|Home | About | Journals | Submit | Contact Us | Français|
ErbB2 protein is essential for the development of Schwann cells and for the normal fiber growth and myelin formation of peripheral nerves. We have investigated the fate of the otocyst-derived inner ear sensory neurons in the absence of ErbB2 using ErbB2 null mutants. Afferent innervation of the ear sensory epithelia shows numerous fibers overshooting the organ of Corti, followed by a reduction of those fibers in near term embryos. This suggests that mature Schwann cells do not play a role in targeting or maintaining the inner ear innervation. Comparable to the overshooting of nerve fibers, sensory neurons migrate beyond their normal locations into unusual positions in the modiolus. They may miss a stop signal provided by the Schwann cells that are absent as revealed with detailed histology. Reduction of overshooting afferents may be enhanced by a reduction of the neurotrophin Ntf3 transcript to about 25% of wild type. Ntf3 transcript reductions are comparable to an adult model that uses a dominant negative form of ErbB4 expressed in the supporting cells and Schwann cells of the organ of Corti. ErbB2 null mice retain afferents to inner hair cells possibly because of the prominent expression of the neurotrophin Bdnf in developing hair cells. Despite the normal presence of Bdnf transcript, afferent fibers are disoriented near the organ of Corti. Efferent fibers do not form an intraganglionic spiral bundle in the absence of spiral ganglia and appear reduced and disorganized. This suggests that either ErbB2 mediated alterations in sensory neurons or the absence of Schwann cells affects efferent fiber growth to the organ of Corti.
v-erb-b2 erythroblastic leukemia viral oncogene homolog 2 (ErbB2) belongs to a family of 4 ligand activated tyrosine kinase receptors (Schlessinger, 2004; Wang et al., 2004). Among ErbB receptors, ErbB2 is unique in being an orphan receptor that must heterodimerize with other receptors to signal. Beyond its clinically important function in cancer (Bianco, 2004; Lemmon, 2003), ErbBs and their ligands, the neuregulins (NRGs) are important in various aspects of survival, proliferation and differentiation of the nervous system. For peripheral nerve development and maintenance, the ligand NRG1 and its receptors, ErbB2 and ErbB3, play key roles in developing neuron-Schwann cell interaction (Adlkofer and Lai, 2000; Michailov et al., 2004). Reduced formation of peripheral nerves, in particular of cranial nerves, has been reported in NRG1 null mutants (Meyer and Birchmeier, 1995; Meyer et al., 1997) and in ErbB2 null mice (Lee et al., 1995; Morris et al., 1999). NRG1 [Heregulin (HRG) or acetylcholine receptor-inducing activity (ARIA)] is expressed in sensory neurons and motoneurons (Bermingham-McDonogh et al., 1997; Michailov et al., 2004), whereas ErbB2 and ErbB3 are expressed in Schwann cells surrounding these fibers in the peripheral nervous system and regulate myelin formation (Corfas et al., 2004; Michailov et al., 2004) and survival of developing Schwann cells (Atanasoski et al., 2006; Lyons et al., 2005; Massa et al., 2005; Young et al., 2005). A dominant negative over-expressor of ErbB4 using the GFP-promoter that drives ErbB4 expression in juveniles (Rio et al., 2002) causes progressive sensory loss in adult mice (Chen et al., 2003), presumably acting as a sink for the limited amount of NRG1 and interfering with ErbB2- and ErbB3-mediated signaling (Corfas et al., 2004). While the functional significance of neuregulin ligands and their receptors for nerve–Schwann cell interaction and Schwann cell survival is thus well established in the peripheral nervous system, specific effects of these receptors and ligands as well as their distribution in the developing and adult ear are less clear.
