SEARCH

SEARCH BY CITATION

Keywords:

  • auditory neurons;
  • development;
  • glial cells;
  • inner ear;
  • neural crest cells;
  • Sox genes

Abstract

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

J. Neurochem. (2010) 114, 1827–1839.

Abstract

The cochleo-vestibular ganglion contains neural crest-derived glial cells and sensory neurons that are derived from the neurogenic otic placode. Little is known about the molecular mechanisms that regulate the tightly orchestrated development of this structure. Here, we report that Sox10, a high-mobility group DNA-binding domain transcription factor that is required for the proper development of neural crest cell derivatives, is specifically expressed in post-migratory neural crest cells in the cochleo-vestibular ganglion. Using Sox10-deficient mice, we demonstrate that this transcription factor is essential for the survival, but not the generation, of the post-migratory neural crest cells within the inner ear. In the absence of these neural crest-derived cells, we have investigated the survival of the otocyst-derived auditory neurons. Surprisingly, auditory neuron differentiation, sensory target innervation and survival are conserved despite the absence of glial cells. Moreover, brain-derived neurotrophic factor expression is increased in the hair cells of Sox10-deficient mice, a compensatory mechanism that may prevent spiral ganglion neuronal cell death. Taken together, these data suggest that in the absence of neural crest-derived glial cells, an increase trophic support from hair cells promotes the survival of spiral ganglion neurons in Sox10 mutant mice.

Abbreviations used
BDNF

brain-derived neurotrophic factor

CVG

cochleo-vestibular ganglion

E

embryonic day

NCC

neural crest cell

NT3

neurotrophin-3

PBS

phosphate-buffered saline

PFA

paraformaldehyde

TdT

terminal deoxynucleotidyl transferase

TUNEL

terminal dUTP nick-end labelling

The development of the mammalian cochlea follows a set of time-regulated processes. After otic placode formation, the otic tissue invaginates and initiates its autonomous developmental program that forms the structures of the inner ear (i.e., the semicircular canals, vestibule and cochlea) (Rubel 1978; Torres and Giraldez 1998; Kiernan 2002). Early in development (embryonic day – E9.5 in mice) a subset of the otic epithelial cells delaminates from the otocyst and forms the neural part of the cochleo-vestibular ganglion (CVG) (D’Amico-Martel and Noden 1983; Rubel and Fritzsch 2002). As this process is occurring, neural crest cells (NCCs) delaminate from the dorsal neural tube. These NCCs migrate extensively to various parts of the embryo where they differentiate into a wide variety of cell types, forming the majority of neurons and glial cells in the peripheral nervous system and the craniofacial skeleton as well as cartilage, neuroendocrine cells, and melanocytes (Le Douarin 1975, 1980; Le Douarin and Dupin 2003). In the mammalian inner ear, NCCs give rise to both the intermediate cells of the stria vascularis (i.e., the secretory epithelium of the cochlea) and to satellite glial cells of the CVG.

Sox genes encode a group of transcription factors that carry a conserved high-mobility group DNA-binding domain (Kamachi et al. 2000) and are divided into subgroups A–J based on their sequence homology (Bowles et al. 2000). Sox genes are expressed by various tissues in a cell type-specific manner and have been demonstrated to be key players in the regulation of embryonic development and the determination of cell fate (Pevny and Lovell-Badge 1997; Wegner 1999; Wegner and Stolt 2005). Sox10, a member of the SoxE subgroup, has been shown to be an important regulator of NCC development (Kelsh and Raible 2002; Hong and Saint-Jeannet 2005). All NCCs express Sox10 upon emigration, but this expression is subsequently restricted (Kelsh and Raible 2002; Wegner 2005; Wegner and Stolt 2005). In mice, loss-of-function mutations in the Sox10 locus cause, cell-autonomous defects in peripheral glial cell and melanocyte development; in contrast homozygous mutants exhibit neonatal lethality associated with a failure in the migration and/or differentiation of multiple NCC derivatives, including glia and autonomic neurons (Herbarth et al. 1998; Southard-Smith et al. 1998; Kim et al. 2003). In humans, Sox10 mutations are associated with Waardenburg-Shah syndrome, which is characterised by enteric nervous system defects, hypopigmentation, and neurosensory deafness (Pingault et al. 1998; Southard-Smith et al. 1998) and related neurocristopathies such as Waardenburg syndrome type 2, Yemenite deaf-blind hypopigmentation syndrome and a multi-syndrome disorder (PCWH; OMIM 609136) (Inoue et al. 2004; Pingault et al. 2010). Although the role of Sox10 in the peripheral nervous system has been extensively studied, there are still large gaps in the understanding of its function in the inner ear. Here, we use Sox10-deficient mice to investigate the function and expression pattern of Sox10 in the spiral ganglion.

Material and methods

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Sox10lacZ mice and genotyping

Sox10lacZ mutant mice (Britsch et al. 2001) and wild-type littermates of C3HeB/FeJ background were obtained from heterozygous crosses. Genotypes were determined from tail biopsies using PCR as previously described (Britsch et al. 2001). Homozygous Sox10-deficient mice (Sox10lacZ/lacZ) die at birth, making it impossible to study postnatal stages.

