Spontaneous Hair Cell Regeneration Is Prevented by Increased Notch Signaling in Supporting Cells

Abstract

During embryonic development, differentiation of cochlear progenitor cells into hair cells (HCs) or supporting cells (SCs) is partially controlled through Notch signaling. Many studies have shown that inhibition of Notch signaling allows SCs to convert into HCs in both normal and drug damaged neonatal mouse cochleae. This mechanism is also implicated during HC regeneration in non-mammalian vertebrates; however, the mechanism of spontaneous HC regeneration in the neonatal mouse cochlea is less understood. While inhibition of Notch signaling can force SCs to convert into HCs and increase the number of regenerated HCs, it is currently unknown whether this pathway is involved in spontaneous HC regeneration observed in vivo. Therefore, we investigated the role of Notch signaling during the spontaneous HC regeneration process using Atoh1-CreER^TM::Rosa26^{loxP-stop-loxP-DTA/+} mice injected with tamoxifen at postnatal day (P) 0 and P1 to ablate HCs and stimulate spontaneous HC regeneration. Expression changes of genes in the Notch pathway were measured using immunostaining and in situ hybridization, with most changes observed in the apical one-third of the cochlea where the majority of HC regeneration occurs. Expression of the Notch target genes Hes1, Hes5, Hey1, HeyL, and Jagged1 were decreased. To investigate whether reduction of Notch signaling is involved in the spontaneous HC regeneration process, we overexpressed the Notch1 intracellular fragment (N1ICD) in cochlear SCs and other non-sensory epithelial cells in the context of HC damage. Specifically, Atoh1-CreER^TM::Rosa26^{loxP-stop-loxP-DTA/+}::Sox10^rtTA::TetO-LacZ::TetO-N1ICD mice were injected with tamoxifen at P0/P1 to stimulate spontaneous HC regeneration and given doxycycline from P0-P7 to induce expression of N1ICD as well as LacZ for fate-mapping. We observed a 92% reduction in the number of fate-mapped regenerated HCs in mice with N1ICD overexpression compared to controls with HC damage but no manipulation of Notch signaling. Therefore, we conclude that increased Notch signaling prevents spontaneous HC regeneration from occurring in the neonatal mouse cochlea. Understanding which components of the Notch pathway regulates regenerative plasticity in the neonatal mouse cochlea will inform investigations focused on stimulating HC regeneration in mature cochlea and eventually in humans to treat hearing loss.

Keywords: CreER; NICD; Notch; Tet-On; cochlea; hair cell regeneration.

FIGURE 1

Hes5-LacZ-positive cells, but not Sox2-positive cells, are reduced during the spontaneous HC regeneration process. Atoh1-DTA::Hes5^LacZ mice were injected with tamoxifen (Tam) at P0–P1 to induce HC death and spontaneous HC regeneration and cochleae were collected at P2, P4, and P6. (A–R) Representative confocal slice images with anti-β-gal antibodies to detect Hes5-LacZ expression (green) and anti-Sox2 antibodies to label all SC nuclei (magenta). (A–C, G–I, M–O) Images are taken from P2, P4, or P6 control samples, respectively. Scale bar = 25 μm. (S) Fewer Hes5-LacZ-positive cells were detected in HC damaged samples at P2 and P4, but not P6. There was no significant difference in the number of Sox2-positive cells in the Atoh1-DTA::Hes5^LacZ cochlea after HC damage compared to undamaged controls. ^∗∗p < 0.01, ^∗∗∗p < 0.001, as determined by a two-way ANOVA with a Sidak’s post hoc test. N = 3–5.

FIGURE 2

Hes5-LacZ-positive cells are reduced in lateral SCs, but not medial SCs. (A) Pillar and Deiters’ cells (red, PCs/DCs) are labeled with S100a1 and inner phalangeal/border cells (yellow, IPhCs/BCs) lie medial to the S100a1 expressing region. Hensen cells (blue, HeCs) express Sox2 and border the lateral edge of the S100a1 region, but do not have Hes5 expression. (B–G”’) Atoh1-DTA::Hes5^LacZ mice were injected with tamoxifen (Tam) at P0–P1 to induce HC death and spontaneous HC regeneration and cochleae were collected at P2, P4, or P6. Representative confocal slice images with anti-βgal antibodies to detect Hes5-LacZ expression (green), anti-Sox2 antibodies to label all SC nuclei (magenta), and anti-S100a1 antibodies to label the cytoplasm of PCs/DCs (red). Scale bar = 25 μm (H) In the PCs/DCs region, fewer Sox2-positive cells were detected at P4 and P6 after HC damage, while a decreased number of Hes5-LacZ-positive cells were detected at all three ages. Only at P2 were there significantly more Sox2-positive/Hes5-LacZ-negative cells detected. (I) There was no significant change in the number of Sox2-positive, Hes5-LacZ-positive, or Sox2-positive/Hes5-LacZ-negative cells in the IPhCs/BCs region. ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001 as determined by a two-way ANOVA with a Tukey’s or Sidak’s post hoc test. N = 4–5.

FIGURE 3

Hes1, Hes5, Hey1, and HeyL are decreased during spontaneous HC regeneration. (A–B’) Hes1, (C–D’) Hes5, (E–F’) Hey1, and (G–H’) HeyL transcripts were investigated during spontaneous HC regeneration using in situ hybridization. Low magnification images show the whole cochlea and high magnification images were taken from the apical turn where the majority of spontaneous HC regeneration occurs. Control cochlea (A,A’,C,C’,E,E’,G,G’) as well as Atoh1-DTA (B,B’,D,D’,F,F’,H,H’) littermates were injected with tamoxifen (Tam) at P0 and P1 to induce HC damage and spontaneous HC regeneration. Samples were collected at P4 and incubated with probes to label mRNA of Notch downstream effectors. Paired samples from the same litter (control and Atoh1-DTA) were processed in parallel and comparisons were only made between the paired samples. While there was some variability in the results obtained among experiments (I), the majority of experiments showed decreased gene expression for all four Notch effectors in Atoh1-DTA cochleae compared to control. Scale bar for (A,B,C,D,E,F,G,H) = 200 μm, Scale bar for (A’,B’,C’,D’,E’,F’,G’,H’) = 20 μm. N = 4–7.

