Developmental and Functional Hair Cell-Like Cells Induced by Atoh1 Overexpression in the Adult Mammalian Cochlea In Vitro

Hair cells (HCs) in the mammalian cochleae cannot spontaneously regenerate once damaged, resulting in permanent hearing loss. It has been shown that Atoh1 overexpression induces hair cell-like cells (HCLCs) in the cochlea of newborn rodents, but this is hard to achieve in adult mammals. In this study, we used a three-dimensional cochlear culture system and an adenoviral-mediated delivery vector to overexpress Atoh1 in adult mouse cochleae. HCLCs were successfully induced from 3 days after virus infection (3 DVI) in vitro, and the number increased with time. HCLCs were myosin7a positive and distinguishable from remnant HCs in a culture environment. Meanwhile, patch-clamp results showed that noninactive outward potassium currents (sustained outward potassium currents) could be recorded in HCLCs and that their magnitude increased with time, similar to normal HCs. Furthermore, transient HCN currents were recorded in some HCLCs, indicating that the HCLCs experienced a developmental stage similar to normal HCs. We also compared the electrophysiological features of HCLCs from adult mice with native HCs and found the HCLCs gradually matured, similar to the normal HCs. Meanwhile, HCLCs from adult mice possessed the same bundles as developmental HCs. However, these HCLCs did not express prestin, which is a special marker for outer hair cells (OHCs), even at 13 DVI. These results demonstrate that Atoh1 overexpression induces HCLC formation in the adult mammalian cochlea and that these HCLCs were functional and experienced a developmental process similar to that of normal HCs.


Introduction
Hair cells (HCs), which serve as inner ear sensory cells, are essential for the transduction of mechanical stimulation into hearing and balance signals in the inner ear. In lower vertebrates, such as birds and zebrafish, lost HCs can be replaced by spontaneous HC regeneration [1][2][3][4]. However, unlike these vertebrates, following acoustic overstimulation, ototoxic drug, trauma, or aging, HC loss in the mammalian inner ear is irreversible and is the primary reason for permanent sensorineural hearing loss [5][6][7][8].
HC reproduction in mammals has made great process in recent years, including the intervention of stem cells, supporting cell (SC) reprogramming, and the direct conversion of nonsensory epithelial cells of the cochlea [9][10][11][12][13][14][15][16]. Atoh1 is a basic helix-loop-helix transcription factor that is both necessary and sufficient for the development and genesis of HCs [17]. Atoh1 is gradually upregulated before HC formation and is downregulated during the HC maturation process [17][18][19]. The absence of Atoh1 in embryonic mice induces complete loss of both HCs and associated SCs [17,19,20]. In contrast, Atoh1 overexpression induces the formation of HCLCs both in vitro and in vivo [10,[21][22][23]. Meanwhile, it appears that HCLCs induced by different approaches remain in the developmental stage, which is a limiting factor for functional recovery [10,22]. A recent study has shown that Atoh1 overexpression significantly enhanced HC regeneration and sustained functional recovery in the adult vestibule [11]. Meanwhile, another recent study also showed that multiple reprogramming, including Atoh1 overexpression, induces HCLCs in the adult cochlea in vitro and in vivo [24]. However, whether HCLCs induced by Atoh1 possess similar electrical physiological function as normal HCs is still unknown.
To address these questions, a human adenovirus serotype 5 (Ad5) vector encoding Atoh1 (Atoh1) and/or RFP (RFP-Atoh1-) or RFP-Atoh1 (RFP-Atoh1+) was used to overexpress Atoh1 in cultured mouse cochlea in vitro, and we assessed the morphological and functional features of HCLCs at different days after viral infection (DVI). Patch-clamp results showed the similarities and differences with respect to electrophysiological features between HCLCs and native HCs at different time points.

