The inner ear contains our balance system (the vestibular system) and our hearing organ (the cochlea). Their sensing units, the hair cells, detect movement or sound waves. A loss of hair cells is a major cause of inner ear disorders, such as dizziness, imbalance and deafness.

When hair cells die, supporting cells that surround them close the ‘wound’ to repair the tissue. In fish, amphibians, reptiles and birds, the supporting cells can replace lost hair cells, but in mammals – including humans – hair cells are unable to regenerate in the cochlea, so hearing loss is permanent. However, previous research has shown that in certain mammals, spontaneous replacement of lost hair cells in the vestibular system can occur, but not enough to lead to a full recovery.

Scientists have been able to convert supporting cells in the vestibular system of mice into hair cells by using either certain chemicals, or by introducing a specific gene into the supporting cells. In the mouse embryo, this gene, called Atoh1, switches on a signalling pathway in the inner ear, through which a non-specialised precursor cell becomes a hair cell. Inducing hair cell regeneration could be a therapy for inner ear disorders. Therefore, Taylor et al. wanted to find out if such procedures would work in inner ear tissue from humans.

The researchers collected intact tissue samples from the vestibular system of patients who had undergone surgery to have a tumour removed, which would normally destroy the inner ear. All existing hair cells were removed so that mainly supporting cells remained. Then, the tissue was either treated with chemicals that increased the production of hair cells or received the gene ATOH1. The results showed that the cells containing the gene were able to develop many features characteristic of hair cells. And a smaller number of hair cells treated with the chemicals also started to develop hair cell-like features. A gene analysis after the ATOH1 transfer revealed a number of active genes known to be markers of hair cells, but also several inactive ones. This suggests that additional factors are necessary for generating fully functional hair cells.

Dizziness and balance disorders present a major health care burden, particularly in the elderly population. Yet, they are often disregarded and overlooked. This study suggests that hair cell regeneration could be a feasible therapy for some forms of balance disorders linked to loss of vestibular hair cells. More research is needed to identify the other factors at play to test if hair cell regeneration in the cochlea could be used to treat hearing impairment.

Loss of the sensory ‘hair’ cells from the cochlea, the mammalian hearing organ, as a consequence of exposure to ototoxic drugs, excessive noise or through ageing, results in permanent hearing loss. More than 40% of those aged over 50% and 70% of those over 70 have a clinically significant hearing loss (Action on Hearing Loss; www.actiononhearingloss.org.uk/your-hearing/about-deafness-and-hearing-loss/statistics.aspx). Hearing loss has also been reported as a risk factor for dementia (Livingston et al., 2017). Loss of hair cells from the vestibular epithelia of the inner ear, results in balance dysfunction causing dizziness and vertigo, significantly under-appreciated disabling conditions. As with hearing loss, the prevalence of vestibular dysfunction increases with age. Dizziness is the most common reason for visits to the GP in those over 75% and 80% of unexplained falls in the elderly are attributable to vestibular dysfunction (Agrawal et al., 2009; Baloh et al., 2001; Department of Health, 1999; Herdman et al., 2000; Pothula et al., 2004). Regeneration of hair cells could potentially offer a therapeutic approach to amelioriate these conditions.

In the sensory epithelia of the inner ear in all vertebrates each hair cell is surrounded and separated from its neighbours by intervening supporting cells. Hair cells derive their name from the organised bundle of projections from the apical poles. They are mechanotransducers that convert motion into electrical signals. Supporting cells play a role in maintaining the physiological environment necessary for hair cell function and survival, and also repair the lesions in the epithelium when hair cells die. In non-mammalian vertebrates, hair cells lost from the auditory or vestibular sensory epithelia are replaced spontaneously by new ones (Collado et al., 2008; Rubel et al., 2013). These nascent hair cells are derived from supporting cells. Initially, new hair cells arise from direct, non-mitotic transdifferentiation (phenotypic conversion) of supporting cells into hair cells (Cafaro et al., 2007; Taylor and Forge, 2005). Other supporting cells re-enter the cell cycle, the daughter cells giving rise to hair and supporting cells (Burns and Stone, 2017; Cafaro et al., 2007; Collado et al., 2008; Rubel et al., 2013). There is no regeneration in the adult mammalian auditory system. However, there is a limited capacity to regenerate hair cells in vivo in the mammalian vestibular system (Forge et al., 1993; 1998; Lopez et al., 1997; Kawamoto et al., 2009). These hair cells arise by direct phenotypic conversion of supporting cells (Li and Forge, 1997; Lin et al., 2011). We recently reported the presence of cells bearing immature hair bundles in the vestibular system of elderly people (Taylor et al., 2015). This suggests that a capacity to regenerate hair cells may exist at a very low level throughout life in humans.

During development, supporting cells and hair cells are derived from the same homogeneous population of precursor cells following a terminal mitotic event (Kelley, 2006). The mosaic patterning of hair and supporting cells develops by lateral inhibition mediated by the Notch signalling pathway (Kiernan, 2013). In cells differentiating as hair cells the basic helix-loop-helix transcription factor Atonal homolog 1 (Atoh1) is transiently expressed and has been shown to be necessary for sensory precursors to differentiate into hair cells (Bermingham et al., 1999; Kelley, 2006; Woods et al., 2004). The nascent hair cells express the Notch ligand Delta1 on their surface which activates the Notch receptor, a transmembrane protein, in adjacent neighbouring cells. The inhibitory activity of Notch prevents the adjacent cell following the same fate, so this cell will not become a hair cell but instead will be a supporting cell. Ligand binding to the Notch receptor triggers the extracellular cleavage of Notch by tumour necrosis factor alpha converting enzyme (TACE) (Kopan and Ilagan, 2009) and intracellular cleavage by γ-secretase (Kiernan, 2013). This releases the Notch intracellular domain which enters the nucleus interacting with several transcription factors to suppress Atoh1 expression thereby inhibiting differentiation as a hair cell and promoting HES/HEY expression, propelling those cells to become supporting cells.

Atoh1 is expressed during hair cell regeneration in chick and zebrafish and several studies have shown that ectopic overexpression of Atoh1 in the organ of Corti or vestibular sensory epithelia of mammals is sufficient to induce generation of cells that express hair cell marker proteins (Zheng and Gao, 2000). These arise by direct transdifferentiation without an intervening mitotic event. Overexpressing Atoh1 or using γ- secretase inhibitors, potentially offer means to induce supporting cells to become hair cells. Both these approaches have been applied to murine vestibular epithelia depleted of hair cells (Lin et al., 2011; Staecker et al., 2007).

We have established a consortium of surgeons throughout the UK to harvest human vestibular epithelia from translabyrinthine operations for the removal of acoustic neuromas (vestibular schwannomas) (Taylor et al., 2015). Here we use the vestibular epithelium collected from such surgeries to examine the capacity of supporting cells to generate new hair cells in adult human inner ear tissue. We have used an adenoviral vector to deliver ATOH1 (adV2- ATOH) to transduce cells in human sensory epithelia from which hair cells have been ablated with the ototoxic agent, gentamicin. We find that significant numbers of cells expressing hair cell markers can be generated. We have also exposed tissue to the γ- secretase inhibitor TAPI1 (TNFα protease inhibitor 1), a TACE inhibitor. Cells expressing hair cell marker proteins are also generated but in fewer numbers that with ATOH1 transduction. However, neither protocol resulted in fully differentiated hair cells. Studies of gene expression following ATOH1 transduction by RNA sequencing confirmed that a significant cascade of downstream effectors is induced by this treatment. However, this induction falls short of complete hair cell conversion but highlights components that may be necessary in completing the conversion events.

