|Home | About | Journals | Submit | Contact Us | Français|
Neuregulin-1 (NRG1) plays a crucial role in axoglial signaling during the development of the peripheral nervous system, however its importance in adulthood following peripheral nerve injury remains unclear. We utilised Single-neuron Labelling with Inducible Cre-mediated Knockout (SLICK) animals, which enabled visualisation of a subset of adult myelinated sensory and motoneurons neurons in which Nrg1 was inducibly mutated by tamoxifen treatment. In uninjured mice, NRG1 deficient axons and the associated myelin sheath were normal and the neuromuscular junction demonstrated normal apposition of pre- and postsynaptic components. Following sciatic nerve crush, NRG1 ablation resulted in severe defects in remyelination: axons were either hypomyelinated or had no myelin sheath. NRG1 deficient axons were also found to regenerate at a slower rate. Following nerve injury the neuromuscular junction was reinnervated, however excess terminal sprouting was observed. Juxtacrine Neuregulin-1 signaling is therefore dispensable for maintenance of the myelin sheath in adult animals but has a key role in reparative processes following nerve injury.
Neuregulin-1 (NRG1) has a key role in the development of the peripheral nervous system. Alternative splicing and differential promoter usage gives rise to at least 15 different NRG1 isoforms which demonstrate different spatiotemporal patterns of expression (Meyer et al., 1997;Falls, 2003;Mei and Xiong, 2008). All isoforms possess an epidermal growth factor (EGF)-like signaling domain, which is essential for mediating biological activity, and can be classified into subgroups according to the structure of their amino termini. Isoforms containing Ig-like domain (types I, II and IV) can either be directly secreted or can be released as soluble proteins from the cell surface following proteolytic cleavage. Isoforms possessing a cysteine-rich domain (CRD, type III isoforms) have two transmembrane domains, require proteolytic cleavage by BACE1 for full activity and signal in a juxtacrine fashion (Willem et al., 2006;Hu et al., 2006).
During development, axonally-derived NRG1 signals via ErbB2 and 3 receptor heteromers expressed on Schwann cells and regulates multiple aspects of Schwann cell differentiation including the survival of Schwann cell precursors, Schwann cell proliferation, motility, axon ensheathment and myelination (Garratt et al., 2000;Taveggia et al., 2005;Michailov et al., 2004;Chen et al., 2003;Fricker et al., 2009;Riethmacher et al., 1997;Woldeyesus et al., 1999;Meyer and Birchmeier, 1995;Morris et al., 1999). Early in vitro studies suggested that NRG1 may also have a role in acetylcholine receptor clustering during development of the neuromuscular junction (NMJ) (Falls et al., 1993). More recent genetic approaches have however indicated that NRG1-ErbB signaling is dispensable for post-synaptic differentiation at the NMJ (Escher et al., 2005;Jaworski and Burden, 2006). NRG1 has also been shown to modulate the behaviour of terminal Schwann cells at the NMJ as it increases survival of these cells following postnatal denervation (Trachtenberg and Thompson, 1996) and promotes process extension (Hayworth et al., 2006).
Because of the key developmental role of NRG1 in Schwann cell development, it has been challenging to study the function of this molecule in adulthood both in relation to maintenance of myelin and also in the response to nerve injury. The peripheral nervous system shows significant ability for repair following either traumatic or immune mediated injury (Chen et al., 2007). The co-ordinated signaling between regenerating axons and Schwann cells is likely to be critical for both axon re-myelination and regeneration; given its developmental functions, NRG1 is an attractive target for fulfilling such a role. In order to investigate the role of axon derived NRG1 signaling we have used ‘SLICK’ mice (Young et al., 2008) in which the enzyme Cre recombinase is inducibly activated, resulting in NRG1 ablation in a subset of myelinated motor and sensory axons which also express Yellow Fluorescent Protein (YFP). We demonstrate that juxtacrine NRG1 signaling is not required for maintenance of the myelin sheath and NMJ, nor for axon integrity, but is essential for multiple aspects of the reparative response following nerve injury including re-myelination, axon regeneration and normal re-innervation of the NMJ.
