Notch pathway inhibitor DAPT enhances Atoh1 activity to generate new hair cells in situ in rat cochleae.

Atoh1 overexpression in cochlear epithelium induces new hair cell formation. Use of adenovirus-mediated Atoh1 overexpression has mainly focused on the rat lesser epithelial ridge and induces ectopic hair cell regeneration. The sensory region of rat cochlea is difficult to transfect, thus new hair cells are rarely produced in situ in rat cochlear explants. After culturing rat cochleae in medium containing 10% fetal bovine serum, adenovirus successfully infected the sensory region as the width of the supporting cell area was significantly increased. Adenovirus encoding Atoh1 infected the sensory region and induced hair cell formation in situ. Combined application of the Notch inhibitor DAPT and Atoh1 increased the Atoh1 expression level and decreased hes1 and hes5 levels, further promoting hair cell generation. Our results demonstrate that DAPT enhances Atoh1 activity to promote hair cell regeneration in rat cochlear sensory epithelium in vitro.

Introduction

Sensory hair cells in the mammalian cochlea convert mechanical stimuli into electrical impulses that allow hearing (Fuchs et al., 2003; Grant et al., 2011; Fujikawa et al., 2014). Loss of hair cells in lower vertebrates causes the surrounding support cells to re-enter the cell cycle and regenerate new hair cells (Corwin and Cotanche, 1988; Ryals and Rubel, 1988); however, lost sensory hair cells in the mammalian organ of Corti do not regenerate spontaneously (Izumikawa et al., 2005; Yang et al., 2012).

The basic helix-loop-helix transcription factor, atonal homolog 1 (Atoh1), plays a critical role in hair cell development and regeneration (Bermingham et al., 1999; Zheng and Gao, 2000; Chen et al., 2002; Kawamoto et al., 2003; Woods et al., 2004). Overexpression of Atoh1 induces hair cell formation in the inner ear of mammals (Izumikawa et al., 2005; Staecker et al., 2007; Gubbels et al., 2008; Huang et al., 2009; Han et al., 2010; Kelly et al., 2012; Pan et al., 2012; Chonko et al., 2013; Atkinson et al., 2014; Luo et al., 2014; Gao et al., 2016). Newly generated hair cells were mainly located in the lesser epithelial ridge and the greater epithelial ridge in rat cochleae (Zheng and Gao, 2000; Yang et al., 2012, 2013; Luo et al., 2014, 2017). In addition, spiral ganglion neurons in sensory regions extended dendrites toward the ectopic hair-cell-like cells (Yang et al., 2012) and formed afferent synapses with ectopic hair-cell-like cells (Luo et al., 2017), which is important for restoration of hearing.

The Notch signaling pathway is vital in determining the fate of hair cells and supporting cells, by lateral inhibition, during inner ear development (Lanford et al., 1999; Zine et al., 2000; Daudet and Lewis, 2005; Yamamoto et al., 2006; Kopan and Ilagan, 2009; Mizutari et al., 2013). Pharmacological inhibition of Notch signaling by γ-secretase inhibitors leads to the generation of supernumerary hair cells in situ in cultured embryonic or neonatal cochleae (Mizutari et al., 2013; Li et al., 2015; Ni et al., 2016; Ren et al., 2016), which mainly involves direct transdifferentiation of supporting cells (Yamamoto et al., 2006; Doetzlhofer et al., 2009; Zhao et al., 2011; Mizutari et al., 2013). A combination of different signals has been shown to promote in situ cochlear hair cell regeneration after neomycin insult in mice (Doetzlhofer et al., 2009; Zhao et al., 2011; Masuda et al., 2012; Chen et al., 2013; Ikeda et al., 2015; Ni et al., 2016; Ren et al., 2016; Walters et al., 2017). However, generation of new hair cells in rat cochlea in situ is difficult due to the low infection efficiency of adenovirus in sensory epithelium. The induction of sufficient numbers of new hair cells in rat cochlea in situ to replace the lost hair cells could facilitate the formation of synapses between new hair cells and the remaining spiral ganglion cells (Zhao et al., 2011; Chen et al., 2013; Ikeda et al., 2015). Stimulating the formation of a sufficient number of hair cells in situ would be a useful way to promote hearing recovery in the future.

