Atoh1 regulation in the cochlea: more than just transcription

Abstract

More than 80% of all cases of deafness are related to the death or degeneration of cochlear hair cells and the associated spiral ganglion neurons, and a lack of regeneration of these cells leads to permanent hearing loss. Therefore, the regeneration of lost hair cells is an important goal for the treatment of deafness. Atoh1 is a basic helix-loop-helix (bHLH) transcription factor that is critical in both the development and regeneration of cochlear hair cells. Atoh1 is transcriptionally regulated by several signaling pathways, including Notch and Wnt signalings. At the post-translational level, it is regulated through the ubiquitin-proteasome pathway. In vitro and in vivo studies have revealed that manipulation of these signaling pathways not only controls development, but also leads to the regeneration of cochlear hair cells after damage. Recent progress toward understanding the signaling networks involved in hair cell development and regeneration has led to the development of new strategies to replace lost hair cells. This review focuses on our current understanding of the signaling pathways that regulate Atoh1 in the cochlea.

1. Introduction

Hair cell formation is dependent on the basic helix-loop-helix (bHLH) transcription factor Atoh1, which is both necessary and sufficient for hair cell differentiation during development (Bermingham et al., 1999). Atoh1 overexpression in the inner ear via genetic or pharmacological manipulation has been shown to result in new hair cell formation in the cochlea and to restore hearing in experimental animal models (Zheng and Gao, 2000; Izumikawa et al., 2005; Gubbels et al., 2008). The transcriptional and post-translational regulations of Atoh1 have been shown to involve several overlapping signaling pathways. Transcriptional regulation of Atoh1 by Wnt or Notch signaling or post-translational regulation by the Huwe1-ubiquitin-proteasome pathway has been shown to sucessfully convert inner ear progenitor cells into hair cells (Shi et al., 2010; 2013; 2014; Mizutari et al., 2013; Bramhall et al., 2014; Cheng et al., 2016). These studies have led to novel approaches for the regeneration of hair cells in the inner ear. This review focuses on recent advances in our understanding of the signaling pathways that regulate Atoh1 in the cochlea.

2. Hearing impairment

Hearing impairment is one of the most prevalent disabilities in industrialized countries, and it is estimated that 360 million people have moderate to profound hearing deficiency globally (WHO, 2017). Most cases of hearing impairment involve sensorineural hearing loss caused by the degeneration or loss of two specific inner ear cell types: cochlear hair cells, which act as the primary mechanoreceptors for sound transduction, and the connected auditory nerve neurons, which transmit electrical signals from the synapsed inner hair cells to the brain (Davis, 1983).

Loss of cochlear hair cells caused by genetic mutations, ototoxic medications, noise overexposure, autoimmune disease, or aging is irreversible in mammals because they have very limited ability to spontaneously regenerate hair cells after loss, and there are currently very few methods to regenerate hearing function after hair cell damage, especially in adult animals (Izumikawa et al., 2005; Mizutari et al., 2013). Vestibular hair cells have a limited regenerative capacity (Forge et al., 1993; Warchol et al., 1993; Wang et al., 2015), but spontaneous hair cell regeneration after damage or hair cell loss has been reported only in the newborn murine cochlea (Bramhall et al., 2014; Cox et al., 2014). For deafness caused by loss of cochlear hair cells, cochlear implants are surgically implanted into the inner ear to stimulate the spiral ganglion. However, cochlear implants depend on the remaining functional spiral ganglion neurons, the loss of which can severely compromise their efficacy (Incesulu and Nadol, 1998). The regeneration or replacement of damaged cochlear hair cells is thus the ultimate goal for the treatment of hearing loss.

Several different approaches for hair cell regeneration have been attempted, including cell therapy, gene therapy, and pharmacological therapy. A few attempts at cell transplantation into the damaged cochlea have been reported, but most have shown extremely limited survival of transplanted cells and the inability of these cells to differentiate into mature hair cell types (Hu et al., 2005; Hu and Ulfendahl, 2006). Gene therapy involving the overexpression of Atoh1 has been shown to be a successful strategy for converting supporting cells into hair cells, thereby restoring hearing loss after damage (Izumikawa et al., 2005; Richardson and Atkinson, 2015). Gene augmentation (Akil et al., 2012; Askew et al., 2015; Pan et al., 2017), as well as emerging CRISPR-Cas9 genome editing methods (Zou et al., 2015; Zuris et al., 2015; Mianné et al., 2016), also provides promising experimental paradigms for treating genetic deafness. However, the complexity of the inner ear anatomy, as well as a lack of ideal vector systems and appropriate delivery methods, has made the delivery of foreign genetic material into the inner ear challenging (Husseman and Raphael, 2009; Shu et al., 2016). Pharmacotherapy using local delivery of small molecule compounds to convert inner ear progenitors into hair cells seems to be a more physiologically relevant approach to regenerating lost hair cells after damage (Mizutari et al., 2013). However, a successful therapy will require a better understanding of the underlying mechanisms controlling hair cell formation and regeneration.