All work carried out conformed to UK Home Office legislation (Scientific Procedures Act 1986). SLICK-A Cre; NRG-1fl/fl were bred by crossing SLICK-A Cre mice (JAX® Mice 007606) with NRG-1fl/fl mice; both colonies were on a C57BL/6 background. The generation and genotyping of mutant mice with floxed alleles of Nrg1 (Nrg1f/f) and SLICKA-Cre mice has previously been described (Fricker et al., 2009;Young et al., 2008;Yang et al., 2001;Brinkmann et al., 2008). Genotyping of SLICK-A Cre; NRG-1fl/fl mice was carried out by using PCR of genomic DNA to detect the presence of the loxP sites. The primers used were: 5′-tttggtggactgggtttctc-3′ and 5′-CTGACTGGCCTTTCTTCCAG-3′, (carried out as follows; heating at 94°C for 2 min, 34 cycles of 94°C for 30 sec, 60°C for 30 sec and 72°C for 50 sec, followed by a final extension at 72°C for 8 min). To detect the SLICKA-Cre construct, ear punches were examined for YFP fluorescence under a fluorescent microscope.
In SLICKA-Cre mice a tamoxifen inducible form of Cre recombinase is simultaneously expressed with yellow fluorescent protein (YFP) in the same cells. Expression is driven by two copies of the Thy1 promoter and is seen in a small population of motor axons and large DRG cells. YFP expression is constitutive whereas Cre activity is drug inducible. 10 week old SLICK-A Cre; NRG-1fl/fl mice were dosed by oral gavage with 0.25 mg per g of body weight with tamoxifen (Sigma T5648) in corn oil, prepared as previously described, (Young et al., 2008) or with corn oil alone vehicle for five consecutive days. After dosing, animals were left for 4 weeks before surgery. In each experiment SLICK-A Cre; NRG-1fl/fl tamoxifen treated animals were compared with SLICK-A Cre; NRG-1fl/fl vehicle treated animals. As a further control, comparison was also made with SLICK-A Cre; NRG-1+/+ tamoxifen treated controls. We included equal numbers of animals of each gender in experimental groups wherever possible.
One month post tamoxifen/vehicle treatment, the right sciatic nerve was exposed and crushed two times, in two different directions 30 s each time, with fine forceps. The lesion site was kept a constant 43 mm from the tip of the third digit, by laying a measured thread over the anatomical trajectory of the sciatic nerve. The crush site was labelled with lamp black. The wound was closed with 5-0 sutures and disinfected. A schematic diagram of the sciatic nerve crush model is shown in Figure 1. The sciatic and tibial nerves and the gastrocnemius muscle were harvested at 10 days, 14 days and 2 months post crush.
Ablation of NRG1 at the mRNA level was demonstrated using in situ hybridization (ISH) as previously described (Fricker et al., 2009). Following pre-hybridization treatments (acetylation, delipidation, dehydration), sections were incubated overnight at 37°C with radioactively (35S-dATP, Perkin-Elmer Life Sciences, Boston, USA) end-labelled (terminal deoxynucleotidyl transferase, Promega, UK) probe in hybridization buffer. The probe sequence was CTGGTGATCGTTGCCAAAACTACGTAATGGCCAGC (Sigma-Aldrich, UK), directed against the βEGF domain of mouse NRG1 mRNA (NM_178591) and its specificity has been previously shown Solutions and materials used during tissue preparation and prehybridization steps were RNAse-free or DEPC-treated. The next day, slides were washed in standard saline citrate (SSC) solutions with increasing stringencies (final wash, 0.2× SSC at 50°C), dehydrated, air-dried, dipped in LM1 autoradiographic emulsion (GE Healthcare, UK) and exposed for 3 weeks before development. Slides were mounted with Vectashield (Vector Laboratories, UK) and signal was visualised on a Zeiss microscope fitted with an epi-fluorescence polarised light block. For quantification of ISH signal in motoneurons, YFP positive cells exhibiting grain density twice background levels were counted as exhibiting a positive hybridization signal. All analyses were performed with the operator blinded to group.
