Disorganized Innervation and Neuronal Loss in the Inner Ear of Slitrk6-Deficient Mice

Slitrks are type I transmembrane proteins that share conserved leucine-rich repeat domains similar to those in the secreted axonal guidance molecule Slit. They also show similarities to Ntrk neurotrophin receptors in their carboxy-termini, sharing a conserved tyrosine residue. Among 6 Slitrk family genes in mammals, Slitrk6 has a unique expression pattern, with strong expression in the sensory epithelia of the inner ear. We generated Slitrk6-knockout mice and investigated the development of their auditory and vestibular sensory organs. Slitrk6-deficient mice showed pronounced reduction in the cochlear innervation. In the vestibule, the innervation to the posterior crista was often lost, reduced, or sometimes misguided. These defects were accompanied by the loss of neurons in the spiral and vestibular ganglia. Cochlear sensory epithelia from Slitrk6-knockout mice have reduced ability in promoting neurite outgrowth of spiral ganglion neurons. Indeed the Slitrk6-deficient inner ear showed a mild but significant decrease in the expression of Bdnf and Ntf3, both of which are essential for the innervation and survival of sensory neurons. In addition, the expression of Ntrk receptors, including their phosphorylated forms was decreased in Slitrk6-knockout cochlea. These results suggest that Slitrk6 promotes innervation and survival of inner ear sensory neurons by regulating the expression of trophic and/or tropic factors including neurotrophins from sensory epithelia.


Introduction
The Slitrk family consists of neuronal transmembrane proteins that control neurite outgrowth [1,2]. Structurally, Slitrks share leucine-rich repeat (LRR) domains located amino-terminal to the transmembrane domain. LRR domains are present in many proteins and mediate protein-protein interactions [3]. LRR domains in the Slitrk family proteins are similar to those in all Slit family proteins, which control axon guidance and branching [4]. Tyrosine residues in the carboxy-terminus are another structural feature of the Slitrk family; these tyrosines are flanked by amino acid sequences similar to those in the carboxy-terminal domain of the Ntrk neurotrophin receptor [5].
The mouse Slitrk family contains 6 genes (Slitrk1-6) [1]. Slitrk1 is required for higher brain functions [6]. However, little is known about the physiological roles of the other family members. Although Slitrk1 through Slitrk5 are expressed broadly throughout the brain, the expression of Slitrk6 is highly restricted to thalamic nuclei [1]. Furthermore, a comprehensive expression analysis of Slitrk6 revealed a strong expression in the auditory and vestibular sensory epithelia of the ear [7]. This unique expression pattern led us to investigate the role of Slitrk6 in inner ear development. Inner ear sensory epithelia contain mechanosensory hair cells that recognize sound as well as linear and angular acceleration for balance [8]. During development, sensory epithelia also play an important role in the development of sensory neurons of the inner ear by releasing diffusible factors that promote survival and outgrowth of sensory neurons [9].
In the present study, we generated Slitrk6-knockout mice and investigated the development of their auditory and vestibular sensory organs. Histological examination and marker labeling studies revealed a reduction in the innervation density in the cochlea of Slitrk6-deficient mice. In addition, the innervation to the posterior crista was lost, reduced, or sometimes misguided in the vestibule. Many sensory neurons within the spiral and vestibular ganglia were lost during development. In the explant co-culture experiment, cochlear sensory epithelia from Slitrk6-knockout mice had less activity in promoting neurite outgrowth of the spiral ganglion neurons. These results indicate that Slitrk6 plays an important role in the sensory neural development of the inner ear by regulating the expression of molecules that promote survival and neurite outgrowth of sensory neurons.