Tissue preparation

Mice were time-mated with the date of the vaginal plug considered as E0.5. Pregnant adult mice were killed by cervical dislocation according to methods approved by the University of Liège Institutional Animal Care and Use Committee (n° 04/467). For histological examination and β-galactosidase staining, embryos were dissected and fixed in 4% paraformaldehyde (PFA) in phosphate buffered-saline (PBS) for 2–4 h at 4°C. After three PBS rinses, the fixed embryos were immersed in 20% sucrose in PBS overnight at 4°C. Heads or cochleae were frozen and embedded in Tissue-Tek (Sakura, Zoeterwoude, The Netherlands), then sectioned at 15 μm, mounted on superfrost slides, and stored at −80°C.

For explant experiments, cochleae were dissected out from the calvaria with watchmaker forceps under a stereomicroscope. Each organ of Corti and spiral ganglion was extracted from surrounding tissues and explanted intact onto the surface of a sterile membrane (Millicell, 12 mm, Millipore, Bedford, MA, USA) in minimum essential medium (Gibco, Gent, Belgium) in a 24-well culture plate (Nunc, Brussels, Belgium). Explants were then fixed for 10 min in PFA in preparation for immunocytochemistry.

β-Galactosidase staining and immunocytochemistry

Detection of β-galactosidase activity followed standard procedures. After fixation overnight in PFA, embryos or cryosections were incubated overnight at 37°C in X-gal staining solution [0.75 mg/mL X-gal, 5 mM K3(CN)6, 5 mM K4(CN)6, 2 mM MgCl2, 0.02% NP-40 (Roche Diagnostics, Vilvoorde, Belgium) and 0.01% Na-deoxycholate in PBS]. Tissues were then washed in PBS and post-fixed overnight in PFA. Sections were counter-stained with eosin.

For immunohistological analyses, head or cochlea sections were washed three times with PBS and blocked for 30 min with gelatin 0.25%–Triton X-100 0.1% (Sigma-Aldrich, Bornem, Belgium) in PBS at 20°C. After being diluted in the same solution, the primary antibodies were incubated overnight at 4°C. The following primary antibodies were used in various combinations: βIII-tubulin (mouse monoclonal IgG2a, clone Tuj1, 1 : 1500, Babco, Richmond, CA, USA and rabbit polyclonal, 1 : 1000, Covance, Cumberland, VA, USA); β-galactosidase (rabbit, 1 : 100, MP Biomedicals Europe, Illkirch, France); brain-derived neurotrophic factor (BDNF; rabbit, 1 : 100, Santa Cruz, Santa Cruz, CA, USA); cleaved caspase 3 (rabbit, 1 : 200, Promega Benelux, Leiden, The Netherlands); Islet1 (mouse, 1 : 2, DSHB, Iowa City, IA, USA); MyosinVI (goat, 1 : 100, Santa Cruz); Nestin (Chicken, 1 : 200, Novus Biologicals, Cambridge, UK); neurotrophin-3 (NT3; rabbit, 1 : 200, Alomone Labs, Jerusalem, Israel); p75 nerve growth factor receptor (p75NGFR) (rabbit, 1 : 100, Millipore, Temecula, CA, USA); pan-Trk (rabbit, SC-11, 1 : 100, Santa Cruz) and Parvalbumin (mouse, 1 : 100, Sigma-Aldrich). All immunostainings were visualised after incubation for 1 h at 20°C using the following reagents (1 : 500): FITC-, RRX-, Cy5-conjugated donkey anti-mouse, donkey anti-rabbit and donkey anti-goat antibodies (Jackson Immunoresearch Laboratories, West Grove, PA, USA) and tetramethyl rhodamine isothiocyanate (TRITC)-phalloidin (1 : 1500, Sigma-Aldrich). Sections were then mounted in Vectashield containing diaminido phenyl indol (DAPI) (HardSet Mounting Medium, Vector laboratories, Burlingame, CA, USA) and examined with a confocal microscope (Olympus Fluoview FV1000, Aartselaar, Belgium). Tissue sections without primary antibodies were used as controls. For BDNF, p75NGFR and Trk labelling, sections were subjected to antigen retrieval before immunostaining via immersion in Dako antigen retrieval solution (Dako Target Retrieval solution, Dako Belgium, Heverlee, Belgium) for 20 min at 95°C. After immersion, sections were slowly cooled down to 20°C.

In situ hybridisation

Sections were air-dried, washed with PBS for 5 min and post-fixed for 10 min with PFA. Sections were then treated with 100 mM triethanolamine pH 8, which was acetylated by adding drops of acetic anhydride 0.25% while rocking 15 min at 20°C. After three washes with PBS-Tween, 0.1%, sections were pre-hybridised with pre-warmed hybridisation buffer (Amresco, Solon, OH, USA) for 60 min at 70°C. Sections were then hybridised with 800 ng/mL Sox10 (pZL1/Sox10) or ErbB3 (ErbB3DR3, a generous gift from Carmen Birchmeier, Max Delbruck Centrum, Berlin, Germany) RNA probes overnight at 70°C. Sections were then washed twice with pre-warmed washing buffer (50% Formamide, 2× SSC, 0.1% Tween 20) for 60 min at 65°C. Sections were then pre-incubated for 1 h at 20°C in Tris-Saline blocking buffer [100 mM Tris pH 7.5, 150 mM NaCl and 10% Normal Goat Serum (Dako)] and then incubated overnight at 4°C with an anti-digoxigenin antibody coupled to alkaline phosphatase (1 : 2000, Roche Applied Science, Vilvoorde, Belgium) in blocking solution. After three washes with Tris–Saline buffer, sections were overlaid with 200 μL filtered NBT/BCIP-Tween 0.1% solution (Sigma-Aldrich, St Louis, MO, USA) for 6–10 h. This reaction was blocked with PBS washes which were followed by post-fixation with PFA for 15 min. Sections were mounted in Aquamount medium (BDH laboratories, Poole, UK).

