent report has shown that proteins such as Arp2, WAVE1 and cortactin that modulate the axonal cytoskeleton are also locally synthesized in axons (Spillane et al. 2012), and it is not known how the synthesis of these proteins is altered in models of SMA. Compartmentalized cultures, in which the axonal mRNA content of motoneurons lacking Smn, or of motoneurons from mouse models from other forms of motoneuron disease are compared, could help to identify groups of mRNAs whose transport is altered under conditions when Smn is deficient or the genes for TDP43 or FUS are mutated.

Concluding remarks

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms used by CNTF signaling for axon maintenance
  5. Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin
  6. Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics
  7. Local protein synthesis in motor axons: function in axon dynamics and plasticity
  8. Defective mRNA processing and axonal defects in SMA
  9. Concluding remarks
  10. Acknowledgement
  11. References

Mutations in genes for TDP-43 (Sreedharan et al. 2008) or FUS (Vance et al. 2009) in patients with familial forms of ALS and the identification of these proteins in neuronal aggregates that are characteristic not only for motoneuron disease but also other neurodegenerative disorders, has focused attention to altered mRNA metabolism in neurodegeneration. Similarly, the role of Smn as a central player in the assembly of U snRNP complexes, and its role in pre-mRNA processing, underlines this aspect. However, it is still unclear why motoneurons are more vulnerable than other types of cells under conditions when these proteins are missing, mutated or structurally altered. Recent reports have suggested that other types of neurons, for example sensory neurons, also contribute to the pathophysiology of motoneuron disease, namely on the basis of altered pre-mRNA splicing of mRNAs that are important for sensory neural function (Imlach et al. 2012; Lotti et al. 2012). Neurite growth in sensory neurons from Smn−/−;SMN2 motoneurons is also defective (Jablonka et al. 2006), but sensory deficiencies remain relatively mild in patients with motoneuron disease (Anagnostou et al. 2005), at least with respect to the somatosensory system. It is conceivable that alterations of synaptic input could alter motoneuron function and plasticity on the axonal site. However, the underlying mechanisms are still not fully understood. Another aspect that also needs to be addressed is the role of compensatory mechanisms, how effective they are and whether they can explain why many types of neurodegenerative disorders, in particular ALS, only become apparent at higher age and not at birth, given that the underlying gene defects are already prevailing at that stage. Such compensatory mechanisms are massive in a mouse model of SMA – i.e. Smn+/− mice (Simon et al. 2010). Despite the loss of more than 50% of spinal motoneurons at the age of 1 year, muscle strength is still preserved on the basis of a compensatory enhancement of functional motor units that can exceed a factor of 3. This compensatory sprouting depends on CNTF, which is produced in high amounts from myelinating Schwann cells. When Smn+/− mice are crossed with Cntf knockouts, no enhancement of motor units and no compensatory sprouting is observed. Thus, the potential effect of local Stat-3 activation by CNTF could play a major role in promoting sprouting and regenerative responses in motoneuron disease. Whether this effect is only mediated via local alterations of the axonal cytoskeleton or whether additional targets, for example axonal mitochondria, contribute to this effect needs to be addressed. Indeed, mitochondrial function can be influenced by Stat-3 (Reich, 2009; Wegrzyn et al. 2009), and alterations in local mitochondria (Zhou & Too, 2011; Akopian et al. 2012; Plucinska et al. 2012) could play a role both in axonal degeneration, but also positively in regenerative responses that mediate plasticity. Thus, local Stat-3 signaling, alterations in axonal mRNA metabolism and mitochondrial function appear to be connected, and it will be interesting to know more about the molecular pathways by which these mechanisms collaborate in axon maintenance and regenerative responses.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Mechanisms used by CNTF signaling for axon maintenance
  5. Local Stat-3 signaling in axons stabilizes microtubules through interaction with stathmin
  6. Local translation of Stat-3 in axons: effects on axon regeneration and potential local effects for axon dynamics
  7. Local protein synthesis in motor axons: function in axon dynamics and plasticity
  8. Defective mRNA processing and axonal defects in SMA
  9. Concluding remarks
  10. Acknowledgement
  11. References