SEARCH

SEARCH BY CITATION

Keywords:

  • FGF;
  • support cells;
  • cochlea;
  • sensory epithelium

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Previous studies have demonstrated the importance of FGF signaling at several stages in the development of the cochlea. At early stages of embryogenesis, Fgfr1, Fgfr2, and several FGFs are critical for both the induction of the otic vesicle and the initial development of the sensory epithelium. At late stages of cochlear development, Fgfr3 is necessary for the development of the tunnel of Corti. To determine the stage of development when Fgfr3 is required, we examined the expression of Fgfr3 and Fgf8 at various developmental stages. We also re-examined the Fgfr3 −/− mouse with additional markers for developing supporting cells. We confirmed the previous analysis of the Fgfr3 −/− mice, indicating that there are deficiencies in support cell differentiation. Specifically, we find that the inner pillar cell never develops, while the outer pillar cell is stalled in its differentiation. In addition, we found an extra row of outer hair cells, and accompanying Deiters' cells, in the apical two thirds of the organ of Corti in the Fgfr3 mutant. Thus, in addition to controlling the fate decision between pillar cells and Deiters' cells, we find that Fgfr3 also regulates the width of the sensory epithelium. Developmental Dynamics 236:525–533, 2007. © 2006 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The organ of Corti is the sensory epithelium mediating auditory function in mammals and resides in the bony cochlea in the inner ear. The organ of Corti contains precisely patterned rows of both inner and outer hair cells (IHCs and OHCs) and several types of supporting cells, including inner phalangeal cells, pillar cells, Deiters' cells, Hensen's cells, and Claudius' cells. The precise arrangement of these cells is critical for normal function, since mutations that affect the development and patterning of the hair cells and supporting cells cause auditory defects. Recent studies have begun to define the molecular signals involved in the patterning of these cell types, and various signaling molecules and transcription factors are known to play important roles (Quint and Steel, 2003; Barald and Kelley, 2004; Dabdoub and Kelley, 2005).

Previous studies have demonstrated the importance of FGF signaling at several stages in the development of the cochlea (Pirvola et al., 1995; Leger and Brand, 2002; Pickles and Chir, 2002). At early stages of embryogenesis, Fgfr1, Fgfr2, and several FGFs, including FGF3, FGF8, FGF10, and FGF19, are critical for either the induction of the otic vesicle or the initial development of the sensory epithelium (Mansour et al., 1993; Ladher et al., 2000, 2005; Pirvola et al., 2000, 2002; Phillips et al., 2001, 2004; Pauley et al., 2003; Wright et al., 2003; Wright and Mansour, 2003). At intermediate stages of development, FGF signaling is also important; tissue-specific deletion of Fgfr1 results in severe defects in the development of the auditory sensory epithelium apparently due to a severe reduction in the precursor cells that give rise to the organ of Corti. In addition, Fgfr1 −/− mice show a selective loss of the OHCs and an increase in the number of cells that differentiate as pillar cells (Pirvola et al., 2002). At late stages of cochlear development, Fgfr3 is necessary for the development of the tunnel of Corti (Colvin et al., 1996). Inhibition of all FGFR signaling with a small molecule SU5402 reduced pillar cell differentiation in explant cultures of mouse cochlea, whereas addition of FGF2 to the cultures induces differentiation of additional pillar cells (Mueller et al., 2002). Finally, mice deficient in sprouty2, a negative regulator of FGF signaling, develop extra pillar cells, at the expense of Deiters' cells (Shim et al., 2005).