Semi-thin sections and transmission electron microscopy

Cochleae from E17.5 embryos were dissected out and fixed for 1 h at 4°C in 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer. The specimens were post-fixed with 1% osmium tetraoxide for 1 h at 4°C in the same buffer and then rinsed with distilled water and dehydrated using a graded series of ethanol solutions and propylene oxide. Cochleae were imbued with epoxypropane and Epon : epoxypropane, and then embedded in Epon. Semi-thin sections (4 μm) were prepared with an ultramicrotome (UCT7, Leica Microsystems, Nussloch, Germany), stained with toluidine blue and photographed on a Zeiss Axioplan microscope equipped with a Zeiss Axiocam camera (Carl Zeiss, Iena, Germany). The spiral ganglion and organ of Corti were identified on semi-thin sections. Ultra-thin sections of 70 nm that contained auditory neurons were obtained and mounted on copper grids. The sections were contrasted using uranyl acetate and lead citrate and observed with a scanning-transmission electron microscope (Joel-electron microscope JEM-100 CX100, JEOL (Europe) BV, Zaventem, Belgium) at 60 kV.

Assays for apoptotic cell death

Apoptotic nuclei were labelled using the terminal dUTP nick-end labelling (TUNEL) method (Gavrieli et al. 1992). Frozen sections were brought to 20°C, rinsed for 15 min in PBS, and treated for 30 min with 0.5% Triton X-100 in PBS to unmask DNA. After a 10 min rinse in PBS, sections were pre-incubated in terminal deoxynucleotidyl transferase (TdT) buffer (30 mM Tris, pH 7.5, 140 mM sodium cacodylate and 1 mM cobalt chloride). Sections were then incubated in TdT buffer containing 300 U/mL TdT (Roche Diagnostics, Vilvoorde, Belgium) and 6 μM biotinylated dUTP (Roche Diagnostics) in a humid chamber for 90 min at 37°C. The reaction was stopped by a 15 min rinse at 20°C in termination buffer (300 mM NaCl, 30 mM sodium citrate). Sections were rinsed in PBS before a 10 min treatment with 2% BSA (bovine serum albumin fraction V, Sigma-Aldrich) in PBS to minimise non-specific staining. After one PBS rinse, biotin-dUTP-labelled sections were incubated with Alexa568-conjugated streptavidin (1 : 1000, Jackson Immunoresearch Laboratories) for 1 h at 20°C. Fluorescent immunohistochemical staining for βIII-tubulin, islet1 and nestin was performed on TUNEL-labelled sections as described above. Nuclei were counter-stained with diamidino phenyl indol (DAPI).

Quantification

For each 15-μm-thick section, the otocyst and CVG were Z-scanned using a 40× objective lens, and the composite of the Z-stack images was analysed. At E10.5, the total number of TUNEL-positive cells within the nestin-positive region was evaluated on every two sections throughout the developing CVG. The totals from each section were summed and multiplied by 2 to obtain the total number of TUNEL-positive cells per CVG. This unbiased method of cell profile counting resembles stereological cell counting as previously described (West et al. 1991; Parrish-Aungst et al. 2007). Sections were coded to ensure that analysis was performed blindly. Values are presented as mean ± SD, and ‘n’ represents the number of animals analysed per condition.

For neuron and TUNEL quantification at E13.5 and E17.5, micrographs were compiled from Z-series scans with an Olympus Fluovieuw confocal microscope. The number of neurons and the ganglion area were determined on every two or three sections of the cochlea using MetaMorph software (Molecular Devices, Toronto, Canada). Quantification is expressed as the total number of cells per volume unit.

Data analysis

Statistical analyses were performed using GraphPad In Stat software (GraphPad Software Inc., San Diego, CA, USA). Student’s t-tests were used for paired comparisons, and other comparisons were analysed using a one-way anova followed by a Tukey’s post-hoc test. Results were considered significant at p < 0.05.

