Hair cell differentiation in chick cochlear epithelium after aminoglycoside toxicity: in vivo and in vitro observations

Abstract

Inner ear epithelia of mature birds regenerate hair cells after ototoxic or acoustic insult. The lack of markers that selectively label cells in regenerating epithelia and of culture systems composed primarily of progenitor cells has hampered the identification of cellular and molecular interactions that regulate hair cell regeneration. In control basilar papillae, we identified two markers that selectively label hair cells (calmodulin and TUJ1 beta tubulin antibodies) and one marker unique for support cells (cytokeratin antibodies). Examination of regenerating epithelia demonstrated that calmodulin and beta tubulin are also expressed in early differentiating hair cells, and cytokeratins are retained in proliferative support cells. Enzymatic and mechanical methods were used to isolate sensory epithelia from mature chick basilar papillae, and epithelia were cultured in different conditions. In control cultures, hair cells are morphologically stable for up to 6 d, because calmodulin immunoreactivity and phalloidin labeling of filamentous actin are retained. The addition of an ototoxic antibiotic to cultures, however, causes complete hair cell loss by 2 d in vitro and generates cultures composed of calmodulin-negative, cytokeratin-positive support cells. These cells are highly proliferative for the first 2-7 d after plating, but stop dividing by 9 d. Calmodulin- or TUJ1-positive cells reemerge in cultures treated with antibiotic for 5 d and maintained for an additional 5 d without antibiotic. A subset of calmodulin-positive cells was also labeled with BrdU when it was continuously present in cultures, suggesting that some cells generated in culture begin to differentiate into hair cells.

Fig. 1.

Anatomy of the chick cochlear epithelium.A, The highly organized array of hair cells spans the width of the chick cochlear epithelium, as shown in this scanning electron micrograph of the lumenal surface of the basal, high-frequency region of the epithelium. Each hair cell has a bundle of stereocilia (white arrow) protruding from its round surface and is surrounded by the lumenal portion of support cells, whose surfaces are rectangular and highly constricted (black arrow).B, The cochlear epithelium is separated from the underlying stroma, or basilar membrane (bm), by a thin basal lamina (thin arrow), as shown in this toluidine blue-stained transverse section of the inferior portion of the epithelium. The tectorial membrane (tm) occupies the lumen above the epithelium. Because of fixation shrinkage of the tectorial membrane, it no longer contacts the hair cells and support cells, as it does in vivo. The stereocilia of hair cells (small shadowed arrow) protrude into the lumen. Only short hair cells are shown here, because they predominate in the basal epithelium. The cell bodies of hair cells are confined to the lumenal surface of the epithelium, whereas support cells span the entire depth of the epithelium. In control basilar papillae, support cell nuclei (straight black arrow) reside below the hair cell nuclei at different levels relative to the basal lamina. Border cells (flared black arrow) and hyaline cells (large shadowed arrow) lie outside the sensory epithelium. C, Three days after a single injection of gentamicin, hair cells in the basal third of the epithelium have been killed, and the support cells expand to occupy the hair cell-free regions. The zone of transition between the basal, damaged region (toward the left) and the apical, undamaged region (toward the right) is shown in this scanning electron micrograph. A swollen, surviving hair cell is indicated by a white arrow. Scale bar, 25 μm.

Fig. 2.

Calmodulin immunolabeling of hair cells in control and drug-damaged chick basilar papillae. A andC are micrographs of the lumenal surfaces of whole mounts of the midbasal region of the basilar papilla; the inferior edge is toward the bottom of the micrograph. Band D are micrographs of transverse sections of a region similar to that in A and C; the inferior edge of the epithelium is toward the left.A, Stereocilia (arrow) and cytoplasm of short hair cells in control basilar papillae strongly labeled with antibodies to calmodulin, but the lumenal surfaces of intervening support cells were unlabeled. B, Antibodies to calmodulin strongly labeled the cytoplasm and stereocilia (shadowed arrow) of short hair cells in control basilar papillae but did not label the nuclei. Tall hair cells (thick black arrow) were labeled more lightly than short hair cells. There was no detectable anti-calmodulin labeling in support cells (support cell nucleus indicated by thin arrow).C, Two days after gentamicin treatment, very few cells were strongly labeled with antibodies to calmodulin in the basal epithelium, where hair cell loss is nearly complete. In this region, the lumenal surface of the epithelium is composed predominantly of support cells and was only lightly labeled with antibodies to calmodulin. D, Sections of the damaged region confirmed that there was no strong anti-calmodulin labeling in any cells that remained after hair cell loss (arrow indicates support cell nucleus). Scale bars: 25 μm for A,C (shown in A); 25 μm forB, D (shown in B).