Cochlear Explant
Cultures and Viral Infection. Adult C57BL/6 mice (JSJ Company, Shanghai, China), ranging from 8 to 10 weeks of age, were used in this study. Mice were decapitated, and the cochleae were carefully dissected out. Then, under a dissecting microscope (Zeiss), a hole was perforated at the apex of the cochlea, and the round window was opened and expanded by gently rotating the tip using a syringe needle. Using a fine-tipped pipette, phosphate buffer solution (PBS, HyClone) was gently flushed through the oval and round windows (as inlets) and the hole at the apex (as an outlet) (Figure 1(a)). Finally, the whole cochlea with the sensory epithelium was cultured in this experiment in Dulbecco's modified Eagle medium as follows: nutrient mixture F-12 (DMEM/F12, Invitrogen) medium containing 10% fetal bovine serum (FBS, Invitrogen) was used on the first day of culture.
The human adenovirus serotype 5 (Ad5) vector (with a final concentration of 1.0 × 10 8 pfu/ml) encoding RFP and/or Atoh1 (RFP-Atoh1+/-) was used in this study. On the next day of culture, the virus diluted in the serum-free culture medium supplemented with B27 (Invitrogen) was added for a 12-hour infection and then replaced with serum-free medium every two days until the end of the culture. Petri dishes were kept in an incubator at 37°C with 5% CO 2 during culture. The moment of adenovirus addition was recorded as day 0, then 1 day after viral infection (1 DVI), and so on.

Tissue Preparation and Immunofluorescence Staining.
After 3 and 7 DVI, specimens were fixed in 4% paraformaldehyde (in PBS) at room temperature (RT) for 30 min followed by 10% ethylenediaminetetraacetic acid (EDTA) at RT for 8 hours to decalcify the bones. The bony capsule of the cochlea was dissected under the microscope. Then, the basilar membrane and associated organ of the Corti were carefully stripped from the modiolus. Next, the tectorial membrane and stria vascularis were torn off. The apical turn of the basilar membrane was prepared for immunofluorescence staining.

Statistical
Analysis. Statistical analyses were performed using GraphPad Prism (v8.0c software, GraphPad, San Diego, CA, USA). Statistical significance was determined using Student's t-tests and two-way ANOVA followed by Bonferroni's multiple comparisons test. Electrophysiological data were analyzed using Igor (HEKA Instruments, USA) and Origin Pro 8.0 (Microcal Software, USA). Nonlinear regression analysis was performed to quantify the  In this study, n values represent the number of biologically independent samples. Data are shown as the mean ± s:d:, and p values less than 0.05 were considered significant: * p < 0:05, * * p < 0:01, and * * * p < 0:001.

Atoh1
Overexpression Induces HCLC Formation in the Sensory Epithelial Region in the Apical Turn of Adult Mouse Cochleae. As shown in Figures 1(a)-1(c), myosin7a+ HCLCs were detected in the sensory region in the Atoh1 overexpression group (RFP-Atoh1+) from 3 DVI at the apical turn. HCLCs were distinguishable from the remnant native HCs. In the control group, myosin7a+ native HCs were shrunken and relatively well arranged at 3 DVI (Figure 1(c)). In the Atoh1 overexpression group, myosin7a+ HCLCs (arrows) were much larger and rounder than the remnant native HCs (Figure 1(b)). At the middle and basal turns, the viral infection rate was low, and only a small amount of SCs was RFP positive with no HCLCs detected (Figures 1(b) and 1(c)).

Atoh1-Induced HCLCs Possess Bundles Similar to Developmental HCs but Do Not Express Prestin in Mature
Cochleae. Phalloidin was stained to detect whether HCLCs possess similar bundles as developmental HCs. According to our results, HC bundles were observed at the apical side of HCLCs (Figures 3(b) and 3(c)). Moreover, we used prestin, a specific motor protein expressed in the OHCs, to identify HCLCs at 7 and 11 DVI. As shown in Figure 3(a), although some shrunken OHC debris still expressed prestin with pyknotic nuclei, HCLCs were prestin negative both at 7 (n = 5) and 11 DVI (n = 3). In the control group, there were no HCLCs and only some scattered HC debris was prestin positive (n = 5) (Figure 3(a)).

Outward K+ and HCN Currents of Atoh1-Induced
HCLCs in Mature Cochleae. A whole-cell patch-clamp recording technique was used to assess electrophysiological properties of HCLCs. Distinguishable among the HCLCs were remnant native HCs and nonsensory epithelial cells in the HC region. For native IHCs, we could see V-shaped bundles in a relative orderly arrangement. However, in native OHCs, the condition was too deteriorated to patch successfully (data not shown). For native IHCs, the condition was better than in native OHCs. For HCLCs, which were round and large, shown in red fluorescence and in better condition, they exhibited more durability in extracellular solutions dur-ing the patch time. Nonsensory epithelial cells in the sensory epithelium are finger shaped.