To increase the infection efficiency of adenovirus in the sensory epithelium, we applied a high concentration of fetal bovine serum and prolonged tissue culture prior to viral infection to increase the width of the sensory epithelium. We successfully generated new hair cells after viral gene transfer of Atoh1 in situ, but the number was insufficient. Concurrent application of the γ-secretase inhibitor N-[N-(3,5-Difluorophenacetyl)-L-alanyl]-S-phenylglycine t-butyl ester (DAPT) with Atoh1 overexpression increased the rate of supporting cell conversion and produced a greater number of new hair cells in situ.

Materials and Methods

Animals

A total of fifty-five one-day-old (P1) Sprague-Dawley rat pups, weighing 5 g and of both sexes, were provided by the Shanghai SLAC Laboratory Animal Co., Ltd. (Shanghai, China). The protocol for animal use was reviewed and approved by the Institutional Animal Care and Use Committee of Eye & ENT Hospital of Fudan University of China. The experimental procedures followed the United States National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1985).

For measuring the width of sensory epithelium in organs of Corti, rat pups were randomized into four groups: neonatal day 1 (n = 5); 5% fetal bovine serum (FBS) for 24 hours (n = 5); 5% FBS for 48 hours (n = 5); 10% FBS for 48 hours (n = 5). The rat pups were randomized into five groups for measuring hair cell numbers or relative mRNA expression levels in sensory epithelium: neomycin-green fluorescent protein (neo-GFP) (n = 5), neo-dimethyl sulfoxide (neo-DMSO) (n = 5), neo-DAPT (n = 5), neo-Atoh1 (n = 5), and neo-Atoh1-DAPT (n = 5).

Cochlear cultures and Atoh1 infection

Rats were sacrificed by CO2 asphyxiation and cochleae dissected. Only the middle-basal turns were explanted onto poly-L-lysine (Sigma, St. Louis, MO, USA) coated glass coverslips in Dulbecco's modified Eagle medium/nutrient mixture F12 (DMEM/F12, Life Technologies, Carlsbad, CA, USA) containing 5% FBS (Life Technologies).

On the second day in vitro, 1.5 mM neomycin was applied for 24 hours. On the third day in vitro, the medium was replaced with one of the following: DMEM/F-12 containing 5% FBS for 24 hours (the 5% FBS for 24 hours group), DMEM/F-12 containing 5% FBS for 48 hours (the 5% FBS for 48 hours group), or DMEM/F-12 containing 10% FBS for 48 hours (the 10% FBS for 48 hours group). On the fifth day in vitro, all of the media were replaced with serum-free DMEM/F-12 supplemented with B27 (Life Technologies). This culture medium was replaced every other day. Ad5-GFP-Atoh1 or Ad5-Atoh1 (Huang et al., 2009; Han et al., 2010; Yang et al., 2012, 2013) was co-incubated with the cochlear explants on the fifth day in vitro for 20 hours, to overexpress Atoh1 in the cochlear segments, and the control group was infected with Ad5-GFP. For the DAPT experiments, DAPT was added to the medium 24 hours after adding Ad5-vector and the culture medium containing DAPT was replaced every other day until the explants were fixed. Culture medium containing 0.1% DMSO was used as a control. The culture dishes were placed in a 37°C incubator containing 5% CO2 in a humid environment.

The final concentration of the Ad5 vector was 1.6 × 108 PFU/mL in serum-free DMEM/F-12 medium. The viruses used in the present experiment have been described in previous reports (Huang et al., 2009; Han et al., 2010; Yang et al., 2012, 2013). The neomycin sulfate stock (10 mg/mL in 0.9% NaCl, purchased from Sigma-Aldrich) was diluted in culture media to 1.5 mM. DAPT (catalogue number 565784, EMD Chemicals, Darmstadt, Germany) was diluted in 100% DMSO to a stock concentration of 10 mM and was used at a working concentration of 10 μM.