These results have led to the following model of the function of FGF signaling in cochlear development. FGF signaling, through the Fgfr1 and Fgfr2, is necessary for the initial induction of the otic vesicle, while FGFR3, expressed in the domain of cells that will give rise to both pillar and Deiters' cells, is activated by a gradient of FGF8 released from the inner hair cells to regulate pillar cell development. The cells with the highest levels of signal differentiate as pillar cells (Shim et al., 2005), while those with a lower level of signal develop as Deiters' cells. This model is consistent with most of the previous findings; however, there is some debate as to whether the FGF signal results in a switch in fate between presumptive Deiters' cells and pillar cells (Shim et al., 2005), or alternatively, whether FGFR3 is necessary for full differentiation of the pillar cells, but does not affect their fate (Mueller et al., 2002). In an attempt to resolve this issue, we have reanalyzed the timing and location of Fgfr3 and Fgf8 expression during cochlear development, and re-examined the Fgfr3 −/− mouse with additional markers for developing supporting cells. We have confirmed the previous analysis of the Fgfr3 −/− mice, that there are deficiencies in support cell differentiation (Colvin et al., 1996). Specifically, we find that the inner pillar cell never develops, as indicated by Prox1 expression, while the outer pillar cell is stalled in its differentiation. In addition, we find that Fgfr3-deficient mice in a CBA/CaJ background show an extra row of outer hair cells, and accompanying Deiters' cells, in the apical two thirds of the organ of Corti. Thus, in addition to previous evidence that FGF signaling controls the fate decision between pillar cells and Deiters' cells, we find that Fgfr3 also regulates the number of OHCs and Deiters' cells.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We analyzed the expression of Fgfr3 and Fgf8 in relation to the expression of Prox1, a support cell marker, and Math1, an early marker of hair cells. Previous reports have described the expression of both Fgfr3 and Fgf8 in the developing cochlea by in situ hybridization and immunohistochemistry (Pirvola et al., 1995, 2002; Shim et al., 2005). Our own in situ analysis confirms and extends these early reports (Fig. 1). At E15, we do not find any expression of Fgfr3 (Fig. 1A). The onset of Fgfr3 is at E15.5 (Fig. 1D,G). At this time, Fgfr3 is expressed at the base of the cochlea in the region of the developing sensory epithelium that contains the Prox1-expressing cells (Bermingham-McDonogh et al., 2006). The expression of Fgfr3 proceeds rapidly to the apex, and by E18.5, Fgfr3 is expressed throughout the length of the cochlea (Fig. 1J,M). The domain of expression of Fgfr3 includes the zone of cells that express Prox1, the developing pillar cells and Deiters' cells, and the outer hair cells (Bermingham-McDonogh et al., 2006).

thumbnail image

Figure 1. Expression patterns of mFgfr3 (A, D, G, J, M), FGF8 (B, E, H, K, N), Math1 (C, F, I, L, O) in the developing cochleas of embryonic mice. Levels (half-turns) of the cochlear duct are numbered from base (1) to apex (4). Arrowheads point to expressing domain. Arrows in M, N, and O point to developing inner hair cells. Brackets present developing outer hair cells. A–C: Adjacent sections of E15.0 cochlea. No expression was detected by Fgfr3 and Fgf8 probes. Math1 was expressed basal turns (1, 2). D–F: Adjacent sections of E15.5 cochlea. Fgfr3 and Fgf8 signals were detected in basal turns (1, 2). Math1 was additionally expressed in the 3rd turn. C–I: Higher magnification of the basal (1st) turn of above panel D–H. J, K: Adjacent sections of E18.5 cochlea. Fgfr3 and Fgf8 expressions reached to the apex. L: Math1 was already expressed in the apex on E16.5. M–O: Higher magnification micrograph of the basal (1st) turn of above panel J–L. Scale bar = 100 μm.

Download figure to PowerPoint

The principal ligand for FGFR3 in the developing cochlea is thought to be FGF8 (Mueller et al., 2002). We, therefore, also compared the expression of Fgf8 to that of Fgfr3 by in situ hybridization. We found that Fgf8 is expressed at E15.5 in the base of the cochlea (Fig. 1E,H) in what appear to be developing inner hair cells. Like Fgfr3, the Fgf8 expression sweeps from base to apex over the next 3 days. At E18.5, Fgf8 is expressed in all turns of the cochlea (Fig. 1K,N). Fgfr3 and Fgf8 are expressed approximately one day after the appearance of Math1. Adjacent sections showing Math1 expression are included for comparison (Fig. 1C,F,I,L,O). Comparing Fgfr3 to Fgf8, it appears that Fgfr3 is expressed very slightly earlier than Fgf8 at any given region of the cochlea. We find that both Fgfr3 and Fgf8 are expressed in the post-natal cochlea. The expression of Fgfr3 at P0 is confined to the support cells and is reduced in intensity (Fig. 2A). At postnatal day 7 (P7), Fgfr3 expression is maintained in the pillar cells but is absent from the Deiters' cells (Fig. 2D, arrowheads). Figure 2F shows the same section as in Figure 2D stained for MyosinVI after the in situ to reveal the hair cells. Fgf8 continues to be expressed specifically in the inner hair cell at both P0 and P7 (Fig. 2B,E, arrows). Figure 2C shows the expression of Math1 in an adjacent section to Figure 2B for identification of the hair cells at P0.

