Hair cell regeneration in the zebrafish lateral line after neomycin-induced death. A–D, Confocal maximum projections of representative neuromasts of H2AZ-GFP transgenic zebrafish. GFP (green) was expressed in all nuclei, whereas mature HCs were labeled with FM1–43FX (red). Scale bar, 20 μm. Larvae 5 dpf (A) were treated with 400 μm neomycin (Neo) for 1 h (B) to kill HCs, rinsed, and left to recover. New HCs were observed at 24 h (C) and increased in number by 48 h (D). E, Graph of HC regeneration as seen with FM1–43FX. Results are graphed as mean percentages of HCs (± 1 SEM) normalized to the control at 5 dpf. n = 6 fish per condition, 7 neuromasts per fish; p < 0.001 (ANOVA). F, Graph of HC regeneration assayed with myosin VI and HCS-1 antibodies. Fish were concurrently labeled and counted for both HC markers. Results are normalized to 5 dpf controls for each neuromast for the specific marker. Error bars represent ±1 SEM (n = 6 fish per condition, 7 neuromasts per fish). G, Number of regenerated HCs correlates with the original size of the neuromast. Mean HC numbers (±1 SEM) for seven neuromasts, as labeled with myosin-VI, at 72 h after neomycin exposure are plotted relative to the 5 dpf control siblings. The line of best fit and correlation coefficient was calculated and drawn. H–L, New HCs mainly originate from proliferative progenitors. Continuous exposure to BrdU after neomycin exposure was used to label all proliferating cells and their progeny. Fish were colabeled for both BrdU (red) and myosin-VI (green) (I–L). H, Stacked graph of mean (±1 SEM) myosin-VI-labeled HCs per neuromast either colabeled with BrdU (red) or BrdU-negative (green). n = 4 fish per condition, 3 neuromasts per fish. I–L, Confocal maximum projections of representative neuromasts continuously exposed to BrdU, collected at 24 and 48 h after initial neomycin exposure. BrdU-positive (asterisks) and BrdU-negative (arrowheads) HCs were observed. Scale bar, 10 μm.
Support cell proliferation temporarily increases after neomycin exposure. A, Experimental protocol for BrdU pulse-fix time course. Larvae were pulsed with 10 mm BrdU for 1 h before fixing. B, C, Time course graphs of SC proliferation showing neuromast cells in S-phase (B) as assayed with BrdU pulse fix, and M-phase (C) as seen with anti-phosphohistone (H3) antibody labeling. A significant transient increase in proliferation occurring after neomycin exposure was seen with both proliferation markers (ANOVA; p < 0.001). Error bars represent ±1 SEM (n = 10 fish per condition; 7 neuromasts per fish). D, Schematic showing the internal (yellow) and peripheral (green) SC subpopulations within a representative neuromast. The schematic was drawn from the confocal image on the right, a neuromast labeled with 1 h of BrdU (red) at 15 h after neomycin treatment and counterstained with SYTOX (green). E, BrdU pulse-fix time course graph of SC proliferation after mock (blue, circle) or neomycin (red, triangle) treatment, subdivided as internal (solid lines, solid shape) or peripheral (dotted lines, open shape) based on neuromast cell position. Error bars represent ±1 SEM (n = 10 fish per condition; 3 neuromasts per fish). F–H, Confocal maximum projections of representative neuromasts at 17 h after mock treatment (F) or neomycin exposure (Post-Neo) (G) incubated with BrdU (1 h) before fixing. A subset of the neomycin-treated BrdU-pulsed fish were then rinsed and left to recover in fresh EM for an additional 31 h (48 h time point) (H). Immunohistochemistry was performed for both BrdU (red) and acetylated tubulin (green), another HC marker. BrdU label was present in a subset of HCs (asterisks) after additional recovery (H), indicating that these new HCs were generated from proliferating SCs at 17 h after neomycin exposure. Scale bar, 20 μm.