Tissue preparation and immunofluorescence analysis

The cochlear explants were fixed with 4% paraformaldehyde for 30 minutes and then treated with 0.1% Triton X-100 plus 10% donkey serum for 30 minutes. This procedure was followed by incubation for 24 hours at 4°C with the following primary antibodies: rabbit-anti-myosin 7A (1:100; Proteus Biosciences Inc., Ramona, CA, USA) and goat-anti-Sox2 (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA). The preparation was washed three to five times with PBS and then incubated with secondary antibodies for 2 hours at 37°C in the dark. The secondary antibodies included donkey anti-mouse/rabbit Alexa Fluor 555 (1:1,000; Jackson ImmunoResearch, West Grove, Philadelphia, PA, USA) and donkey anti-mouse/rabbit/goat (H + L) Alexa Fluor 647 (1:1,000; Jackson ImmunoResearch). The specimens were incubated with rhodamine phalloidin (1:1,000; Life Technologies) for 1 hour at room temperature after the secondary antibody was washed out. The specimens were visualized with a Zeiss LSM 510 confocal laser scanning microscope (Zeiss, Germany). All of the images were digitally processed and evaluated using Adobe Photoshop CS5 software. Images were acquired with a pixel size of 0.03 µm × 0.03 µm × 0.30 µm following Nyquist sampling, with no pixel saturation to ensure that no structural information was lost.

Measurement of the specimens

We measured the distance from the outermost Sox2-immunoreactive cells to the innermost Sox2-immunoreactive cells in organ of Corti area 24 and 48 hours post FBS culture to determine the width of the sensory epithelium region. Each group included at least five different cochlear explants, and each explant was sampled in 10 randomly chosen areas.

To evaluate the number of GFP- and/or myosin 7A-immunoreactive cells, at least 10 cochleae in each experimental group were examined. Only the cells located in the mid-basal turn of the organ of Corti were counted. We counted the myosin 7A-immunoreactive cells in each 200 μm segment along the length of the cochlear sensory epithelium and randomly sampled five areas in each cochlea. Myosin 7A-immunoreactive cells were counted 8 days after the virus infection. When calculating the infection rate, we counted the Sox2-immunoreactive cells and GFP-Sox2 double-immunoreactive cells in the sensory epithelium. We pooled the number of Sox2-immunoreactive cells in the sensory epithelium in five areas of each cochlea. When calculating the conversion rate, we counted the GFP-immunoreactive cells and GFP-myosin 7A double-immunoreactive cells in the sensory epithelium. We pooled the number of GFP-immunoreactive cells in the sensory epithelium in five areas of the cochlea.

BrdU-incorporation assay to test cell proliferation

To test the proliferation of supporting cells, BrdU (Sigma) was used at a final concentration of 10 μM. We added the BrdU when we first changed the medium after Atoh1 infection and kept it in the medium until the end of the culture period. BrdU incorporation was detected at 7 days post infection using an mouse-anti-BrdU antibody at a concentration of 1:400 (Sigma). The cultured preparations were incubated with 2 M HCl for 30 minutes before permeabilization using Triton-X 100.

RNA extraction and quantitative real-time reverse transcriptase-polymerase chain reaction (qRT-PCR)

For qRT-PCR, RNA was purified from four explanted cochlea of each group 72 hours after Ad5-GFP-Atoh1 infection using an RNeasy Plus Micro extraction Kit (Code No. 74034; Qiagen, Germany) and reverse-transcribed using a High Capacity RNA-to-cDNA kit (Code No. RR036A; TaKaRa, Japan). The qRT-PCR was performed in duplicate using a PreMix SYBR Green kit (Code No.RR420A; TaKaRa, Japan) on a 7500HT Fast Real-Time PCR System (Applied Biosystems, Waltham, MA, USA). All qRT-PCR reactions were performed in triplicate and relative quantification of gene expression was analyzed using the 2−ΔΔCT method (Livak and Schmittgen, 2001) with the house keeping gene β-actin as the endogenous reference.