thumbnail image

Figure 2. Expression patterns of mFgfr3 (A, D), Fgf8 (B, E), Math1 (C), and MyoVI (F) in the organ of Corti of neonatal mice. Arrows point to inner hair cells, arrowheads point to developing pillar cells. Outer hair cells are presented by brackets and Deiters' cells by d. A–C: Adjacent sections of postnatal day 0 organ of Corti. D–F: Postnatal day 7 organ of Corti. D and E are adjacent. D and F are the same section. Scale bar = 100 μm in A–C and 20 μm in D–F.

Download figure to PowerPoint

As noted above, the loss of Fgfr3 has been previously shown to result in the failure of pillar cell differentiation. We found that this is also true for mice on the CBA background. Sections through the cochlea as early as postnatal day 15 show the failure of opening of the tunnel of Corti (data not shown). At postnatal day 0 (P0), the failure of pillar cell differentiation is clear from the lack of separation between the inner and outer hair cells (Fig. 3A,B). Labeling the sections with antibodies against p75 NTR confirms the lack of pillar cell differentiation: wild type cochleas show strong p75 NTR labeling in the pillar cells (Fig. 3C), while cells in the same location relative to the inner and outer hair cells in cochleas from Fgfr3-deficient mice are not labeled for p75 NTR (Fig. 3D). The nerve fibers are still labeled with p75NTR in the mutant. We used an additional marker, S100A1, that labels developing inner hair cells, inner phalangeal cells, and Deiters' cells, but does not label the pillar cells in wild-type (Fig. 3E, arrow). In the littermate Fgfr3 mutant animal, the S100A1 immunostained cells are juxtaposed. This suggests that the pillar cells have not developed, and Deiters' cells may be in their place (Fig. 3F, arrowhead).

thumbnail image

Figure 3. Sections through cochlea from Fgfr3 −/− and littermate control at post-natal day 0, showing labeling for Prox1 (A, B, red); p75 (C, D, red), and S100A1 (E, F, red). All sections were also labeled with Myosin VIIA to reveal the hair cells (green). IHC indicates the inner hair cell and the arrows and arrowheads point to the location of the pillar cells. Asterisk indicates the spiral vessel.

Download figure to PowerPoint

To further analyze the development of pillar cells in the Fgfr3 mutant animals, we used Prox1 immunofluorescent labeling. Prox1 is an early marker of the developing pillar and Deiters' cells (Bermingham-McDonogh et al., 2006). When we compared the Prox1 labeling in whole mounts of cochlea from wild type and Fgfr3 −/− mice, we found that cochlea from the Fgfr3 −/− mice have a very different appearance. In the apical two thirds of the cochlea, the cochlea has 5 rows of Prox1-labeled cells (Fig. 4H), while in the basal third of the cochlea, there are only 4 rows of Prox1-labeled cells (Fig. 4I). Figure 4I shows the transition from 4 rows to 5 rows of Prox1-immunoreactive cells. In order to determine which type of support cell is missing from the basal part of the cochlea, we can rely on the levels of expression of Prox1. In wild type animals, Prox1 labeling is strongest in the outer pillar cell and the outermost Deiters' cell (Fig. 4G, arrowheads). The inner pillar cell can be further identified by its characteristic elongated nucleus (Fig. 4G, IP). In the basal part of the mutant cochlea, there appears to be only a single row of pillar cells (Fig. 4I, PC?) and 3 rows of Deiters' cells, but at the transition, it looks as if there is an additional row of Deiters' cells. This additional row of Deiters' cells continues to the apex, although it is not continuous and there are abrupt transitions from regions of 3 rows to regions with 4 rows of Deiters' cells that may run for several hundred microns. Thus, on the basis of Prox1 labeling, in the Fgfr3 −/− mouse, there appears to be a loss of one row of pillar cells throughout the cochlea, and in the apical two thirds and additional row of Deiters' cells.

thumbnail image

Figure 4. Whole mounts of organs of Corti from Fgfr3 −/− (B,C,E,F,H,I) and littermate controls (A,D,G), labeled for Prox1 (red) and MyosinVIIA (green). OHC, outer hair cell; IHC, inner hair cell; DC, Deiter cell; OP, outer pillar cell; IP, inner pillar cell. Arrow in C and I points to abrupt transition from 4 outer hair cell and Deiters' cells to 3 of each cell type.