Utricles harvested from vestibular schwannoma surgeries were collected in medium and transported to the laboratory as previously described (Taylor et al., 2015). Tissue samples were examined and assessed before incubation in medium for ca. 18 hr prior to exposure to gentamicin or as a control in medium alone. Samples were used only when an intact epithelium was clearly visible under microscopic examination. Several harvested samples were not suitable for experimental use being damaged on excision or transfer. The age of the patients ranged between 17 and 81 (mean 50.6, median 51) but there was no selection or exclusion of samples on the basis of age or gender of the donor.

Untreated utricles

In utricles fixed and processed immediately following harvesting, hair cells, labelled for myosin VIIa, a hair cell marker, were present across the entire epithelium (Figure 1A), but, as reported previously (Taylor et al., 2015), the density of hair cells varied and in some samples, particularly those from older individuals, there were very few. SEM showed the characteristic hair bundles of vestibular hair cells with stereocilia of graded height but there was considerable variability in the morphology of the bundles and the height of the longest stereocilia in each bundle (Figure 1B). Short microvilli covered the apical surfaces of supporting cells. Thin sections of untreated utricles revealed a bilayer of cells with rounded hair cell nuclei in the more apical region closer to the luminal surface, while the more irregularly shaped supporting cell nuclei were located close to the underlying basement membrane (Figure 1C,D). Hair cells survived in vitro in most untreated samples maintained in explant culture for 28 days although in some cultures incubated for this period, condensed, mis-shapen remnants also labelled positively for myosin VIIa suggesting a possible incipient deterioration in the cultures by this time (Figure 1D). This defined a period of 21–22 days for an optimal total time of incubation in subsequent experiments, a period of sufficient length to cover that over which spontaneous regeneration of hair cells occurs in the vestibular organs in vivo in guinea pigs (Forge et al., 1993, 1998), chinchillas (Lopez et al., 1997) and mice (Kawamoto et al., 2009).

Figure 1

Download asset Open asset

Gentamicin treatment results in extensive hair cell loss

To ablate hair cells we exposed utricles to the ototoxic aminoglycoside antibiotic gentamicin. At 24 hr following 48 hr exposure to 2 mM gentamicin, few intact hair cells remained. Apoptotic death of hair cells in the body of the epithelium was evident by positive immunolabelling for activated caspase 3 (not shown) and in thin sections for TEM by pyknotic nuclei or marginated chromatin, with apoptotic bodies inside supporting cells (Figure 2A,B). Some hair cells were also seen to be extruded from the epithelium (Figure 2C). Both apoptosis, with apoptotic bodies phagocytosed by supporting cells, and extrusion of hair cells from the epithelium have been observed to occur in the vestibular sensory epithelia in vivo in animals treated with aminoglycoside (Li et al., 1995). With loss of hair cells, supporting cells closed the lesions (Figure 2D). The number of hair cells in cultures treated with gentamicin (N = 16) was assessed by counting myoVIIa positive cells in two groups: an early timepoint, 2–4 days post-gentamicin (dpg) (N = 5); and a late timepoint, 11–21 dpg (N = 11). At the earlier stage there was a mean of 5.26 ± 1.48 per 10000 µm². The hair cell bodies that persisted in the first few days after gentamicin exposure were scattered across the epithelium (Figure 2E). They were always rounded in shape and most contained an actin-rich rod-like inclusion structure (Figure 2E,F,G) indicative of pathology; they are reminiscent of the ‘cytocaud’ observed in damaged guinea pig vestibular hair cells (Kanzaki et al., 2002), and in mice appear in damaged hair cells destined for phagocytosis by supporting cells (Bucks et al., 2017). In tissue cultured for longer periods, up to 21 days post gentamicin treatment, there were very few hair cells (Figure 2H, ca. 1.4 ± 0.31 per 10000 µm²). The surfaces of almost all cells across the epithelium were of similar appearance (Figure 2I), with no surface projections or other structural specialisations, except for dispersed short microvilli, and with a polygonal outline, features characteristic of supporting cells following loss of hair cells. There was no evidence of spontaneous regeneration of hair cells. Thus, the prolonged high dose exposure to gentamicin ablated the majority of hair cells resulting in an epithelium composed predominantly of supporting cells.

Figure 2

Download asset Open asset

Adenoviral transduction of human vestibular epithelium

To promote hair cell generation by expression of exogenous ATOH1 transduction we used a second generation multiply-depleted replication-incompetent adenoviral vector, Ad2, previously shown to transduce human hair cells and supporting cells in vitro (Kesser et al., 2007). The vector carried the genes for green fluorescent protein (GFP) and atonal homologue 1 (ATOH1) independently driven by CMV promoters. After ablation of hair cells by incubation with gentamicin, utricles (and some cristae) were thoroughly rinsed with medium and incubated for up to 24 hr with the viral vector in serum-free medium. Expression of GFP at 4 days after transduction showed that supporting cells in the human tissue could be efficiently transduced, and delineated variable shapes of supporting cells revealing some with thin basally directed projections (Figure 3A). Labelling for SOX2, which is expressed by supporting cells, showed that GFP was co-expressed in cells with nuclei positively labelled for SOX2 (Figure 3B). Expression of ATOH1 could be detected by immunolabelling in some cells expressing GFP both in utricles (Figure 3C) and in cristae (Figure 3D), but not all GFP expressing cells also expressed ATOH1. Likewise, myosin VIIa was also expressed in some cells expressing GFP but not others (Figure 3E). By 17 days after incubation with the virus (tissue cultured for a total of 20 days) many myosin VIIa-positive cells were apparent (Figure 3F). These myosin VIIA-positive cells were often tightly packed together, contacting each other in some regions of the tissue (Figure 3G) and of variable shapes, rarely rounded but often elongated and some with thin basally directed projections (Figure 3H.) Only some cells expressed both GFP and myosin VIIA (Figure 3G,H) but in many that did, GFP was less strongly expressed than in cells expressing GFP alone (Figure 3G), perhaps suggesting that GFP may be downregulated over time as cells differentiate. The number of cells expressing myoVIIa (14.54 ± 3.68, N = 7) was significantly greater than the number in control tissue from which hair cells had been ablated and then maintained for an equivalent period of time without further treatment (Figure 4).