In situ hybridization data suggest the presence of NRG1 in spiral and vestibular sensory neurons (Morley, 1998). However, the NRG1 gene encodes over 15 isoforms of which only some share the EGF-domain (Falls, 2003). Exactly which of these 15 isoforms is expressed in the sensory neurons requires further analysis. Immunocytochemical and in situ studies suggest the presence of ErbB1, 2, 3 and 4 in the adult inner ear with a somewhat differential distribution in hair cells and supporting cells that may vary with the species studied (Hume et al., 2003; Matsunaga et al., 2001; Stankovic et al., 2004; Zhang et al., 2002). Involvement of these ligands and their receptors in proliferation of neonatal vestibular hair cells and supporting cells has been suggested (Hume et al., 2003; Montcouquiol and Corwin, 2001), but no direct test involving ErbB2 or ErbB3 knock down in vitro or in vivo has been performed. In agreement with data from other peripheral nerves, a reciprocal signal between spiral neurons and Schwann cells has been suggested (Hansen et al., 2001), which may relate to the development of acoustic Schwannoma (Hansen et al., 2004). In addition, loss of NRG1 interaction with ErbB2 and ErbB3 has been related to adult sensory neuron death in dominant negative ErbB4 overexpressing transgenic mice (Stankovic et al., 2004), suggesting effects in the ear that closely parallel those in other parts of the peripheral nervous system (Chen et al., 2003; Corfas et al., 2004).
We used ErbB2 null mutant mice to test whether inner ear innervation depends on Schwann cells. We also wanted to test whether supporting cells have molecular similarities to glial cells (Rio et al., 2002) using the proteolipid protein-enhanced green fluorescent protein (PLP-eGFP) mouse (Mallon et al., 2002).
Previous investigations have shown that the absence of neurotrophins will result in innervation alteration of both afferents (sensory neuron fibers) and efferents (modified facial branchial motoneuron fibers; Karis et al., 2001) from this age onwards (Fritzsch et al., 2004), a stage when all sensory neurons are postmitotic (Matei et al., 2005; Ruben et al., 1967).
We started our investigation of the effects of ErbB2 abrogation at E13.5 embryos by injecting lipophilic tracers into the brainstem second order auditory (cochlear) and vestibular nuclei. Such injections resulted in labeled fibers extending to all six sensory epithelia of the ear in wild type and ErbB2 null mice (Fig. 1A, B). Closer examination of the vestibular projection showed a somewhat less precise targeting as more fibers seemed to diverge away from the vestibular epithelia. Comparison of the cochlear projection indicated two differences: (1) spiral ganglia were not found close to the basal turn of the organ of Corti in ErbB2 null mice, and (2) spiraling fibers, present in wild type littermates, were absent in ErbB2 mutants (Fig. 1A, B). Interestingly, the cochlear projecting (spiral) neurons were in the center of the cochlear modiolus and not near the organ of Corti.
From this stage onward, we could label selectively the afferent fibers to the cochlea and vestibular sensory epithelia with injections into the cochlea and vestibular nuclei (labeled in red in all figures). Simultaneous injections into the medial vestibular nucleus near the internal facial genu labeled efferent fibers to the cochlea and vestibular epithelia as well as vestibular afferents with dye (labeled green in Fig. 1). These injections revealed the following differences between ErbB2 null mutants and wild type littermates (Fig. 1C–F).
Wild type littermates showed only afferent and efferent fibers in the modiolus and an assembly of spiral ganglion neurons in a spiral ganglion that paralleled the organ of Corti (Fig. 1D). Only short radial fibers emanated from the spiral ganglion to the organ of Corti (arrow in Fig. 1D). In contrast, spiral ganglion cells were not found in the position of the spiral ganglion in erbB2 null mice but apparently filled the much enlarged center of the modiolus (Fig. 1C). Long radial fibers emanated from these sensory neurons located in the modiolus. Despite the altered position of the spiral neurons and the increased length of radial fibers in ErbB2 null mice, numerous afferent fibers extended to the organ of Corti (Fig. 1C, D).
Analysis of efferent fibers (Fig. 1E, F) also showed differences. Wild type efferent fibers extended along the spiral ganglion cells and started to form the intraganglionic spiral bundle within which single efferent fibers spiral from the base of the organ of Corti toward the apex (Fig. 1F). In contrast, no intraganglionic spiral bundle was found in ErbB2 null mice (Fig. 1E). Instead efferent fibers extended along and in parallel to the elongated afferent radial fibers from the modiolus to the organ of Corti (Fig. 1C, E).