Animals were deeply anesthetized with pentobarbitone and transcardially perfused with 5 ml saline followed by 25 ml paraformaldehyde (4% in 0.1 M phosphate buffer). L4 DRGs and gastrocnemius muscle were post fixed in paraformaldehyde (4% in 0.1 M phosphate buffer) for 2hrs and then transferred to 20% sucrose overnight (4°C). Tissue was then mounted in OCT embedding compound on dry ice and stored at −80°C. Transverse sections of the DRGs were cut at 10 μm on a cryostat onto SuperFrost Ultra Plus® slides. Longitudinal sections of gastrocnemius were cut on a freezing microtome at 100 μm into a 24 well plate and stored in PBS containing 0.1% sodium azide. Whole tibial nerves were mounted on SuperFrost Ultra Plus® slides in Vectashield® mounting medium (Vector Labs).
DRG cell profiles were visualised by staining for either anti-β-Galactosidase (5 Prime→3 Prime), IB4 Isolectin B4 biotin conjugate (BSI-B4, 10 μg/ml Sigma-Aldrich L2140), polyclonal sheep calcitonin gene related peptide (CGRP, Biomol CA1137, 1:800), or monoclonal mouse anti-neurofilament 200 clone N52 (NF200, 1:500 Sigma N0142). Secondary antibodies used were: ExtrAvidin®-FITC (Sigma E 2761, 1:500), anti–sheep Cy3 (Stratech 713-166-147, 1:400), and donkey anti-mouse FITC (Stratech 715-095-150, 1:200). For analysis, YFP signal was enhanced using either rabbit anti-GFP (Invitrogen A11122, 1:1000) or chicken anti-GFP (Abcam 13970-100, 1:1000) and the secondary antibodies Alexafluor 488 goat anti-rabbit (Invitrogen A11070, 1:1000) and Alexafluor 488 goat anti-chicken (Invitrogen A11039, 1:1000). Neuromuscular junctions were visualised using α-bungarotoxin Alexafluor 647 conjugate (Invitrogen B35450, 1:1000) or α-bungarotoxin tetramethylrhodamine conjugate (Invitrogen T1175, 1:1000). Terminal Schwann cells were visualised with S100 staining using rabbit S100 polyclonal antibody (Dako Z0311, 1:400) and secondary Alexafluor 546 goat anti-rabbit (Invitrogen A11035, 1:1000). Axonal myelin was visualised using rat mAb to myelin basic protein (Abcam 7349-2, 1:200) and secondary Alexafluor 546 goat anti-rat (Invitrogen A11051, 1:1000). Incubation of primary antibodies was overnight and was preceded by incubation with normal donkey serum (Chemicon S30, 1:10, 30mins). All reagents were diluted in PBS containing 0.2% Triton X-100 and 0.1% sodium azide.
Immunofluorescence was visualized under a Zeiss Imager.Z1 microscope or a confocal Zeiss LSM 700 laser scanning microscope. Photographs were taken using the Axio Cam and AxioVision LE Rel. 4.2 or the LSM Image Browser software for image analysis. Three intact representative muscle sections were selected from the central third of the gastrocnemius muscle and analyzed. Counts were taken of total and YFP-axon innervated neuromuscular junctions live on a Zeiss Imager.Z1 microscope and the experimenter was blinded to the animal group. Terminal neuromuscular junction sprouts were analyzed from confocal photomicrographs, 50 images were taken from one section per animal and the sprouts were measured using Image J software (NIH). The operator was also blinded to each group for this analysis.