Figure 3

Download asset Open asset

Figure 4

Download asset Open asset

Morphological assessment

In explants exposed to ATOH1, either with or without TAPI1, following the ablation of hair cells with gentamicin, SEM (3 samples of each condition) showed the apical surface of many cells across the epithelium had bushy microvillar projections that were noticeably longer than the microvilli of neighbouring cells (Figure 5A). Some cells had a central protrusion within the cluster of microvilli, reminiscent of the position of the kinocilium seen in immature hair bundles during development of the sensory epithelia and hair cell regeneration (Denman-Johnson and Forge, 1999; Forge et al., 1998; Tilney et al., 1992a; Tilney et al., 1992b) (Figure 5B,C). The projections were present on cells that expressed GFP and labeled for myosin VIIa and were composed of actin (Figure 5D), but no cells showed organized hair bundles, and in whole mount samples prepared for immunolabelling, very little labelling for espin, a known actin bundling protein expressed along the length of stereocilia (Zheng et al., 2000), could be detected. In thin sections, cells with prominent elongated microvilli were identified. They were also evident in samples that had first been examined by SEM (two samples) (Figure 5D,E) as well in samples which had been embedded after immunolabelling for myosin VIIa (two samples) (Figure 5F,G). These cells showed some features resembling hair cells: they were generally cylindrical in shape with a rounded nucleus located towards the luminal pole and, when their profiles were followed through the serial sections, they were not in contact with the basement membrane, although some cells had thinner elongated basal projections similar to that seen with myosin VIIA labelling and expected of a cell converting from a supporting cell (Li and Forge, 1997; Taylor and Forge, 2005). However, these cells did not possess cuticular plates - the actin meshwork that forms a platform beneath the stereocilia in differentiating hair cells - nor were there any synaptic specializations such as synaptic ribbons (Taylor et al., 2015). Thus, it appeared that while cells expressing hair cells markers could be generated, those cells did not differentiate fully as hair cells.

Figure 5

Download asset Open asset

Gene expression changes with Ad2-ATOH1-GFP transduction

To investigate alterations in gene expression after Ad2-ATOH1-GFP transduction of human supporting cells, we measured transcriptome changes by RNA-seq. We compared gentamicin-treated tissue transduced with Ad2-ATOH1-GFP cultured for 5 days, versus control tissues only treated with gentamicin (cultured for 2 days, 8 days and 14–18 days). Three separate comparisons were performed between the Ad2-ATOH1-GFP transduced tissues and the three different controls to limit the variability between control samples and derive a consistent set of gene expression changes. Based upon this stringent comparison, a total of 494 genes were significantly differentially expressed (−2 >= fold change >= 2 and p<0.05 across all three comparisons). The expression of 441 out of these 494 genes exhibited the same direction of change in all three comparisons. Of these, 53 were down-regulated by ATOH1 transduction whereas 388 were upregulated. ATOH1 was among the most significantly overexpressed genes in tissues transduced with ATOH1 (70 to 286-fold < p < 0.05).

Within the 441 significantly differentially expressed genes we were interested in identifying key regulatory genes that showed consistent changes in gene expression. Including ATOH1, a total of 18 transcription factors (TFs) and five known chromatin modifiers (CMs, one gene SATB2 has both activities) fall into this class. These are shown in Table 1. To validate results from the RNA-sequencing, quantitative RT-PCR was conducted on five of the genes listed in Table 1. These were CITED, IRF9, SNAI1, EP300 and HDAC9. In all cases the trends for fold-changes were the same as in RNA-seq and in most cases they were very close in actual values (Figure 6).

Figure 6

Download asset Open asset

Table 1

GENE	Description	Ab T	Ab C	LOG2	Funct
ATOH1	atonal bHLH transcription factor 1	37.8	0.3	6.9	TF
RAX	retina and anterior neural fold homeobox	0.7	0.1	2.5	TF
ZNF296	zinc finger protein 296	17.1	3.0	2.5	TF
POU4F3	POU class 4 homeobox 3	3.4	0.8	2.1	TF
CITED4	Cbp/p300 interacting transactivator	56.6	13.7	2.0	TF
GTF2IRD2	GTF2I repeat domain containing 2	2.7	0.7	2.0	TF
SNAI1	snail family transcriptional repressor 1	30.7	8.9	1.8	TF
ZNF775	zinc finger protein 775	17.8	5.9	1.6	TF
MEF2B	myocyte enhancer factor 2B	2.1	0.7	1.6	TF
SOX12	SRY-box 12	10.1	3.6	1.5	TF
ZNF784	zinc finger protein 784	3.7	1.3	1.4	TF
ZNF837	zinc finger protein 837	1.8	0.7	1.4	TF
IRF9	interferon regulatory factor 9	137.3	53.5	1.4	TF
ZNF692	zinc finger protein 692	36.7	15.1	1.3	TF
GTF2H3	general transcription factor IIH subunit 3	9.3	21.5	−1.2	TF
ZNF382	zinc finger protein 382	0.9	2.2	−1.2	TF
EP300	E1A binding protein p300	10.9	26.3	−1.3	CM
ZFHX4	zinc finger homeobox 4	1.0	2.4	−1.3	TF
ASH1L	ASH1 like histone lysine methyltransfer	5.1	12.8	−1.3	CM
ATRX	ATRX, chromatin remodeler	7.0	22.4	−1.7	CM
SATB2	SATB homeobox 2	0.8	3.0	−1.9	TF/CM
HDAC9	histone deacetylase 9	5.7	20.9	−1.9	CM
TCF19	transcription factor 19	0.7	0.2	1.8	TF
HES1	hes family bHLH transcription factor 1	78.1	24.1	1.7	TF
ZNF467	zinc finger protein 467	4.3	1.5	1.5	TF
GSC	goosecoid homeobox	0.8	0.3	1.5	TF
SALL1	spalt like transcription factor 1	9.7	3.9	1.3	TF
KCNIP3	potassium voltage-gated channel protein 3	10.1	4.3	1.3	TF
THAP3	THAP domain containing 3	30.6	13.2	1.2	TF
MSX1	msh homeobox 1	8.4	3.7	1.2	TF
SCAND1	SCAN domain containing 1	58.9	26.8	1.1	TF
ING2	inhibitor of growth family member 2	13.2	6.1	1.1	CM
MLXIP	MLX interacting protein	8.8	4.2	1.1	TF
ZNF652	zinc finger protein 652	11.4	5.7	1.0	TF
PATZ1	POZ/BTB and AT hook containing zincfinger	18.1	9.3	1.0	TF/CM
RARG	retinoic acid receptor gamma	18.0	9.2	1.0	TF
GLI4	GLI family zinc finger 4	13.1	6.8	1.0	TF
THAP8	THAP domain containing 8	7.5	4.3	0.8	TF
ING5	inhibitor of growth family member 5	17.5	10.4	0.8	CM
TAF6	TATA-box binding protein associated factor	42.6	26.1	0.7	TF
ZNF768	zinc finger protein 768	47.2	29.3	0.7	TF
CREBBP	CREB binding protein	11.6	21.8	−0.9	TF

It is interesting to note from Table 1 that only four of the TFs and all of the CMs exhibit downregulation after ATOH1 transduction. The vast majority of the TFs are upregulated. Among the upregulated TF genes, six are of particular note (ATOH1, POU4F3, RAX, ZNF837, GTF2IRD2 AND MEF2B) since they are induced from close to zero to a significant level of transcript abundance (abundance levels are shown in Table 1). The remaining TFs appear to already be present at significant abundance levels in gentamicin-treated controls.