3. Atoh1, a bHLH transcription factor, is required for hair cell differentiation and regeneration after damage

bHLH transcription factors orchestrate cell fate commitment and specification during development (Lo et al., 1991; Ross et al., 2003), and a homologous group of “proneural” bHLH transcription factors plays a central role in neurogenesis. Several proneural bHLH transcription factors, including Atoh1, Neurog1, Neurod1, and Ascl1, contribute to the development of inner ear neurons and hair cells (Bertrand et al., 2002; Fritzsch, 2003; Tiveron et al., 2003; Jahan et al., 2013).

Atoh1, the mammalian homolog of the Drosophila gene atonal (also known as Math1 for “mouse atoh1” and HATH1 for “human atoh1”) (Ben-Arie et al., 1996), is a bHLH transcription factor that is both necessary and sufficient for the formation and differentiation of inner ear hair cells. Animal studies have shown that Atoh1-knockout mice fail to form cochlear and vestibular hair cells (Bermingham et al., 1999) as a result of progenitor cell apoptosis (Chen et al., 2002). In gain-of-function studies, overexpression of Atoh1 in mice by viral delivery led to ectopic hair cell formation in organ of Corti explants (Zheng and Gao, 2000). In vivo and in utero experiments in mice confirmed that forced upregulation of Atoh1 leads to the formation of extra hair cells (Woods et al., 2004; Izumikawa et al., 2005; Staecker et al., 2007; Gubbels et al., 2008; Liu et al., 2012; 2014; Atkinson et al., 2014), but very little is known about how cells regulate the level of Atoh1, especially in the cochlea.

4. Atoh1 level is critical for hair cell development and regeneration in mice

Proper spatiotemporal expression of Atoh1 is key for the differentiation and viability of hair cells. Atoh1 is expressed in mice in the inner ear progenitors before they have committed to the hair cell fate. It is first detected in the prosensory cells of the basal cochlear turn of the embryonic cochlea at embryonic day (E) 13.5. The expression of Atoh1 surges at the base of the cochlea by E14.5 and starts to appear in the apical turn by E17.5. Atoh1 expression begins to decline immediately after birth in the basal turn and by postnatal day 4 in hair cells in the apical turn (Yang et al., 2010; Pan et al., 2012).

Atoh1 affects not only hair cell differentiation, but also hair cell viability. Embryonic reduction of Atoh1 led to significant cochlear hair cell loss and abnormal hair cell bundle formation in a transgenic mouse model (Pan et al., 2012). Temporally, Atoh1 is critical for the survival of hair cell progenitors at the base of the cochlea in a 72-h window from E13.5 to E16.5, and deletion of Atoh1 outside this developmental window does not affect survival (Cai et al., 2013). Although Atoh1 overexpression in the supporting cells leads to ectopic hair cell formation, transgenic mouse models have shown that persistent Atoh1 overexpression in the cochlear hair cells leads to hair cell death and eventual hearing loss (Liu et al., 2012). A similar phenotype was also seen when Atoh1 protein degradation mechanisms were disrupted (Cheng et al., 2016).

5. Atoh1 is widely expressed

In addition to hair cell formation, Atoh1 plays important roles in the differentiation and formation of several other cell types. Atoh1 knockout induces cell loss in several tissues, including cerebellar granule cells (Ben-Arie et al., 1997; Flora et al., 2007; Forget et al., 2014), retrotrapezoid nucleus (RTN) neurons in the medulla (Wang et al., 2005; Huang et al., 2012; Ruffault et al., 2015), dorsal commissural interneurons in the spine (Miesegaes et al., 2009), intestinal goblet cells (Yang et al., 2001; VanDussen and Samuelson, 2010), and Merkel cells in the skin (Ben-Arie et al., 2000; Fröhlich et al., 2009; Morrison et al., 2009; Maksimovic et al., 2014; Wright et al., 2015). The distribution of Atoh1 in cells other than inner ear hair cells and the associated phenotypes in loss-of-function assays are listed in Table 1.