A pre-embedding immune-electron microscopy method was applied. Animals were deeply anesthetized with pentobarbitone and transcardially perfused with 5 ml saline followed by 25 ml 4% paraformaldehyde, in 0.1M phosphate buffer. The sciatic nerve was dissected and pinned straight on a gelatine coated dish. Nerves were postfixed in 4% paraformaldehyde, 2% glutaraldehyde, 0.1 M phosphate buffer 4°C overnight before being embedded in 10% gelatine (Sigma 48724); once the gelatine had set, blocks containing each nerve were cut and stored in 4% paraformaldehyde at 4°C overnight, after which the paraformaldehyde was replaced with PBS/0.1% azide. Transverse sections of sciatic nerves were cut using a vibratome (Leica VT 1000 S) at 100 μm. Sections were stored in PBS/0.1% azide at 4°C.
Sections from 2 mm distal to the crush site were taken for immuno-EM. For YFP staining, gelatine sections of sciatic nerve were washed in PBS/0.1% bovine serum albumin (BSA Sigma A9576), incubated for 30 mins with 10 mg/ml NaBH4, then washed again in PBS/0.1% BSA before incubation for 20 mins with 0.3% H2O2 in PBS. Sections were then blocked with normal donkey serum (Chemicon S30, 1:10, 30 mins) before incubating overnight with rabbit anti-Green fluorescent protein (GFP) in PBS (Invitrogen A11122, 1:1000). Sections were washed and then incubated with anti-rabbit-biotin in PBS (1:400, Vector Laboratories BA1000, 3hrs), sections were washed before incubation for 30 mins with Vectastain ABC kit (Vector Laboratories Ltd. PK-6100). Sections were washed and DAB staining was performed according to manufacturer’s instructions with a development time of 10 mins (DAB Peroxidase Substrate Kit Vecto Laboratories Ltd. SK-4100). Sections were stored in 0.1 M phosphate buffer at 4°C.
Stained sections were osmicated, dehydrated and embedded in epoxy resin (TAAB® Embedding Materials). Sections of 1 μm thickness were cut on a microtome and stained with toluidine blue before being examined on a light microscope. Ultrathin sections were cut on an ultramicrotome and stained with lead uranyl acetate by the Centre for Ultrastructural Imaging, Kings College London. Sections were mounted on unsupported 100 mesh grids and sections were visualised on a Hitachi H7600 transmission electron microscope.
For analysis, photographs of positively stained axons only from each animal were taken at a magnification of 15,000. All positively stained axons in the field of view on one mesh grid were photographed and analyzed and at least 125 axons were analyzed per experimental group. Total counts of myelinated axons were carried out using a semi-thin cross-section of the whole sciatic nerve at the same level. G-ratios of positively stained axons were measured per animal using AxioVision LE Rel. 4.2 software.
Differences between SLICK-A Cre; NRG-1fl/fl tamoxifen-treated animals and vehicle-treated controls were determined using the Students t test for comparison of two groups, or One-Way ANOVA or Two-Way ANOVA using the Tukey post hoc test for more than two groups. Results are reported as mean values ± SEM. Cumulative frequency of axonal sprout length were compared statistically using the Kolmogorov-Smirnov test.
SLICK-A mice co-express a tamoxifen inducible form of Cre recombinase and yellow fluorescent protein (YFP) in a subpopulation of motor and sensory neurons (Young et al., 2008). In the SLICK-A line the proportion of motoneurons which express YFP depends on the spinal level and muscle examined (Young et al., 2008). At the level of the lumbar spinal cord 33.14% ± 2.00 (n=4) of motoneurons (identified by Choline-Acetyl Transferase imunostaining, ChAT) expressed YFP (Supplemental Fig. 1). In the SLICK-A mouse a small percentage (1.7%±0.3) of DRG cell profiles were also YFP positive and all of these DRG cells were also immunostained for phosphorylated Neurofilament heavy chain (a marker for DRG cells with large diameter myelinated axons, Supplemental Fig. 2) but did not express CGRP or bind IB4 (markers of small diameter DRG cells with unmyelinated axons, data not shown).