19 putative HC markers (ACTC1, ATOH1, CCDC60, CES2, ENO2, EPS8L2, HIST3H2A, JAG2, OBSCN, ODF3B, PKN3, POU4F3, RABL2A, RAX, SYTL1, SYT7, TAS1R1, UBXN11, UNC5A) and six putative downstream targets of ATOH1 (EPS8L2, JAG2, OBSCN, PKN3, POU4F3, UNC5A) (Cai et al., 2015), plus 4 markers of HC bundles (EPS8L2, GPI, PFKL, UBA7) (Shin et al., 2013) show significant upregulation in the ATOH1-treated samples. MYO7A, is also overexpressed in ATOH1-transduced tissues compared to all three controls but did not reach statistical significance. These markers of HC maturation are listed in Supplementary file 1. A large number of additional established hair cell markers (>150) exhibit consistent upregulation in all three comparative analyses, but failed to pass the statistical threshold of significance (see Supplementary file 1). Overall, of the 1375 putative HC markers we searched for within the upregulated gene expression dataset, a total of 335 were detectable and upregulated. These data support the contention that ATOH1 upregulation is driving towards an incomplete program of HC differentiation.

Enrichment analysis using ToppGene Suite (Chen et al., 2009) showed that GO biological processes such as muscle filament sliding and muscle system process were significantly enriched (FDR < 0.01). These processes included actin, myosin and troponin genes which suggests that hair cell bundles might be forming in the ATOH1-transduced tissues. However, most of these genes encoded relatively low abundance transcripts within the total dataset. Interestingly, there was also a significant enrichment of genes with TF binding sites for TCF3 and MEF2A/B transcription factors (FDR < 0.01). TCF3 is a known ATOH1 co-factor (Masuda et al., 2012) and is overexpressed in the ATOH1-transduced tissues compared to all three controls but failed to reach statistical significance. MEF2A is not differentially expressed within our data but, as noted above, MEF2B is induced three-fold from a basal level in ATOH1-transduced tissues. These may represent possible potentiators of ATOH1 function.

This unbiased transcriptomic analysis further supports the premise that manipulation of cell fate with gene modification enhances the conversion of supporting cells to a transcriptional signature that overlaps with known markers of the hair cell phenotype.

The objective of the study was to determine whether supporting cells in the vestibular sensory from adult humans can be induced to generate cells with hair-cell like features. As we illustrated previously (Taylor et al., 2015) the number of hair cells normally present in the vestibular sensory epithelia of humans may vary widely mainly due to ageing (and genetic variations that likely affect how susceptible to the effects of ageing different individuals might be). Consequently, the initial goal was to try to ablate as many hair cells as possible to create conditions that were essentially the same for all samples, with almost no hair cells present, and generate an epithelium composed predominantly of supporting cells. This was the starting point to test the capacities of supporting cells and also to remove possible confounds presented by any remaining hair cells. This condition also would mimic something similar to what we (Taylor et al., 2015) and others (Wright, 1983) have shown to be the situation in the vestibular sensory epithelia of elderly people who might be the ones to benefit from a regenerative strategy were one available. The prolonged (48 hr), high dose (2 mM) gentamicin treatment we used was designed to achieve this condition and the results indicate that this was largely accomplished.

Previous work has suggested that supporting cells in the mature utricular maculae of mice can be induced to convert towards a hair cell fate (Lin et al., 2011; Staecker et al., 2007). The present results demonstrate, not only from expression of hair cell marker proteins but also from analysis of expressed genes, that a similar plasticity of the supporting cell phenotype is retained in the vestibular tissues of humans. In our work, we made no distinction between samples on the basis of age but it has been suggested that age may influence the ability of supporting cells to regenerate lost hair cells. In mammals, a capacity of post-mitotic supporting cells in immature early postnatal (mouse) tissue to respond to regenerative signals or to generate new hair cells is dramatically reduced in the mature tissue of pre-weaner juveniles or young adults (Burns and Stone, 2017; Gu et al., 2007). The number of cells in the mature vestibular system that have a ‘stem cell’-like capacity to form spheroids also declines with age (Oshima et al., 2007). However, there is nothing to suggest that in the mature inner ear the activities of supporting cells are affected by age. In the present case, all our material was taken from adults the majority of whom were at the older end of the age range (mean 50.4 years, median 51). Yet in all cases where it was examined, supporting cells retained their capacities to effect repair of the epithelium and maintain tissue integrity. Loss of hair cells was accompanied by concurrent closure of the lesion by expansion of supporting cells, initially with the formation of characteristic scar formations (Meiteles and Raphael, 1994; Li et al., 1995; Taylor et al., 2015). The supporting cells also removed apoptotic bodies efficiently and acted to extrude damaged hair cells, in a manner very reminiscent of the similar processes occurring in vivo in animals following lethal damage to vestibular hair cells. In our earlier work we also showed through dye loading experiments the continuing viability of the supporting cell syncytium in tissue taken from the same population of patients as in this study, and in which there was significant hair cell loss. (Taylor et al., 2015). The retention of these capacities and activities seemingly regardless of age or ex vivo maintenance would point to the health of the tissue used in this work and also to the likelihood that the capacity of supporting cells to be converted is maintained throughout life in people. This is consistent with our earlier observations (Taylor et al., 2015) of occasional cells bearing immature hair bundles in the utricular maculae of elderly patients where most mature hair cells had been lost ‘naturally’, suggesting a very low level of spontaneous hair cell generation may be ongoing in the human vestibular sensory epithelia. Interestingly, it has been suggested that in birds, the capacity to regenerate hair cells from the supporting cell population continues unabated throughout life (Krumm et al., 2017).

We found that viral vector-mediated transduction with ATOH1 resulted in generation of highly significant numbers of myosin VIIA positive cells but the use of TAPI1 alone was not as successful. Since the ages of the donors of the tissues used in the TAPI1 experiments was not significantly different from those that were transduced with ATOH1 age is not a factor underlying the difference in the number of cells expressing the hair cell marker protein generated by the treatment regimes. TAPI1 has been shown to be an inhibitor of the Notch pathway that can induce hair cell regeneration in adult mouse utricles (Lin et al., 2011) and we used TAPI1 at a concentration the same as that used in that study. Other, γ-secretase inhibitors (GSIs) DAPT (Lin et al., 2011) and LY411575 (Mizutari et al., 2013) have also been used to regenerate hair cells in murine inner ear tissues. We performed a small number of studies with both of these GSIs, again using doses reported in the literature to be effective in mouse tissue, but neither produced significant numbers of cells expressing myosin VIIa. However, the number of samples examined was insufficient to draw any definitive conclusion as to their efficacy in inducing generation of hair cells. It may also be that higher concentrations of agents that inhibit the Notch pathway are required in the human tissue than in mouse and this may warrant further investigation.