Comparison of Electrophysiological Properties of HCLCs from Adult Mice with Normal HCs.
To compare electrophysiological properties of HCLCs and normal HCs, outward K+ currents of HCLCs and OHCs in the developmental stage were measured. Although recorded currents from these cells seemed similar to HCLCs, the mean membrane capacitance (Cm) was different, as shown in Figures 5(c) and 5(d). First, the sustained outward K+ currents of the inner and outer HCs of wild-type P9 mice were recorded, which were 2:09 ± 0:57 nA (n = 5) and 1:84 ± 0:21 nA (n = 5), respectively ( Figure 5(c)). On 3 DVI, the mean amplitude of the current density of HCLCs showed no statistical difference from the inner and outer HCs. Another two groups (5 and 7-9 DVI) were higher than IHCs in P9 mice (p < 0:05); however, they still showed no statistical significant compared to OHCs. Second, the K+ currents of the inner and outer HCs in adult (P60) cochleae were also recorded, which were 2:73 ± 0:64 nA (n = 5) and 2:24 ± 0:28 nA (n = 5), respectively ( Figure 5(d)). The mean amplitude of current density in the three groups (3, 5, and 7-9 DVI) was higher than the HCs in the adult cochlea (p < 0:01).

Discussion
In recent years, great progress has been made in HC regeneration in neonatal and adult cochleae [10,11,16,[24][25][26][27]. Recent studies have shown that sustained Atoh1 4 Neural Plasticity overexpression enhances the maturation of regenerated HCs in the damaged vestibular system [11,12,28]. HCs induced by Atoh1 in neonatal mammals are reminiscent of "ancestral" HCs that have commonalities with vestibular HCs [11,29]. However, initial Atoh1 overexpression levels are correlative with mature levels and arrangement of HCLCs [30]. Higher initial Atoh1 overexpression levels induce more mature HCLCs according to electrophysiological results [30]. In this study, Atoh1 overexpression induced HCLCs in adult mice in vitro. On 3 DVI, several HCLCs were myosin7a positive in the sensory region, especially in the IHC region. Some HCLCs overlapped with RFP fluorescence, and some did not, consistent with previous studies [30,31], due to Atoh1 expression level decreasing once HCLCs become mature. Meanwhile, HCLCs were distinguishable from remnant HCs. HCLCs exhibited a much rounder, bigger, and fresher appearance. Remnant HCs seemed to shrink and be in a dying state. The number of HCLCs increased from 3 DVI to 7 DVI. HCLCs at 7 DVI still failed to express prestin, a marker of mature OHCs. This is likely because the HCLCs were not mature enough, or perhaps the HCLCs were IHC-like cells. The HCLCs possessed bundles at the apical aspect similar to the developmental HCs, indicating incomplete maturity.
Compared to ectopic HCLCs induced from the neonatal inner ear, HCLCs from adult mice were primarily in the sensory region, especially the IHC region. Ectopic HCLCs induced from the neonatal inner ear are primarily from the nonsensory region, including the LER (lesser epithelial region) and GER (greater epithelial region) [29,[31][32][33]. This is mostly due to the LER and GER cells being stretched in the culture environment, which facilitates viral infection. For the adult cochlear culture, we reserved the cochlear bone, and the sensory region was the infection area. The IHC region in adult sensory epithelium presents a relatively loose structure, facilitating adenovirus infection. In the control group, no HCLCs were detected. Remnant HCs died in the culture environment, and cell debris was also myosin7a positive and even prestin positive. HCLCs in situ may directly build connections with the tectorial membrane and associated neurons, which may lead to eliciting mechanical transduction for hearing restoration.