Primers used for genotyping are listed as follows:

Statistical analysis

All data are presented as the mean ± SEM. Statistical analyses were conducted using Microsoft Excel and GraphPad Prism 6.0 software (GraphPad Software, USA). One-way analysis of variance followed by least a significant difference post hoc test was used to analyze the data; P < 0.05 was considered statistically significant.

Results

Atoh1 overexpression mainly induced the formation of ectopic hair cells

The explants were cultured in medium containing 5 % FBS for 24 hours before adenovirus infection. Eight days after Ad5-GFP-Atoh1 infection, GFP-Myosin 7A double immunoreactive cells were observed to be restricted to the lesser epithelial ridge, which was external to the outer hair cells (). The greater epithelial ridge was also a main target of the adenovirus (data not shown), which was consistent with the findings of previous studies (Zheng and Gao, 2000; Woods et al., 2004; Yang et al., 2012, 2013). For the GFP alone control, no myosin 7A-GFP double-immunoreactive cells were detected in the lesser epithelial ridge (), suggesting that the formation of these hair cells was dependent on Atoh1 expression. The GFP-Atoh1 virus vector failed to infect the sensory epithelium, thus no new hair cells were found in situ (, white bracket). We also used neomycin to destroy the original hair cells to increase the virus infection rate, but this was not effective. Although the original hair cells were killed by neomycin, none of the supporting cells were immunoreactive for GFP after virus infection (, white bracket), indicating that cells in sensory epithelium were hard to infect with adenovirus.

Figure 1

Atoh1 overexpression induced ectopic hair cell formation in the lesser epithelial ridge.

Representative images of myosin7A staining and immunofluorescence were observed using a Zeiss confocal microscope. We cultured cochlear explants of neonatal rats in medium containing 5% fetal bovine serum for 24 hours then infected them with Ad5-GFP-Atoh1. (A) Eight days after Atoh1 infection, GFP-Myosin 7A double immuno-reactive cells were observed to be restricted to the lesser epithelial ridge. (B) Magnified view of (A): White arrows point to the newly generated hair cells. (C) Eight days after infection with GFP alone, no myosin 7A-GFP double-immunoreactive cells could be detected in the lesser epithelial ridge. The virus vector failed to infect the sensory epithelium (white brackets) and no Myosin 7A-immunoreactive cells were found in situ. Green: GFP-immuno-reactive cells; blue: myosin7A-immunoreactive hair cells. Scale bars: 100 μm in A, C; 50 μm in B. Atoh1: Atonal homolog 1; Ad5: adenovirus type 5; GFP: green fluorescent protein.

Figure 2

Adenovirus infection rate in the organ of Corti depends on width of the sensory epithelium.

Representative images of myosin7A and Sox2 staining and immunofluorescence were observed using a Zeiss confocal microscope. (A) The organ of Corti was well organized when samples were fixed immediately after dissection. (B) Cochlear explants were cultured in medium containing 5 % FBS for 24 hours after neomycin insult and were then infected with Ad5-GFP. (C) Cochlear explants were cultured in medium containing 5 % FBS for 48 hours after neomycin insult and were then infected with Ad5-GFP. (D) Cochlear explants were cultured in medium containing 10 % FBS for 48 hours after neomycin insult and were then infected with Ad5-GFP. Many GFP-immunoreactive cells were observed in the sensory epithelium. (E) The bar graph shows the width of sensory epithelium. (F) The bar graph shows the infection rate of Ad5; *P < 0.05, vs. 5 % FBS for 24 hours (mean ± SEM, n = 3, one-way analysis of variance followed by least significant difference post hoc test). Green: GFP-immunoreactive cells; blue: myosin7A-immunoreactive hair cells; red: sox2-immunoreactive supporting cells. Scale bars: 50 μm in A–D. White brackets: Sensory epithelium; white arrows: supporting cells. FBS: Fetal bovine serum; Ad5: adenovirus type 5; GFP: green fluorescent protein; h: hours.