Cells expressing myosin VIIA+ did not form organised hair bundles. The emergence of numerous cells bearing bushy bundles of microvilli in tissues in which Myo VIIa positive cells were generated but not in tissue maintained in culture for similar period but in which only hair cell ablation had occurred, further confirms phenotypic changes to supporting cells. These apical surface characteristics bear a striking resemblance to ‘hair-cell-like’ cells generated from human embryonic stem cells as described by Ronaghi et al. (2014) as well as those defined as hair cells generated in the mouse organ of Corti by genetic co-manipulation of Atoh1 and P27^Kip1 (Walters et al., 2017). However, it would appear that additional factors or further manipulations are required to generate fully differentiated hair cells, a contention supported by our findings that several genes that have been reported to be part of the cascade that follows Atoh1 expression during the initial embryonic formation of hair cells are not expressed. A similar conclusion is drawn from a recent report (Walters et al., 2017) that Atoh1-based hair cell regeneration therapies in mouse cochleae may be enhanced by genetic manipulation of p27^kip1,Gata3 or Pou4f3. From our transcriptional analysis we have identified transcription factors and chromatin modifiers that may play roles in the ATOH1 induced cascade. However, the most glaring absentee from this cascade is GFI1 which is believed to lie downstream of ATOH1 and POU4F3 (Hertzano et al., 2004) in the (still ill-defined) TF network that specifies a mature hair cell. In mice lacking Gfi1 ‘hair cells’ are generated during development but they show morphological abnormalities and fail to form hair bundles (Wallis et al., 2003). Also, in embryoid bodies derived from mouse stem cells the co-expression of Gfi1 and Pou4F3 as well as Atoh1 was required to generate cells with hair bundle-like structures (Costa et al., 2015). Since there was significant up-regulation of POU4F3 expression following the ATOH1 transduction of the human utricular maculae, it would point to GFI1 expression as an interesting candidate gene necessary for inducing hair cell differentiation. Gfi1 has been reported to act as a transcriptional regulator acting through chromatin modification (Duan et al., 2005). Identification of genes regulated by Gfi1 would also be of interest in a search for factors to induce regeneration of functional hair cells.

The expression of GFP showed that the supporting cells in the human utricular macula can be transduced with quite high efficiency with the Ad2 viral vector, consistent with prior reports (Kesser et al., 2007). Efficient transduction of the supporting cells in the human utricle can also be accomplished with a synthetic AAV vector Anc80L65 (Landegger et al., 2017). This indicates that other genetic manipulations of human vestibular sensory epithelia are possible and could potentially form the basis for other therapeutic interventions. One known receptor for adenovirus is αvβ5 integrin (Lyle and McCormick, 2010; Wickham et al., 1993) and in other work (Taylor, Hussain and Forge, in preparation) using RT-PCR screening we have found both these integrin subunits to be expressed in the human vestibular epithelia although as yet we do not know if they are partnered. Nevertheless, this may be the basis of the efficient transduction with adenoviral vectors. However, some cells that expressed GFP did not appear to express ATOH1 and some cells that expressed ATOH1 did not show expression of GFP. One reason for this maybe that the two genes were driven by independent CMV promoters so may not have always been expressed together or they were expressed at differing levels and sometimes below the level of detection.

The generated cells expressing Myo VIIa, as well as the cells bearing bushy microvillar bundles, were often closely clustered together, sometimes seemingly in contact with each other. The pattern of distribution of hair and supporting cells that is normally present in the utricular macular where each hair cells is surrounded and separated from it neighbours by intervening supporting cells was not restored. This may imply that the signalling pathways through which differentiating hair cells instruct their immediate neighbours to adopt a different, supporting cell, fate from themselves may not be properly active in these conditions where hair cells are induced.

All sequencing data from all of these samples have been deposited in NCBI GEO (accession number: GSE109320).

1. 1. Forge A
   2. Taylor RR
   3. Filia A
   4. Paredes U
   5. Asai Y
   6. Holt JR
   7. Lovett M

   (2018) Regenerating hair cells in human vestibular sensory epithelia

   Publicly available at the NCBI Gene Expression Omnibus (accession no. GSE109320).

   http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE109320

1. 1. Agrawal Y
   2. Carey JP
   3. Della Santina CC
   4. Schubert MC
   5. Minor LB

   (2009) Disorders of balance and vestibular function in US adults: data from the national health and nutrition examination survey, 2001-2004

   Archives of Internal Medicine 169:938–944.

   https://doi.org/10.1001/archinternmed.2009.66

   • PubMed
   • Google Scholar
2. 1. Baloh RW
   2. Enrietto J
   3. Jacobson KM
   4. Lin A

   (2001) Age-related changes in vestibular function: a longitudinal study

   Annals of the New York Academy of Sciences 942:210–219.

   https://doi.org/10.1111/j.1749-6632.2001.tb03747.x

   • PubMed
   • Google Scholar
3. 1. Bermingham NA
   2. Hassan BA
   3. Price SD
   4. Vollrath MA
   5. Ben-Arie N
   6. Eatock RA
   7. Bellen HJ
   8. Lysakowski A
   9. Zoghbi HY

   (1999) Math1: an essential gene for the generation of inner ear hair cells

   Science 284:1837–1841.

   https://doi.org/10.1126/science.284.5421.1837

   • PubMed
   • Google Scholar
4. 1. Bucks SA
   2. Cox BC
   3. Vlosich BA
   4. Manning JP
   5. Nguyen TB
   6. Stone JS

   (2017) Supporting cells remove and replace sensory receptor hair cells in a balance organ of adult mice

   eLife 6:e18128.

   https://doi.org/10.7554/eLife.18128

   • PubMed
   • Google Scholar
5. 1. Burns JC
   2. Stone JS

   (2017) Development and regeneration of vestibular hair cells in mammals

   Seminars in Cell & Developmental Biology 65:96–105.

   https://doi.org/10.1016/j.semcdb.2016.11.001

   • PubMed
   • Google Scholar
6. 1. Cafaro J
   2. Lee GS
   3. Stone JS

   (2007) Atoh1 expression defines activated progenitors and differentiating hair cells during avian hair cell regeneration

   Developmental Dynamics 236:156–170.

   https://doi.org/10.1002/dvdy.21023

   • PubMed
   • Google Scholar
7. 1. Cai T
   2. Jen HI
   3. Kang H
   4. Klisch TJ
   5. Zoghbi HY
   6. Groves AK

   (2015) Characterization of the transcriptome of nascent hair cells and identification of direct targets of the Atoh1 transcription factor

   Journal of Neuroscience 35:5870–5883.

   https://doi.org/10.1523/JNEUROSCI.5083-14.2015

   • PubMed
   • Google Scholar
8. 1. Chen J
   2. Bardes EE
   3. Aronow BJ
   4. Jegga AG

   (2009) ToppGene suite for gene list enrichment analysis and candidate gene prioritization

   Nucleic Acids Research 37:W305–W311.

   https://doi.org/10.1093/nar/gkp427

   • PubMed
   • Google Scholar
9. 1. Collado MS
   2. Burns JC
   3. Hu Z
   4. Corwin JT

   (2008) Recent advances in hair cell regeneration research

   Current Opinion in Otolaryngology & Head and Neck Surgery 16:465–471.

