The Atoh1-lineage gives rise to hair cells and supporting cells within the mammalian cochlea

Abstract

The organ of Corti, located within the mammalian cochlea, contains a precise mosaic of hair cells (HC) and supporting cells (SC), the patterning of which is critical for auditory function. Progenitors of HCs and SCs are found in the same post-mitotic region of the cochlear duct during early stages of cochlear development, and both HCs and SCs are absent in mice lacking the transcription factor Atoh1. Based on existing data, Atoh1 is thought to be the earliest determinant of HC fate, and to have a cell-autonomous role in HC differentiation, but the lineage of Atoh1-positive cells within the cochlear duct remains unclear. To address this issue, we used an inducible Atoh1(Cre⁎PR) allele to permanently mark Atoh1-expressing cells at different developmental time points. We found that up to 30% of cells from the Atoh1-lineage develop as SCs, and that the number of Atoh1-positive SCs decreases both spatially and temporally in a pattern consistent with ongoing commitment. Modulation of Notch signaling, necessary for formation of the HC-SC mosaic, changes the percentage of cells from the Atoh1-lineage that develop as either HCs or SCs. The HC-SC ratio is also affected by morphogenesis of the cochlea, as inhibiting the outgrowth of the cochlear duct increases the number of Atoh1-lineage cells that develop as SCs. Our results demonstrate that the Atoh1-lineage is established early in cochlear development, but also show that expression of Atoh1 does not absolutely result in commitment to a HC fate.

Figure 1. In vivo induction of Atoh1^Cre*PR labels both hair cells and supporting cells

A. Surface view of the base of the cochlear epithelium from an Atoh1^lacZ knock-in reporter at P0. Atoh1 is only found in HCs. Scale bar (same for (B, D), 20 μm). B–C. β-galactosidase activity in in vivo induced Atoh1^Cre*PR; R26R^lacZ cochleae. RU486 was injected daily into pregnant females from 9.5–17.5 dpc, then cochleae were dissected and stained at P0. B. Low magnification view, similar to (A). Few cells in the OC are labeled. C. High magnification of a β-galactosidase-positive OHC (arrow). Scale bar (same for (E–G), 10 μm). D. Low magnification view of an Atoh1^Cre*PR; R26R^tdTomato cochlea. Cre*PR activity was induced with a single injection of RU486 at 14.5 dpc. Both IHCs (arrows) and OHCs (open arrows) are tdTomato-positive. E. High magnification view of tdTomato-positive cells from the Atoh1-lineage. IHCs and OHCs are labeled (arrows), as well as one SC, identifiable as a Dieters’ cell (arrowhead) based on its morphology and position between OHCs. F, G. Z-stack confocal projections. F. One IHC and one OHC, identified by position and phalloidin-labeled stereocilia bundles (green), are tdTomato-positive (arrows). A Dieters’ cell is also labeled by tdTomato (arrowhead), and has a strongly Sox2-positive nucleus (blue). G. tdTomato positive OHC (arrow), and inner phalangeal cell (arrowhead). MyosinVI (green) labels all HCs.

Figure 2. Atoh1-expressing cells develop as both hair cells and supporting cells

Atoh1^Cre*PR; R26R^tdTomato cochlear explants established at E13, induced for the first 24 hours in culture and maintained for 6 DIV. A–C. Low magnification views. A. Most HCs along the length of the explants are tdTomato-positive (magenta). B. MyosinVI (MyoVI) labeling of HCs in the same cochlear explant (green). C. Merge of (A) and (B) shows overlap (white) of tdTomato and MyoVI labeling. Scale bar, (A–C), 200 μm. D–F. Higher magnification views labeled as in (A–C). Nearly all HCs are tdTomato-positive, although MyoVI labeling indicates a few HCs that are not (arrows). In addition, some tdTomato-positive cells have a SC morphology and are negative for MyoVI (open arrowheads). Scale bar, (D–F), 25 μm. G. The average percentage of total HCs (MyoVI-positive) also expressing tdTomato in cochlear explants treated with the indicated concentrations of RU486. H. The percentage of total tdTomato-positive cells that develop as SCs following treatment with the indicated concentrations of RU486. I. Z-stack confocal projection of cells from a similar explant as in (A–F). Sox2 is strongly expressed in the nuclei of supporting cells. Open arrowheads indicate tdTomato-labeled supporting cells with bright Sox2-positive nuclei and lumenal projections. Scale bar, 10 μm.