Slitrk6 Expression during Inner Ear Development
First, we carried out in situ hybridization analysis (Fig. 1) to know the Slitrk6 transcripts distribution in the course of inner ear development. Slitrk6 transcripts are first detected at embryonic day (E)8.5 in the otic placode, which invaginates to form the otic vesicle [7]. In the E10.5 otic vesicle, Slitrk6 transcripts were strongly expressed in the ventromedial and laterodorsal regions (Fig. 1A), which give rise to the cochlear and vestibular sensory epithelia [8,10,11]. At E15.5, Slitrk6 expression marked the presumptive organ of Corti (Fig. 1C). In addition, we detected relatively weak expression of Slitrk6 in a narrow region of the spiral ganglion near the developing cochlear sensory epithelium from E12.5 to E15.5 (Fig. 1C). In the postnatal day (P)1 cochlea, Slitrk6 transcripts were detected strongly in supporting cells and weakly in both inner and outer hair cells (Fig. 1D, D'). Furthermore, Slitrk6positive signals were detected in vestibular sensory epithelia, including ampullary cristae, saccular and utricular maculae (Fig. 1B, E, F, and data not shown). Slitrk6 transcripts were densely located at the lumenal surface of the sensory epithelium, where hair cells localize (Fig. 1F'). The hybridization signal was also detected weakly in supporting cells located at the basal cell layer of the vestibular sensory epithelium (Fig. 1F'). We also found weak expression of Slitrk6 in vestibular ganglion neurons near the sensory epithelium from E11.5 to E14.5, comparable to what was observed in spiral ganglion neurons (Fig. 1B).
Next, we examined the localization of the Slitrk6 protein using an anti-Slitrk6 antibody. Sltrk6-immunopositive signals were detected early in the otic vesicle ( Fig. 2A, B), and later in cochlear and vestibular sensory epithelia where Slitrk6 transcripts were distributed ( Fig. 2C-L, Fig. S1, S2). In the presumptive organ of Corti, Slitrk6-immunopositive signals were strongly detected by E14.5 (Fig. 2C, D), and then confined to the supporting cell types at later stages of embryonic development and at the newborn (Fig. 2E, F, Fig. S2A-F). In the vestibular sensory epithelia, the signal was strong at the lumenal surface ( Fig. 2G-L). However, double labeling with a hair cell marker calretinin indicated that the strong signals were in the non-hair cell region, presumably in the lumenal processes of the supporting cells ( Fig. S2G-I). Transient expression of Slitrk6 in the spiral and vestibular ganglion neurons was also confirmed at the protein level, and positive signals were partly overlapped with neurofilament immunostaining (Fig. 2C, D, G, H). The cell type preferences of the Slitrk6 expression in the cochlear and vestibular organs are comparable to each other.

Generation of Slitrk6-Null Mutant Mice
To analyze the role of Slitrk6 in inner-ear development, we generated Slitrk6-deficient mouse lines. A targeting vector was designed to replace the open reading frame (ORF) of Slitrk6 with a phosphoglycerol kinase (PGK)-neo expression cassette flanked by a loxP sequence (Fig. 3A). We isolated 5 independent embryonic stem (ES) clones with homologous recombination, and 2 ES clones . Slitrk6 transcripts are found in the ventromedial (arrow) and laterodorsal (arrowhead) regions of the otic vesicle at E10.5 (A). At E15.5, the expression of Slitrk6 mRNA marks the region of the developing organ of Corti (C). In addition, a faint positive signal is seen in the nascent spiral ganglion neurons adjacent to the sensory epithelium (arrowhead in C). Slitrk6 mRNA is still expressed in the organ of Corti at P1 (D), and higher magnification of the organ of Corti reveals that Slitrk6 transcripts are localized densely in supporting cells and weakly in inner and outer hair cells (D'). Slitrk6 mRNA is also detected in vestibular sensory epithelia, including ampullary cristae and utricular macula (B, E, F). Higher magnification of the utricular epithelium reveals that Slitrk6 transcripts are densely located at the lumenal layer of the sensory epithelium (F'). A faint positive signal is also observed in the nascent vestibular ganglion neurons adjacent to the sensory epithelium (arrows in B). Anterior, posterior, dorsal and ventral directions are indicated by arrows in (A). AC, anterior crista; IHC, inner hair cell; OHC, outer hair cell; OV, otic vesicle; PC, posterior crista; SAG, statoacoustic ganglion; SG, spiral ganglion; U, utricle. Scale bars: A, B, C, D, D', F' 50 mm; E, F, 100 mm. doi:10.1371/journal.pone.0007786.g001 yielded chimeric mice capable of transmitting the disrupted allele (+neo) through the germline. Subsequently, the PGK-neo cassette was removed by crossing the heterozygous mice with mice transgenic for the Cre recombinase gene under the control of the cytomegalovirus immediate-early enhancer/chicken b-actin hybrid (CAG) promoter [12], which express Cre recombinase in their zygotes (Dneo, Fig. 3B). Ablation of Slitrk6 mRNA was confirmed by RT-PCR (Fig. 3C). Heterozygous mating produced Slitrk6 +/+ , Slitrk6 +/2 , and Slitrk6 2/2 pups at the expected Mendelian frequencies. Both male and female mice grew without showing any external abnormalities and were fertile. The inner ear phenotypes described in the present study were observed in two mouse lines derived from two independent ES clones. Thus, the phenotypes are considered to correctly reflect the effects of the inserted mutation.