   https://doi.org/10.1097/MOO.0b013e32830f4ab5

   • PubMed
   • Google Scholar
10. 1. Costa A
    2. Sanchez-Guardado L
    3. Juniat S
    4. Gale JE
    5. Daudet N
    6. Henrique D

    (2015) Generation of sensory hair cells by genetic programming with a combination of transcription factors

    Development 142:1948–1959.

    https://doi.org/10.1242/dev.119149

    • PubMed
    • Google Scholar
11. 1. Daudet N
    2. Gibson R
    3. Shang J
    4. Bernard A
    5. Lewis J
    6. Stone J

    (2009) Notch regulation of progenitor cell behavior in quiescent and regenerating auditory epithelium of mature birds

    Developmental Biology 326:86–100.

    https://doi.org/10.1016/j.ydbio.2008.10.033

    • PubMed
    • Google Scholar
12. 1. Davies S
    2. Forge A

    (1987) Preparation of the mammalian organ of Corti for scanning electron microscopy

    Journal of Microscopy 147:89–101.

    https://doi.org/10.1111/j.1365-2818.1987.tb02821.x

    • PubMed
    • Google Scholar
13. 1. Denman-Johnson K
    2. Forge A

    (1999) Establishment of hair bundle polarity and orientation in the developing vestibular system of the mouse

    Journal of Neurocytology 28:821–835.

    https://doi.org/10.1023/A:1007061819934

    • PubMed
    • Google Scholar
14. 1. Department of Health

    (1999)

    Provision of Adult Balance Services: A Good Practice Guide

    Provision of Adult Balance Services: A Good Practice Guide.

    • Google Scholar
15. 1. Duan Z
    2. Zarebski A
    3. Montoya-Durango D
    4. Grimes HL
    5. Horwitz M

    (2005) Gfi1 coordinates epigenetic repression of p21Cip/WAF1 by recruitment of histone lysine methyltransferase G9a and histone deacetylase 1

    Molecular and Cellular Biology 25:10338–10351.

    https://doi.org/10.1128/MCB.25.23.10338-10351.2005

    • PubMed
    • Google Scholar
16. 1. Forge A
    2. Li L
    3. Corwin JT
    4. Nevill G

    (1993) Ultrastructural evidence for hair cell regeneration in the mammalian inner ear

    Science 259:1616–1619.

    https://doi.org/10.1126/science.8456284

    • PubMed
    • Google Scholar
17. 1. Forge A
    2. Li L
    3. Nevill G

    (1998) Hair cell recovery in the vestibular sensory epithelia of mature guinea pigs

    The Journal of Comparative Neurology 397:69–88.

    https://doi.org/10.1002/(SICI)1096-9861(19980720)397:1<69::AID-CNE6>3.0.CO;2-G

    • PubMed
    • Google Scholar
18. 1. Gu R
    2. Montcouquiol M
    3. Marchionni M
    4. Corwin JT

    (2007) Proliferative responses to growth factors decline rapidly during postnatal maturation of mammalian hair cell epithelia

    European Journal of Neuroscience 25:1363–1372.

    https://doi.org/10.1111/j.1460-9568.2007.05414.x

    • PubMed
    • Google Scholar
19. 1. Hart SN
    2. Therneau TM
    3. Zhang Y
    4. Poland GA
    5. Kocher JP

    (2013) Calculating sample size estimates for RNA sequencing data

    Journal of Computational Biology 20:970–978.

    https://doi.org/10.1089/cmb.2012.0283

    • PubMed
    • Google Scholar
20. 1. Herdman SJ
    2. Blatt P
    3. Schubert MC
    4. Tusa RJ

    (2000)

    Falls in patients with vestibular deficits

    The American Journal of Otology 21:847–851.

    • PubMed
    • Google Scholar
21. 1. Hertzano R
    2. Montcouquiol M
    3. Rashi-Elkeles S
    4. Elkon R
    5. Yücel R
    6. Frankel WN
    7. Rechavi G
    8. Möröy T
    9. Friedman TB
    10. Kelley MW
    11. Avraham KB

    (2004) Transcription profiling of inner ears from Pou4f3(ddl/ddl) identifies Gfi1 as a target of the Pou4f3 deafness gene

    Human Molecular Genetics 13:2143–2153.

    https://doi.org/10.1093/hmg/ddh218

    • PubMed
    • Google Scholar
22. 1. Kanzaki S
    2. Beyer LA
    3. Canlon B
    4. Meixner WM
    5. Raphael Y

    (2002) The cytocaud: a hair cell pathology in the waltzing Guinea pig

    Audiology and Neuro-Otology 7:289–297.

    https://doi.org/10.1159/000064447

    • PubMed
    • Google Scholar
23. 1. Kawamoto K
    2. Izumikawa M
    3. Beyer LA
    4. Atkin GM
    5. Raphael Y

    (2009) Spontaneous hair cell regeneration in the mouse utricle following gentamicin ototoxicity

    Hearing Research 247:17–26.

    https://doi.org/10.1016/j.heares.2008.08.010

    • PubMed
    • Google Scholar
24. 1. Kelley MW

    (2006) Regulation of cell fate in the sensory epithelia of the inner ear

    Nature Reviews Neuroscience 7:837–849.

    https://doi.org/10.1038/nrn1987

    • PubMed
    • Google Scholar
25. 1. Kesser BW
    2. Hashisaki GT
    3. Fletcher K
    4. Eppard H
    5. Holt JR

    (2007) An in vitro model system to study gene therapy in the human inner ear

    Gene Therapy 14:1121–1131.

    https://doi.org/10.1038/sj.gt.3302980

    • PubMed
    • Google Scholar
26. 1. Kesser BW
    2. Hashisaki GT
    3. Holt JR

    (2008) Gene transfer in human vestibular epithelia and the prospects for inner ear gene therapy

    The Laryngoscope 118:821–831.

    https://doi.org/10.1097/MLG.0b013e318164d0aa

    • PubMed
    • Google Scholar
27. 1. Kiernan AE

    (2013) Notch signaling during cell fate determination in the inner ear

    Seminars in Cell & Developmental Biology 24:470–479.

    https://doi.org/10.1016/j.semcdb.2013.04.002

    • PubMed
    • Google Scholar
28. 1. Kim D
    2. Pertea G
    3. Trapnell C
    4. Pimentel H
    5. Kelley R
    6. Salzberg SL

    (2013) TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions

    Genome Biology 14:R36.

    https://doi.org/10.1186/gb-2013-14-4-r36

    • PubMed
    • Google Scholar
29. 1. Kopan R
    2. Ilagan MX

    (2009) The canonical Notch signaling pathway: unfolding the activation mechanism

    Cell 137:216–233.

    https://doi.org/10.1016/j.cell.2009.03.045

    • PubMed
    • Google Scholar
30. 1. Krumm B
    2. Klump G
    3. Köppl C
    4. Langemann U

    (2017) Barn owls have ageless ears

    Proceedings of the Royal Society B: Biological Sciences 284:20171584.