Download figure to PowerPoint

The presence of an extra row of outer Deiters' cells suggests that the width of the sensory epithelium might be extended in the Fgfr3-deficient mice. We, therefore, labeled cochlea from Fgfr3 −/− and wild type littermates with an antibody to myosin VIIa, a hair cell marker. We found that in the Fgfr3 −/− mice, those regions of the cochlea with an extra row of Deiters' cells also had an extra row of outer hair cells (Fig. 4B,C). The extra row of hair cells appeared precisely at the transitions from 3 to 4 rows of Deiters' cells. The extra row of hair cells was present as early as E17.5 (data not shown). A quantitative analysis of the cochleas of the Fgfr3 mutant animals, heterozygotes, and wild type littermates was carried out at postnatal day 3. We found that the length of the dissected cochleas was not different between the wild type and mutant animals (Fgfr3+/+: 5.1+/−1 mm; Fgfr3+/−: 5.8 +/− 0.3 mm; Fgfr3 −/−: 5.6 +/− 0.4 mm) indicating that the loss of Fgfr3 had no effect on the convergent extension responsible for establishing the overall cochlear length. We then measured the length of the organ of Corti that contained an extra row of hair cells and Deiters' cells as a percentage of the total length over which the assessment was made of the dissected organ (b+c in Fig. 5A). At this early postnatal time, the most apical part of the cochlea has not yet attained the well-ordered rows, thus this region was excluded from the analysis (Fig. 5, region a). The quantitative analysis confirms that approximately two thirds of the cochlea had an extra row of hair cells and Deiters' cells (Fig. 5B).

thumbnail image

Figure 5. A: P3 organ of Corti that has been stained as a whole mount for analysis of the number of extra hair cells/Deiters' cells. Region a indicates the most apical region that was not included in the analysis, because at this age the rows are not well established in the apex. Region b indicates the area of the cochlea where an extra row of hair cells/Deiters' cells were found. Region c indicates the region that contained no extra row. B: The quantitation of this analysis; there is a significant number of extra cells in the Fgfr3+/− (het) and a dramatic increase in the Fgfr3 −/− (homo) with 66% of the area analyzed showing an expansion of the sensory epithelia to include 1 extra hair cell and its accompanying Deiter. ANOVA analysis of the data indicates a significant difference for +/+ versus −/−, P < 0.01, but the difference between +/+ and +/− does not attain significance. N = 6 for each genotype.

Download figure to PowerPoint

We also analyzed cochleas from mature animals with SEM and found that the extra row of hair cells was present in animals as old as four weeks (Fig. 6). Figure 6 shows scanning electron micrographs taken approximately 180 degrees from the most apical part of the cochlea for both wildtype and Fgfr3 −/− mouse cochleas. In the wild-type littermate there is a single row of inner hair cells and three outer rows of hair cells; the tunnel of Corti is developed with a well-defined space between the inner and outer hair cells (Fig. 6A). By contrast, in the Fgfr3 −/− littermate, there is no separation between the inner and outer hair cells and there is one row of inner hair cells but four rows of outer hair cells (Fig. 6B). The hair cells in the Fgfr3 −/− mice appear to have normal stereocillia bundles. Figure 6C shows the transition area between the region that contains four rows of OHCs and that containing three rows of OHCs. The image looks similar to that we obtained with immunofluorescent labeling (Fig. 4C–I). Analysis of the Fgfr3 mutant cochleas with TEM confirmed the presence of four outer hair cells, and associated Deiters' cells (Fig. 7A). In addition, there appears to be an undifferentiated single pillar cell, which has failed to extend a process to the lumenal surface of the cochlea (Fig. 7A,B, asterisk). We identify this as a pillar cell because of its dense cytoplasm. The cell to the left in this picture resembles the cell on the other side of the inner hair cell and thus we identify this as a phalangeal cell. The inner hair cell is surrounded by afferents and efferents (Fig. 7A,B) and we can observe synaptic specializations (Fig. 7C, arrows). Consequently, the severe hearing loss with a 50–60 dB threshold shift across all frequencies (N=8, data not shown) is probably accounted for by a mechanical problem due to pillar defects.