Table 1.

Atoh1 expression and its function in cell types other than hair cells

Cell type	Function of Atoh1	Phenotype in loss-of-function assays	Reference
Cerebellar granule cells	Maturation and proliferation	Missing external germinal layer (EGL) of the developing cerebellum	Ben-Arie et al., 1997; Wang et al., 2005; Flora et al., 2007; Forget et al., 2014
Retrotrapezoid nucleus (RTN) neurons of the medulla	Migration of neurons and establishment of essential connections with the pre-Bötzinger complex	RTN neurons fail to develop and establish connections with the respiratory rhythm-generating center in mammals	Rose et al., 2009; Huang et al., 2012; Ruffault et al., 2015
D1 commissural interneurons of the spinal cord	Ventral migration and fate determination of D1 interneurons	Progenitors adopting roof plate or D2 interneuron fates	Miesegaes et al., 2009
Intestinal goblet cells	Secretory lineage fate determination	Failure to form intestinal secretory cells	Yang et al., 2001; Shroyer et al., 2007; VanDussen and Samuelson, 2010
Merkel cells	Specification of Merkel cells	Absence of Merkel cells	Ben-Arie et al., 2000; Morrison et al., 2009; van Keymeulen et al., 2009; Maksimovic et al., 2014

Atoh1 expression has also been reported to be associated with tumorigenesis. Atoh1 acts as a tumor suppressor gene in the colon and skin, as it antagonizes tumor formation and growth, and many colorectal cancer and Merkel cell carcinoma patients show genetic mutations of Atoh1 (Bossuyt et al., 2009). However, Atoh1 acts as an oncogene in medulloblastoma (Zhao H. et al., 2008; Flora et al., 2009; Ayrault et al., 2010), indicating its multiple roles in differentiation and proliferation in different cell types.

6. Signaling pathways that regulate Atoh1

6.1. Notch signaling pathway

Notch signaling plays a key role in Atoh1 regulation and inner ear development (Jarriault et al., 1995; Lanford et al., 1999; Brooker et al., 2006). The major role of Notch signaling is to regulate cell fate decisions through lateral inhibition during development. It acts as a mediator between prosensory cells during development, leading these cells to differentiate into hair cells and supporting cells. Thus, Notch signaling is utilized to establish the structural pattern of hair cells and supporting cells in the cochlea (Kelley, 2003).

The inner ear is derived from the otic placode, which is recognized initially as a thickening near the hindbrain at E8 in mice. The otic placode invaginates by E10.5 to form an otocyst that contains prosensory cells and ganglion neuroblasts. The determination of hair cell or supporting cell fate is influenced by Notch signaling, and developing hair cells express Atoh1 along with the Notch ligands Jag2 and Delta1. These ligands bind to Notch1 in adjacent cells and induce the release of Notch intracellular domain (NICD) from their membranes. Upregulation of inhibitory bHLH transcription factors such as Hes1 and Hes5 follows after the increase of NICD, and expression of Atoh1 is blocked, leading to inhibition of hair cell fate and development of supporting cells (Kelley, 2003; 2006).

Pharmacological inhibition of Notch signaling by γ-secretase inhibitors increases the number of cochlear hair cells in neonatal cochlear explants (Yamamoto et al., 2006; Takebayashi et al., 2007; Bramhall et al., 2014; Li et al., 2015; Maass et al., 2015) and increases the differentiation of inner ear stem cells into hair cells (Jeon et al., 2011). In an in vitro lineage-tracing study involving genetically marking supporting cells with a fluorescent reporter, treatment with γ-secretase inhibitors after aminoglycoside-induced hair cell damage has been shown to promote hair cell regeneration through the direct transdifferentiation of supporting cells in the cochlea (Bramhall et al., 2014). An in vivo study also showed that direct round-window application of γ-secretase inhibitors resulted in recovery of cochlear function after noise-induced hair cell damage. This was evidenced by increased Atoh1 expression and direct conversion of supporting cells into hair cells, as measured by a similar genetic tagging model (Mizutari et al., 2013).

The responsiveness of the cochlea to γ-secretase inhibitors has been shown to be position-dependent and age-dependent. The apical region of the cochlea exhibits a stronger response in terms of transdifferentiation of supporting cells into hair cells compared to the basal region, and the response declines rapidly with age (Mizutari et al., 2013; Bramhall et al., 2014; Maass et al., 2015).