SLICK-A Cre; ROSA 26 reporter mice demonstrated clear induction of Cre activity (assessed by ß-galactosidase expression) in YFP-expressing motoneurons and DRG cells following tamoxifen but not vehicle administration in the adult (Supplemental Fig. 2). SLICK-A Cre mice were crossed with mice homozygous for the conditional Nrg1 allele (Nrg1f/f) which have loxP sites flanking the sequences encoding the essential EGF domain (Meyer and Birchmeier, 1995;Yang et al., 2001;Li et al., 2002). In SLICK-A Cre; NRG-1fl/fl animals, we used in situ hybridization and a probe directed against the β isoform of the EGF domain to confirm the absence of EGF-domain mRNA sequences in YFP-labeled sensory and motoneurons following administration of tamoxifen but not vehicle (Fig. 2). After blinded analysis of control animals, 94% (49/52) of YFP positive motoneurons expressed β EGF domain mRNA, however this was detectable in only 9% (9/98) of YFP positive motoneurons in SLICK-A Cre; NRG-1fl/fl animals following tamoxifen administration.
To determine whether juxtacrine NRG1 signaling was required for maintenance of the myelin sheath in the naive state, immuno-electron microscopy was used to identify YFP positive axons within the sciatic nerve of SLICK-A Cre; NRG-1fl/fl animals. 12 weeks following tamoxifen or vehicle administration, myelinated fiber morphology was normal (Fig. 3), and there were no significant differences in G-ratio (0.63±0.01 versus 0.62±0.007, mean ± SE, tamoxifen versus vehicle treatment) or axon diameter (4.7±0.3 μm versus 4.3±0.34 μm, tamoxifen versus vehicle). There was no difference in the number of YFP axons in the sciatic nerve of vehicle or tamoxifen-treated SLICK-A Cre; NRG-1fl/fl mice (136±22 versus 149±14, mean ± SE, tamoxifen versus vehicle). Even at 22 weeks following treatment with tamoxifen myelinated fiber morphology remained normal (Fig. 3) and there were no significant differences in G-ratio (0.63±0.02 versus 0.6±0.002 tamoxifen versus vehicle) or axon diameter (5.0±0.04 μm versus 4.9±0.4 μm, tamoxifen versus vehicle).
In order to investigate a potential function of NRG1 in axon remyelination and regeneration, we analyzed conditional mutant mice after peripheral nerve injury (Fig. 1). At 8 weeks following sciatic nerve crush, control animals (either vehicle treated SLICK-A Cre; NRG-1fl/fl or tamoxifen treated SLICK-A Cre; NRG-1+/+) showed highly effective remyelination (Fig.4 A and Supplemental Fig. 3). In contrast YFP positive axons of conditional NRG1 mutant (tamoxifen treated SLICK-A Cre; NRG-1fl/fl) mice demonstrated markedly deficient remyelination: a significant proportion of YFP-axons (38.6%±3.3) were ensheathed by Schwann cells but had failed to elaborate a myelin sheath (including axons of up to 5 μm in diameter Fig. 4 C,E). In tamoxifen treated SLICK-A Cre; NRG-1fl/fl animals the axons that had been remyelinated displayed a significantly thinner myelin sheath, with an average G ratio of 0.84±0.02 (versus 0.72±0.007 and 0.68±0.007 for vehicle treated SLICK-A Cre; NRG-1fl/fl and tamoxifen treated SLICK-A Cre; NRG-1 +/+ respectively p<0.005 Fig. 4 B, D). In contrast to the marked changes in G-ratio, the axon diameter of regenerated YFP positive axons was unaffected (3.78 μm±0.11 and 3.63 μm±0.18 vehicle and tamoxifen treated SLICK-A Cre; NRG-1fl/fl respectively, n = 4).
Further evidence of deficits in remyelination was apparent when NRG1 deficient YFP labeled motor axons were visualized within the gastrocnemius muscle. Immunohistological analysis of tamoxifen treated SLICK-A Cre; NRG-1fl/fl animals 8 weeks post crush revealed YFP positive axons along which myelin basic protein (MBP) was absent over long segments, equivalent to the length of multiple internodes (Fig.5).