Figure 3. tdTomato-labeled cells express markers of supporting cells

Atoh1^Cre*PR; R26R^tdTomato cochlear explants established at E13, treated with 10 nM RU486 for the first 24 hours in culture, and maintained for 6 DIV. tdTomato-positive cells are shown in red, and HCs are marked with MyoVI or MyoVIIa (blue). Open arrowheads indicate tdTomato-labeled cells with SC morphology. A. Shallow confocal Z-stack projection of the lumenal surface of a cochlear explant. Three tdTomato-positive, MyoVI-negative cells are shown. B. F-actin-rich stereocilia bundles present on HCs are stained with phalloidin (green). C. Merge of (A) and (B). The tdTomato-labeled cells lack stereocilia bundles (open arrowheads). D. A single MyoVIIa-negative cell is present between the IHC and OHC rows (open arrowhead). E. Pillar cells are labeled with anti-p75^NTR (green). F. Merge of (D) and (E). p75^NTR labeling coincides with tdTomato staining. D′–F′. Z-stack confocal reconstructions from the images shown in (D–F). The XY location of the Z-stack is indicated by the brackets in (D–F). The p75^NTR and tdTomato-positive, MyoVIIa-negative cell has characteristic PC location and morphology. G. Several tdTomato-positive, MyoVIIa-negative cells in the OHC or PC region (open arrowheads), or IHC region (arrow) are shown. H. Anti-Prox1 staining in the nuclei of SCs in the OHC and PC regions (green). I. Merge of (G) and (H). tdTomato-labeled cells found near OHCs and PCs are Prox1-positive (open arrowheads), but a labeled SC adjacent to the IHCs is Prox1-negative (arrow). J. Z-stack confocal reconstruction showing two tdTomato-positive SCs. K. S100A1 marks the IHCs, PCs, and Dieters’ cells (green). L. Merge of (J) and (K). The SC next to the IHCs is S100A1- negative (arrow), but the tdTomato-positive outer PC is S100A1-positive (open arrowhead).

Figure 4. Cell fate specification occurs in a basal to apical gradient

A. Low magnification view of an Atoh1^Cre*PR; R26R^tdTomato cochlear explant established at E15 and maintained for 4 DIV. RU486 was added for the first 24 hours of culture. The positions along the cochlear length at which cells were counted are indicated. Scale bar, 200 μm. B. Image taken at 5% of the cochlear length, measured from the base. With few exceptions (arrows), MyoVIIa-positive HCs are also tdTomato-positive. Few tdTomato-labeled SCs are seen, and those present are mostly found near the rows of OHCs (OSCs, open arrowheads). Only one ISC is observed in this field (closed arrowhead). Scale bar (same for C), 25 μm. C. Image from the 95% position. Many more labeled ISCs (closed arrowheads) are visible, along with several tdTomato-positive OSCs (open arrowheads). Arrows indicate unlabeled HCs. D. The percentage of total tdTomato-positive cells that develop as SCs at each position along the cochlear length, from base to apex, at E15. More labeled SCs are found at apical positions. E. Relative percentages of total tdTomato-labeled SCs that develop as either ISCs or OSCs, based on position along the duct. The percentage of ISCs increases toward the apex.

Figure 5. HCs along the length of the cochlea are Atoh1-positive at E13

A–B. Images from an Atoh1^Cre*PR; R26R^tdTomato cochlear explant established at E13, treated with 1 nM RU486 for 1 hour, and then maintained for an additional 6 DIV. A. Image from the 25% position. The highest percentage of tdTomato-labeled cells is found in this mid-basal region. B. Image from the 95% position. Some OHCs are found in the Atoh1-lineage, even after only a 1 hour RU486 treatment at E13 (arrows). C. The average percentage of MyoVIIa-positive HCs also labeled with tdTomato in explants treated with 1 nM RU486 for either 1 or 4 hours at E13. Differences among the percentage of labeled cells at each position were not statistically significant from one another.

Figure 6. Atoh1 expression in hair cells decreases over developmental time

A. The average percentage of MyoVIIa-positive HCs that are also positive for tdTomato in cochlear explants treated with 100 nM RU486 for 24 hours at E13, E15, E17, or P0. B. Same data as in (A), but shown as measured at each position along the cochlear length, from base to apex. More tdTomato-positive cells are found at the apex. C–D. Low magnification images from an Atoh1^Cre*PR; R26R^tdTomato explants treated with RU486 at E17 (C) and P0 (D), stained for MyoVIIa (green) and tdTomato (magenta). E–G. High magnification views of the explant in (D). E. At 5% of the cochlear length from the base, no tdTomato-positive cells are observed. F. A few tdTomato-labeled cells are present at the 50% position. G. Most but not all (arrows) MyoVIIa-positive HCs are also tdTomato-labeled at the 95% position. A single SC is tdTomato-positive (open arrowhead).