Increasing the width of the sensory epithelium increased the adenovirus infection rate in the organ of Corti

The middle-basal turns of the cochlear explants were treated with 1.5 mM neomycin to kill all of the normal hair cells, as described in the methods section, and then were infected with Ad5-GFP. In samples that were fixed 8 days after virus infection, few myosin 7A-immunoreactive cells were observed (Figure –), and Sox2-immunoreactive nuclei in supporting cells were present at lower levels (Figure –), indicating that treatment with 1.5 mM neomycin for 24 hours could destroy all of the normal hair cells of the middle-basal turn in culture. When the cultured explants were incubated with medium containing 5% FBS for 48 hours before adenovirus infection, GFP-immunoreactive cells were also seldom detected in the sensory epithelium (). When we increased the FBS concentration from 5% to 10% and prolonged the FBS application from 24 to 48 hours, many GFP-immunoreactive cells were observed in the sensory epithelium ().

We next investigated whether the adenovirus virus infection rate in the sensory epithelium depends on the width of sensory epithelium. In the 5% FBS for 24 hours group, the average width of organ of Corti was similar to that in the non-cultured explants (P > 0.05). There were no GFP-immunoreactive cells in the sensory epithelium. The average width of the organ of Corti in the 5% FBS for 48 hours group was significantly wider than that in the no culture group (P < 0.05), and 28 ± 12% of the cells in the sensory epithelium were immunoreactive for GFP. In the 10% FBS for 48 hours group, the width of the sensory epithelium was significantly wider than that in the other groups (P < 0.05). The supporting cells in the group were fully extended and 88 ± 10% of the cells in the sensory epithelium were immunoreactive for GFP. These results implied that increasing the FBS concentration and prolonging the duration of culture increased the width of sensory epithelium and this also increased the adenovirus infection rate.

DAPT enhanced Atoh1 transcription to increase the quantity of new hair cells in situ

To determine whether Atoh1 could induce the formation of new hair cells in situ, 1.5 mM neomycin was applied to kill almost all of the normal inner and outer hair cells and Ad5-GFP or Ad5-GFP-Atoh1 was applied after 48 hours of culture in 10% FBS medium. After infection with only Ad5-GFP, few (Figure , ) myosin 7A-immunoreactive cells were detected in the sensory epithelium and no myosin 7A-immunoreactive cells were detected in the lesser epithelial ridge. After infection with Ad5-GFP-Atoh1, both the sensory epithelium and lesser epithelial ridge contained myosin 7A-immunoreactive cells (Figure , ). Thus, the supporting cells in the sensory epithelium formed new hair cells when induced by Atoh1.

Figure 3

DAPT enhances Atoh1 activity to generate more new hair cells in situ.

Representative images of myosin7A and Sox2 staining and immunofluorescence were observed using a Zeiss confocal microscope. Neomycin (1.5 mM) was applied to destroy almost all of the normal hair cells and Ad5-GFP or Ad5-GFP-Atoh1 was applied after 48 hours of culture in 10 % FBS medium. (A) Cochlear explants were infected by Ad5-GFP after neomycin insult. No myosin 7A-immunoreactive cells were found in the sensory epithelium. (B) Cochlear explants were treated with Ad5-GFP and DAPT. No myosin 7A-immunoreactive cells were detected in the sensory epithelium. (C) Cochlear explants were infected by Ad5-GFP-Atoh1 after neomycin insult. Several GFP/myosin 7A double immunoreactive cells were observed in the sensory epithelium. (D) Cochlear explants with concurrent application of Ad5-GFP-Atoh1 and DAPT after neomycin insult. Numerous GFP/myosin 7A double immunoreactive cells were found in the sensory epithelium. (A1–D1) Magnified view from A–D. (E) The number of new hair cells in situ in the different groups (n = 10). (F) Quantitative polymerase chain reaction results for the different groups; *P < 0.05, vs. neo-GFP or the neo-DMSO group (mean ± SEM, one-way analysis of variance followed by the least significant difference post hoc test, n = 3 per group). Green: GFP-immunoreactive cells; blue: myosin7A-immunoreactive hair cells; red: sox2-immunoreactive supporting cells. Scale bars: 100 μm in A–D, 20 μm A1–D1. White arrows in C–D: New hair cells. Atoh1: atonal homolog 1; FBS: fetal bovine serum; Ad5: adenovirus type 5; GFP: green fluorescent protein; neo: neomycin.