Cochlear Innervation Defects in the Absence of Slitrk6
The gross morphology of the Slitrk6-deficient inner ear was normal. In addition, light microscopic analysis of hematoxylin and eosin (H&E)-stained sections showed that the general organization of the organ of Corti and the size of cochlear sensory epithelium in the Slitrk6-deficient mice was not obviously different from those of wild-type littermates (Fig. S3A, B, E, F).
To initially characterize the innervation patterns of the inner ear in wild-type and mutant mice, we selectively labeled the afferent innervation of the cochleae at P0 by applying DiI to the alar plate of the brainstem, where the fibers terminate centrally. In contrast to the dense, evenly spaced innervation by radial fibers of the wild-type cochlea (Fig. 4A, C), the radial fibers in the Slitrk6knockout mice were fasciculated with large spaces between the bundles (Fig. 4B, D). In Slitrk6-knockout mice, the innervation defects were distributed evenly throughout the cochlea, and all the cochlear turns were affected to the same degree ( Fig. 4A-D).
To label the efferent fibers, we then placed DiI crystals in the olivo-cochlear efferent bundle near the floor plate [13,14]. Selective DiI tracing of the olivo-cochlear efferents in the cochleae revealed that the efferent fibers were also abnormally fasciculated into fewer radial bundles, showing innervation defects similar to afferent fibers (Fig. 4E, F). Efferent fibers use afferent fibers as a scaffold during pathfinding, and the track of the efferents closely follows the afferent innervation pattern [15,16]. Therefore, the observed cochlear efferent innervation defects are likely to be secondary to the afferent innervation defects.
We next examined the developmental changes of the cochlea innervation defects between E16.5 and P7, using whole-mount surface preparations of the cochlea labeled with anti-neurofilament antibody to visualize the innervation pattern ( Fig. 5A-F). The radial fiber projection to the organ of Corti was strongly reduced in the mutant mice, leaving larger gaps between radial bundles than in the wild-type mice ( Fig. 5A-F). At P7, a higher magnification view revealed that the type II spiral fibers that extend through the tunnel of Corti to the three rows of outer hair cells were reduced in number in Slitrk6-knockout mice (Fig. 5E', F'). Myelin staining by osmium tetroxide revealed that the defects in the innervation pattern were still prominent in Slitrk6-deficient mice at P28 (Fig. 5G, H).

Vestibular Innervation Defects in the Absence of Slitrk6
We next examined the abnormalities in the vestibular region of Slitrk6-deficient mice. Light microscopic examination revealed that the sensory hair cells in all of the vestibular regions (ampullary cristae and saccular and utricular maculae) appeared normal with regard to their number, shape, and tissue architecture (Fig. S3C,  D). However, the sensory epithelia in saccular macula and . The spiral ganglion neurons adjacent to the sensory epithelium are immunolabeled with Slitrk6 in a pattern that partly overlaps with neurofilament (arrows in C). At P0, Slitrk6 expression is confined to the supporting cells within the area of the organ of Corti (E, F). The neurofilament-labeled fibers project below the hair cells. In the macula of the sensory epithelium from an E15.5 saccule, Slitrk6 is uniformly and heavily distributed in the lumenal layer of the epithelium (I, J). In the sensory epithelium from E15.5 anterior crista, Slitrk6 expression is highly observed at the lumenal layer of the striolar region of the crista (K, L). At E12.5, faint Slitrk6-positive signals are also observed in the vestibular ganglion neurons adjacent to the sensory epithelium (arrows in G). DAPI-stained nuclei (blue) are seen in all panels of the merged images (B, D, F, H, J, L). AC, anterior crista; HC, horizontal crista; IHC, inner hair cell; OHC, outer hair cell; OV, otic vesicle; SAG, statoacoustic ganglion; SG, spiral ganglion. Scale bars, 50 mm. doi:10.1371/journal.pone.0007786.g002 posterior crista were slightly smaller than those of wild-type littermates at E16.5, but not at E13.5 ( Fig. S3E-G).

Slitrk6-Deficient Mice
To examine whether the number of neurons in the spiral and vestibular ganglia is affected by the deletion of Slitrk6, we calculated the number of neurons in each ganglion and its volume. In the spiral ganglion, neither the number of neurons nor ganglion volume was affected at E13.5; however, both were significantly reduced at E16.5, and at P0, they decreased to about a half of those of wild-type mice (Fig. 7A, B). Cell death in the spiral ganglion was significantly higher in Slitrk6-deficient mice at E16.5 than in the wild-type (Fig. 7C, G, K, Fig. S4A, B). Although the incidence of cell death was markedly reduced at P0 relative to the incidence at E16.5, it was still significantly higher in Slitrk6deficient than in wild-type mice. In contrast to the change in the spiral ganglion, in the vestibular ganglion both the number of neurons and the volume began to decrease from E13.5, and their values were about 75% of those of wild-type mice from E16.5 to P0 (Fig. 7D, E). The incidence of cell death in the vestibular ganglion was significantly higher in Slitrk6-knockout mice at E13.5 and E16.5 than in wild-type mice (Fig. 7F, I, M, Fig. S4C, D). At P0, H&E-stained sections showed no apparent differences between wild-type and knockout mice in cell size and organization in either the spiral or vestibular ganglion (Fig. 7H, J, L, N). Reduction in both the number of neurons and volume in vestibular ganglion was prominent in inferior vestibular ganglion that gives rise to posterior crista innervation (Fig. S5) [17,18]. This result suggests that the neuronal loss within the vestibular ganglion is largely due to the loss of neurons that innervate the posterior crista.