Quantitative RT-PCR

Total RNA was extracted from cochlear explants dissected from E17.5 embryos. Two cochlear explants from a single animal were processed in each sample. Three animals were tested for each genotype. RNA extraction was performed by adding 1 mL TRIzol® (Invitrogen, Merelbeke, Belgium) to the sample, which was then carefully crushed and incubated for 5 min at 20°C. Chloroform (200 μL) was added to each sample, and the mix was vortexed for 15 s and kept at 20°C for 15 min. After centrifugation at 12 000 g for 15 min at 4°C, the clear supernatant (containing the RNA) was spared. The RNA was precipitated by adding 500 μL of isopropanol to each sample, which was then incubated at 20°C for 10 min and then centrifuged at 12 000 g for 10 min. The RNA pellets were washed with 500 μL of 75% ethanol in diethylene pyro carbonate (DEPC)-treated water and then centrifuged at 6000 g for 5 min. The pellets were air dried and suspended in 30 μL of diethylene pyro carbonate (DEPC)-water. All RNA samples were then subjected to treatment with DNase I (Roche Diagnostics) at 20°C for 15 min. Synthesis of cDNA was performed starting beginning with total RNA, which was reverse-transcribed with SuperScript II or III Reverse Transcriptase (Invitrogen) according to the manufacturer’s instructions. This process was also performed without reverse transcriptase to confirm the absence of contaminating genomic DNA in all samples. The resulting cDNA was used for quantitative PCR with the Faststart Universal SYBR Green Master (Roche Diagnostics). Thermal cycling was performed on an Applied Biosystems 7900HT Fast Real-Time PCR detection system (Applied Biosystems, Foster City, CA, USA). Optimal annealing temperature for the primers was determined to be 60°C and 40 cycles. The quantity of each mRNA transcript was measured and expressed relative to β2-microglobulin (B2M). Relative expression of mRNA was determined by calculations of threshold cycle (Ct) and was presented as the fold change compared to wild-type mice. All reactions were run in triplicate and the mean Ct for each triplicate set was calculated. The fold change was expressed using the 2−ΔΔCt method (Livak and Schmittgen 2001). The primers used for BDNF and NT3 detection were designed by Stankovic and Corfas (Hear Res 2003).

Results

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Sox10 expression pattern in the developing cochleo-vestibular ganglion

To investigate in vivo the role of Sox10 during inner ear development, we used Sox10 knock-in LacZ mice, which have the Sox10 coding sequence replaced with the LacZ reporter gene (Britsch et al. 2001). On whole-mount embryos, β-galactosidase expression was detected during inner ear development beginning at the formation of the otic placode at E8.5 (data not shown). As soon as the CVG begins forming, strong LacZ expression appears (E10.5) (Fig. 2a) and persists throughout the development of the CVG (Fig. 1a–c) until adulthood (data not shown). In addition, in situ hybridisation experiments performed on E17.5 mouse embryos showed the accumulation of Sox10 mRNA throughout the entire cochlear duct and the spiral ganglion (Fig. 1d–e). In order to identify cells that expressed Sox10 in the spiral ganglion, we first performed double immunolabelling using antibodies against β-galactosidase (red) and βIII-tubulin, an early pan-neuronal marker (green), on postnatal day 1 (P1) heterozygous cochleae (Fig. 1f–h). The βIII-tubulin-positive cells were never positive for β-galactosidase immunoreactivity. We next examined the expression of nestin, an intermediate filament protein that is expressed in a wide variety of cells including peripheral glial cells (Friedman et al. 1990; Jessen and Mirsky 2005). In the spiral ganglion, perfect co-expression of β-galactosidase and nestin was detected (Fig. 1i–k). The co-localization of β-galactosidase with nestin rather than βIII-tubulin indicates that Sox10 is expressed in glia in the spiral ganglion.

image

Figure 2.  Cochleo-vestibular ganglion development in the Sox10lacZ/lacZ mice. (a–d) Transverse cochlear sections stained with X-gal and counter-stained with eosin from Sox10lacZ/+ (a, c) and Sox10lacZ/lacZ (b, d) mouse embryos. The β-galactosidase-expressing cells are present in the cochleo-vestibular ganglion (CVG) at E10.5 (b) and are totally absent of the spiral ganglion in the Sox10lacZ/lacZ mouse (dotted area) at E17.5 (d). (e–m) Confocal images of cochlear duct transverse sections in E12.5 wild-type (e, h and k), Sox10lacZ/+ (f, i and l) and Sox10lacZ/lacZ (g, j and m) mice labelled with βIII-tubulin (green) and p75NGFR (red) antibodies. Auditory neurons are present in the CVG (dotted area) of the Sox10lacZ/lacZ mouse embryo (g, m). Whereas post-migratory neural crest cells (NCCs) that are p75NGFR-positive were detected in the CVG of the wild-type (h, k) and Sox10lacZ/+ (i, l) mice, no p75NGFR-positive cells were present in the CVG of Sox10lacZ/lacZ mice (j, m). Scale bar (in m) = 200 μm in (a–d) and 75 μm in (e–m).

Download figure to PowerPoint

image

Figure 1.  Sox10 expression in the developing spiral ganglion. (a) E11.5 and (b–c) E17.5 Sox10lacZ/+ cochleae stained with X-gal and counter-stained with eosin. (a) Inner ear transverse section showing Sox10 expression in the entire epithelium of the otic vesicle (ov) and in part of the cochleo-vestibular ganglion (CVG) at E11.5. (b–c) β-galactosidase is expressed in the cochlear duct and in the spiral ganglion at E17.5. (d–e) In situ hybridisation on cochlear transverse section from E17.5 wild-type mouse showing Sox10 mRNAs in the spiral ganglion and in the cochlear duct. (f–k) Confocal images of inner ear transverse sections in P1 Sox10lacZ/+ mouse immunostained with β-galactosidase (red, f, i), βIII-tubulin (green, g) and nestin (green, j) antibodies. (h) and (k) are merged images. In the spiral ganglion, β-galactosidase is not expressed in the auditory neurons stained with βIII-tubulin (h) but is expressed in the glial cells (k) stained with nestin. Scale bar (in k) = 200 μm in (a), 300 μm in (b), 75 μm in (c), 300 μm in (d), 75 μm in (e) and 50 μm in (f–k).