We also applied DAPT alone after the normal hair cells were destroyed by neomycin; however, no myosin 7A-immunoreactive cells were detected (per 200 µm), indicating that DAPT alone could not induce the formation of new hair cells (Figure , ).

To determine whether DAPT could enhance Atoh1 activity and induce the formation of more new hair cells in situ, we combined DAPT application with Atoh1 infection and then counted the myosin 7A-immunoreactive cells in the sensory epithelium. When DAPT was applied after Atoh1 infection, the number of myosin 7A-immunoreactive cells per 200 µm in the sensory epithelium (Figure , ) was two times higher than without DAPT (P < 0.05). The results demonstrated that DAPT enhanced the activity of Atoh1, leading to the formation of many more new hair cells, by increasing the rate of conversion of supporting cells to hair cells.

Meanwhile, we detected the transcriptional levels of Atoh1, hes1 and hes5 in the different groups using qRT-PCR. DAPT significantly increased the Atoh1 expression level and decreased hes1 and hes5 levels in the DAPT alone and Atoh1-DAPT groups (P < 0.01; ). The Atoh1 expression level in the DAPT group was significantly lower than in the adenovirus-Atoh1 group, which was lower than in the adenovirus-Atoh1-DAPT group (P < 0.01; ).

DAPT combined with Atoh1 induced new hair cell formation via transdifferentiation

To investigate whether DAPT or Atoh1 affected the cell cycle of the supporting cells or promoted the proliferation of the supporting cells to generate new hair cells in situ, we administered BrdU from the time of DAPT and/or Ad5-GFP-Atoh1 application until the end of the culture, after neomycin insult. For the control group, we administered BrdU from the time Ad5-GFP infection after neomycin insult (). There were no Sox2-BrdU double immunoreactive cells in the sensory epithelium, suggesting the supporting cells did not proliferate in the culture conditions. We also observed no Sox2-BrdU double-immunoreactive cells or myosin-BrdU double-immunoreactive cells in the sensory epithelium when there was new hair cell formation in situ (Figure , ). This indicates that DAPT enhanced the Atoh1-mediated induction of new hair cell formation via transdifferentiation without prior proliferation, which differs from the process that occurs in lower vertebrates.

Figure 4

New hair cell formation in situ via transdifferentiation.

Representative images of BrdU and Sox2 staining and immunofluorescence were observed using a Zeiss confocal microscope. Neomycin (1.5 mM) was applied to destroy normal hair cells. The explants were cultured in medium containing 10% fetal bovine serum for 48 hours. (A) BrdU applied after neomycin insult. (B) BrdU applied with Ad-Atoh1. (C) BrdU applied with Ad-Atoh1 and DAPT. There were no Sox2-BrdU double immunoreactive cells in sensory epithelium in A–C, indicating that there was no proliferation. (A1–C1) Magnified views from A–C merged. Blue: Sox2-immunoreactive cells; green: myosin7A-immunoreactive cells; red: BrdU-immunoreactive cells. Scale bars: 100 μm in A–C, 20 μm in A1–C1. White arrows in B, C: New hair cells. Neo: Neomycin; Ad: adenovirus.

Compared with the cells in lesser epithelial ridge, the supporting cells were quiescent even in the high-FBS medium for a prolonged duration of culture (Figure –); thus, none of the supporting cells in the sensory epithelium were immunoreactive for BrdU. Further, Atoh1 expression did not force the supporting cells in the sensory epithelium to proliferate, as none of the supporting cells in the sensory epithelium was immunoreactive for BrdU ().