Decrease in the Neurite Outgrowth Activity in Cochlear Sensory Epithelia of Slitrk6-Deficient Mice
Inner ear sensory epithelia are known to release diffusible factors that promote survival and outgrowth of statoacoustic ganglion neurons [9]. Thus we examined whether Slitrk6-deficient sensory epithelia promote the survival and outgrowth of ganglion neurons by using culture experiments. In the explant culture, spiral ganglion neurons of both wild-type and Slitrk6-knockout mice could extend neurites toward wild-type sensory epithelia (Fig. 8A, B, E). However, sensory epithelia of knockout mice weakly attracted the neurites of both wild-type and knockout spiral ganglia ( Fig. 8C-E). These results suggest that there is a reduction in the expression of trophic and/or tropic factors in Slitrk6deficient sensory epithelia.

Decreased Expression of Neurotrophins and Their Receptors in the Slitrk6-Deficient Inner Ear
Among the molecules that promote survival and outgrowth of inner ear sensory neurons, neurotrophins are best characterized by multiple experiments. Brain-derived neurotrophic factor (Bdnf) and Neurotrophin-3 (Ntf3) are two neurotrophins that are produced in the developing sensory epithelia, and gene knockout studies revealed that they promote the innervation and support the survival of sensory neurons [19,20]. Receptors for Bdnf (Ntrk2) and Ntf3 (Ntrk3) are expressed in sensory neurons, and mice deficient in these receptors also show reduced innervation and the loss of sensory neurons that mirror the ligand defects [21]. Therefore, we analyzed the expression of these molecules in E14.5 inner ear (which included the cochlear and vestibular sensory epithelia and the spiral and vestibular ganglia) by using real-time PCR. The expression of Bdnf and Ntf3 mRNA was significantly decreased in Slitrk6-deficient inner ear, whereas the expression of Ntrk2 and Ntrk3 was not affected (Fig. 9A). In situ hybridization of Bdnf and Ntf3 did not show any differences in the distribution of these mRNAs between wild-type (data not shown). We further examined the expression of Ntrk receptors using protein samples extracted from E14.5 cochleae (which included the cochlear sensory epithelium and spiral ganglion). We used cochleae at this  developmental stage because neuronal loss in the spiral ganglion might affect the results at later stages. The expression level of both Ntrk2 and Ntrk3 proteins was significantly decreased in Slitrk6-deficient cochleae (Fig. 9B, C). In addition, the amount of phosphorylated form of Ntrk receptor was decreased in knockout mice (Fig. 9B, C). However, their spatial localization was not altered in the knockout mice (Fig. 9D-U). This suggests that Slitrk6 deletion interferes with the level of neurotrophin mRNA but affects Ntrks predominantly at the post-transcriptional level.

