A. Section of P2 mouse cochlea. Pillar and Hensen cells (arrows), as well as Claudius cells (bracket) are p75+ (red); hair cells are Atoh1-GFP+ (green).
B. Section of P2 mouse cochlea, stained with anti-EGFR (red) and DAPI (cyan) to reveal nuclei. Arrows and brackets are the same as in (A). Size bar: 50 microns.
C. Expression levels of mRNA transcripts for four ErbB receptors measured in p75+ FACS isolated supporting cells with QPCR. Gene expression levels are shown as fold change, the relative values in expression of p75+ sorted cells over the unpurified parent population.
D. Expression levels of mRNA transcripts for 8 ErbB family ligands measured in freshly isolated supporting cells with QPCR: epidermal growth factor (Egf), transforming growth factor α (Tgfα), heparin-binding EGF (Hbegf), neuregulin-1 (Nrg1), neuregulin-2 (Nrg2), neuregulin-3 (Nrg3), neuregulin-4 (Nrg4), and betacellulin (Btc). Amphiregulin and eregulin did not amplify from P2 cochlea. Fold expression normalizes the level of transcripts detected to unpurified cells.
E. EGFR protein expression on supporting cells cultured for 1 hour (red). Top inset shows a higher magnification image of a cell (in grayscale, to better show detail). Bottom inset shows an identically exposed control without primary antibody. Although a no-antibody control is not conclusive evidence of specificity, the presence of EGFR staining in vivo (Figure 2B) and the proliferation of supporting cells in EGF (Figure 2F), taken with the absence of staining in the no-antibody control suggest the antibody staining is specific as previously shown (Hume et al., 2003; Stankovic et al., 2004).
F. Cochlear supporting cells take up BrdU (green). These cells were pulsed with BrdU from 20 to 24 hours in culture and fixed immediately.
G. Cochlear supporting cells enter G2/M, as revealed by pH3 staining (green), and by chromatin morphology (inset). These cells were fixed after 30 hours culture.
H. Cochlear supporting cells pulsed with BrdU from 20 to 40 hours in culture. 78.0 ± 3.9% of cells are BrdU-labeled (green).
I. Same field as (E), stained with DAPI. Top and bottom insets correspond as well.
J. Same field as (F), stained with DAPI.
K. Same field as (G), stained with DAPI. Size bar 25 μm.
L. Same field as (G), stained with DAPI.
M. Time course of the onset of cell cycle re-entry. BrdU uptake is first observed at 16 hours in vitro, and pH3+ cells at 24 hours in vitro. BrdU bars indicate s.e.m., and pH3 bars indicate range.
N. AG1478, an EGFR inhibitor, blocks BrdU uptake. Supporting cells were pulsed with BrdU from 20 to 24 hours or to 40 hours, as indicated, in either 20 ng/ml EGF (control, green bars), 1 μM AG1478 (gray bars), no EGF (brown bars), and 1 μM AG1478 alone (purple bars). These conditions measure initial cell cycle re-entry (by 24 hours, EGF compared to EGF + AG1478, n=10, p=10⁻¹⁰ by t-test) and overall cell cycle re-entry (by 40 hours, EGF compared to EGF + AG1478, n=8, p=10⁻⁷ by t-test). No EGF and 1 μM AG1478 alone were each performed 3 times. Bars represent s.e.m.
O. Dose response of AG1478. Cultures were pulsed with BrdU from 20 to 24 hours with varying doses of AG1478. n=3 or more cultures per time point. Bars represent s.e.m.

A. LY294002, a PI3K catalytic subunit inhibitor, blocks BrdU uptake. Mouse supporting cells were pulsed from 20 to 24 hours or from 20 to 40 hours, as indicated in either control (green bars) or 1 μM LY294002 (orange bars). These conditions measure initial cell cycle re-entry (by 24 hours, n=13, p=10⁻¹⁴, t-test) and overall cell cycle re-entry (by 40 hours, n=3, p=10⁻⁶, t-test).
B. Dose response of LY294002. Three or more mouse supporting cell cultures were pulsed with BrdU from 20 to 24 hours with varying doses of LY294002.
C. PI3K activity is also required for supporting cell proliferation in the chick. 3–12 organs were treated with gentamycin, followed by varying concentrations of LY294002, and proliferation was quantified in the region shown in Fig. 1. All error bars represent s.e.m.

A. Diagram of the cultured chick basilar papilla (BP). Chick BP images are from the region indicated by the box: distal of the midpoint, in the neural half.
B. Undamaged cultured chick BP, incubated for 48 hours, the last half with with EdU. Hair cells are revealed by MYO6 (green) and no proliferation is observed (EdU, red).
C. Chick BP cultured for 24 hours with 1 μM streptomycin followed by 24 hours with EdU; stained identically to (B). Hair cell loss (green) is nearly complete and proliferation is observed (red).
D. Same field as (B), showing the supporting cell nuclear layer (DAPI, blue) and EdU (red). No proliferation in the supporting cell layer. Size bar 10 μm.
E. Same field as (C), showing co-localization of EdU (red) and supporting cell nuclei (blue).
F. Diagram of the mouse organ of Corti. Images (E–H) are taken from the approximate region indicated.
G. Undamaged mouse organ of Corti, incubated for 48 hours, the last half with EdU. Hair cells are revealed by Atoh1-GFP fluorescence (green), and no proliferation (red) is observed in the sensory region.
H. Mouse organ of Corti cultured for 24 hours with 1 μM gentamycin followed by 24 hours with EdU; stained identically to (G). Hair cell loss (Atoh1-GFP, green) is nearly complete, but no proliferation (EdU, red) is observed.
I. Same field as (G). Mouse supporting cells, revealed by PROX1 immunostaining (blue) do not incorporate EdU (red). Size bar 25 μm.
J. Same field as (H). PROX1 immunostaining (blue) shows that surviving mouse supporting cells do not proliferate (EdU, red) after hair cell ablation.