FIGURE 4

Jag1 labeling is reduced in lateral, but not medial, SCs during spontaneous HC regeneration. Atoh1-DTA mice were injected with tamoxifen (Tam) at P0–P1 to induce HC death and spontaneous HC regeneration and cochleae were collected at P2, P4, or P6. (A,R) Representative maximum projection images with anti-Jag1 antibodies to detect Jag1 expression (white) and anti-S100a1 (red) antibodies to mark the pillar and Deiters’ cells region. After HC damage, both S100a1 and Jag1 labeling was reduced in the outer pillar and Deiters’ cell region, while expression is maintained in medial SCs (D–F,J–L,P–R) and in all SCs in control samples (A–C,G–I,M–O). Scale bar = 25 μm. N = 3.

FIGURE 5

The combination of CreER/loxP and Tet-On genetic systems allow gene manipulation in distinct cell types without cross-reactivity. Atoh1-CreER^TM::Rosa26^{loxP-stop-loxP-tdTomato}::Sox10^rtTA::TetO-LacZ mice were injected with tamoxifen (Tam) at P0 and P1 to induce tdTomato expression. Doxycycline (Dox) was administered to nursing mother in the diet and each pup received a dox injection at P1 to induce LacZ expression. Cochleae were collected at P7 and stained with antibodies against myosin VIIa (white) to label HCs and β-gal (green) to label LacZ-positive cells. (A–C) Representative confocal slice images show expression of endogenous tdTomato fluorescence (red) detected in HCs similar to Cox et al. (2014). β-gal expression (green) was detected in SCs and other non-HCs similar to Walters and Zuo (2015). (D–G) Optical cross-sections showing expression of tdTomato in HCs and β-gal in SCs and all non-HCs. Scale bar = 50 μm. N = 3.

FIGURE 6

Increased expression of N1ICD driven by Sox10^rtTA leads to increased expression of Notch downstream effectors. Sox10^rtTA-negative, Sox10^rtTA::TetO-N1ICD, and Sox10^rtTA::TetO-LacZ::TetO-N1ICD mice were administered doxycycline (Dox), in the diet to nursing mother, as well as an injection to each pup at P1. Cochleae were frozen at P7 and mRNA transcripts were measured. (A) N1ICD expression was increased in Sox10^rtTA::TetO-N1ICD, and Sox10^rtTA::TetO-LacZ::TetO-N1ICD cochlea, compared to Sox10^rtTA-negative controls. There was no difference in the expression of N1ICD when TetO-LacZ was also present. (B) Expression level of Notch effectors Hes1, Hes5, Hey1, and HeyL were increased in Sox10^rtTA::TetO-LacZ::TetO-N1ICD cochlea, compared to Sox10^rtTA-negative controls. ^∗∗∗p < 0.001 as determined by a one-way ANOVA with a Dunnett’s post hoc test or a Student’s t-test. N = 10.

FIGURE 7

Fewer spontaneously regenerated HCs were observed in the presence of N1ICD overexpression. Representative confocal slice images taken from the apical turn of the cochlea of (A–H) Atoh1-DTA::Sox10^rtTA::TetO-LacZ and (I–P) Atoh1-DTA::Sox10^rtTA::TetO-LacZ::TetO-N1ICD mice that were injected with tamoxifen (Tam) at P0/P1 to induce CreER-mediated HC loss and induce spontaneous HC regeneration, as well as administered doxycycline (Dox) in the diet to nursing mother with a dox injection given to pups at P1 to induce expression of N1ICD and the LacZ reporter in SCs and all non-HCs. Cochlea were collected at P7 and stained with antibodies against myosin VIIa (red), Sox2 (white), and β-gal (green). Fewer fate-mapped regenerated HCs (LacZ-positive/myosin VIIa-positive cells, white arrows) were observed when N1ICD was overexpressed in combination with HC damage compared to samples with HC damage and normal Notch signaling. (D–H,L–P) High magnification images from boxes in (C,K) showing LacZ-positive/Sox2-positive/myosin VIIa-positive regenerated HCs. Scale bar = 25 μm. (Q) Quantification of LacZ-positive HCs from the images presented in (A–P) and the two additional controls. Comparison of the number of LacZ-positive HCs between the group with spontaneous HC regeneration and normal Notch signaling (blue, Atoh1-DTA::Sox10^rtTA::TetO-LacZ mice) and the experimental group with N1ICD overexpression as well as HC damage (orange, Atoh1-DTA::Sox10^rtTA::TetO-LacZ::TetO-N1ICD), showed a ∼92% decrease in the number of LacZ-positive, regenerated HCs. (R) LacZ-positive HCs after HC damage with physiological Notch signaling (blue, Atoh1-DTA::Sox10^rtTA::TetO-LacZ) decreased in an apical to basal gradient while LacZ-positive HCs were only detected in the base when N1ICD was overexpressed without HC damage (green, Sox10^rtTA::TetO-LacZ::TetO-N1ICD) mice. Segment 1 is the most apical region and segment 6 is the most basal. N = 3–4. ^∗p < 0.05, ^∗∗∗p < 0.001 as determined by a two-way ANOVA with a Tukey’s post hoc test. N = 4.