To assess long range axon regeneration, YFP labeled axons were visualized in whole mount preparations of the tibial nerve at either 10, 14 days or 8 weeks following sciatic nerve crush (Fig. 6). Compared to axons in uninjured tibial nerve, the regenerating axons were thinner and growth cones could be identified (Supplemental Fig. 4). At 10 days following sciatic nerve crush YFP positive axons from tamoxifen treated SLICK-A Cre; NRG-1fl/fl animals demonstrated a significant reduction in regeneration distance from the crush site compared to control (either vehicle treated SLICK-A Cre; NRG-1fl/fl or tamoxifen treated SLICK-A Cre; NRG-1 +/+ animals p<0.05, two way ANOVA, Fig. 6 and Supplemental Fig. 4). At 14 days post crush although the YFP positive axons of tamoxifen treated SLICK-A Cre; NRG-1fl/fl mice had regenerated a further distance through the sciatic nerve this was still significantly reduced compared to vehicle treated animals (tamoxifen versus vehicle treatment of SLICK-A Cre; NRG-1fl/fl, p<0.05, two way ANOVA). At 8 weeks post crush, the axons had fully regenerated through the tibial nerve, and no significant difference between the groups was observable (Fig. 6).
We examined the consequences of NRG1 ablation on the neuromuscular junction (NMJ). In the gastrocnemius muscle of SLICK-A mice 30% of NMJs are innervated by YFP positive motoneurons. In SLICK-A Cre; NRG-1fl/fl mice 12 weeks following tamoxifen administration, the structure of the NMJ was normal (Fig. 7 A). The re-innervation of gastrocnemius NMJs was then assessed at 10, 14 days and 8 weeks post sciatic nerve crush. At 10 days the level of NMJ reinnervation was low in both groups and did not significantly differ between them (1.4±0.5% vs 2.5±0.6% of gastrocnemius NMJs were innervated by YFP positive axons at 10 days post crush in the tamoxifen versus vehicle treatment groups of SLICK-A Cre; NRG-1fl/fl animals, p=0.2). At 14 days post crush NMJ reinnervation was more established; the proportion of NMJs which were reinnervated by YFP positive axons was not significantly different comparing tamoxifen to vehicle treated SLICK-A Cre; NRG-1fl/fl animals (20.5%±3.2 versus 18.2%±3.2 of gastrocnemius NMJs respectively). However at this time a greater proportion of the NRG1 deficient axons demonstrated partial rather than full innervation (the proportion of NMJs which were partially innervated by YFP positive axons was 10.8% ±1.25 versus 6.8% ±1.49 tamoxifen versus vehicle treated SLICK-A Cre; NRG-1fl/fl, p<0.05). At 8 weeks post crush, the number of NMJs innervated by YFP positive axons had returned to the pre-injury level in all groups and were fully innervated (30.8±5.5% versus 22.8±3.4% tamoxifen versus vehicle SLICK-A Cre; NRG-1fl/fl, no significant difference). Although in the tamoxifen treated SLICK-A Cre; NRG-1fl/fl group the NMJ was fully reinnervated at this time point, there was a much higher rate of terminal sprouting (Fig. 7 D,E). These sprouts were associated with the processes of terminal Schwann Cells (tSCs) as revealed by immunostaining with S100 antibodies.
In this study we have used ‘SLICK’ mice to ablate NRG1 conditionally in a sub-set of adult sensory and motoneurons. We have found that axons lacking NRG1 were able to maintain a myelin sheath but were severely impaired in remyelination after nerve injury. Such axons also showed a slower rate of regeneration and although they reinnervated their targets, the neuromuscular junction demonstrated excessive terminal sprouting.