Fig. 3.

Specificity of anti-calmodulin antibodies.A, The specificity of the Sigma anti-calmodulin antibody was confirmed by Western blot analysis. Homogenates of the mature chick cochlear duct were processed through SDS-PAGE, blotted onto PVDF membrane, and probed with the anti-calmodulin antibody. Lane 1 contained purified bovine calmodulin as a standard, andlane 2 contained chick cochlear duct samples. Prestained molecular weight standards: 112, 84, 53.2, 34.9, and 20.5 kDa. The 28.7 kDa standard ran indistinguishably as a smear with the 34.9 kDa standard; we arbitrarily designated the smear as 34.9 kDa.B, Cochleae immunolabeled with the UBI anti-calmodulin antibody displayed a pattern identical to that of the Sigma primary antibody (Fig. 2A), strongly labeling the cytoplasm and stereocilia (black arrow) of hair cells. Scale bar, 25 μm.

Fig. 4.

TUJ1 immunolabeling of hair cells in control and drug-damaged chick basilar papillae. A andC are micrographs of the lumenal surfaces of whole mounts of the midbasal region of the basilar papilla; the inferior edge is toward the bottom of the micrograph. Band D are micrographs of transverse sections of a region similar to that in A and C; the inferior edge of the epithelium is toward the left.A, The lumenal surfaces of short hair cells (arrow), but not the intervening support cells, in control basilar papillae labeled with the TUJ1 antibody.B, The cytoplasm of short hair cells (stereocilia indicated by thick black arrow) in control basilar papillae was strongly labeled with the TUJ1 antibody. Tall hair cells were more lightly labeled than short hair cells (data not shown). There was no detectable labeling with TUJ1 in the nucleus, cuticular plate, or stereocilia of any hair cells. Neural elements (small shadowed arrow) were labeled, but support cells were not (support cell nucleus is indicated by thin black arrow).C, Two days after gentamicin treatment, there were no TUJ1-labeled cells present in the basal epithelium, where hair cell loss is nearly complete. The only TUJ1 labeling was detected in the neural elements (arrow) that persist after hair cell loss. D, Sections of the damaged region confirmed that TUJ1 labeling was confined to the neural elements (thick arrow) in the region of complete hair cell loss (thin arrow indicates support cell nucleus). Scale bars: 25 μm forA, C (shown in A); 25 μm for B, D (shown inB).

Fig. 5.

Cytokeratin immunolabeling of support cells in control and drug-damaged chick basilar papillae. A andC are micrographs of the lumenal surfaces of whole mounts of the midbasal region of the basilar papilla; the inferior edge is toward the bottom of the micrograph. Band D are micrographs of transverse sections of a region similar to that in A and C; the inferior edge of the epithelium is toward the left.A, In control cochleae, antibodies to cytokeratins intensely labeled the lumenal surfaces of support cells (straight black arrow) but not hair cells (shadowed arrow points to hair cell nucleus). In the basal end, strong cytokeratin immunolabeling was restricted to the lumenal surfaces of support cells along the inferior edge of the epithelium (straight black arrow); support cells in the middle and superior regions demonstrated weaker or no labeling (flared black arrow). In the apical end, support cells throughout the entire width of the epithelium were strongly cytokeratin-positive (data not shown). B, Intense cytokeratin labeling was present along the lumenal surfaces of support cells in the basal region of control sensory epithelia (thick black arrow). Very little labeling occurred in the cytoplasm of support cells (thin black arrow points to support cell nucleus). No labeling occurred in hair cells (shadowed arrow indicates hair cell stereocilia). C, Two days after gentamicin treatment, the lateral edges of support cells (arrow) across the entire width of the damaged epithelium labeled with anti-cytokeratin antibodies. Cytokeratin labeling in the undamaged, apical end of drug-treated cochleae resembled controls (data not shown). D, At 2 d after gentamicin treatment, cytokeratin labeling in the damaged sensory epithelium was heaviest along the apicolateral borders of support cells (arrow). Scale bars: 25 μm for A,C (shown in A); 25 μm forB, D (shown in B).

Fig. 6.