Download figure to PowerPoint

Cochleo-vestibular ganglion development in Sox10-deficient mice

In the developing CVG, a similar β-galactosidase staining was observed between Sox10lacZ/+ and Sox10lacZ/lacZ E10.5 mice (Fig. 2a and b), suggesting that post-migratory NCCs are present in both genotypes. However, no β-galactosidase staining was detected later during development in the spiral ganglion of the Sox10lacZ/lacZ mice (Fig. 2d). These results were confirmed with double immunolabelling using βIII-tubulin antibody and glial lineage markers such as p75NGFR (Fig. 2e–m) and nestin antibodies (Figure S1). Whereas p75NGFR- and nestin-positive cells were present in the CVG of the wild-type and Sox10lacZ/+ mice, no glial cells could be detected in the CVG of the Sox10lacZ/lacZ mouse embryos. These results showed that Sox10 is necessary for glial cell survival and specification in the CVG. Because ErbB3 receptors are specifically present on peripheral glial cells (Hansen et al. 2001), we performed ErbB3 in situ hybridisation on cochlear sections from E13.5 and E17.5 mice. In the wild-type spiral ganglion, a strong level of ErbB3 expression was observed at both ages (Fig. 3a, d, g and j), while in situ hybridisation signals were largely reduced in heterozygous sections (Fig. 3b, e, h and k) and were absent in Sox10lacZ/lacZ spiral ganglia (Fig. 3c, f, i and l).

image

Figure 3.  ErbB3 expression in the developing cochleo-vestibular and spiral ganglion. (a–l) ErbB3 mRNA expression in the cochlea of E13.5 (a–f) and E17.5 (g–l) wild-type (a, d, g and j), Sox10lacZ/+ (b, e, h and k) and Sox10lacZ/lacZ (c, f, i and l) mouse embryos. ErbB3, a well-described indirect target of Sox10, is expressed in the glial cell precursors of the wild-type and Sox10lacZ/+ CVG and spiral ganglion. ErbB3 expression is largely reduced in the Sox10lacZ/+ mouse embryo relative to the wild-type mouse and is completely absent in the Sox10lacZ/lacZ CVG and spiral ganglion. Scale bar (in l) = 150 μm in (a–c), 50 μm in (d–f), 150 μm in (g–i) and 50 μm in (j–l).

Download figure to PowerPoint

Post-migratory neural crest cell death in the absence of Sox10

Because glial cell precursors in Sox10lacZ/lacZ embryos were completely absent in E12.5, we monitored CVG cell death to elucidate the function of Sox10 in CVG gliogenesis. We performed TUNEL analyses accompanied by a co-labelling with nestin and βIII-tubulin antibodies on transverse sections of E10.5 embryos. At E10.5, nestin-positive cells were present in the developing CVG of both wild-type and Sox10 mutant embryos (Fig. 4c–f). Although apoptotic figures were detected in the wild-type CVG (Fig. 4a and e), cell death was considerable in the CVG of Sox10 mutant embryos (Fig. 4b and f) and was increased almost 3-fold in Sox10lacZ/lacZ CVG relative to wild-type (Fig. 4g). Apoptosis was prominent in undifferentiated NCCs expressing nestin but was not found in differentiating βIII-tubulin-positive neurons (Fig. 4e–f). Double immunolabelling with antibodies for β-galactosidase and another neuronal marker, islet1, confirmed these results (Figure S2).

image

Figure 4.  Sox10 influences the survival but not the appearance of neural crest derived-cells in the cochleo-vestibular ganglion. (a–f) Confocal images of E10.5 inner ear sections from wild-type and Sox10lacZ/lacZ mice labelled for apoptosis with TUNEL (red, a–b) and immunostained with anti-nestin (green, c–d) and βIII-tubulin (grey, e–f). (e)–(f) are merged images. Nestin-positive cells derived from the NCCs are present in the developing CVG of each genotype. (g) Quantification of cell death within the nestin-positive area (dotted area). NCC death is increased in the Sox10lacZ/lacZ mouse embryos compared to wild-type littermates. Results are expressed as mean ± SD (Student’s t-test: *p < 0.05; n = 3). Scale bar = 200 μm in (a–f).

Download figure to PowerPoint

Embryonic auditory neurons survive and develop in the absence of Sox10

After analysing cell death, we then investigated the development, differentiation and survival of spiral ganglion neurons in the absence of Sox10. Immunofluorescence analyses on cochlear explants at E17.5 demonstrated the presence of βIII-tubulin-expressing cells in Sox10lacZ/+ (data not shown) and Sox10lacZ/lacZ spiral ganglia. In both wild-type and Sox10lacZ/lacZ mice, immunofluorescent staining for βIII-tubulin showed that auditory neurons sent their projections to the hair cells of the organ of Corti, which were labelled with phalloidin (Fig. 5a–d). At higher magnification, these neuronal projections reached both inner and outer hair cells (Figure S3). This finding suggests that embryonic inner ear innervation does not require peripheral glial cells. In order to confirm the absence of glial cells and the presence of well-developed neurons, we analysed semi-thin sections of wild-type and Sox10lacZ/lacZ spiral ganglia (Fig. 3e–f). These analyses confirmed the absence of satellite glia in E17.5 Sox10lacZ/lacZ spiral ganglion relative to wild-type.