thumbnail image

Figure 6. Scanning electron micrographs of adult cochleas from Fgfr3 −/− and littermate controls. A: A scan from a wildtype cochlea taken 180° from the apex. B: The equivalent position in a mutant cochlea clearly demonstrating an extra row of outer hair cells and a complete lack of tunnel formation. C: The region towards the base in a mutant animal where there is an abrupt transition from 4 rows of outer hair cells to the normal 3 rows. Scale bars = 10 μm.

Download figure to PowerPoint

thumbnail image

Figure 7. Transmission electron micrographs of an Fgfr3 −/− cochlea. A: The 4 outer hair cells (O) and their accompanying Deiter cells (d). B: The inner hair cell (I) surrounded by inner phalangeal cells (P) closely apposed to the first outer hair cell and a single outer pillar-like cell (asterisk), which has failed to extend a process to the luminal surface. C: The base of an inner hair cell (I) is normally surrounded by afferent (a) and efferent (e) endings. The arrows point to synaptic specializations. Scale bar = 2 μm in A and 1 μm in C.

Download figure to PowerPoint

The other two mouse mutants that give rise to extra outer hair cells are mutants in Hes5 and Sprouty2 (Zine et al., 2001; Shim et al., 2005). We, therefore, looked at the expression of both of these genes in the Fgfr3 mutant animals to ask if changes in these genes are causing the extra hair cell phenotype. We saw no changes in the expression pattern of Sprouty2 or Hes5 (Fig. 8 E–H). We also found no change in the expression of Fgfr1 in the cochlea of the Fgfr3 mutant animal (Fig. 8C and D). It is interesting that the extra support cell under the extra hair cell expresses Hes5 (Fig. 8J and I).

thumbnail image

Figure 8. Expression patterns of Fgfr3, Fgfr1, Sprouty2, and Hes5 in wild type (left) and Fgfr3 knockout (right) mouse. A,C,E: Basal turn of E15.5 cochlea of wild type littermate. B,E,F: Basal turn of E15.5 cochlea of mutant. Fgfr3 expression is not detected in the mutant cochlea (open arrow) but expression of both Fgfr1 and Sprouty2 are unchanged. Hes1 expression in E17.5 cochlea of wild type (G) and Fgfr3 mutant (H). Insets: Myo6 staining of the same sections. Row of inner hair cells is indicated by arrow, arrowheads indicate rows of outer hair cells. Scale bar = 100 μm.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

The results of our analysis of the Fgfr3-deficient mouse cochlea demonstrate that this receptor is important in regulating the width of the sensory epithelium, as well as in the previously documented role in pillar cell differentiation. We have found that in mice deficient for this receptor, the cochlea has an extra row of outer hair cells and associated Deiters' cells. We have also more carefully defined the time of onset of Fgfr3 and Fgf8 throughout the cochlea and we have found that both the receptor and its ligand are nearly contemporaneous. Lastly, we have characterized the effects of loss of the Fgfr3 on pillar cell development, and find that while both the inner and outer pillar cells fail to fully differentiate in the mutant, the inner pillar cell appears to be more sensitive and fails to label with the early support cell marker, Prox1.

Previous studies have reported that pillar cell differentiation requires Fgfr3 (Colvin et al., 1996). Moreover, the addition of a putative ligand for this receptor, FGF2, to cultures of cochlea, caused extra pillar cell development (Mueller et al., 2002). In addition, loss of Sprouty2, a negative regulator of FGF signaling, results in extra pillar cells (Shim et al., 2005). These results have led to a model in which the level of FGF signaling is a critical regulator of support cell diversity: FGF8 released by the inner hair cell activates FGFR3 in the neighboring support cells to induce the closest ones to develop as pillar cells; those cells more than two cell diameters away from the ligand receive a lower level of signal and so remain as Deiters' cells. Our data lend further support for this model, by demonstrating the close temporal and spatial expression patterns of Fgf8 and Fgfr3. In addition, we have found that the FGF signal is critical for the very early stages of inner pillar cell development. In these cells, the cells closest to the source of FGF8, an activated FGFR3, may be necessary for the maintenance of Prox1 expression; in the Fgfr3-deficient mouse, the inner pillar cell is either lost altogether or fails to express Prox1. The outer pillar cell seems to be somewhat less dependent on Fgfr3, since in the mutant animals, the outer pillar cells are still recognizable by their Prox1 expression, even though they fail to express p75NTR or develop their characteristic morphology.