Discussion
In the present study, we developed Slitrk6-knockout mice and analyzed the development of their auditory and vestibular sensory organs. The overall innervation density of the cochlea was markedly reduced in the Slitrk6-deficient mice, and the innervation to the posterior crista in the vestibule was reduced, disoriented, or lost. Many sensory neurons in both the spiral and vestibular ganglia died during late embryonic development.
Slitrk6 is predominantly expressed in the sensory epithelia, which release trophic and/or tropic factors for sensory neurons. Especially, supporting cells of the organ of Corti in which Slitrk6 is strongly expressed, have been shown to play an important role in the survival of cochlear sensory neurons [22,23]. The present results suggest that Slitrk6 participates in the transmembrane signal transduction processes that enhance the expression of such factors in sensory epithelia. Among these factors, neurotrophins are best understood by numerous studies [19,20,21]. In the developing inner ear, Ntf3 is expressed predominantly in the supporting cell and Bdnf in the hair cell lineages, and both Ntrk2 and Ntrk3 are expressed in the sensory neurons [22,24,25]. Mice deficient in one of these neurotrophins or Ntrk receptors show reduced innervation and loss of sensory neurons (Fig. S6) [16,[26][27][28][29]. Slitrk6-deficient inner ear showed mild but significant decrease in the expression level of both Bdnf and Ntf3 (Fig. 9A). In addition, we found the decrease in the expression of both Ntrk2 and Ntrk3 at protein level, and the amount of phosphorylated form of Ntrk receptors also decreased in Slitrk6-deficient cochlea (Fig. 9B, C). These results suggest a disturbance in the neurotrophin-Ntrk signaling in Slitrk6-deficient inner ear. Indeed, there are some similarities in the inner ear phenotypes of Slitrk6and neurotrophin/Ntrk-deficient mice (Fig. S6, Table S1). Although the phenotypes of the inner ear with a partial reduction of both Bdnf-Ntrk2 and Ntf3-Ntrk3 signalings are still unknown, Bdnf heterozygous mutant mice display an intermediate decrease in size and neuronal number of the vestibular ganglion [28]. Thus, it is likely that quantitative changes in neurotrophin signaling contribute to the development of the inner ear phenotypes of Slitrk6knockout mice.
Together with the co-culture experiment (Fig. 8), the reduction of the neurotrophins from cochlear sensory epithelia may be primarily responsible for the spiral ganglion innervation defects. This idea is also supported by a result that isolated spiral ganglia of Slitrk6-knockout mice can extend their neurites almost the same as those of wild-type mice in the presence of Ntf3 (Fig. S7). On the other hand, the sizes of the sensory epithelia in saccular macula and posterior crista were slightly lower in Slitrk6-deficient mice compared to those of wild-type littermates in late embryonic stages (Fig. S3E, F). This result indicates that Slitrk6 play some unidentified roles in the development of sensory epithelia per se. However, we believe that the innervation defects are not due to the sensory epithelial size reduction, because the innervation defect to the posterior crista was observed as early as E13.5 when the size reduction in the epithelium was not apparent.
The results in the present study indicate the involvement of neurotrophin-Ntrk signaling in the inner ear defects of Slitrk6knockout mice. However, the decrease in the expression of neurotrophins in Slitrk6-knockout mice was relatively small compared to their dramatic phenotype. Culture experiments identified several candidate molecules (TGF-b, LIF, MCP-1, erythropoietin and Fgf2) that affect the survival and outgrowth of inner ear sensory neurons [30][31][32][33]. Slitrk6 may regulate the expression of some of those or other unidentified trophic and/or tropic factors that promote neuronal survival and neurite extension in addition to neurotrophins. Decrease in the expression of such factors may also affect the dynamics of Ntrk neurotrophin receptors.
Several genes are known to be critical for sensory neural development in the mouse inner ear, including NeuroD1, Brn3a (also called as Pou4f1), and ErbB [34][35][36][37][38]. Mice lacking NeuroD1 or Brn3a display defects in posterior crista innervation similar to those in Slitrk6-knockout mice, as well as loss of the inner ear sensory neurons [34][35][36]. The expression of Ntrk neurotrophin receptors is , and incidence of cell death (C, F) in the spiral (A-C) and vestibular (D-F) ganglia. Both the number of neurons and volume of the spiral ganglion of Slitrk6-knockout mice were significantly lower than those of wild-type mice at E16.5, and they decreased to about a half of those of wild-type mice at P0 (A, B). Cell death in the spiral ganglion significantly increased in Slitrk6-deficient mice at E16.5 and P0 (C). Both the number of neurons and the volume of the vestibular ganglion were significantly lower in the knockout than in the wild-type from E13.5, and both were about 75% of those of wild-type mice at E16.5 and P0 (D, E). Cell death in the vestibular ganglion was significantly increased in knockout mice at E13.5 and E16.5 (F). Measurements of Slitrk6deficient vestibular ganglia were made in inner ears without posterior crista innervation. n = 5 to 7 in the spiral ganglion and n = 3 to 5 in the vestibular ganglion. Mean + SD, * p,0.05, ** p,0.01, Student's t-test. H&E-stained spiral ganglia at E16.5 (G, K) and P0 (H, L) and vestibular ganglia at E13.5 (I, M) and P0 (J, N) of Slitrk6 +/+ (G-J) and Slitrk6 2/2 mice (K-N). Pyknotic cells are frequently observed in the spiral ganglion of Slitrk6-deficient mice at E16.5 (arrowheads in G, K). In the vestibular ganglion, the number of pyknotic cells is significantly increased at E13.5 in mutant mice (arrowheads in I, M). At P0, there are no apparent differences between wild-type and knockout mice in cell size and organization in either the spiral or vestibular ganglia (H, J, L, N). Scale bars: 50 mm. doi:10.1371/journal.pone.0007786.g007 decreased in NeuroD1and Brn3a-knockout mice, suggesting that neurotrophin-Ntrk receptor signaling is involved in the innervation defects and neuronal loss in these mutant mice [34][35][36]. Mice lacking ErbB function also show a marked decrease in the number of spiral ganglion neurons and disorganized inner ear innervation [37,38]. ErbB2 and ErbB3 are highly expressed in the supporting cells of the organ of Corti and regulate the expression of neurotrophins, especially Ntf3. However, we could not detect any significant alteration in the expression of NeuroD1, Brn3a, ErbB2, or ErbB3 in the E14.5 cochleae (Fig. S8), in contrast to the reduction of neurotrophin receptors (Fig. 9B, C). Therefore, these genes may not function downstream of Slitrk6 during inner ear development.
Besides the strong expression in the sensory epithelia, Slitrk6 is also expressed in ganglion neurons weakly and transiently. However, isolated spiral ganglia of Slitrk6-knockout mice can extend their neurites almost the same as those of wild-type mice in a tissue culture experiment (Fig. 8A, B, E, Fig. S7). Although we cannot exclude the possibility that Slitrk6 expressed in ganglion neurons play some roles in target recognition or axon pathfinding, we consider that the observed abnormalities of Slitrk6-knockout mice are mainly attributed to the loss of Slitrk6 in sensory epithelia. Interestingly, previous studies have also reported the transient expression of neurotrophins in the inner ear ganglia, similar to the expression of Slitrk6 [22,25]. The neurons that transiently express neurotrophins are cells delaminating from the sensory epithelia [39], and are postulated to have a role in guiding neurites from sensory neurons back to the area they delaminated from [40]. On the basis of its spatial distribution, we believe that Slitrk6 is localized in the delaminating cells, where it could participate in the neuronal organization of the inner ear and provides guidance as shown in the derailed pathfinding of posterior crista projections.
In conclusion, our results indicate that Slitrk6 is involved in the survival and innervation of sensory neurons in the inner ear, at least in part by modulating neurotrophin-Ntrk signaling. The molecular mechanisms involved in auditory and vestibular axon pathfinding are beginning to be unraveled [41,42]. The results presented here provide new insight into the mechanisms of innervation in the developing inner ear. Further investigation of the molecular mechanisms of Slitrk6 function will increase our knowledge of sensory neural development of the inner ear.