A. Expression levels of transcripts for four ErbB receptors and the supporting cell marker Sox2, normalized to chicken β-actin. The value 1 indicates 1% of chicken β-actin mRNA. Blue bars indicate levels in epithelia from the base of the basilar papilla from control animals. Similar levels are seen in tissue obtained from 4 day streptomycin-treated animals (red), consistent with the idea that regenerating cells express these receptors. The supporting cell marker Sox2 is shown for comparison. All error bars indicate s.e.m.; n=3.
B. Experimental design for measuring proliferation in chick basilar papilla supporting cells. Control experiments receive DMSO instead of AG1478. Quantified fields were similar in location to Fig. 1.
C. P5 chick basilar papilla organ culture, treated with 1 μM gentamycin to kill hair cells, and incubated with BrdU. Anti-MYO6 antibody (green) demonstrates hair cell loss; BrdU uptake (pink) reveals proliferating supporting cells, and DAPI (blue) taken at the same optical plane shows supporting cell nuclei. (C') and (C”) show BrdU and DAPI channels from the same field. Size bar 15 μm.
D. High-power image of nuclei from the same field, demonstrating mitotic figures (arrows). Size bar 5 μm.
E. P5 chick basilar papilla organ culture treated with 1 μM gentamycin and 10 μM AG1478, and stained as in (C). Hair cell loss (green) and supporting cell survival (blue) are both unchanged by the treatment (cf. with C) but BrdU incorporation is significantly reduced. (E') and (E”) show BrdU and DAPI channels in the same field.
F. High-power image of nuclei from the field in (E); nuclear morphology is normal but no mitotic figures are evident.
G. Quantification of BrdU+ cells as a function of AG1478 concentration. Each bar represents the average ± s.e.m. of 5–12 organs. Significant reductions are seen with 1 and 10 μM (p=0.005, and p=0.0006, respectively, t-tests).

A. Quantification of the relative fraction of p27^Kip1+ cells with time after plating. The percent of p27^Kip1+ cells decreases from 88% to 30.9% ± 2.0% over the first 24 hours (n=11, p=4*10⁻⁸ ANOVA and t-test). Between 3 and 11 cultures were quantified per time point.
B. Quantification of p27^Kip1 mRNA in purified supporting cells over time in culture. 3 cultures were assessed by QPCR in triplicate for levels of p27^Kip1 and L19 messages for each time point, and fold was calculated relative to the mRNA levels in an aliquot of the same cells lysed immediately after purification.
C–K. Purified supporting cells were cultured with or without inhibitors as indicated, pulsed with BrdU from 20 to 40 hours (green, C–E) to label all dividing cells. Cultures were counterstained with anti-p27^Kip1 (red). P27^Kip1 immunoreactivity is shown together with BrdU (C–E), to illustrate a lack of co-localization, and alone (F–H), to illustrate relative levels of staining. DAPI was used to reveal nuclei (white, I–K). Non-dividing cells expressing p27^Kip1 appear red in the upper panels. Size bar: 50 μm.
L. Quantification of the non-dividing p27^Kip1+ cells over time in cultures with or without inhibitors. In control cultures, significantly fewer non-dividing p27^Kip1+ cells are seen at both 24 hours and 40 hours (n=7, p=10⁻⁸, t-test). At 40 hours, significantly more p27+ G1 cells are seen in AG1478 (47.1% ± 5.0, n=7, p=10⁻⁴, t-test) and in LY294002 (40.0% ± 10.0, n=3, p=0.0008, t-test) compared to control cultures. Bars represent averages of 3–11 cultures per time point and condition. All error bars represent s.e.m.

A–C. Purified supporting cells from wild-type mice were cultured with or without inhibitors as indicated and pulsed with BrdU from 20 to 40 hours (green) to label all dividing cells.
D–F Same fields as above, stained with DAPI.
G–L. Same experiment as (A–F), except supporting cells were isolated from p27^Kip1 knockout mice.
M. Quantification of cultures presented in (A–L). Significantly more cells from mutant animals re-enter the cell cycle compared to wild-type in AG1478 only (KO: n=3, p=0.002, t-test). Bars represent an average of 3 cultures per condition, ± s.e.m.
N. Summary diagram of signaling mechanisms deduced from data presented here. In quiescent supporting cells, p27^Kip1 maintains the cells in a post-mitotic state. Once activated after dissociation, EGFR and PI3K signaling causes p27^Kip1 down-regulation (dotted lines) to promote proliferation. PI3K signaling is also required independently of p27^Kip1 (see I, M).