Surprisingly, we also found that the Fgfr3 −/− animals had an extra row of both hair cells and underlying Deiters' cells throughout most of the apical two thirds of their cochleas. This phenotype was not reported in an earlier analysis of Fgfr3-deficient mice, though these animals were a different strain. It is not clear why this difference exists between mutant animals of the different strains, though many mutant phenotypes are dependent on background. We do not know the details of the mechanism underlying the extra hair cell phenotype in the Fgfr3 mutant animals or why this phenotype is not seen in more basal regions of the cochlea. We examined the possibility that the loss of Fgfr3 causes a failure of Sprouty2 expression in the Deiters' cells, and an additional receptor, such as Fgfr1 or Fgfr2 might then have a broader range of activation, recruiting additional hair and Deiters' cells to the as-yet undifferentiated epithelium in the lateral region of the developing organ of Corti. However, our results show normal expression of Sprouty2. In addition, we found normal expression of Hes5, another gene whose absence causes extra outer hair cells (Zine et al., 2001).

To explain our results and those of Shim et al. (2005), we propose the following model. In the absence of Fgfr3 expression by the developing pillar and Deiters' cells, there is an excess of FGF8 ligand that can now diffuse more laterally activating another FGF receptor and leading to a recruitment of undifferentiated epithelial cells into the sensory epithelium. Based on in situ patterns, we propose that the FGF receptor that is being activated is FGFR1. This model (Fig. 9) of excess of FGFR1 activation leading to extra outer hair cells fits with the dramatic reduction in outer hair cells described by Pirvola and colleagues when Fgfr1 is knocked out in the developing inner ear (Pirvola et al., 2002). These authors suggested that FGFR1 regulates the numbers of hair cells and support cells that are formed in a dose-dependent manner (Pirvola et al., 2002). Interestingly, they also observe a less severe effect basally mirroring what we describe here where in the basal turn of the cochlea there are normal numbers of outer hair cells and associated Deiters' cells. The elimination of either Sprouty 2 (an FGFR inhibitor) or FGFR3 (a potential “sink” for the FGF 8) could lead to a similar over-activation of FGFR1, thus allowing for the observation of extra hair cells in both situations. Loss of FGFR3 did not lead to an increase in expression of Fgfr1 in the support cells. Thus, we feel that the extra activation of FGFR1 occurs in the cells already expressing Fgfr1 at the lateral edge of the sensory epithelium. Although Sprouty expression is thought to be downstream of FGFR activation, in the cochlea it would appear that it is not downstream of FGFR3 activation. In the developing cochlea, Sprouty 2 is expressed prior to the expression of Fgfr3 and at postnatal day 5, the most highly expressing Fgfr3 cells, the pillar cells, do not express Sprouty 2 (Shim et al., 2005).

thumbnail image

Figure 9. Model of sensory epithelium development. A gradient of FGF8 is indicated by the purple triangle. In the mutant animal, there are no FGFR3 receptors to bind the FGF8 and, thus, there is more ligand available to the FGFR1 receptors at the lateral edge of the developing sensory epithelium. We speculate that it is the increase in activation of the FGFR1 receptors that lead to an increase in outer hair cells and Deiters' cells in the mutant animal.