E16.5 ErbB2 null mice showed an abundance of fibers to the cochlea and beyond (Fig. 2B). Indeed, ErbB2 null mice had more fibers reaching the level of the organ of Corti than 2 day older wild type animals in which we found only a limited growth of fibers to the outer hair cells (Fig. 2A). Focusing through the confocal stacks of images at different focal planes showed that the vast majority of fibers did not actually enter the organ of Corti but stayed below the basilar membrane. These fibers could be traced to the lateral wall were they coalesced into large bundles that spiraled toward the apex (LW in Fig. 2B). These data show that afferent fiber growth is more profound to the endorgans in ErbB2 null mice than in littermates (Fig. 2A, B). However, the normal pattern of innervation is disrupted and fibers are extending outside the organ of Corti.
All six sensory epithelia were found to be innervated by both afferent and efferent nerves at E18.5 in wild type and ErbB2 null mice. Although these epithelia were innervated, the pattern of innervation was altered, with the cochlea epithelia having the most pronounced alteration in the pattern of innervation. As already observed at E14. 5, the spiral ganglion was no longer present in its usual position (Fig. 3B, C), but resided instead in an expanded central part of the modiolus in ErbB2 null mice (Fig. 3A, E). As a consequence of the unusual position of these neurons, much elongated radial fibers extended from the modiolus to the organ of Corti in ErbB2 null mice (Fig. 3A, E) compared to short radial fibers from the prominent spiral ganglion to the organ of Corti in wild type mice (Fig. 3B, F). Closer examination at a higher magnification showed that the radial fibers in the ErbB2 null mice were less organized as demonstrated by several fibers crossing between radial bundles (Fig. 4A, B). Overall, the ErbB2 null mutants showed an apparent reduction of radial fiber density to the organ of Corti (Fig. 3A, B, E, F). These data suggest that in ErbB2 null mice the spiral ganglion neurons have migrated beyond their normal location of the spiral ganglion near the osseous spiral lamina and are now in the center of the modiolus.
In the inner ear of the wild type E18.5 embryos, the efferent fibers turn into the intraganglionic spiral bundle (IGSB) inside the spiral ganglion (Figs. (Figs.3D3D and and4D).4D). The IGSB gives rise to the densely spaced, short radial fiber bundles (Fig. 4D). In the ErbB2 null embryos, fibers project as disorganized radial bundles of efferent fibers directly to the organ of Corti without formation of an IGSB (Figs. (Figs.3C3C and and4C).4C). This disorganization may be due to the abnormal position of the spiral ganglion which is normally present closer to the organ of Corti (Fig. 4A, B). Additionally, fewer and more disorganized efferent fibers can be labeled in the ErbB2 null mutant embryos compared to the control fibers (Figs. 3C, D and 4C, D). Previous studies demonstrated that the efferent fibers rely on the afferent fibers for guidance to the organ of Corti (Farinas et al., 2001; Tessarollo et al., 2004). Therefore, it appears that the improper migration of the spiral ganglion and aberrant afferent projections results in disorganization in the efferent fibers (Fig. 4A–D) comparable to the afferent fibers.