Calmodulin is expressed in early differentiating hair cells in vivo. A–C, Micrographs of transverse sections of the midbasal region of the basilar papilla (the inferior edge is toward the left). Dshows a whole mount of the midbasal region of the basilar papilla; the neural, superior edge of the epithelium is toward thetop. A, Calmodulin immunoreactivity reappeared in the previously damaged region by 4 d after gentamicin injection. All calmodulin-positive cells seemed to extend from the basilar membrane (bm) to the lumen (lu) or to have processes that extended adlumenally toward the basilar membrane. Labeled cells had nuclei in either the adlumenal, support cell layer (thin black arrow) or the lumenal, hair cell layer (thick black arrow). Many of the cells with lumenal nuclei had short immature stereocilia (shadowed arrow). B, By 7 d after gentamicin injection, most calmodulin-positive cells (black arrow) in the damaged region had lumenal nuclei and bundles of short stereocilia (shadowed arrow). C, By 10 d after gentamicin injection, all calmodulin-positive cells in the recovered epithelium seemed to be relatively mature hair cells with columnar or bucket-shaped cell bodies (black arrow) and staircase-shaped bundles of stereocilia (shadowed arrow). D, The organization of the basal, previously damaged region was nearly restored at 10 d. Calmodulin antibodies labeled the cytoplasm (thick black arrow) and stereocilia (thin black arrow) of regenerated hair cells. Because of the curvature of the epithelium, some hair cells are seen directly from above, whereas others are viewed from the side. Scale bars: 25 μm for A–C (shown inC); 50 μm for D. All immunoreactions shown were conducted with the Sigma anti-calmodulin antibody.

Fig. 7.

Calmodulin and TUJ1 antibodies co-label early regenerating hair cells. Confocal image of double-labeled cells in the basal region of the basilar papilla at 4 d after gentamicin.A, Numerous cells in the regenerating region were labeled with antibodies to calmodulin. Most of the labeled cells had morphological profiles characteristic of early differentiating hair cells (arrow indicates one example). B, All of the cells that were calmodulin-immunoreactive also labeled with the TUJ1 antibody. One double-labeled cell is indicated by thearrow. Neural elements were labeled with the TUJ1 antibody (arrowheads) but not with the anti-calmodulin antibody. Scale bar, 10 μm.

Fig. 8.

Time course of calmodulin immunoreactivity in differentiating hair cells. A, Two hours after³H-thymidine injection, ³H-thymidine incorporation occurred in calmodulin-negative cells (arrow) located in the damaged region. Most³H-thymidine-positive cells had nuclei in the support cell (adlumenal) layer. The lumen (lu) and basilar membrane (bm) are shown. No double-labeled cells were detected at this time. B, Twenty-four hours after³H-thymidine injection, numerous calmodulin-negative/³H-thymidine-positive cells (arrow) were again detected in the damaged region, but at this time their nuclei were located in both the adlumenal and lumenal layers. C, At 24 hr after BrdU injection, a few cells that were labeled for calmodulin and BrdU (thick arrow) were detected in the damaged region. Such cells resembled early differentiating hair cells, because they had a process that descended toward the basal lamina. Thin arrowpoints to BrdU-labeled cell that is not labeled with antibodies to calmodulin. D, At 48 hr after BrdU injection, cells with calmodulin-positive cytoplasm and BrdU-labeled nuclei (thick arrow) still seemed relatively immature, because they were elongated and possessed no detectable stereocilia.Thin arrow points to BrdU-labeled cell that isnot labeled with antibodies to calmodulin.E, At 72 hr after ³H-thymidine injection, all double-labeled cells seemed more mature than at 48 hr. Most double-labeled cells had rounded cell bodies (thick arrows), and some had short immature stereocilia (shadowed arrow). Thin arrow points to³H-thymidine-labeled cell that is notlabeled with antibodies to calmodulin. F, By 240 hr after ³H-thymidine injection, most double-labeled cells (thick arrow) had the shape of mature hair cells and distinct stereocilia bundles (shadowed arrow).Thin arrow points to ³H-thymidine-labeled cell that is not labeled with antibodies to calmodulin. Scale bar (shown if F): 25 μm. All immunoreactions shown were conducted with the Sigma anti-calmodulin antibody.

Fig. 9.