image

Figure 5.  Embryonic auditory neurons can survive and develop without Sox10. (a–d) Confocal images of whole-mount explants of E17.5 embryos showing double labelling with βIII-tubulin (green) and phalloidin (red). Auditory neurons (green) are present on the entire length of the organotypic explant and send their projections to the hair cells of the organ of Corti (stereocilia stained with phalloidin) in the wild-type (a–b) and in the Sox10lacZ/lacZ (c–d) mouse embryos. (e–f) Semi-thin sections focus on the spiral ganglion of the wild-type (e) and Sox10lacZ/lacZ (f) mouse embryos at E17.5. Two types of cells can be identified in the wild-type spiral ganglion: the auditory neurons with a big nucleus and the satellite glial cells, which are darker cells with less cytoplasm and morphological heterogeneity (e). In the Sox10lacZ/lacZ spiral ganglion (f), there are only auditory neurons. (g–h) Transmission electron microscopy micrographs. (g) shows the two types of cells in the spiral ganglion of the wild-type mouse embryo; (h) shows the presence of auditory neurons alone and a completely normal morphology in the Sox10lacZ/lacZ spiral ganglion. Scale bar (in h) = 200 μm in (a–d), 20 μm in (e–f), 2 μm in (g–h).

Download figure to PowerPoint

Next, comparative ultrastructural analyses were performed to evaluate the fine structure of the spiral ganglion at E17.5 (Fig. 5g–h). The wild-type spiral ganglion exhibited two cell types: (i) the auditory neurons, which are large ganglionar cells with a lightly coloured circular nucleus containing one or two nucleoli and (ii) the glial cells, darker cells with a polymorphic nucleus containing condensed chromatin. In contrast, an ultrastructural examination of Sox10lacZ/lacZ mouse spiral ganglion confirmed the absence of satellite glial cells (Fig. 5h). In addition, no signs of suffering (such as autolysis, cellular shrinkage, chromatin condensation, or disappearance of organelles) were observed in neurons despite the absence of glial cells.

To examine the role of Sox10 in auditory neuron survival, we evaluated the neuronal density in the spiral ganglia of wild-type, Sox10lacZ/+ and Sox10lacZ/lacZ mouse embryos at both E13.5 and E17.5. As assessed with βIII-tubulin immunostaining, no significant difference in the overall neuronal density between the three genotypes was observed at either age (Fig. 6a–b). To determine whether the absence of glial cells can cause a loss of neurons through programmed cell death, we performed TUNEL staining combined with βIII-tubulin immunolabelling (Fig. 6c–g). No significant difference in the number of apoptotic nuclei in the spiral ganglia of wild-type, Sox10lacZ/+ and Sox10lacZ/lacZ mice was observed at E13.5 (Fig. 6c). In contrast, significantly fewer TUNEL-positive neurons were found at E17.5 in the spiral ganglion of Sox10lacZ/lacZ mice relative to wild-type and Sox10lacZ/+ mice (Fig. 6d). These results were confirmed using cleaved caspase 3 immunolabelling (data not shown).

image

Figure 6.  Quantitative analysis of neurons in the spiral ganglion. Quantification of the spiral ganglion neuronal density (a–b) and quantification of neurons undergoing apoptosis per μm3 (c–d). (a–b) Number of auditory neurons per μm3 at E13.5 (a) and E17.5 (b) in the spiral ganglion. Neuronal density in the spiral ganglion is not significantly different between wild-type, Sox10lacZ/+ and Sox10lacZ/lacZ mouse embryos; this similarity demonstrates that Sox10 is not necessary for embryonic auditory neuron development and survival. (c–g) TUNEL assays to assess cell death in the spiral ganglion at E13.5 and E17.5. (c–d) Number of TUNEL-positive auditory neurons per μm3 at E13.5 (c) and E17.5 (d) in the spiral ganglion. At E13.5, cell death is not increased in the spiral ganglion despite the absence of Sox10; at E17.5, cell death is even decreased in the Sox10lacZ/lacZ spiral ganglion. (e–g) Confocal micrographs of cochlear sections at E17.5 showing double labelling for TUNEL-positive cells (red) and βIII-tubulin (green) in wild-type (e), Sox10lacZ/+ (f) and Sox10lacZ/lacZ (g) mouse embryos. (anova followed by a Tukey’s post-hoc test: *p < 0.05; n = 3). Scale bar = 60 μm in (e–g).

Download figure to PowerPoint

Hair cell-derived BDNF is up-regulated in the absence of Sox10

The signalling pathway of the two neurotrophins, BDNF and NT3, and their corresponding receptors, TrkB and TrkC, is responsible for the survival of developing cochlear neurons (Fritzsch et al. 1997, 2002; Rubel and Fritzsch 2002). To determine wether this signalling pathway is still active in Sox10 mutant mice, we searched for the presence of TrkB and TrkC proteins in auditory neurons using pan-Trk and βIII-tubulin antibodies. Double immunostaining on E17.5 cochlear sections did not show any differences in the distribution of Trk proteins between Sox10lacZ mutant mice and wild-type (Fig. 7).

image

Figure 7.  Expression of neurotrophin receptors in the E17.5 spiral ganglion. (a–i) Confocal micrographs of cochlear sections at E17.5 showing double labelling for βIII-tubulin (green, a–c) and panTrk (red, d–f) in wild-type (a, d, g), Sox10lacZ/+ (b, e, h) and Sox10lacZ/lacZ (c, f, i) mouse embryos. PanTrk, which labels high-affinity neurotrophin receptors (i.e., TrkA, B and C), is present in all auditory neurons independently of the genotype. Scale bar = 75 μm in (a–i).