Afferent fibers interact with surrounding Schwann cells to regulate myelin formation and certain aspects of pathfinding (Corfas et al., 2004; Michailov et al., 2004; Morris et al., 1999). Afferent fiber bundles to the sensory epithelia are surrounded by Schwann cells, which start to form myelin sheaths in the mouse ear as early as E18.5 (Fig. 5). Previous work using PLP in situ hybridization suggested that the adult Schwann cells of the ear are unique in that they do not express PLP except for the ganglion close to the proximal VIIIth nerve root (Knipper et al., 1998). We used a recently available PLP-eGFP expressing mouse line (Mallon et al., 2002) in which eGFP is expressed under the control of the mouse PLP promoter and contain an additional mRNA stability region that is found in the 3′ upstream region of the endogenous mouse PLP gene. In these PLP-eGFP transgenic mice, all Schwann cells of the ear were strongly PLP-eGFP positive, except for a small population of cells near the VIIIth nerve root, which were negative (Fig. 5A–C). All sensory epithelia contained eGFP positive cells (Fig. 4D–F) in wild type embryos. Specifically, in the cochlea, only supporting cells that surrounded the inner and outer hair cells were positive for eGFP. However, some of these supporting cells were not positive after fixation and a patchy distribution with various intensities of label or even apparent absence at E18.5 was observed (Fig. 5D). Lipophilic tracers were used to label the nerve fibers (Fig. 5B, E). These data demonstrated that Schwann cells were found around all vestibular and spiral ganglion neurons as well as along all nerve fibers (Fig. 5C, F). Schwann cell process accompanied nerve fibers to the habenula perforate, the entering holes into the organ of Corti (Fig. 5D–F, inserts). In the organ of Corti, fibers extended between eGFP positive supporting cells surrounding inner and outer hair cells (Fig. 5E, F). These data demonstrated that Schwann cells and eGFP positive supporting cells of the sensory epithelia surround the afferent and efferent fibers at E18.5 or earlier. Such cells can provide a cellular basis for afferent and efferent fibers to grow along and be supported by trophic factors in addition to those known to be released from hair cells (Farinas et al., 2001).
In mice null for ErbB2 and ErbB3, Schwann cells are absent in the peripheral nervous system. This absence of Schwann cells suggests an important role for these receptors in Schwann cell development (Morris et al., 1999) in addition to maturation (Corfas et al., 2004; Lyons et al., 2005; Michailov et al., 2004) but not in adults (Atanasoski et al., 2006). These results are consistent with our data as we could not detect PLP-eGFP cells along the peripheral nerves of the inner ear of ErbB2 null embryos. The only eGFP positive cells identified were scattered in the center of the spiral modiolus (Fig. 6A, C) and the supporting cells of the sensory epithelia. Nevertheless, nerve fibers reached all sensory epithelia and clearly reached the hair cells in both the cochlea and the vestibular epithelia, albeit in a disorganized manner (Fig. 6A, C).
Comparison between wild type and ErbB2 null embryos demonstrated fewer fibers in the cochlea and vestibular sensory epithelia of ErbB2 null mice, many of which with aberrant projections (Figs. (Figs.66 and and7).7). For example, many fibers did not enter the cochlea but stayed on the scala tympani side of the basilar membrane (Fig. 7A, arrow). Other fibers extended beyond the hair cell region of the cochlea and reached the lateral wall which also displayed some PLP-eGFP positive cells (Fig. 6A). Overall, fibers did not form a regular lattice of fiber projections to the cochlea in distinct radial bundles but rather showed fibers which crossed paths, particularly in the region adjacent to the organ of Corti (Fig. 7A) where all fibers enter the cochlea in ErbB2 wild type mice (Fig. 7B). In addition, ErbB2 null mice had expression of PLP-eGFP in almost all supporting cells of the cochlear and vestibular region, with particularly high levels of eGFP expression in the border cells of the organ of Corti, leaving very few cells unlabeled (Fig. 7C, D).
Given the disorganization of afferent fibers in the ear, we wanted to know whether central projections of vestibular and cochlear fibers are at least grossly normal. NeuroVue dye tracing revealed grossly normal projections to cochlear and vestibular nuclei (Fig. 6E, F). However, the labeled fiber density of central projections was markedly reduced compared to wild type animals, proportional to the apparent reduction at the periphery. To analyze further these effects and separate ear-specific form, brain-specific effects will require generation of ear-specific ErbB2 conditional null mice.
We next examined the histology of the organ of Corti to verify that development of the sensory epithelia is grossly normal and that Schwann cells are indeed absent as suggested by the virtually complete lack of PLP-eGFP positive cells along the nerve fibers of ErbB2 null mice (Fig. 7A, B). Epoxy resin sections showed that the spiral ganglion cells were indeed further away from the organ of Corti in the ErbB2 null mice (Fig. 8A, B). Higher magnification of the spiral ganglion showed both neurons and Schwann cells in the E16.5 wild type (Fig. 8D) but only neurons in the ErbB2 null mice (Fig. 8C, E). Moreover, large empty spaces were present in the spiral ganglion of ErbB2 null mice and many apoptotic cells were distributed throughout the ganglion (Fig. 8C, E). Whether these apoptotic cells are neurons or Schwann cells remains unknown.
Previous work has shown that inner ear innervation is critically dependent in the embryo on two neurotrophins, Bdnf and Ntf3 (Fritzsch et al., 2004). Of these two neurotrophins, Ntf3 is predominantly expressed in E13–18 embryos in supporting cells of the cochlea and some vestibular organs (utricle and saccule; Farinas et al., 2001). In contrast, Bdnf is predominantly expressed in hair cells in E13–18 embryos with only a limited expression in supporting cells, particularly near the apex of the cochlea (Fritzsch et al., 2005a). Recent experiments have shown that over-expression of a dominant negative ErbB4 in juvenile supporting cells can reduce Ntf3 to the extent that spiral neurons die in adult mice whereas Bdnf is only mildly affected (Stankovic et al., 2004). In order to determine if Bdnf or Ntf3 expression was altered in the ErbB2 null embryos, the presence of Bdnf and Ntf3 in the ear of wild type and ErbB2 null embryos at E18.5 was determined using RT-PCR analysis. No significant reduction of Bdnf mRNA was observed, while Ntf3 was significantly reduced to 25% compared to wild type embryos (Fig. 9). Overall, these effects are similar to observations in adult mice overexpressing a dominant-negative ErbB4 (Stankovic et al., 2004) suggesting that similar molecular changes occurred at the level of the supporting cells of the inner ear sensory epithelia in ErbB2 null embryos.
In summary, our data demonstrate that ErbB2 null mutants have aberrant projections of afferents, and aberrant migration of spiral neurons; both effects are possibly related to the absence of Schwann cells that may provide a stop and guidance signal. ErbB2 null mice display a reduction and disorganization of both afferent and efferent fibers in the much longer trajectory from the center part of the modiolus to the organ of Corti. In contrast, wild type mice have much shorter and straight trajectories of afferent and efferent fibers coming off the spiral ganglion and IGSB, respectively, that are adjacent and parallel to the organ of Corti. However, despite the absence of Schwann cells, there is an obvious innervation of all sensory epithelia in the ErbB2 null mice.
The inner ear sensory neurons of the vestibular and spiral ganglion develop from the inner ear and require the bHLH gene Neurogenin1 for their formation (Ma et al., 2000; Rubel and Fritzsch, 2002) and the bHLH gene Neurod1 for proper migration and maturation (Kim et al., 2001) comparable to the neural crest-derived cranial ganglia. However, despite this molecular similarity in early development, the inner ear sensory neurons differ from all other cranial neurons in that they send a dendrite back to their area of delamination, the ear (Pauley et al., 2005). Inner ear sensory neurons are also unique in that they express two neurotrophin receptors, trkB and trkC (Farinas et al., 2001), whereas other peripheral sensory neurons require only a single trk receptor (Huang and Reichardt, 2001). Here we demonstrate a role for the ErbB2 gene in proper migration and pathfinding of spiral sensory neurons and show that Schwann cells are not necessary for initial targeting of afferent fibers, although efferent fibers are reduced and disorganized in ErbB2 null mice. Inner ear sensory neurons differ in this respect from other cranial ganglion neurons (Morris et al., 1999).
Schwann cells in the PNS start myelin formation around birth. Our data show that these cells express the PLP-eGFP transgene at this stage. In contrast, myelin in the CNS does not express PLP prior to 10 days postnatal in the rat (Milner et al., 1985), consistent with our observation that no PLP-eGFP positive cells are found in the proximal part of the vestibular or cochlear nerve in embryos (Fig. 5). This seemingly contrasts with a report that shows strong expression of PLP only in the proximal part of the vestibular–cochlear nerve in adult rats (Knipper et al., 1998). However, it is possible that PLP expression is both spatially and temporally dynamic and the apparent differences in distribution reported here as compared to the previous study (Knipper et al., 1998) may reflect differential changes of PLP expression in postnatal mice rather than real expression differences.
We show histologically as well as by the absence of PLP-eGFP positive cells that all Schwann cells are absent in ErbB2 null ears, consistent with previous reports on other peripheral nerves (Lyons et al., 2005; Massa et al., 2005; Morris et al., 1999). Interestingly, PLP expression in the inner ear sensory epithelia supporting cells is not only present but appears to be upregulated (Figs. (Figs.66 and and7).7). Most importantly, whereas only some supporting cells of the organ of Corti of ErbB2 wild type littermates show expression of PLP-eGFP transgene, all cells are strongly positive for PLP-eGFP in ErbB2 null littermates (Figs. (Figs.66 and and7).7). These data suggest that absence of ErbB2 alters the maturation of these cells with respect to PLP expression. Consistent with these effects of ErbB2 on gene regulation in supporting cells, our data on Ntf3 expression suggest that expression of this neurotrophin, known to be expressed in supporting cells of the ear in embryos (Farinas et al., 2001), is affected (Fig. 9). Despite all these cellular and molecular deficiencies, afferent and efferent fibers nevertheless reach the sensory epithelia in large numbers. However, there is an apparently transient massive overshooting of afferents fibers (Fig. 2) that may be corrected by the concurrent wave of apoptosis in the ganglia (Fig. 8).
During ear development, spiral and vestibular neurons delaminate from the ear and migrate to the vestibular and spiral ganglia, respectively, and project back with their dendrites (Fritzsch, 2003; Pauley et al., 2005; Rubel and Fritzsch, 2002), in part via attraction to hair cells and the neurotrophin they release (Fritzsch et al., 2004; Tessarollo et al., 2004). The spiral ganglia migrate to the peripheral part of the modiolus inside the ear whereas the vestibular ganglia migrate outside the ear to form the vestibular ganglion between the ear and brainstem. Vestibular sensory neuron migration depends in part on Neurod1 (Kim et al., 2001). Previous work has already indicated that ErbB receptors and their ligands are involved in neuronal migration (Rio et al., 1997) and migration of breast tumor cells (Bianco, 2004; Lemmon, 2003) and our present data support this idea for the ear spiral neurons. In the absence of ErbB2-mediated signaling, spiral neurons migrate further into the center of the modiolus and toward the vestibular ganglion, behaving much like vestibular ganglion sensory neurons. This phenotype implies that some interaction with the substrate, possibly mediated by the neural crest-derived Schwann cells, provides a stop signal for spiral neurons to stay at the periphery of the modiolus in Rosenthal’s canal. In the absence of such a stop signal, spiral neurons continue their migration deeper into the center of the modiolus. The formation of longer radial fibers emanating from the modiolar center in ErbB2 null mice (Figs. (Figs.33 and and4)4) is a consequence of the altered position of spiral ganglion neurons in these mutants. Conditional ear-specific ErbB2 null mutations are needed to analyze whether the additional migration is mediated by signals intrinsic to the neurons or is mediated by the neural crest-derived Schwann cells. Ear-specific conditional ErbB2 null mice should not affect Schwann cell development and thus could help narrow down the cellular basis for this effect.
Our data suggest that a profound fiber outgrowth that over-shoots the organ of Corti in ErbB2 mutants (Fig. 2) is followed by a near complete loss of these overshooting fibers (Fig. 7), possibly through apoptosis (Fig. 8). But even at E18.5 some disorganized afferent fibers that overshoot the organ of Corti remain (Figs. (Figs.6A6A and and7A).7A). Other studies have found unusual afferent fibers that pass below the organ of Corti when Ntf3 is replaced by Bdnf (Tessarollo et al., 2004). As in these transgenic mice, we found that the fibers that extended beyond the organ of Corti and toward the striola had failed to enter the organ of Corti through the habenula perforata. The habenula perforata is near the border cells that are strongly positive for PLP-eGFP in ErbB2 wild type and null littermates (Figs. 6A, B and 7A, B). It is possible that the overall reduction in Ntf3 expression (Fig. 9) as well as the likely specific reduction of Ntf3 in these cells does not allow proper navigation of spiral neuron processes into the organ of Corti.