Animals
Animal experiments were approved by the Animal Experiment Committee of the RIKEN Brain Science Institute. The mice were maintained by the Laboratory Animal Facility, Research Resource Center, RIKEN Brain Science Institute.

Generation of Slitrk6-Null Mutant Mice
Slitrk6-null mutant mice were generated as described previously [6,43]. Briefly, to construct the Slitrk6 targeting vector, overlapping Slitrk6 genomic clones were isolated from a phage library made from mice of the 129SV strain (Stratagene, La Jolla, CA). The targeting construct contained the 1.8-kb 59 and 5.3-kb 39 homology regions, and the 3.3-kb fragment containing the ORF of Slitrk6 was replaced with a floxed PGK-neo expression cassette (Fig. 3A). E14 ES cells were transformed with the targeting construct by electroporation and selected with G418. Drugresistant clones were screened by Southern blot analysis using PstIand SpeI-digested genomic DNA hybridized, respectively, with a 0.7-kb 59 genomic fragment corresponding to a genomic sequence outside of the sequence used in the targeting vector and a 0.6-kb PstI PGK-neo probe (Fig. 3A). Chimeric mice were generated by Figure 8. Cochlear sensory epithelia of Slitrk6-knockout mice have less activity in promoting neurite outgrowth of spiral ganglion neurons. Spiral ganglia (SG) and sensory epithelia (SE) were dissected from E14.5 wild-type and Slitrk6-deficient mice and the explants were cultured. Neurites were visualized by neurofilament immunostaining (green). There were hardly any detectable processes at the beginning of the explant culture. Spiral ganglia of wild-type (A) and Slitrk6-deficient (B) mice can strongly extend their neurites to sensory epithelia obtained from wildtype mice, whereas sensory epithelia of Slitrk6-knockout mice weakly attract neurites of spiral ganglia obtained from both wild-type (C) and knockout mice (D). Scale bar, 400 mm. (E) Measurement of neurite outgrowth (%) of spiral ganglia. The graphs represent mean + SEM. Neurite outgrowth of spiral ganglia was reduced in the presence of Slitrk6-deficient sensory epithelia compared to the presence of wild-type or heterozygous sensory epithelia. * p,0.05, Mann-Whitney's U-test. doi:10.1371/journal.pone.0007786.g008 Figure 9. Expression of neurotrophins and their receptors was decreased in Slitrk6-deficient inner ear. (A) Real-time PCR analysis of Bdnf, Ntf3, Ntrk2, and Ntrk3 in E14.5 inner ear (includes cochlear and vestibular sensory epithelia, and spiral and vestibular ganglia). The graphs depict mean + SD of 3 independent analyses. The mRNAs of Bdnf and Ntf3 were significantly decreased in Slitrk6-deficient inner ear, whereas those of Ntrk2 and Ntrk3 were not. * p,0.05, Student's t-test. (B) Western blot of protein samples extracted from E14.5 cochlea (includes cochlear sensory epithelium and spiral ganglion). (C) Both Ntrk2 and Ntrk3 proteins in the cochlea were significantly decreased in Slitrk6-knockout mice. Furthermore, phosphorylated form of Ntrk (p-Ntrk) protein was decreased in the knockout mice. Amounts of a neuronal marker (bIII tubulin) and a sensory epithelium marker (Myosin VIIa) were comparable between the wild-type and knockout mice. The graphs represent mean + SD of 3 independent analyses. * p,0.05, Student's t-test. Immunohistochemistry for Ntrk2 (green; D, G), Ntrk3 (green; J, M), p-Ntrk (red; P, S), and bIII tubulin (red; E, H, K, N and green; Q, T) in mid-turn cochlea of Slitrk6 +/+ (D-F, J-L, P-R) and the injection of targeted ES cells into C57BL/6J blastocysts. To excise the PGK-neo cassette, mice with germline transmission of the transgene were first mated with CAG-Cre mouse. The correct excision of the PGK-neo cassette was confirmed by Southern blot analysis (Fig. 3B). The mice carrying the mutated Slitrk6 allele were backcrossed to C57BL/6J for more than 4 generations before analysis. Genotyping of progeny was performed by Southern blot or PCR analysis of genomic DNA extracted from tails. The PCR primers used were Slitrk6S (59-CAGGAAGCATTCTCTTC-TCTGTGA-39), Slitrk6WTAS (59-CATTGGTGACTGGGACT-GTAGA-39), and Slitrk6KOAS (59-GGAAGCTCAGAACACT-TACCTG-39).

Generation of an Anti-Slitrk6 Antibody
A polyclonal anti-Slitrk6 antibody was raised in a rabbit against peptides corresponding to the carboxy-terminal region of mouse Slitrk6 (KNEYFELKANLHAEPDYLEVLEQQT). Peptides were synthesized and conjugated to keyhole limpet hemocyanin through a cysteine added to the N-terminus of the peptide. After immunization by conventional methods, antiserum was obtained and the antibody was purified by affinity chromatography with the immunized peptide. The antibody specificity was confirmed by the absence of Slitrk6 immunopositive signals in the Slitrk6-deficient inner ear (Fig. S1, Fig. 9B).

Morphological Analyses
In situ hybridization analysis for Slitrk6 was performed on paraformaldehyde-fixed paraffin sections obtained from wild-type mice as described previously [1,7].
Immunohistochemistry was performed on paraffin or frozen sections and whole mounts. Rabbit polyclonal anti-Slitrk6 and anti-phospho-Ntrk (phosphorylated Tyr490 in NTRK1 and its corresponding residues in Ntrk2 and Ntrk3, Cell Signaling Technology, Beverly, MA), goat polyclonal anti-Ntrk2 (R&D Systems, Minneapolis, MN) and anti-Ntrk3 (R&D Systems), and mouse monoclonal anti-neurofilament (NF, Sigma, St. Louis, MO) and anti-bIII tubulin (Promega, Madison, WI) antibodies were used as primary antibodies. Alexa488 or 594-conjugated antirabbit, -goat, and -mouse IgG (Molecular Probes, Eugene, OR) were used as the secondary antibodies. To visualize inner and outer hair cells, whole mounts were stained with phalloidintetramethylrhodamine B isothiocyanate (TRITC) (Sigma). For the visualization of myelinated nerve fibers, relatively mature inner ears were impregnated with osmium tetroxide.
To reveal the general pattern of innervation, nerve fibers were visualized with a lipophilic tracer. DiI (1,19-dioctadecyl-3,3,39,39tetramethylindocarbocyanine perchlorate; Molecular Probes) crystals were placed into the brainstem alar plate to label eighth nerve afferent fibers to the inner ear, or into the olivo-cochlear efferent bundle near the floor plate to label the efferent fibers [13,14]. After an appropriate diffusion time of 3 to 5 days at 37uC, whole inner ears or cochleae were then dissected out and viewed as wholemount surface preparations. All the specimens mentioned above were examined with an Axioscop 2 plus (Carl Zeiss, Göttingen, Germany), MZ 16 FA (Leica Microsystems, Wetzlar, Germany), or FV1000 confocal laser microscope (Olympus, Tokyo, Japan).
For morphometrical analyses, inner ears from wild-type and Slitrk6-deficient mice were fixed in paraformaldehyde, embedded in paraffin, and serially sectioned at 6 mm. Every fourth section was stained with H&E. Comparisons of neuronal numbers in the spiral and vestibular ganglia were made by counting the number of neurons with clear nucleoli. No correction was made for split nucleoli. Dying neuronal cells that could be identified by pyknotic nuclei were counted in the same way. The volume of each ganglion was calculated by measuring the area occupied by neurons in each section using AxioVision software (Carl Zeiss).

Explant Culture of Cochlear Sensory Epithelium and Spiral Ganglion
Sensory epithelium and spiral ganglion were dissected from E14.5 cochlea. A cochlear sensory epithelium and a spiral ganglion were randomly chosen and placed in three-dimensional collagen gel (Cell Matrix; Nitta Gelatin, Osaka, Japan) about 500 mm apart each other. Then they were cultured in neurobasal medium (Invitrogen, Carlsbad, CA) containing 2 mM L-glutamine, B27 supplement (Invitrogen) for 48 hours. Neurites were visualized by the immunostaining for neurofilament (Sigma). Neurite outgrowth was analyzed with ImageJ (http://rsb.info.nih.gov/ij/). Area occupied by the neurofilament-positive neurites were measured and relative to the area of the ganglion body (%) is presented.