Expression of atoh1a, notch3, and deltaA are elevated in neuromasts during HC regeneration. Whole-mount RNA in situ hybridizations were performed to assess atoh1a (A–F), notch3 (G–L), and deltaA (M, N) expression within regenerating neuromasts. A–R, Bright-field images of representative O1 neuromasts are shown. In sibling mock-treated controls, moderate levels of atoh1a (A, B) and notch3 (G, H) were observed at 5 dpf (12 h) and decreased over time. Little deltaA expression (M, N) was detected, which also decreased slightly with time. During HC regeneration, all three transcripts were upregulated between 12 and 24 h (C–E, I–K, O–Q), correlating with the peak of SC proliferation. Expression returned to control levels by 48 h (F, L, R). FISHs of notch3 (S, T) and deltaA (U, V) were performed on regenerating neuromasts 17 h after neomycin treatment to evaluate expression localization. Samples were immunostained with myosin-VI to label HCs (green) and counterstained with DAPI (blue) to visualize cell nuclei. Confocal images of a surface view projection (S, U) and an optical section (T, U) are shown for representative neuromasts. Both notch3 and deltaA were not expressed in HCs (S–V). notch3 was mainly expressed throughout the internal SCs, with little to no expression in the peripheral cells (S, T). Expression of deltaA was localized in a few cells adjacent to the HCs within the internal SC population (U, V), which we hypothesize to be HC precursors. Scale bar, 10 μm.
Notch inhibition results in excess regenerated hair cells. A, Graph of HC regeneration after 400 μm neomycin (Neo) exposure as seen with FM1–43FX in larvae with 50 μm DAPT incubation, a Notch inhibitor, or 0.5% DMSO vehicle control. Results are graphed as mean percentages of HCs (±1 SEM) normalized to the control at 5 dpf (n = 8 fish per condition; 6 neuromasts per fish). B–E, Confocal maximum projections of representative neuromasts under continuous BrdU incorporation for 48 h, with and without 50 μm DAPT. Hair cells were labeled with myosin-VI (green), and BrdU-labeled cells are red. Scale bar, 10 μm. F, Stacked graph of average myosin-VI-labeled HCs per neuromast either colabeled with BrdU (red) or are BrdU-negative (green). Excess HCs with 50 μm DAPT incubation all contained the BrdU label, indicating mitotic origin. Error bars represent ±1 SEM (n = 4 fish per condition; 3 neuromasts per fish).
Proliferation of support cell subpopulations is differentially affected by Notch inhibition. A, BrdU pulse-fix time course graph of SC proliferation in control and regenerating neuromasts with 50 μm DAPT or 0.5% DMSO vehicle incubation. B, BrdU pulse-fix time course graph after 400 μm neomycin treatment with 50 μm DAPT (triangle) or 0.5% DMSO (circle) vehicle incubation. BrdU-labeled SCs were subdivided as internal (solid lines, solid shape) or peripheral (dotted lines, open shape) based on neuromast cell position. DAPT incubation increased the number and duration of internal SC proliferation, whereas peripheral cells decreased in proliferation. C, BrdU pulse-fix time course graph of mock-treated control siblings with 50 μm DAPT (triangle) or 0.5% DMSO (circle) vehicle incubation. Both internal (solid lines, solid shape) and peripheral (dotted lines, open shape) SC proliferation decreased with DAPT incubation compared with the vehicle control, but effects were difficult to interpret because of the small numbers of BrdU-positive cells observed. Error bars represent ±1 SEM (n = 10 fish per condition, 7 neuromasts per fish for all 3 graphs) (A–C).
Working models of HC regeneration. A, In undamaged neuromasts, local signaling between mature HCs and internal SCs keeps cells quiescent. This process is not Notch dependent, and the inhibitory signal from the mature HCs is still unknown. After neomycin-induced death, internal SCs (IS) divide and give rise to two postmitotic HC precursors (HP). These HC precursors then differentiate into two new HCs. A second signal from the HC precursors signals to the adjacent internal SCs to inhibit proliferation and limit HC production. This second inhibitory signal is via Delta-Notch, which regulates and limits the number of regenerated HCs. B, When Notch signaling is inhibited with DAPT during regeneration, an increased number of internal SCs enter the cell cycle caused by no inhibition, resulting in an overproduction of HCs.