6.2. Wnt/β-catenin signaling pathway

The Wnt pathway plays a critical role in patterning and cell fate specification in the early development of the inner ear (Stevens et al., 2003; Gregorieff and Clevers, 2005; Riccomagno et al., 2005; Ohyama et al., 2006). In canonical Wnt signaling, Wnt binds to receptors on the cell membrane surface to disrupt the destruction complex consisting of axin, adenomatosis polyposis coli (APC), glycogen synthase kinase 3β (GSK-3β), and Disheveled (Dvl). This prevents β-catenin from being degraded in the cytoplasm. β-Catenin then translocates into the nucleus and subsequently forms nuclear β-catenin/Tcf complexes that drive the expression of Wnt target genes. In non-canonical Wnt signaling, Wnt proteins can work in a β-catenin-independent manner via the planar cell polarity (PCP) and Wnt/calcium pathways (Jansson et al., 2015).

Wnt/β-catenin signaling induces patches of hair cells and supporting cells in the auditory sensory epithelium and is involved in establishing or maintaining the distinction between sensory domains that contain hair cells and non-sensory domains in the inner ear (Stevens et al., 2003). Overexpression of β-catenin increases Atoh1 expression in neuroblastoma cells and neural progenitor cells. This upregulation is due to the interaction of β-catenin with the 3'-enhancer of the Atoh1 gene (Shi et al., 2010). Conditional knockout of β-catenin inhibits hair cell formation from sensory progenitors, while constitutive upregulation of β-catenin expands sensory progenitors and results in the formation of extra hair cells (Shi et al., 2012; 2014; McLean et al., 2017).

The Notch and Wnt/β-catenin signaling pathways interact with each other. As mentioned earlier, β-catenin increases Atoh1 expression through interaction with the 3' enhancer of Atoh1, and this accounts for the effect of Notch inhibition on Atoh1. Shi et al. (2010) found that inhibition of Notch signaling increases β-catenin expression, while simultaneous inhibition of β-catenin abolishes the increase in Atoh1 expression caused by Notch inhibition. Thus, β-catenin expression is required not only for increased expression of Atoh1, but also for the mitotic generation of hair cells in the cochlea after Notch inhibition (Li et al., 2015). Interestingly, a combination of Notch inhibition and Wnt signaling activation seems to have synergistic effects on the proliferation and transdifferentiation of progenitors or supporting cells in the neonatal cochlea, especially in the basal turn that usually shows only limited ability to form new hair cells after birth (Ni et al., 2016; McLean et al., 2017).

6.3. Post-translational regulation of Atoh1: the CK1-Huwe1-ubiquitin-proteasome pathway

Increasing evidence has shown that Atoh1 gene expression is tightly regulated by overlapping pathways, some of which have been described above. Post-translational regulation of Atoh1, however, is only now beginning to be understood. The ubiquitin-proteasome pathway (UPP) is the major post-translational regulatory system in eukaryotes, and it plays an important role in the development and physiology of eukaryotic cells (Tai and Schuman, 2008). The UPP has been implicated in the regulation of stem cell differentiation and lineage commitment via proteolytic degradation of key regulatory proteins. Specific protein substrates are conjugated by polyubiquitylation in an adenosine triphosphate (ATP)-dependent process and then targeted by the proteasome, the machinery by which proteins are degraded. The system is highly selective and tightly regulated. It not only degrades misfolded or damaged proteins, but is also essential for regulating cell-signaling pathways (Naujokat and Šarić, 2007).

The UPP consists of three key enzymes: a ubiquitin-activating enzyme (E1), a ubiquitin-conjugating enzyme (E2), and a ubiquitin ligase (E3). E3 catalyzes polyubiquitin (and sometimes monoubiquitin) chain formation by transferring ubiquitin that has been activated by E1 and E2 onto the lysine residues of specific substrates. Huwe1 (HECT, UBA, and WWE domain containing 1) is a large 4371-amino acid HECT-domain E3 ubiquitin ligase that is involved in proteasomal degradation of several protein substrates, including Atoh1 in the cerebellum and in the cochlea (Forget et al., 2014; Cheng et al., 2016). Huwe1 binds to Atoh1 and forms a lysine 48-cojugated polyubiquitin chain on Atoh1 as a signal for proteasomal degradation (Cheng et al., 2016). As a result, overexpression of Huwe1 decreases Atoh1 levels, while inhibiting Huwe1 increases the half-life of the Atoh1 protein.