The expression of the SLICK transgene in a sub-population of sensory and motor neurons enabled us to probe the role of juxtacrine NRG1 signaling in the maintenance of the myelin sheath. There is evidence that a transcriptional program maintains the Schwann cell myelinating phenptype in adulthood. Krox20 is a key transcriptional regulator of myelination during development (Topilko et al., 1994) and conditional ablation of Krox20 in adult Schwann cells results in rapid myelin breakdown accompanied by the expression of genes such as Sox2 associated with Schwann cell dedifferentiation (Decker et al., 2006;Le et al., 2005). The fact that in vitro axon contact promotes Krox20 expression by Schwann cells (Murphy et al., 1996) and that loss of axonal contact in vivo results in downregulation of Krox20 (and other myelin related genes, (Zorick et al., 1996)) might suggest that NRG1 is required for continued Krox20 expression and myelin maintenance. We did not find however that juxtacrine NRG1 signaling was required for myelin maintenance. 22 weeks following tamoxifen administration to SLICK-A Cre; NRG-1fl/fl animals, YFP axon and myelin morphology remained normal. The exact half-life of the NRG1 protein is unknown, so the actual period of NRG1 deficiency may be less than the full 22 weeks. Because we only ablated NRG1 in a subset of axons, neighboring axons or even Schwann cells (Rosenbaum et al., 1997) might release Ig-containing NRG1 isoforms which could signal in a paracrine fashion and theoretically compensate for the absent juxtacrine NRG1 signaling in tamoxifen treated SLICK-A Cre; NRG-1fl/fl animals. Conditional ablation of ErbB2, the receptor that participates in NRG1 signal transduction, in Schwann cells of mature mice does not result in any change in myelinated fiber morphology up to 2 months following tamoxifen administration, supporting the notion that ongoing NRG1-ErbB signalling is not required for myelin maintenance (Atanasoski et al., 2006).
Traumatic nerve injury results in a sequence of cellular changes termed Wallerian degeneration: distal axonal segments degenerate, myelin breaks down, Schwann cells dedifferentiate/proliferate and macrophages phagocytose debris (Chen et al., 2007). Subsequently, axons regenerate into bands of Bungner (Schwann cells within tubes of basal lamina), Schwann cells differentiate into a myelinating phenotype (Zorick et al., 1996) axons are remyelinated and saltatory conduction is re-established. Of note, remyelinated fibers have a thinner myelin sheath (Schroder, 1972) and shorter internodal distance (Minwegen and Friede, 1985).
Given its key developmental role, we investigated whether NRG1 provides an axoglial signal promoting remyelination. It has previously been shown that nerve injury results in increased expression of the ErbB2 and 3 receptors within Schwann cells of the distal nerve (Carroll et al., 1997;Cohen et al., 1992;Kwon et al., 1997). Changes in the expression of Neuregulin-1 isoforms after injury are complex. Expression of the type III isoform, the isoform responsible for myelination, in sensory and motoneurons decreases initially and then returns to normal levels as axons reinnervate peripheral targets (Bermingham-McDonogh et al., 1997). In contrast the expression of isoforms containing Ig-like domains, increases at the site of injury (Carroll et al., 1997).
We observed at 2 months following sciatic nerve crush, axons lacking NRG1 had significantly thinner myelin sheaths and 40% had completely failed to elaborate a myelin sheath despite being ensheathed by Schwann cells. While visualizing motor axons, it was clear that long lengths of NRG1 deficient axons were amyelinated. The type III isoform of NRG1 requires proteolytic cleavage by BACE1 for full activity, and reduced remyelination of axons has also been observed in BACE−/− mice (Hu et al., 2008). Not all aspects of the Schwann cell response to axon injury are however dependent on NRG1-ErbB signaling. In a previous study in which the ErbB2 receptor was conditionally ablated in adult Schwann cells, no effect on Schwann cell proliferation distal to a site of nerve transaction was noted (Atanasoski et al., 2006). Given the small proportions of axons in which we ablated NRG1 in this study, we did not analyze Schwann cell proliferation here.