Download figure to PowerPoint

Our study further highlights the multiple, interrelated roles of FGF signaling in the development of the cochlea. The patterning of the hair and support cells during cochlear development requires precise spatial and temporal activation of FGF receptors. At the present time, we know little about the downstream molecular events regulated by FGF signaling in the cochlea, but it is clear that a better understanding of this pathway will be key to unraveling the mechanisms that control cell fate and differentiation in this tissue.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Animals

Animals were housed in the Department of Comparative Medicine and all procedures were carried out in accordance with the guidelines of the animal care committee at the University of Washington. Embryos were obtained from timed pregnant matings of either Swiss-Webster or CBA mice purchased from Harlan (Indianapolis). We used the staging system of Theiler (1989) to accurately stage the embryos at the time of harvest (http://genex.hgu.mrc.ac.uk/Atlas/intro.html). For the postnatal animals, P0 is defined as the day of birth. The strain background for the Fgfr3 mutation was changed to CBA/CaJ by crossing into CBA/CaJ for 8 generations. The animals were genotyped following a protocol from David Ornitz (Washington University) using DNA obtained from tail clips and amplified with the following primers: 5′-GGG CTC CTT ATT GGA CTC GC-3′, 5′-TGC TAA AGC GCA TGC TCC AGA CTG C-3′; 5′-AGG TAT AGT TGC CAC GAT CGG AGG G-3′. The wildtype Fgfr3 allele yields a 200-bp product while the disrupted allele gives a 322-bp amplification product.

Histology/In Situ Hybridization

Embryos were harvested from timed pregnant Swiss Webster females at E15, E15.5, E16.5, and E18.5. Whole heads were fixed in a modified Carnoy's solution for 6 hr at room temperature. The samples were then dehydrated in 100% ethanol overnight at 4°C and were embedded in paraffin. Inner ear tissues were also harvested from P0 and P7 Swiss Webster mice, and were fixed in a modified Carnoy's solution, and dehydrated and embedded in paraffin. P7 inner ear tissues were decalcified by 0.5 M EDTA for 12 hr at 4°C before dehydration. Midmodiolar sections (6 μm thick) were collected on SuperFrost slides and air dried for 12 hr. The plane of sectioning was oriented to produce midmodiolar sections through the cochlea. At least three animals were examined at each time point. Digoxigenin-labeled probes; FGF8 (gift from Gail Martin, UCSF); FGFR1 (2,526 bp containing nts X-Y were sub-cloned into pCRII plasmid); FGFR3 (430 bp, probe was a gift from D. Ornitz, Washington University); Math1 (2,144 bp clone purchased from OpenBiosystems Inc. Huntsville, AL, clone ID 6530849) were applied to sections, which had been deparaffinized and rehydrated. After hybridization overnight at 68°C, slides were washed twice in 50% formamide, 1×SSC at 68°C, then in PBST. Nonspecific antibody binding was blocked by incubation in PBST with 10% goat serum, 2% blocking reagent (Roche Biochemicals, Indianapolis, IN) for 60 min. AP-conjugated anti-digoxigenin antibody (Roche) was diluted in this solution to 1/2,000 and applied. The following day, slides were rinsed in PBST before equilibration in 100 mM NaCl, 50 mM MgCl2, 100 mM Tris (pH 9.5), and 0.1% Tween-20 (NTMT). For a blue precipitate, NBT/BCIP were used in NTMT. After color development, sections are fixed with 4% paraformaldehyde for 30 min, rinsed with PBST, and coverslipped in FluoromountG (Southern Biotechnology, Birmingham, AL). Some sections were stained with rabbit anti-Myosin VI after in situ according to the method described below.

Immunofluorescence

Inner ear tissues were harvested from P3 mice and fixed in 4% paraformaldehyde for 2 hr at room temperature. Then, the tissues were washed in several changes of phosphate-buffered saline (PBS, pH 7.4) for over 1 hr, cryoprotected in successive changes of increasing sucrose concentrations, embedded in OCT, and 10-μm-thick-sections were collected. For whole-mount staining, the cochlea ducts were dissected out from P3 mice. To expose the organ of Corti, the anlage of striavascularis was removed using fine forceps. The tissue was fixed for 2 hr at room temperature in 4% paraformaldehyde. Sections or whole mounts were washed in PBS/0.1% Triton X-100 (PBST), blocked with 10% normal goat serum and 2% nonfat dry milk in PBST for 60 min, and then incubated overnight (sections) or 24 hr (whole-mount) at 4°C with primary antibody diluted in PBST containing normal goat serum and nonfat dry milk. The following day, the tissues were rinsed with PBST, incubated for 90 minutes (sections) or overnight (whole-mount) with a fluorescent-conjugated secondary antibody, and then rinsed with PBST, coverslipped in FluoromountG. Primary antibodies used were as follows: rabbit anti-Prox1 (Chemicon, Temecula, CA; AB5475) used at 1:1,000 dilution; guinea pig anti-Myosin VIIA (gift from Stefan Heller, Stanford University) used at 1:2,000 dilution; mouse anti-S100A1 (Dako, Carpinteria, CA) used at 1:300; rabbit anti-Myosin VI (Proteus BioSciences Inc., Ramona, CA) used at 1:400 dilution. Secondary antibodies used were goat anti-mouse Alexa 594, goat anti-rabbit Alexa 594, and goat anti-guinea pig Alexa 488, all from Molecular Probes (Eugene, OR) and diluted1:400 for sections, 1:800 for whole-mounts. Images of sections were captured on a Zeiss Axioplan microscope using a Photometrics CoolSnap CCD camera and Slidebook software. Confocal images were captured on a Zeiss LSM Pascal confocal microscope. Images were further processed using ImageJ (1.32) and Photoshop 7.0 (Adobe) software.