RNA Extraction and Real-Time RT-PCR Analyses
Inner ears from 5 to 6 E14.5 wild-type and Slitrk6-deficient mice were pooled, and total RNA was extracted with an RNeasy Protect Kit (QIAGEN, Hilden, Germany). Reverse transcription was performed with Superscript II reverse transcriptase (Invitrogen). Real-time PCR was performed with SYBR Green Real-time PCR Master Mix (Applied Biosystems, Foster City, CA) on an ABI Prism 7900HT (Applied Biosystems). The primers used for detection of Bdnf and Ntf3 were designed by Stankovic and Corfas [44], and those for Ntrk2 and Ntrk3 were as follows: Ntrk2, sense: 59-AGCCTTCTCCAGGCATCGT-39, antisense: 59-TCCGGGT-CAACGCTGTTAG-39, and Ntrk3, sense: 59-TGTAGTTT-CTGGCGGATTTTCTT-39, antisense: 59-AGCACGGAGCC-CACATAGTC-39. The intensity relative to G3PDH was calculated, and the fold change relative to the relative intensity in wildtype mice is presented.

Supporting Information
Table S1 Comparison of the phenotypes of Slitrk6-knockout mice and those of neurotrophin/Ntrk knockout mice.  A, B). In both the wild-type and Slitrk6deficient mice, one row of inner hair cells and three rows of outer hair cells can be identified in all turns, and the differentiation of these cells in the mutant animals appears normal (A', B'). The organization of the anterior crista and utricle also appears normal in Slitrk6-deficient mice (C, D). Higher magnification views of the anterior crista (C', D') and utricle (C', D') are not clearly different between wild-type and knockout mice. Scale bars, 50 mm. Sizes of sensory epithelia in wild-type (+/+) and Slitrk6-deficient (2/2) at E13.5 (E; +/+, N = 5; 2/2, N = 5) and E16.5 (F; +/+, N = 7; 2/ 2, N = 7). The sizes were measured in every fourth of serially prepared H&E-stained sections (thickness: 6 mm) that include the entire sensory epithelia. The sizes are indicated as the summed lumenal surface lengths of the sensory epithelia in the serial sections. A common arbitrary unit is used between (E) and (F). * p,0.05, ** p,0.01, Student's t-test. (G) Densities of the hair cells in the vestibular sensory epithelia at E16.5. The cell densities were measured by counting Myosin VIIa immuno-stained cells per unitary length (1 mm) in the serial sections. There were no significant differences in the hair cell densities between wild-type and Slitrk6-deficient mice. AC, anterior crista; Co, cochlea; HC, horizontal crista; IHC, inner hair cell; OHC, outer hair cell; PC, posterior crista; S, saccular macula; SG, spiral ganglion; U, utricle.  [21,26,27]. In the cochlea, Ntf3and Ntrk3-null mutants predominantly lose spiral ganglion neurons in the basal turn (D), and neuronal loss in Bdnfand Ntrk2-null mutants is most obvious in the apex (C) [16,21,28,29]. In the Slitrk6-knockout mice, neuronal loss was more pronounced in the cochlea than in the vestibule, but Slitrk6deficient cochleae showed no phenotypic gradient in neuronal loss along the baso-apical axis (B). AC, anterior crista; HC, horizontal crista; PC, posterior crista; S, saccule; SG, spiral ganglion; U, utricle; VG, vestibular ganglion. Found at: doi:10.1371/journal.pone.0007786.s007 (0.69 MB PDF) Figure S7 Spiral ganglion neurons of Slitrk6-deficient mice can extend neurites identical to those of wild-type mice in the presence of neurotrophin. Spiral ganglia from E14.5 wild-type (+/+) and Slitrk6-knockout (2/2) mice were embedded in the collagen gel (Nitta Gelatin) and cultured in neurobasal medium (Invitrogen) containing 2 mM L-glutamine, B27 supplement (Invitrogen) and 20 ng/ml of Neurotrophin-3 (+Ntf3, PeproTech EC, London, UK) for 48 hours. Neurites were visualized by neurofilament immunostaining (Sigma, green). Spiral ganglion neurons of Slitrk6knockout mice (B) can extend their neurites as those of wild-type mice (A). Scale bar, 200 mm. (C) Measurement of neurite outgrowth (%) of spiral ganglia. The graphs represent mean + SEM. Neurite outgrowth of spiral ganglia was not significantly different between wild-type and Slitrk6-deficient mice. Found at: doi:10.1371/journal.pone.0007786.s008 (0.69 MB PDF) Figure S8 Amounts of proteins that are known to mediate inner ear sensory neural development in Slitrk6-knockout mice. (A) Western blot analysis of proteins extracted from E14.5 cochlea (includes cochlear sensory epithelium and spiral ganglion). Goat polyclonal anti-NeuroD1 (Santa Cruz Biotechnology), mouse monoclonal anti-Brn3a (Santa Cruz Biotechnology) and anti-ErbB3 (Thermo Fisher Scientific, Fremont, CA), and rabbit polyclonal anti-ErbB2 (Thermo Fisher Scientific) and anti-b-actin (Sigma) antibodies were used as primary antibodies. (B) The graphs represent the mean + SD of 3 independent analyses. There were no significant differences in the expression of NeuroD1, Brn3a, ErbB2, or ErbB3 between wild-type (+/+) and Slitrk6 knockout (2/2) mice. Found at: doi:10.1371/journal.pone.0007786.s009 (0.69 MB PDF)