Download figure to PowerPoint

Interestingly, BDNF expression was slightly increased in the hair cell cytoplasm of both heterozygous and homozygous Sox10lacZ mice relative to the wild-type littermates (Fig. 8a–f), whereas NT3 expression remained unchanged (Fig. 8g–l). In addition, quantitative RT-PCR performed on cDNA samples prepared by reverse transcription from E17.5 cochlear explant RNA extracts confirm these results (Fig. 8m–n) and suggest that an increase production and supply of BDNF by hair cells would support the survival of auditory neurons in the absence of glial cells in Sox10lacZ mutant embryos.

image

Figure 8.  Neurotrophin expression in the organ of Corti. (a–l) Confocal micrographs of organ of Corti sections at E17.5 showing double labelling for neurotrophin with either BDNF (red, a–f) or NT3 (red, g–l) antibodies and double labelling for a hair cell marker (green), with either myosinVI (d–f) or parvalbumin (j–l) antibodies in wild-type (a, d, g and j), Sox10lacZ/+ (b, e, h and k) and Sox10lacZ/lacZ (c, f, i and l) mouse embryos. BDNF expression is more intense in the hair cells of the Sox10lacZ/+ (b and e) and Sox10lacZ/lacZ (s) organ of Corti, whereas NT3 is conserved in the supporting cells (g–l). (m–n) BDNF (m) and NT3 (n) expression of gene in E17.5 wild-type compared to Sox10lacZ/+ and Sox10lacZ/lacZ cochlear ducts by quantitative RT-PCR. The level of expression of BDNF is significantly increased in the Sox10lacZ/+ and Sox10lacZ/lacZ mouse relative to wild-type littermates while there is no statistically significant difference in the level of expression of NT3. Results are expressed as mean ± SD (anova followed by a Tukey’s post-hoc test: *p < 0.05 and **p < 0.01; n = 3 independent experiments for BDNF and NT3 expression). * represents hair cell. Scale bar = 30 μm in (a–l).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

In the present study, we used Sox10lacZ mutant mice to analyse the potential role of Sox10 in the survival, migration and specification of spiral ganglion cells. The spiral ganglion is a structure composed of cells from various origins: the auditory neurons that are of placodal origin and the glial cells that are NCC derivatives (Rubel and Fritzsch 2002; Alsina et al. 2009). Our experiments showed that the absence of Sox10 does not affect the appearance nor migration of early neural crest derivatives in the CVG (i.e., glial cell precursors). Indeed, previous studies on the peripheral nervous system revealed that Sox10 is not necessary for the formation of NCC derivatives (Paratore et al. 2002). Later in development, the absence of Sox10 depleted the pool of NCC-derived glial precursors, through cell death and resulted in a lack of mature glial cells in the Sox10lacZ/lacZ CVG, a result also observed in other peripheral ganglia (Southard-Smith et al. 1998; Kapur 1999; Britsch et al. 2001; Paratore et al. 2001). These results confirmed that CVG non-neuronal cells are derived from NCCs and not from otic placode progenitors.

During the embryonic period, we also found neuronal precursors that differentiate into healthy auditory neurons in the absence of Sox10. This result indicates that auditory neuron differentiation and survival do not depend upon Sox10 expression and glia differentiation. It is worth noting that CVG neurons are derived from the otic placode, whereas other peripheral neurons originate from the neural crest. As already described in the enteric and dorsal root ganglion neurons, the progressive loss of neurons could be attributed to a direct defect in the survival and the maintenance of neural crest multipotent stem cells (Sonnenberg-Riethmacher et al. 2001; Kim et al. 2003; Haldin and Labonne 2010). In contrast, the embryonic spiral ganglion neurons differentiate and survive in the inner ear despite the absence of glial cells. Indeed, cranial sensory neurons are derived from both the embryonic placode and the neural crest, while auditory neurons come exclusively from the otic placode. Therefore, the differentiation and survival of auditory neurons are independent of neural crest-derived stem cells.

In the peripheral nervous system, neuronal defects are also typically attributed to a loss of peripheral glial cell-derived trophic support (Britsch et al. 2001; Sonnenberg-Riethmacher et al. 2001). Indeed, glial cells contribute to the mechanisms that underlie neuronal survival, growth and regeneration (Fallon 1985; Kleitman et al. 1988a,b; Whitlon et al. 2009). As in Sox10lacZ/lacZ mutant mice, the survival of peripheral glial cell precursors in ErbB2−/− mice is impaired, leading to a complete absence of glial cells after E13.5 (Morris et al. 2006). In addition, this study has found that the spiral ganglion is located in an unusual position and has observed numerous apoptotic neurons at E16.5. However, our study indicated that the determination, proliferation and differentiation of auditory neurons are unaffected in the embryonic Sox10lacZ mutant mice. Moreover, we observed a significant decrease in neuronal apoptosis in Sox10lacZ/lacZ mutant mice relative to wild-type mice. Taken together, these results showed for the first time that mammalian neuronal differentiation and survival do not require the presence of glial cells. This result is in accordance with the recent finding that the numbers of stato-acoustic ganglion neurons was not affected in Sox10 mutant zebrafish (Dutton et al. 2009).