Neural Plasticity
In the cochlea, Atoh1-induced HCLC formation occurs primarily through direct transdifferentiation due to the limitations of the proliferation of SCs [34]. In our study, there were no HCLCs detected in the control group, indicating HCLC formation depends on Atoh1 overexpression, and there was no HC regeneration in the adult cochlea, distinct from the adult vestibule.
From embryonic development to maturation, HCs undergo an extremely complicated maturation process [7,29,[35][36][37][38]. In addition to a series of related protein expression, the expression of outward K+ and HCN currents in HCs also goes through developmental changes. In the developmental stage, cochlear HCs express nondeactivated outward K+ current (sustained outward K+ current) and do not express fast deactivated outward K+ current (transient outward K+ current), and the amplitude increases with the time as HCs mature [35,37,39,40]. Meanwhile, cochlear OHCs and IHCs transiently express HCN currents [36,39].
A previous study reported that after Atoh1 overexpression, a fast deactivated outward potassium rectifier (I K ) first appeared in most of the ectopic HCs, followed by sustained outward K+ current appearing later, and the proportion of cells with the sustained current increased over time as the HCLCs matured [29]. This means that HCLCs with no deactivated outward K+ current are more mature than HCLCs with fast-deactivated outward potassium. Similarly, HCN current first appeared in most of the ectopic HCs, then disappeared, and the proportion of cells with HCN current decreased over time as the HCLCs matured [29], indicating that HCLCs with HCN current are more immature. Thus, these ectopic HCLCs seem primordial. Another study reported that initial Atoh1 overexpression levels define the maturity level and arrangement of HCLCs, with higher initial Atoh1 overexpression levels inducing more mature HCLCs according to electrophysiological results, which process the higher proportion of sustained outward K+ current and lower proportion of HCN current [25].
For HCLCs induced by Atoh1 in adult cochlea, no transient outward current was recorded, and all of the HCLCs expressed sustained outward K+ current, distinct from  We also compared HCLCs from adult cochleae at different time points with normal inner and outer HCs at P9 and P60. For mouse IHCs, a very small outward K+ current debuts at E14.5. Later, spontaneous action potentials (SAPs) occur at E17.5, and the magnitude of outward K+ currents increases steadily until the onset of hearing (about P12-14) [35,37,41]. Therefore, the magnitude of outward K+ currents in IHCs of P9 cochlea is on the rise. Judging from our research results, the maturity of HCLCs on 3 DVI was similar to IHCs in P9 cochlea. However, at 5 and 7-9 DVI, the HCLCs from adult cochlea were more mature than IHCs in the P9 cochlea ( Figure 5(c)). Maturation of OHCs is achieved at about P8, a little earlier than in IHCs. In our study, the magnitudes of outward K+ currents in HCLCs at all time points were similar to normal OHCs in P9 cochlea. Furthermore, HCLCs from adult cochlea expressed sustained outward K+ current without the transient outward currents observed in native HCs. In general, these results demonstrated that outward K+ currents in HCLCs process a developmental regularity and synchrony, the same as normal HCs.
HCN currents, also known as inward rectifier (Ih) currents, have been largely recorded in immature HCs and regenerated ectopic HCs [36,38,41]. Similar to IK1 current in developing HCs, the function of the HCN current is setting the resting potential and regulating potential repolarization [36]. The inwardly transient currents will disappear abruptly when the HCs reach maturity, between approximately P12 and P14 in IHCs and earlier in OHCs [35,36]. In our study, transient expression of HCN currents from 3 to 9 DVI mimicked the developmental process as that found in normal HCs. Additionally, the time point that HCN currents disappeared was on 9 DVI, about one week earlier than normal

Neural Plasticity
HCs and ectopic HCLCs induced from the neonatal inner ear, strongly indicating that functional development of the HCLCs was occurring relatively faster. The cell state of remnant IHCs and OHCs degenerated in the culture environment. For remnant IHCs in the culture, only small sustained outward current could be recorded. For the OHC remnants in the culture environment, we were unable to even patch the cells.

Conclusions
In this study, HCLCs were induced in adult mouse cochleae by Atoh1 overexpression, and they were distinguishable in the sensory region. HCLCs expressed sustained outward K+ currents, and the magnitude increased with time similar to the developmental HCs. They also transiently expressed HCN channel currents similar to HCs, mimicking the developing stage. Meanwhile, HC bundles were detected in HCLCs in a developmental process, but they did not express prestin. These findings indicate that HCLCs from adult mice induced by Atoh1 overexpression undergo a similar developmental process as that of normal HCs, thereby providing guideline for future research into hearing restoration.

Data Availability
All data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no competing interests.