It is well established that the response of Schwann cells following nerve injury promotes axon regeneration (Chen et al., 2007). This is likely to be mediated through multiple mechanisms including the expression of neurotrophic factors, extracellular matrix components and removal of myelin components that are inhibitory to axonal outgrowth. In axons lacking NRG1, axonal regeneration was slowed at 10 and 14 days post sciatic nerve crush but at 8 weeks there was no difference from control. Altered interactions with Schwann cells could have a role in the slowed regeneration of NRG1 deficient axons, for instance a reduction in the synthesis of factors that promote axonal outgrowth. Addition of soluble NRG1 to the culture media promotes axon outgrowth from cultured superior cervical ganglion neurons; this effect is indirect and occurs via the release of growth promoting agents from Schwann cells in response to NRG1 (Mahanthappa et al., 1996). Impaired myelination per se is unlikely to cause the slowed axon regeneration. In the control animals, myelination had not commenced in the majority of YFP axons in the sciatic nerve at ten days post crush, a time when there was a clear difference in regeneration of these axons into the tibial nerve (unpublished observations). NRG1 might promote axon outgrowth in a cell autonomous fashion by ‘back signaling’ of transmembrane NRG1 isoforms. ErbB binding to NRG1 has been shown in certain model systems to result in gamma-secretase dependent cleavage, releasing the intracellular domain that is targeted to the nucleus where it modulates gene transcription (Bao et al., 2003). It was recently reported that axon outgrowth in developing cortical neurons can be modulated by expression of the type III isoform of NRG1, which appeared to be independent of ‘back signaling’ and was mediated by an as yet unidentified signaling system (Chen et al., 2010).
Recent analysis of mutant mice suggest that NRG1 signaling is dispensable for NMJ formation (Escher et al., 2005;Jaworski and Burden, 2006). We did not observe changes in NMJ structure following NRG1 ablation in adult motoneurons. At 2 weeks following sciatic nerve crush, there was no difference in the total number of NMJs reinnervated by axons lacking NRG1, although a greater number were partially rather than fully reinnervated. The mild phenotype in neuromuscular reinnervation compared to the clear reduction in long range axon regeneration may correlate with a greater capacity of NRG1 deficient axons to sprout locally once they reach their targets. 8 weeks after injury, the NMJ was fully reinnervated in all groups, but a much higher level of terminal sprouting was noted in axons lacking NRG1. Terminal sprouting is a phenomenon which is triggered by denervation or inhibition of synaptic transmission (e.g. from botulinum toxin) at the NMJ (Kang et al., 2003). We found, as others have previously described, that such terminal sprouts follow terminal Schwann cell processes (Son and Thompson, 1995). These specialized non-myelinating Schwann cells cap the NMJ and have been shown to respond to NRG1-ErbB signaling. Treatment with exogenous NRG1 rescues terminal Schwann cells from apoptosis following neonatal axotomy (Trachtenberg and Thompson, 1996) and expression of constitutively active ErbB2 in Schwann cells results in proliferation and process extension from these cells (Hayworth et al., 2006). Our findings suggest that axonally-derived NRG1 is clearly dispensable for terminal Schwann cell process extension. The excess terminal sprouting which we observed may not be a direct consequence of the absence of NRG1 at the NMJ itself but instead secondary to the grossly impaired remyelination of the motor axons. Increased NMJ terminal sprouting has previously been described in a mutant mouse model of a demyelinating neuropathy, EGR2 I268N, as a consequence of conduction block (Baloh et al., 2009).
In conclusion, during adulthood NRG-1 has a key role in mediating the signaling between axons and Schwann cells required for effective nerve repair. Our data indicate that therapeutics based on the inhibition of ErbB signaling in oncology (Lurje and Lenz, 2009) and BACE1 in Alzheimer’s disease (Vassar et al., 2009), will not impair normal nerves, but will significantly interfere with the regenerative response to nerve injury.
FRF is a BBSRC CASE PhD student part funded by Acorda Therapeutics Inc. DLHB is a Wellcome Intermediate Clinical Scientist (Grant No. 077074/z/05/z). CT is a PhD student of the Wellcome Trust funded London Pain Consortium. ANG and CB acknowledge the support of the German Science Foundation (SFB 665).