Role of the hair cell-derived neurotrophins

The survival and maintenance of spiral ganglion neurons depend largely on peptide neurotrophic factors such as the neurotrophins: BDNF and NT3 (Ernfors et al. 1992, 1995; Pirvola et al. 1992; Ylikoski et al. 1993; Wheeler et al. 1994). Neurons are exposed to neurotrophins via paracrine (secreted from their synaptic partners) or autocrine (from themselves) mechanisms (Davies and Wright 1995). This mechanism also occurs in spiral ganglion neurons, which are exposed to BDNF released from the hair cells and NT3 released from the supporting cells (Farinas et al. 2001; Agerman et al. 2003). The glia associated with neuronal somata or axons provide another paracrine source of neurotrophins (Davies 1998). In this study, we reported an increase in BDNF expression in the hair cells of the Sox10lacZ/+ and Sox10lacZ/lacZ mutant mice relative to wild-type littermates; while the level of NT3 also tented to increase in the organ of Corti, although without statistical significance. In addition, we observed that Trk receptors are expressed similarly in spiral ganglion neurons independently of the mouse’s genotype. Because the loss of neurotrophic support from the hair cells is considered a major contributing factor to the degeneration of spiral ganglion neurons in the deaf ear (Ernfors et al. 1994, 1995), these results suggest that Sox10lacZ/lacZ auditory neurons could survive and develop via the production of hair cell-derived BDNF. In addition, the overproduction of BDNF by hair cells in Sox10lacZ/lacZ mice could explain the reduced neuronal cell death that was observed at E17.5 relative to their wild-type littermates. This increased BDNF expression as a compensatory mechanism has already been shown in other neuronal cell types (Bustos et al. 2009). Nevertheless, further studies are required to establish the exact role of Sox10 in late spiral ganglion neuron survival using conditional ear-specific knock-out mice to overcome the embryonic lethal recessive phenotype of Sox10 mutant mice.

In conclusion, our data indicate that Sox10 plays an essential role in the maintenance of cochlear glial cells but does not affect neurons. Surprisingly, we found that a lack of glial cells in the cochlea does not disrupt the development and survival of auditory neurons, a finding that distinguishes the cochlea from the rest of the peripheral nervous system. Evidence from animal studies has indicated that prolonged degeneration of spiral ganglion neurons will compromise hearing. Understanding how spiral ganglion neurons develop and survive may lead to the development of new therapies for sensorineural deafness following otic neuron degeneration.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

This work was supported by the National Fund for Scientific Research (FNRS, Belgium), the Léon Frédéricq Founds and the Deutsche Forschungsgemeinschaft. We thank Dr. G. Moonen (University of Liege, Belgium) for his helpful advice and comments. We acknowledge Dr. C. Birchmeier (Max Delbruck Centrum, Germany) for the ErbB3 RNA probe. We are also grateful to P. Ernst-Gengoux and A. Brose for their technical support and expertise. B. Malgrange is a Research Director, L. Nguyen a Research Associate and I. Breuskin a Postdoctoral Researcher for the FNRS.

References

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Figure S1. Nestin expression in the cochleo-vestibular ganglion. (a–i) Confocal images of cochlear transverse sections in E13.5 wild-type (a, d and g), Sox10lacZ/+ (b, e and h) and Sox10lacZ/lacZ (c, f and i) mice labelled with anti-nestin (red, a–c) and anti-βIII-tubulin (green, d–f). (g)–(i) are merged images of each staining. Auditory neurons are well present in the cochleo-vestibular ganglion (CVG) (dotted area) of the Sox10lacZ/lacZ mouse embryo (f, i). Whereas nestin-positive cells are detected in the CVG of the wild-type (a) and Sox10lacZ/+ (b) mice, none appears in the one of the Sox10lacZ/lacZ mouse (c). Scale bar = 75 μm in (a–i).

Figure S2. Auditory neurons do not undergo apoptosis during early embryonic development in Sox10 mutant mice. (a–g) Confocal images of CVG sections in E10.5 wild-type (a), Sox10lacZ/+ (b, d and f) and Sox10lacZ/lacZ (c, e and g) mice labelled for apoptosis with TUNEL (a–c) and immunostained with anti-islet1 (green, a–c and f–g) and anti-β-galactosidase (d–e). (f)–(g) are merged images. TUNEL-positive cells are only detected in the area of β-galactosidase-positive cells. Scale bar = 100 μm in (a–g).

Figure S3. Hair cell innervation pattern in the absence of Sox10. (a–f) Confocal images of organ of Corti sections in E17.5 wild-type (a, c and e) and Sox10lacZ/lacZ (b, d and f) mice immunostained with anti-βIII-tubulin (c–d) and labelled with a DAPI nuclear counter-staining (a–b). (e)–(f) are merged images. Neuron projections labelled with βIII-tubulin reach the inner and outer hair cells in the wild-type and the Sox10LacZ/LacZ mice. * represents hair cell. Scale bar = 30 μm in (a–f).

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
JNC_6897_sm_FigS1.tiff8225KSupporting info item
JNC_6897_sm_FigS2.tiff12564KSupporting info item
JNC_6897_sm_FigS3.tiff6863KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.