A “degron” is a sequence element or modification that is sufficient to be recognized or targeted by an E3 ligase to promote ubiquitylation (Varshavsky, 1991). Atoh1 is found to be enriched in serine residues in the C-terminus, and many of the serines are conserved among different species. Casein kinase 1 (CK1), a serine/threonine selective protein kinase, phosphorylates Atoh1 at its serine-enriched C-terminal region, specifically serine 334, to form a phosphodegron that enhances the Atoh1-Huwe1 interaction and the subsequent ubiquitylation of Atoh1 (Cheng et al., 2016). The overexpression of CK1 increases the degradation of Atoh1, while inhibition of CK1 stabilizes Atoh1.

Interestingly, other sites, including other conserved serines at the C-terminus, have been reported to contribute to Atoh1 stability. Forget et al. (2014) found that sonic hedgehog (SHH) controls the phosphorylation of serine 328 and serine 339 of Math1 through protein phosphatase 2A (PP2A) activity to protect Atoh1 from degradation. Tsuchiya et al. (2007) reported that Atoh1 protein is expressed in normal colon tissue, but it is degraded by GSK3β in cancer tissue through phosphorylation of serine 54 and serine 58 of Hath1 (the equivalent of serine 52 and serine 56 of Math1).

The biological roles of Huwe1 in development and neurogenesis have recently been described (Zhao X. et al., 2008; 2009; D'Arca et al., 2010; Dominguez-Brauer et al., 2016; Urbán et al., 2016). Atoh1 degradation by Huwe1 is required for proper neuronal migration and differentiation in the cerebellum (Forget et al., 2014) because uncontrolled Atoh1 expression in the cerebellum leads to early postnatal lethality and interferes with cerebellar development (Helms et al., 2001). Spatiotemporal control of the Atoh1 protein level by Huwe1 is essential for cochlear hair cell fate determination and survival (Cheng et al., 2016), and conditional knockout of Huwe1 in the cochlear supporting cells at the embryonic or early postnatal stage leads to extra hair cells in the cochlea. However, hair cell death is observed when Huwe1 is conditionally knocked out in the cochlear hair cells, and this phenotype appears to be correlated with an increase in Atoh1 protein expression. This is in line with a previous study showing that Atoh1 overexpression in the hair cell causes hair cell death (Liu et al., 2012). Thus, the termination of Atoh1 activity after hair cell differentiation and maturation seems to be critical for normal cell patterning in the cochlea. Although several substrates of Huwe1 have been found to be associated with cell apoptosis, including Mcl-1, protein phosphatase 5 (PP5), p53, Myc, and cdc6 (Chen et al., 2005; Hall et al., 2007; D'Arca et al., 2010; Kurokawa et al., 2013), further studies are required to unravel the mechanisms of hair cell death underlying Huwe1 deletion and Atoh1 overexpression.

7. Concluding remarks

Atoh1 is a key transcription factor in the development and regeneration processes of the inner ear. Several signaling pathways have been shown to be involved in the regulation of its expression in the cochlea, including the Notch, Wnt, and ubiquitin-proteasome pathways. The studies reviewed in this article have shown that proper spatiotemporal control of Atoh1 expression is essential for inner ear development, not only at the transcriptional level, but also at the post-translational level (concluded in Fig. 1). Forcing Atoh1 overexpression or manipulating Atoh1-enhancing signaling pathways is one of most efficient strategies known for regenerating damaged hair cells. Incorporating knowledge of how Atoh1 is regulated will lead to a better understanding of cochlear development and to more insights into therapeutic strategies for hair cell regeneration.

Fig. 1.

Summary of Atoh1 regulation

Wnt signaling activates Atoh1 expression through interactions between β-catenin and the 3' enhancer of Atoh1. Overexpression of Notch signaling decreases β-catenin and decreases Atoh1 expression, while inhibition of Notch leads to upregulation of β-catenin and increased Atoh1 expression. At the post-translational level, CK1 binds to and phosphorylates Atoh1 to facilitate interaction with the E3 ubiquitin ligase Huwe1, which orchestrates ubiquitin transfer events from E1 (ubiquitin-activating enzyme) and E2 (ubiquitin-conjugating enzyme). This in turn leads to the polyubiquitylation and proteasomal degradation of Atoh1. Figure adapted and modified from Shi et al. (2010) and Cheng et al. (2016)