Figure 7. ATOH1 protein expression

A. Section from the base of an E14 cochlea. Three ATOH1-positive nuclei are present (red, arrows). B–F, I. Sections from cochleae at the indicated ages, stained with antibodies against ATOH1 (red) and MyoVI (green). F-actin stained by phalloidin is shown in blue. B, C. E16 cochlea. Strong staining for the ATOH1 antibody is present in all HCs (also positive for MyoVI), at both the apex and the base. D–F. P0 sections. D. Low magnification view of two turns of the cochlea. MyoVI staining marks HCs in the apex and base (arrows). Non-specific background staining from the ATOH1 antibody is present on the surface of the stria vascularis (open arrowheads). E. Apical section from (D). Strong ATOH1 signal is present in the OHCs (bracket), but slightly weaker staining is found in the IHC (arrow). F. ATOH1 staining is almost absent in a section from the base at P0 in both the IHC and OHCs. G, H. In situ hybridizations for Atoh1 transcript. Atoh1 mRNA is still present at the apex (G) and base (H) of a P0 cochlea. I. ATOH1 protein is not detectable at the apex of the cochlea by P5. J. ATOH1 protein (red) in the mid-base of an E16 Atoh1^+/+ cochlea. K. Same section as in (J), also showing MyoVI staining (blue), weak Sox2 staining in the HC nuclei, and strong Sox2 staining in the SC nuclei (green). L. In a section from an E16 Atoh1^−/−cochlea, no reactivity for either ATOH1 (red) or MyoVI (blue) is visible. Sox2-positive cells remain (green). M. Relative expression levels of Atoh1 and MyoVI in the cochlear epithelium at E13, E15, P0, and P5 as determined by qPCR. Values are presented as relative to E15. Error bars represent SEM. *p<0.01.

Figure 8. Modulation of Notch signaling alters the fates of Atoh1-positive cells

A–D. Atoh1^Cre*PR; R26R^tdTomato cochlear explants established at E13 and maintained for 6 DIV. RU486 was added for the first 24 hours of culture. All images are from the 25% of cochlear length position. Explants are stained for tdTomato (magenta) and anti-MyoVIIa (green). Open arrowheads indicate tdTomato-positive SCs, and arrows show non-tdTomato labeled HCs. A. Control explant. B. Explant treated with 50 μM DAPT from 3 to 6 DIV. Some tdTomato, MyoVI-positive HCs are adjacent to other HCs with no apparent SCs in between; more tdTomato-negative, MyoVI-positive HCs are also observed (arrows). Few tdTomato-labeled SCs are present (open arrowheads). C. Explant treated with 5 μg/ml control IgG2A Fc. D. Explant treated with 5 μg/ml DLL1-Fc protein, beginning after 1 DIV. E. Percentages of all tdTomato-labeled cells exhibiting a SC morphology, and percentages of all MyoVIIa-positive cells also labeled with tdTomato. Data are from paired explants from multiple experiments, and expressed as means ± SEM. DLL1-Fc data represent the average from all explants treated with 2.5 to 10 μg/ml DLL1-Fc. *p <0.05, ** p <0.005, *** p <0.00001. F. Average change in the % of tdTomato-labeled cells with a SC morphology in DLL1-treated explants vs. control Fc-treated. The increase in labeled SCs is significant for all concentrations of DLL1 except 1.25 μg/ml (p<0.05), but no dose response is evident.

Figure 9. Inhibition of cochlear outgrowth leads to change in fates of Atoh1-positive cells

A, B. Basal halves of E13 explants treated with 10 nM RU486 for 24 hrs. and maintained for 6 DIV; stained for tdTomato (magenta) and MyoVIIa (green). The control (A) shows normal extension and patterning while the paired explant treated with 10 μM blebbistatin beginning after 1 DIV is markedly shortened in length and disruption in cellular patterning is apparent. Dashed lines indicate position of high magnification images in (C, D). C, D. Open arrowheads indicate tdTomato- positive SCs and arrows show tdTomato-negative HCs as in previous figures. C. Control explant. HCs and SCs are present in ordered rows, and the separation of IHCs and OHCs is obvious. D. Blebbistatin-treated explant. The rows of HCs are disorganized, with both regions of closely packed HCs and gaps where no HCs are present. There is no discernible separation between IHCs and OHCs. Several tdTomato-positive SCs are visible (open arrowheads). E. Percentages of all tdTomato-labeled cells exhibiting a SC morphology, and percentages of all MyoVIIa-positive cells also labeled with tdTomato. Data are from paired explants from multiple experiments, and expressed as means ± SEM. *p <0.05, ** p<0.01 F. Relative expression levels of Hes1 and Hes5 in blebbistatin-treated cochlear explants. Explants were established and treated as in (A–D), and relative gene expression levels determined by qPCR after 6 DIV. Hes5 expression is significantly increased relative to control at 10 μM blebbistatin (p <0.05). Error bars represent SEM.