    https://doi.org/10.1098/rspb.2017.1584

    • PubMed
    • Google Scholar
31. 1. Ku YC
    2. Renaud NA
    3. Veile RA
    4. Helms C
    5. Voelker CC
    6. Warchol ME
    7. Lovett M

    (2014) The transcriptome of utricle hair cell regeneration in the avian inner ear

    Journal of Neuroscience 34:3523–3535.

    https://doi.org/10.1523/JNEUROSCI.2606-13.2014

    • PubMed
    • Google Scholar
32. 1. Landegger LD
    2. Pan B
    3. Askew C
    4. Wassmer SJ
    5. Gluck SD
    6. Galvin A
    7. Taylor R
    8. Forge A
    9. Stankovic KM
    10. Holt JR
    11. Vandenberghe LH

    (2017) A synthetic AAV vector enables safe and efficient gene transfer to the mammalian inner ear

    Nature Biotechnology 35:280–284.

    https://doi.org/10.1038/nbt.3781

    • PubMed
    • Google Scholar
33. 1. Li L
    2. Forge A

    (1997) Morphological evidence for supporting cell to hair cell conversion in the mammalian utricular macula

    International Journal of Developmental Neuroscience 15:433–446.

    https://doi.org/10.1016/S0736-5748(96)00102-5

    • PubMed
    • Google Scholar
34. 1. Li L
    2. Nevill G
    3. Forge A

    (1995) Two modes of hair cell loss from the vestibular sensory epithelia of the guinea pig inner ear

    The Journal of Comparative Neurology 355:405–417.

    https://doi.org/10.1002/cne.903550307

    • PubMed
    • Google Scholar
35. 1. Lin V
    2. Golub JS
    3. Nguyen TB
    4. Hume CR
    5. Oesterle EC
    6. Stone JS

    (2011) Inhibition of Notch activity promotes nonmitotic regeneration of hair cells in the adult mouse utricles

    Journal of Neuroscience 31:15329–15339.

    https://doi.org/10.1523/JNEUROSCI.2057-11.2011

    • PubMed
    • Google Scholar
36. 1. Livingston G
    2. Sommerlad A
    3. Orgeta V
    4. Costafreda SG
    5. Huntley J
    6. Ames D

    (2017)

    Dementia prevention, intervention and care

    Lancet Commissions.

    • Google Scholar
37. 1. Lopez I
    2. Honrubia V
    3. Lee SC
    4. Schoeman G
    5. Beykirch K

    (1997) Quantification of the process of hair cell loss and recovery in the chinchilla crista ampullaris after gentamicin treatment

    International Journal of Developmental Neuroscience 15:447–461.

    https://doi.org/10.1016/S0736-5748(96)00103-7

    • PubMed
    • Google Scholar
38. 1. Lyle C
    2. McCormick F

    (2010) Integrin alphavbeta5 is a primary receptor for adenovirus in CAR-negative cells

    Virology Journal 7:148.

    https://doi.org/10.1186/1743-422X-7-148

    • PubMed
    • Google Scholar
39. 1. Martin M

    (2011) Cutadapt removes adapter sequences from high-throughput sequencing reads

    EMBnet.journal 17:10–12.

    https://doi.org/10.14806/ej.17.1.200

    • Google Scholar
40. 1. Masuda M
    2. Pak K
    3. Chavez E
    4. Ryan AF

    (2012) TFE2 and GATA3 enhance induction of POU4F3 and myosin VIIa positive cells in nonsensory cochlear epithelium by ATOH1

    Developmental Biology 372:68–80.

    https://doi.org/10.1016/j.ydbio.2012.09.002

    • PubMed
    • Google Scholar
41. 1. Meiteles LZ
    2. Raphael Y

    (1994) Scar formation in the vestibular sensory epithelium after aminoglycoside toxicity

    Hearing Research 79:26–38.

    https://doi.org/10.1016/0378-5955(94)90124-4

    • PubMed
    • Google Scholar
42. 1. Mizutari K
    2. Fujioka M
    3. Hosoya M
    4. Bramhall N
    5. Okano HJ
    6. Okano H
    7. Edge AS

    (2013) Notch inhibition induces cochlear hair cell regeneration and recovery of hearing after acoustic trauma

    Neuron 77:58–69.

    https://doi.org/10.1016/j.neuron.2012.10.032

    • PubMed
    • Google Scholar
43. 1. Oshima K
    2. Grimm CM
    3. Corrales CE
    4. Senn P
    5. Martinez Monedero R
    6. Géléoc GS
    7. Edge A
    8. Holt JR
    9. Heller S

    (2007) Differential distribution of stem cells in the auditory and vestibular organs of the inner ear

    Journal of the Association for Research in Otolaryngology 8:18–31.

    https://doi.org/10.1007/s10162-006-0058-3

    • PubMed
    • Google Scholar
44. 1. Pothula VB
    2. Chew F
    3. Lesser TH
    4. Sharma AK

    (2004) Falls and vestibular impairment

    Clinical Otolaryngology and Allied Sciences 29:179–182.

    https://doi.org/10.1111/j.0307-7772.2004.00785.x

    • PubMed
    • Google Scholar
45. 1. Ronaghi M
    2. Nasr M
    3. Ealy M
    4. Durruthy-Durruthy R
    5. Waldhaus J
    6. Diaz GH
    7. Joubert LM
    8. Oshima K
    9. Heller S

    (2014) Inner ear hair cell-like cells from human embryonic stem cells

    Stem Cells and Development 23:1275–1284.

    https://doi.org/10.1089/scd.2014.0033

    • PubMed
    • Google Scholar
46. 1. Rubel EW
    2. Furrer SA
    3. Stone JS

    (2013) A brief history of hair cell regeneration research and speculations on the future

    Hearing Research 297:42–51.

    https://doi.org/10.1016/j.heares.2012.12.014

    • PubMed
    • Google Scholar
47. 1. Scheffer DI
    2. Shen J
    3. Corey DP
    4. Chen ZY

    (2015) Gene expression by mouse inner ear hair cells during development

    Journal of Neuroscience 35:6366–6380.

    https://doi.org/10.1523/JNEUROSCI.5126-14.2015

    • PubMed
    • Google Scholar
48. 1. Shin JB
    2. Krey JF
    3. Hassan A
    4. Metlagel Z
    5. Tauscher AN
    6. Pagana JM
    7. Sherman NE
    8. Jeffery ED
    9. Spinelli KJ
    10. Zhao H
    11. Wilmarth PA
    12. Choi D
    13. David LL
    14. Auer M
    15. Barr-Gillespie PG

    (2013) Molecular architecture of the chick vestibular hair bundle

    Nature Neuroscience 16:365–374.

    https://doi.org/10.1038/nn.3312

    • PubMed
    • Google Scholar
49. 1. Staecker H
    2. Praetorius M
    3. Baker K
    4. Brough DE

    (2007) Vestibular hair cell regeneration and restoration of balance function induced by math1 gene transfer

    Otology & Neurotology 28:223–231.

    https://doi.org/10.1097/MAO.0b013e31802b3225

    • PubMed
    • Google Scholar
50. 1. Taylor RR
    2. Forge A

    (2005) Hair cell regeneration in sensory epithelia from the inner ear of a urodele amphibian