Recent work has shown that expression of a dominant negative form of ErbB4 under the control of a GFAP promoter reduces Ntf3 expression to about 25% of the control values and causes sever loss of approximately 80% of spiral ganglion neurons in juvenile mice (Stankovic et al., 2004). GFAP is expressed in the supporting cells of the ear (Rio et al., 2002), the same cells in which we report here PLP-eGFP expression (Fig. 7). Interestingly, we find a comparable reduction in Ntf3 expression levels to 25% and also show an apparent reduction in fiber density, likely related to spiral neuron loss through apoptosis (Fig. 8). Owing to the unusual position of spiral ganglion neurons, we need to label them distinctly from vestibular neurons using a marker such as Gata3 (Karis et al., 2001; Lawoko-Kerali et al., 2002), to count them independent of vestibular ganglion neurons. In addition, quantification requires newborn mice as we find numerous apoptotic cells in the spiral and vestibular ganglion of ErbB2 null mice. This suggests that cell death may continue in newborn mice, possibly eliminating all overshooting fibers.
The overall reduction in fiber density and disorganization of afferents is also obvious in all vestibular epithelia (Fig. 7). However, no specific losses of fibers to distinct epithelia are apparent, thus indicating that the phenotype is different from a Bdnf null phenotype and more closely resembles the Ntf3 null phenotype. Closer examination of inputs from the cerebellum in postnatal animals (Maklad and Fritzsch, 2003) is needed to clarify in more detail the degree of disorganization. We are currently raising conditional mutants with a targeted deletion of ErbB2 in the ear to conduct such studies.
Previous work has demonstrated that efferent fibers largely follow afferents to their sensory epithelia no matter what the molecular basis of the deficit (Farinas et al., 2001; Fritzsch et al., 2004). We show here that efferent fibers do indeed home toward the sensory epithelia in ErbB2 null mice but do so in apparently reduced numbers (Figs. (Figs.1,1, ,3,3, and and4).4). Most important is the clear absence of the intraganglionic spiral bundle, a well known feature of efferent trajectories to the cochlea (Figs. (Figs.1,1, ,3,3, and and4).4). Nevertheless, fibers reach the organ of Corti, albeit disorganized and reduced. These data suggest that the absence of Schwann cells affects the fiber growth of efferents as much as that of afferents. We suggest that the reduction of Ntf3 levels to only 25% of wild type alters afferents and reduces their ability to act as a substrate for efferents to grow along. However, it cannot be ruled that low levels of expression of neurotrophin receptors exist in efferent neurons thus mediating a more direct effect of the reported Ntf3 reduction. Targeted deletion of erbB2 only in the ear, maintaining the neural crest-derived Schwann cells, is on its way to clarify this issue.
Breeding and genotyping of ErbB2 null mice and PLP-eGFP mice followed previously established procedures (Mallon et al., 2002; Morris et al., 1999). ErbB2 heterozygous mice carrying the gene to rescue the heart phenotype were crossed with mice carrying the PLP-eGFP reporter gene to allow direct visualization of the effects of ErbB2 absence on the degree of differentiation of Schwann cells and other cells that express PLP. Timed mating was used to determine the date of conception by the appearance of a vaginal plug as embryonic day 0.5 (E0.5). Embryos were collected from the mother at embryonic days 13.5, 14.5, E16.5 or 18.5. E13.5, E14.5 and E16.5 embryos were fixed by immersion in 4% paraformaldehyde in 0.1 M Sorenson’s buffer (pH 7.4). E18.5 embryos were fixed using perfusion with the same fixative. A section of the tail was obtained for PCR analysis of genomic DNA to establish genotype using primers previously described (Morris et al., 1999).
Ears from E13.5 (N = 8), E14.5 (N = 6), E16.5 (N = 6) and 18.5 (N = 12) ErbB2 mutants and a similar number of wild type and heterozygous littermates were successfully analyzed using lipophilic tracers in fixed tissue. Briefly, different colored NeuroVue tracers (Fritzsch et al., 2005b) were injected to directly image the inner ear afferents and efferents as well as expression of eGFP under control of the PLP promoter. Fixed heads were bisected. Filter strips with dye were inserted into the bisected brains near the crossing of the olivo-cochlear bundle to label efferents selectively. For afferents, dye was inserted into the cerebellum (nodulus and uvula) or the cochlear nuclei to label auditory and/or vestibular afferents. After appropriate diffusion times (Fritzsch et al., 2005b; Tessarollo et al., 2004), ears were dissected in cold buffer and mounted flat in glycerol. Ears were viewed in a Nikon E800 using a BioRad 2000 confocal system or a Zeiss LSM 510 confocal system.
