Adult mammals cannot regenerate cochlear hair cells, so hair-cell death causes permanent hearing loss. Deng and Hu (2020) showed that local delivery of the DNA methyltransferase inhibitor 5-azacytidine (5-aza) regenerates hair cells in chemically deafened adult mouse cochleae, and that the new hair cells derive at least in part from Sox2-positive supporting cells. The study left three issues explicitly unresolved: whether depleting supporting cells to make hair cells damages the organ of Corti and whether any deficit is compensated by supporting-cell cell-cycle reentry; whether the regenerated hair cells are functionally competent; and by what molecular mechanism DNA demethylation drives the supporting-cell-to-hair-cell transition. A 2022 follow-up by the same group largely settled basic functional competence (regenerated cells express prestin, espin, and ribbon-synapse proteins, and auditory thresholds partially recover), and a 2023 methylome study established that age-dependent DNA hypermethylation of hair-cell genes silences the supporting-cell regenerative program. What remains open is the precise causal mechanism of 5-aza action in the adult cochlea in vivo, the fate and compensation of the supporting-cell pool after conversion, and whether the program can be driven to full functional maturation.

Background

The organ of Corti is the sensory epithelium of the mammalian cochlea and contains two cell types arranged in a strict mosaic: receptor hair cells that transduce sound, and supporting cells that separate and mechanically support them. In adult mammals, hair cells killed by noise, ototoxic drugs, or aging are never functionally replaced, causing permanent sensorineural hearing loss; in birds, fish, and amphibians, neighboring supporting cells replace lost hair cells by direct conversion (transdifferentiation) or by division. Sox2 is a transcription factor whose expression in the mature cochlea is restricted to supporting cells, which makes Sox2-driven fluorescent reporter mice a way to pre-label supporting cells and track their fate. 5-azacytidine (5-aza) is a nucleoside analog that inhibits DNA methyltransferases (DNMTs, the enzymes that maintain cytosine methylation at CpG dinucleotides), thereby causing passive DNA demethylation and re-expression of methylated (silenced) genes. The Deng–Hu group showed in 2019 that round-window injection (local delivery across the membrane-covered opening into the cochlea) of 5-aza into cochleae of adult mice deafened with kanamycin/furosemide generates new outer hair cells expressing Myosin VIIa, Pou4f3, and Myosin VI, and that Dnmt1 mRNA drops after treatment. In the 2020 source study, Sox2-EGFP reporter mice were used to show that the new hair cells derive from Sox2-positive supporting cells, with the conversion occurring roughly 5–7 days after 5-aza delivery. Functional hearing is conventionally measured by auditory brainstem response (ABR), the sound-evoked scalp potential whose threshold tracks hair-cell and neural function, and by distortion-product otoacoustic emissions (DPOAE), sounds generated by outer hair cell electromotility (mediated by the motor protein prestin); hair bundles and ribbon synapses are assayed by immunolabeling for espin (an actin-bundling protein of the hair bundle) and for CtBP2 and GluR2 (presynaptic ribbon and postsynaptic glutamate-receptor markers). In parallel work, Nguyen et al. (2023) profiled the cochlear DNA methylome and showed that supporting cells progressively hypermethylate promoters of hair-cell genes — including binding sites of Atoh1, the transcription factor necessary for hair-cell fate — during postnatal maturation, that this hypermethylation replaces earlier repression by Polycomb complexes (developmental repressors that mark genes with the inhibitory H3K27me3 histone modification), and that active demethylation by TET enzymes (the ten-eleven-translocation family of dioxygenases that actively demethylate DNA) is required for neonatal supporting cells to transdifferentiate. This gives a mechanistic rationale for why removing methylation might unlock hair-cell fate in supporting cells, but it was established in neonatal tissue and does not itself explain the 5-aza adult in vivo paradigm. The open question therefore concerns the mechanism, completeness, and tissue-level consequences of the demethylation-induced conversion in the adult deafened cochlea.

Problem Statement

For the 5-azacytidine/DNA-demethylation route to hair-cell regeneration in the deafened adult mammalian cochlea, three linked questions from the source study remain open after the 2022 demonstration of partial functional recovery: (1) Molecular mechanism — by what causal chain does DNMT inhibition and DNA demethylation convert adult Sox2-positive supporting cells into hair cells, i.e. which loci are demethylated (for example at the Atoh1 regulatory region and other hair-cell genes), which DNMT(s) must be inhibited, and is demethylation causally sufficient and necessary for the conversion in adult supporting cells in vivo? (2) Supporting-cell pool fate — does the conversion measurably deplete the supporting-cell population of the organ of Corti, and if so, is the deficit compensated by supporting-cell cell-cycle reentry or not at all? (3) Completeness — can the program be driven beyond the currently observed compromised state (abnormal bundle polarity, subnormal ribbon-synapse counts, at most ~30 dB ABR recovery) to full morphological and physiological maturation of the regenerated hair cells? The three questions are linked for the adult deafened cochlea: demethylation-caused conversion is one program, and its mechanism, cell-population consequences, and functional outcome are coupled.

References

  1. Deng X, Hu Z. Generation of Cochlear Hair Cells from Sox2 Positive Supporting Cells via DNA Demethylation. International Journal of Molecular Sciences 2020;21(22):8649. doi:10.3390/ijms21228649. https://doi.org/10.3390/ijms21228649
  2. Deng X, Liu Z, Li X, Zhou Y, Hu Z. Generation of new hair cells by DNA methyltransferase (Dnmt) inhibitor 5-azacytidine in a chemically-deafened mouse model. Scientific Reports 2019;9:7997. doi:10.1038/s41598-019-44313-0. https://doi.org/10.1038/s41598-019-44313-0
  3. Deng X, Hu Z. Hearing Recovery Induced by DNA Demethylation in a Chemically Deafened Adult Mouse Model. Frontiers in Cellular Neuroscience 2022;16:792089. doi:10.3389/fncel.2022.792089. https://doi.org/10.3389/fncel.2022.792089
  4. Deng M, Hu Z. 5-Azacytidine induces sensory hair cell marker expression in human inner ear cells in vitro. Human Cell 2026;39(3):46. doi:10.1007/s13577-026-01361-9. https://doi.org/10.1007/s13577-026-01361-9
  5. Nguyen JD, Llamas J, Shi T, Crump JG, Groves AK, Segil N. DNA methylation in the mouse cochlea promotes maturation of supporting cells and contributes to the failure of hair cell regeneration. Proceedings of the National Academy of Sciences USA 2023;120(33):e2300839120. doi:10.1073/pnas.2300839120. https://doi.org/10.1073/pnas.2300839120
  6. Smith-Cortinez N, Tan AK, Stokroos RJ, Versnel H, Straatman LV. Regeneration of Hair Cells from Endogenous Otic Progenitors in the Adult Mammalian Cochlea: Understanding Its Origins and Future Directions. International Journal of Molecular Sciences 2023;24(7):7840. doi:10.3390/ijms24097840. https://doi.org/10.3390/ijms24097840
  7. McGovern MM, Cox BC. Hearing restoration through hair cell regeneration: A review of recent advancements and current limitations. Hearing Research 2025;461:109256. doi:10.1016/j.heares.2025.109256. https://doi.org/10.1016/j.heares.2025.109256
  8. Gunewardene N, Lam P, Song J, Nguyen T, Mendez Ruiz S, Wong RCB, Wise AK, Richardson RT. Extent of genetic and epigenetic factor reprogramming via a single viral vector construct in deaf adult mice. Hearing Research 2025;457:109170. doi:10.1016/j.heares.2024.109170. https://doi.org/10.1016/j.heares.2024.109170