³H-thymidine labeling in the regenerating cochlear epithelium. The number of thymidine-labeled cells in the basal 300 μm of the cochlea that were calmodulin-positive or calmodulin-negative was counted, and the location of the nucleus of each cell in either the lumenal or adlumenal layer was determined, at 2, 24, 48, 72, and 240 hr after gentamicin treatment in vivo. Error bars represent SEM.n = 3 basilar papillae per time point.

Fig. 10.

Regeneration of calmodulin-positive cells. The percentage of the total ³H-thymidine-labeled cells that were calmodulin-positive was determined at 2, 24, 48, 72, and 240 hr after gentamicin treatment in vivo. Error bars represent SEM. n = 3 basilar papillae per time point.

Fig. 11.

Cultures of the chick cochlear sensory epithelium. A, Preplating epithelial explants were recognizable as being derived from the cochlear epithelium, because hair cell stereocilia (shadowed arrows) projected from their surfaces. B, Sections of the tissue that remained in the cochlear duct after removal of the epithelium, counterstained with toluidine blue, revealed that all sensory epithelial cells had been removed, and the dissection process did not disrupt the basal lamina (thin black arrows). Hyaline cells (thick black arrow) also remained with the cochlear duct and are not isolated with this technique. C, Explants of the cochlear epithelium attach and spread on the coverslips by 2 d after plating. A 3 d culture is shown here. D, By 11 d after plating, support cells have proliferated and spread to cover most of the culture well. Scale bars: 25 μm for A, B; 100 μm for C, D (shown inD).

Fig. 12.

Complete hair cell loss is triggered in epithelial cultures by aminoglycoside antibiotics. Explants of the cochlear epithelium were labeled with cell-specific antibodies after 2 d of culturing with or without streptomycin. A, C, and E show control cultures, and B, D, and F show drug-treated cultures.A, In control cultures, anti-calmodulin antibodies labeled the cytoplasm and stereocilia (shadowed arrows) of hair cells but not support cells (black arrow).B, In streptomycin-treated cultures, calmodulin immunoreactivity was completely lost (arrow points to a support cell nucleus). C, In control cultures, rhodamine phalloidin labeled hair cell stereocilia (white arrow), the intercellular junctions between hair cells and support cells, and support cell microvilli (not evident in this figure). D, In drug-treated cultures, no stereocilia were labeled with rhodamine phalloidin, suggesting that all hair cells were killed. The cells that remained after hair cell loss were labeled with rhodamine phalloidin only at their cell/cell junctions (straight white arrow) and microvilli (fuzzy white material at end of curved white arrow), which is characteristic of support cells.E, In control cultures, antibodies to cytokeratins labeled the lumenal surfaces of support cells (black arrow) but not hair cells (stereocilia are indicated byshadowed arrows). F, In drug-treated cultures, antibodies to cytokeratin labeled the lateral borders and cytoplasm (black arrow) of support cells. Scale bars, 50 μm. All immunoreactions shown were conducted with Sigma anti-cytokeratin or anti-calmodulin antibodies.

Fig. 13.

Cell proliferation in epithelial cultures. BrdU incorporation into support cells in control and streptomycin-treated cultures was studied at different times after plating following a pulse/fix BrdU-labeling paradigm. A, BrdU labeling (black arrow) in a drug-treated culture at 3 din vitro. Scale bar, 25 μm. B, BrdU-labeled support cells were quantified at 2, 3, 5, 7, 9, and 11 d after plating. Error bars represent SEM.n = 2 culture wells per time point.

Fig. 14.

Calmodulin-positive and TUJ1-positive cells reappear in long-term cultures of the cochlear epithelium.A, Rhodamine phalloidin labeled the lateral borders (white arrowhead) and microvilli (fuzzy material at end of curved white arrow) of cells in long-term cultures. No stereociliary bundles were detectable with rhodamine phalloidin labeling. B, Cells in long-term cultures label with cytokeratin antibodies. C, Calmodulin-positive cells with fusiform cell bodies (black arrows) or round cell bodies (shadowed arrows) were present. D, TUJ1-positive cells (black arrows) with shape and size similar to the calmodulin-positive cells were detected.E, Digitized image showing calmodulin-positive cells in 10 d culture; arrows indicate cells also labeled with BrdU (shown in F). F, Digitized image derived from fluorescence confocal microscopy of BrdU-labeling from the same field as E;arrows point to cells that are double-labeled. Scale bars: 100 μm for A, B (shown inA); 50 μm for C–F (shown inC). In all examples, the Sigma anti-calmodulin antibody was used to detect calmodulin.