We also traced the connections from the ear to the brain using lipophilic tracers as previously described (Fritzsch et al., 2005b; Maklad and Fritzsch, 2003). We inserted Neurovue maroon into the saccule and Neurovue red into the posterior canal in E18.5 fixed embryos. After a diffusion time of approximately 10 days, the brains were dissected, embedded in gelatine and sectioned at 100 μm thickness, as previously described (Gurung and Fritzsch, 2004). Stacks of images were taken with a Zeiss LSM 510 confocal microscope using the proper excitation and emission filters. Images were combined to create plates using CorelDraw software.
In order to confirm our findings with the lipophilic tracers, we also used immunocytochemistry to image nerve fibers to the ear (Farinas et al., 2001). To achieve this, we dissected the sensory epithelia of E16.5 and E18.5 ErbB2 null and littermate control mice fixed in 4% PFA. Sensory epithelia preparations with attached nerves were defatted, blocked and incubated in primary antibody for tubulin (Sigma) followed by an Alexa 647 conjugated secondary antibody. Preparations were mounted whole in glycerol and imaged using the z axis drive of the Zeiss LSM 510 confocal system to obtain stacks of images. These stacks were collapsed into single images to render the distribution of fibers in a single focal plane.
We investigated further aberrations in development, in particular the location of sensory neurons and presence or absence of Schwann cells, using thin plastic sections. For this technique, E16.5 and E18.5 ears were fixed first in 4% PFA, followed by fixation in 0.5% OsO4 in O.1 M Sorensen phosphate buffer. Ears were dehydrated, embedded in epoxy resin, sectioned at 1 μm thickness using glass knives and lightly counterstained. Coverslipped sections were imaged using a Nikon E800 microscope, high-resolution cameras and Metamorph software.
Since our data suggested the presence and overgrowth of sensory fibers, we investigated the expression of Ntf3 and Bdnf using Q-PCR to understand whether fiber growth is disregulated as a function of neurotrophin expression changes. E18.5 ErbB2 null and littermate control mice were euthanized, and the inner ears immediately removed, frozen in liquid nitrogen and ground to a powder using a mortar and pestle. Total RNA was extracted using a PowerGen 700 tissue homogenizer (Fisher Scientific) and Trizol reagent (Invitrogen) according to the manufacturer’s protocol. The RNA was further purified using RNeasy Midi Columns (Qiagen). The quantity and purity of the RNA were determined by obtaining the OD260 and OD260/OD280 ratio, and by an RNA 6000 Pico Assay (Aligent) using a Bioanalyzer (Agilent). Real-time PCR was performed with MGB-labeled probes and primers created using Assays-by-Design (Applied Biosystems) and samples were run in quadruplicate. Sequences spanning exons were selected preferentially for probes and primers by Assays-by-Design (applied Biosystems, Foster City, CA). Real-time PCR was performed using the TaqMan One-Step RT-PCR Kit per manufacturer’s instructions (Applied Biosystems) on an ABI Prism 7000. The relative efficiency of amplification of the selected gene versus the control GAPDH was plotted to ensure that the slope of total RNA versus ΔCT was less than 0.05. The 2−ΔΔCT was used to determine relative gene expression.
This work was supported by a grant from NIH (RO1 DC005590; BF) and NASA (NAG 2-1611; BF), COBRE (1P20RR018788-01; AM, BF, LH) and LB692 (BF, LH). This investigation was conducted in a facility constructed with support from Research Facilities Improvement Program Grant Number 1 C06 RR17417-01 from the National Center for Research Resources, National Institutes of Health. We acknowledge the use of the confocal microscope facility of the NCCB, supported by EPSCoR EPS-0346476 (CFD 47.076).