Discussion

Previous studies have shown that Atoh1 overexpression via adenovirus vector infection is a powerful method of inducing hair cell formation in mammals (Zheng and Gao, 2000; Kawamoto et al., 2003; Woods et al., 2004; Izumikawa et al., 2005; Gubbels et al., 2008; Huang et al., 2009; Zhao et al., 2011; Chen et al., 2013). Adenovirus-mediated Atoh1 overexpression has mainly focused on the lesser epithelial ridge and the greater epithelial ridge, where it induces ectopic hair cell regeneration (Bermingham et al., 1999; Kelly et al., 2012; Liu et al., 2012; Jin et al., 2013; Luo et al., 2017). The supporting cells in the sensory region of the rat cochlea are tightly organized and difficult to transfect, thus new hair cells were hardly produced in situ in rat cochlear explants (Bermingham et al., 1999; Luo et al., 2014, 2017). Inducing new hair cell formation in sensory epithelium after hair cell loss is an important approach to restore hearing, so the low adenovirus infection rate and low hair-cell transformation rate in the sensory epithelium are the problems that need to be solved.

In the present study, we chose the middle-basal turns of the cochlea and applied 1.5 mM neomycin for 24 hours to destroy almost all of the original hair cells. By enhancing the FBS concentration to 10% and prolonging the culture duration to 48 hours before virus infection, the width of sensory epithelium in the rat cochleae was significantly enhanced. Consequently, the adenovirus infection rate in the sensory region was greatly improved. The adenovirus serotype 5 vector that encoded Atoh1 infected the sensory region in situ in rat cochleae. We successfully generated new hair cells after viral gene transfer of Atoh1 in situ, but the number was insufficient.

When we added 10 μM DAPT in combination with Ad5-GFP-Atoh1 infection after neomycin treatment, many more GFP/myosin 7A double-immunoreactive cells were present than in cochleae without DAPT treatment. DAPT, a γ-secretase inhibitor, can block the Notch signaling pathway and reduce Hes1 and Hes5 expression, thereby eliminating their suppression of Atoh1 expression (Lanford et al., 1999; Zine et al., 2000; Zine et al., 2001; Zine and de Ribaupierre, 2002; Daudet and Lewis, 2005; Doetzlhofer et al., 2009; Mizutari et al., 2013). Concurrent application of DAPT with Atoh1 overexpression increased the rate of supporting cell conversion and produced a greater number of new hair cells in situ, which is a transdifferentiation process (Yamamoto et al., 2006; Kopan and Ilagan, 2009; Mizutari et al., 2013). Quantitative PCR results showed that DAPT enhanced the Atoh1 transcript levels as well as inhibiting Hes1 and Hes5 transcript levels. Thus, DAPT treatment significantly enhanced Atoh1 activity and greatly improved the rate of new hair cell formation through the conversion of supporting cells to hair cells, producing many more hair cells in situ.

The application of 10 μM DAPT alone could not induce new hair cell formation in the sensory epithelium and lesser epithelial ridge. According to our quantitative PCR results, although DAPT application increased the Atoh1 expression level, it was insufficient to induce hair cell formation. Thus, hair cell formation is Atoh1 level-dependent and Atoh1 expression is vital for hair cell formation.

In conclusion, our in vitro study demonstrated that increasing the width of the sensory epithelium enhanced the adenovirus infection rate and supporting cells possess the ability to transdifferentiate into hair cells in situ. DAPT alone did not induce hair cell formation in situ or in the lesser epithelial ridge, as the Atoh1 expression level was not high enough. With Atoh1 infection, DAPT further enhanced the Atoh1 expression level to increase the conversion rate and promote the generation of more hair cells via transdifferentiation, which will lay foundation for the hair cell regeneration in vivo. Thus, the synergetic effect of Atoh1 overexpression and γ-secretase inhibitors is a promising approach for hearing recovery in the future.