    The Journal of Comparative Neurology 484:105–120.

    https://doi.org/10.1002/cne.20450

    • PubMed
    • Google Scholar
51. 1. Taylor RR
    2. Jagger DJ
    3. Saeed SR
    4. Axon P
    5. Donnelly N
    6. Tysome J
    7. Moffatt D
    8. Irving R
    9. Monksfield P
    10. Coulson C
    11. Freeman SR
    12. Lloyd SK
    13. Forge A

    (2015) Characterizing human vestibular sensory epithelia for experimental studies: new hair bundles on old tissue and implications for therapeutic interventions in ageing

    Neurobiology of Aging 36:2068–2084.

    https://doi.org/10.1016/j.neurobiolaging.2015.02.013

    • PubMed
    • Google Scholar
52. 1. Tilney LG
    2. Cotanche DA
    3. Tilney MS

    (1992a)

    Actin filaments, stereocilia and hair cells of the bird cochlea. VI. how the number and arrangement of stereocilia are determined

    Development 116:213–226.

    • PubMed
    • Google Scholar
53. 1. Tilney LG
    2. Tilney MS
    3. DeRosier DJ

    (1992b) Actin filaments, Stereocilia, and hair cells: how cells count and measure

    Annual Review of Cell Biology 8:257–274.

    https://doi.org/10.1146/annurev.cb.08.110192.001353

    • PubMed
    • Google Scholar
54. 1. Trapnell C
    2. Williams BA
    3. Pertea G
    4. Mortazavi A
    5. Kwan G
    6. van Baren MJ
    7. Salzberg SL
    8. Wold BJ
    9. Pachter L

    (2010) Transcript assembly and quantification by RNA-Seq reveals unannotated transcripts and isoform switching during cell differentiation

    Nature Biotechnology 28:511–515.

    https://doi.org/10.1038/nbt.1621

    • PubMed
    • Google Scholar
55. 1. Wallis D
    2. Hamblen M
    3. Zhou Y
    4. Venken KJ
    5. Schumacher A
    6. Grimes HL
    7. Zoghbi HY
    8. Orkin SH
    9. Bellen HJ

    (2003) The zinc finger transcription factor Gfi1, implicated in lymphomagenesis, is required for inner ear hair cell differentiation and survival

    Development 130:221–232.

    https://doi.org/10.1242/dev.00190

    • PubMed
    • Google Scholar
56. 1. Walters BJ
    2. Coak E
    3. Dearman J
    4. Bailey G
    5. Yamashita T
    6. Kuo B
    7. Zuo J

    (2017) In Vivo Interplay between p27^Kip1, GATA3, ATOH1, and POU4F3 converts Non-sensory cells to hair cells in adult mice

    Cell Reports 19:307–320.

    https://doi.org/10.1016/j.celrep.2017.03.044

    • PubMed
    • Google Scholar
57. 1. Warchol ME
    2. Stone J
    3. Barton M
    4. Ku J
    5. Veile R
    6. Daudet N
    7. Lovett M

    (2017) ADAM10 and γ-secretase regulate sensory regeneration in the avian vestibular organs

    Developmental Biology 428:39–51.

    https://doi.org/10.1016/j.ydbio.2017.05.014

    • PubMed
    • Google Scholar
58. 1. Warde-Farley D
    2. Donaldson SL
    3. Comes O
    4. Zuberi K
    5. Badrawi R
    6. Chao P
    7. Franz M
    8. Grouios C
    9. Kazi F
    10. Lopes CT
    11. Maitland A
    12. Mostafavi S
    13. Montojo J
    14. Shao Q
    15. Wright G
    16. Bader GD
    17. Morris Q

    (2010) The GeneMANIA prediction server: biological network integration for gene prioritization and predicting gene function

    Nucleic Acids Research 38:W214–W220.

    https://doi.org/10.1093/nar/gkq537

    • PubMed
    • Google Scholar
59. 1. Wickham TJ
    2. Mathias P
    3. Cheresh DA
    4. Nemerow GR

    (1993) Integrins alpha v beta 3 and alpha v beta 5 promote adenovirus internalization but not virus attachment

    Cell 73:309–319.

    https://doi.org/10.1016/0092-8674(93)90231-E

    • PubMed
    • Google Scholar
60. 1. Woods C
    2. Montcouquiol M
    3. Kelley MW

    (2004) Math1 regulates development of the sensory epithelium in the mammalian cochlea

    Nature Neuroscience 7:1310–1318.

    https://doi.org/10.1038/nn1349

    • PubMed
    • Google Scholar
61. 1. Wright A

    (1983) The surface structures of the human vestibular apparatus

    Clinical Otolaryngology 8:53–63.

    https://doi.org/10.1111/j.1365-2273.1983.tb01672.x

    • PubMed
    • Google Scholar
62. 1. Zheng JL
    2. Gao WQ

    (2000) Overexpression of Math1 induces robust production of extra hair cells in postnatal rat inner ears

    Nature Neuroscience 3:580–586.

    https://doi.org/10.1038/75753

    • PubMed
    • Google Scholar
63. 1. Zheng L
    2. Sekerková G
    3. Vranich K
    4. Tilney LG
    5. Mugnaini E
    6. Bartles JR

    (2000) The deaf jerker mouse has a mutation in the gene encoding the espin actin-bundling proteins of hair cell stereocilia and lacks espins

    Cell 102:377–385.

    https://doi.org/10.1016/S0092-8674(00)00042-8

    • PubMed
    • Google Scholar

Author details

1. Ruth Rebecca Taylor

   UCL Ear Institute, University College London, London, United Kingdom

   Contribution

   Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   Ruth.r.taylor@ucl.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7359-1604

2. Anastasia Filia

   National Heart and Lung Institute, Imperial College London, London, United Kingdom

   Contribution

   Formal analysis, Investigation, Methodology, Writing—review and editing

   Competing interests

   No competing interests declared

3. Ursula Paredes

   UCL Ear Institute, University College London, London, United Kingdom

   Contribution

   Investigation

   Competing interests

   No competing interests declared

4. Yukako Asai

   F.M. Kirby Neurobiology Center, Boston Children's Hospital, Boston, United States

   Contribution

   Resources, Investigation

   Competing interests

   No competing interests declared

5. Jeffrey R Holt

   F.M. Kirby Neurobiology Center, Boston Children's Hospital, Boston, United States

   Contribution

   Resources, Investigation, Methodology, Writing—review and editing

   Competing interests

   No competing interests declared

6. Michael Lovett

   National Heart and Lung Institute, Imperial College London, London, United Kingdom

   Contribution

   Data curation, Formal analysis, Investigation, Methodology, Writing—review and editing

   Competing interests

   No competing interests declared

7. Andrew Forge

   UCL Ear Institute, University College London, London, United Kingdom

   Contribution

   Conceptualization, Formal analysis, Funding acquisition, Investigation, Methodology, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   a.forge@ucl.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-0995-0219

• 3,892

  views

• 386

  downloads

• 39

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.