Electron Microscopy

Mice were deeply anesthetized with an i.p. injection of Nembutal, decapitated, and both left and right temporal bones quickly isolated. Tissue was immediately immersed in 2% paraformaldehyde, 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer + 0.001% CaCl2 (pH 7.4) at 4°C. The bulla was opened, cochlea identified, and the stapes was removed from the oval window. A small hole was made in the apex of the cochlea. The cochlea was then perfused with the same fixative via the oval window. This procedure was performed sequentially on both cochleas. Cochleas were held in fixative on a shaker for 1 hr at 25°C and then held overnight in fresh fixative at 4°C. Following 3× (10 min.) washes in 0.1 M sodium cacodylate buffer (pH 7.4), cochleas were placed in 1% osmium tetroxide (in the same buffer) on a shaker, light protected, for 30 min at 4°C. Following 3 washes (in the same buffer), cochleas were dehydrated through graded ethanols: 35, 70, 95, 100, 100%, 10 min each step. At this point, one cochlea from each animal was processed for scanning electron microscopy (SEM) while the opposite cochlea was processed for transmission electron microscopy (TEM).

For SEM, cochleas were critical point dried in an Autosamdri 814 (Tousimis, Rockville, MD), mounted on specimen stubs with colloidal silver, and allowed to dry overnight under vacuum. The sensory epithelium was exposed by removal of the bony capsule using a temporal bone drill and subsequent dissection of the stria vascularis using a fine wire probe. Specimens were sputter-coated with a Hummer VI-A (Anatech LTD, Denver, NC), and viewed in a JSM 6300F (JEOL, Peabody, MA) scanning electron microscope. For TEM, specimens were processed through 2 changes of propylene oxide (5 min each), and then placed in a 1:1 mixture of propylene oxide and Eponate 12 epoxy resin (Ted Pella, Redding, CA) for 2 hr on a shaker. Cochleas were then transferred to pure Eponate 12 resin and held overnight on a shaker at 25°C. Cochleas were placed in fresh Eponate 12 resin and held under vacuum for 1 hr. The specimens were subsequently embedded in silicone molds and polymerized for 24 hr at 60° C. Embedded cochleas were bissected (mid-modialar) using a rotary diamond saw and turns were excised with a razor blade, oriented and glued on blank resin stubs to reveal cross-sections of the organ of Corti by semi- and ultrathin sectioning. Semi-thin (1 μm) sections were mounted on subbed slides and stained with 1% Toluidine blue in 1% sodium borate and examined in a light microscope. Subsequent ultrathin sections (90 nm) were mounted on 200 mesh Athene thin-bar grids, stained with uranyl acetate and lead citrate, and examined in a JEM 1200EX (JEOL, Peabody, MA) transmission electron microscope.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We are grateful to Catherine Ray and Andrew Madsen for technical assistance and Linda Robinson for mouse husbandry. We also thank Dr. David Ornitz (Washington University) for providing us with the original Fgfr3 mutant mice and Dr. Stefan Heller for providing the Myosin7a antisera. We thank Dr. Remy Pujol for help with the EM analysis and for helpful discussions and critical comments on the manuscript. We thank Drs. Ed Rubel and Tom Reh for their support, Drs. Rubel, Reh, Tempel, and Hume for critical comments on the manuscript, and members of the Reh and Rubel lab for helpful discussions.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES