Sensorineural hearing loss (SNHL) afflicts over 278 million individuals worldwide and nearly half of all individuals over the age of 65 [1-3]. This multifactorial disease results from gene mutations, ototoxic drugs, environmental insults, or aging, and is often irreversible. Hearing loss imparts a substantial financial cost to society and bears an emotional and quality of life burden to the affected individual along with his/her family [1,4,5]. Current treatments with cochlear implants require an invasive surgical procedure without faithfully recapitulating normal hearing. Cochlear implants do not represent a perfect cure but rather a crutch to obtain some hearing capability [1, 6-12]. Prevention of hearing loss and regenerative medicine are two alternative pathways toward a possible perfect cure [13-15]. While prevention of hearing loss through protection of hair cells is ideal, it is not always possible [16-19], leaving regenerative medicine the sole ex post facto option for a permanent and restorative treatment of hearing loss. However, regeneration of damaged sensory epithelia in the mammalian inner ear is complicated as differentiated adult inner ear neurosensory tissue cannot re-enter the cell cycle and be used to replace neighboring cells, as is spontaneously accomplished in non-mammalian systems. This absence of proliferative capacity necessitates the use of exogenous stem cells as a medium for hair cell restoration, unless possibilities are discovered to jumpstart adult neurosensory epithelia proliferation.

The organ of Corti is a highly specialized sense organ with polarized and highly ordered inner and outer hair cells positioned in a stereotypic pattern on the basilar membrane surrounded by multiple, specialized supporting cells. Inner hair cells are innervated by approximately 95% of all sensory neurons and play the major role in sound perception while outer hair cells serve to fine-tune the hearing sensitivity. SNHL results primarily from damage to the inner hair cells in the organ of Corti. Regenerative treatment must therefore be aimed at replacing damaged organ of Corti inner hair cells; this can be done by either regenerating an entirely new organ of Corti or by replacing only the damaged hair cells, on a one-by-one basis. As inner hair cells die, morphologic changes occur to the organ of Corti including eventual loss of afferent neurons, making replacement of lost hair cells on a one-by-one basis problematic, in particular, years after loss of hair cells and subsequent organ of Corti dedifferentiation. In the absence of hair cells and key neurotrophins, originating from organ of Corti cells, spiral ganglion neurons are progressively lost [20,21]. The organ of Corti's central projections [20] form a complex tonotopic map that may not be capable of being faithfully recapitulated in a regenerated organ of Corti as brain plasticity declines with age [22-26]. If central projections cannot be retained in patients suffering from SNHL for years, not only must the individual hair cells be reformed, but so too must the neurons. In such cases, regeneration of the entire organ of Corti (with its delaminating sensory neurons) may provide the only feasible therapeutic option. This approach requires an initial jumpstart of the organ of Corti developmental process using stem cells transplanted into the cochlea, after which the context within the cochlea may suffice to activate neurosensory cell developmental programs to form a replacement organ of Corti. However, it is unlikely that an adult cochlea will provide the contextual environment needed as developmental signaling has long been turned off [27]. Therefore, stepwise cell fate restriction of transplanted stem cells using molecular signals required for organ of Corti induction, specification, and neurosensory cell formation is needed. Alternatively, the development program for the organ of Corti must be re-initiated to reform through proliferation and differentiation the lost organ. Additionally, applications of neurotrophic factors may be needed to facilitate a functional connection between the brain and the inner ear.

In patients suffering from short-term hearing loss, regeneration of damaged portions of the organ of Corti on a one-by-one basis may be possible as sensory neurons have not undergone apoptosis and the organ of Corti has not dedifferentiated. This therapeutic option requires a similar molecular understanding of neurosensory cell development but in addition requires exact targeting of prosensory cells to regions of damaged sensory epithelia while retaining proper orientation, structure, and functionality of healthy neighboring tissue. Transdifferentiation of neighboring cells cannot be used in mammals as they would deplete supporting cell populations as supporting cells are currently unable to re-enter the cell cycle [28-38]. In vitro attempts using stem cells have been initiated to reconstitute hair cells in the murine system and have been shown to be capable of differentiating into hair cell like cells [33,39-46]. However, despite great advances, it is apparent that we do not quite understand the molecular basis of hair cell development in vitro or in vivo or have the ability to control the placement of stem cells and/or hair cell precursors to damaged tissue.

Evolutionarily, the inability of mammalian neurosensory cell re-entry into the cell cycle may be due to the increased cellular diversity and complexity of the mammalian cochlea compared to non-mammalian species, where spontaneous proliferation does occur to restore lost hair cells [4,45,46]. For example, avian systems have two cell types (one supporting cell and one hair cell with graded size change) whereas mammalian cochleae have two distinct hair cell types (inner and outer) in a unique topographic distribution and at least six morphologically distinct supporting cell types in topographic restricted distribution in the organ of Corti [2]. Molecularly, a number of changes make cell re-entry difficult. In mice, hair cell precursors exit the cell cycle in an apex to base gradient from embryonic day 12.5 (E12.5) to E14.5 [47,48]. After cell cycle exit, re-entry is permanently prohibited by an increased concentration of cell cycle inhibitors, including p27^Kip1 (now CDKN1) [49], an increased phosphorylation burden of the tumor suppressor retinoblastoma (pRB) [50], and a corresponding relative decrease in proto-oncogene (including Myc) levels. This evolutionary loss of proliferation may have evolved to protect the delicate inner ear from additional proliferation which would disrupt the existing stereotyped organization and has been shown to be impossible without causing cell death in adult hair cells [27,51]. In mitotically quiescent hair cell precursors, a bidirectional wave of Atoh1, originating in the middle-base portion of the cochlea [52], specifies hair cell fate [53-57] and stabilizes the cell fate of surrounding cells to differentiate as supporting cells [58-60] (Figure 1).

Non-mammalian transdifferentiation and spontaneous cell cycle re-entry are seemingly incompatible with the mammalian system as far we understand it. Ideally, differentiated mammalian cells can be forced to re-enter the cell cycle. Forced expansion of proliferation through manipulation of both pRB and especially p27^Kip1 lasts until P14 in mice but human restoration is needed much later than P14 mouse equivalent. Furthermore, hair cells formed during this time are not viable. While cell specification [47-49,52-57] may render cells incapable of cell cycle entry by changing their molecular signature and enacting epigenetic changes [61-64], inhibiting tumor suppressors and increasing endogenous levels of proto-oncogene may reinitiate the cell cycle for the generation of new hair cells and supporting cells. Important to this is a greater understanding of Myc, implicated in both hair cell precursor proliferation and stem cell maintenance [65-69]. Clearly, a more complete understanding of proliferative control in the mammalian ear, including the regulatory network of miRs, may provide insights into the ability of neurosensory cells to re-enter the cell cycle. If the cell cycle could be reinitiated in adult and senescent ears to form either an asymmetrically dividing supporting cell or a symmetrically dividing inner ear stem cell, this could provide ample precursors available for transdifferentiation and/or cell fate re-specification (Figure 1). Until we possess the ability to restart proliferation in the adult ear, transplantation of stem cells seems to provide the sole avenue to hair cell regeneration and for enhancing our understanding of the molecular steps to turn a naïve cell into a hair cell. Some reports suggest the presence of an inner ear stem cell in mammals but the usefulness or potential for regeneration of these cells have not been studied. While a stem cell population in the ear would provide the ideal medium to manipulate hearing restoration, the same principles that would be applied to an ear specific stem cell population can be used with exogenous stem cells for the same purpose. Indeed, lessons learned through manipulation of stem cells in vitro may prove useful for the manipulation of ear stem cells in vivo. In conclusion, the apparent inability for adult mammalian neurosensory cells to re-enter the cycle makes stem cells the only viable therapeutic option for patients with SNHL.

In summary, both replacement of the entire organ of Corti and the replacement of lost hair cells on a one-by-one basis require the use of stem cells as adult mammalian inner ear neurosensory cells cannot be forced to re-enter the cell cycle as far as we understand it currently. As a result, the potential use of stem cells as therapy for hearing loss has blossomed [1,70-73]. Because of the varied cell types necessary for permanent hearing loss restoration (i.e. inner ear hair cells, sensory neurons), using either multipotent or pluripotent stem cells transplanted to the cochlea followed by incremental inductive cues provides a reasonable therapeutic option. Further progress to a perfect and permanent cure demands first, an understanding of stem cell sources and differentiation potentials; second, a comprehensive knowledge of the minimally essential genes required for ear development, organ of Corti formation, and hair cell maintenance; and last, the technological savvy to safely and non-invasively introduce stem cells, their differentiated derivatives, and/or the appropriate genetic therapy into a damaged cochlea.

Stem cells are self-renewing cells with variable differentiation potentials capable of both symmetric and asymmetric division. Stem cells are categorically divided into four groups depending on their potency: unipotent, multipotent, pluripotent, totipotent (Figure 2). The potency, or ability of a stem cell to differentiate into a specific cell type, is a property of intracellular gene regulation, epigenetics, and contextual interactions (the so-called stem cell niche) [74-80]. Stem cells exist throughout the body, but the skin, with both hair follicles and fibroblasts, are an easily accessible and non-invasive source of both multipotent and pluripotent stem cells. Depending on the potency of the stem cell, which ranges from the ability to make every cell type (totipotent) to the ability to make just a single cell type (unipotent), different stepwise cell fate restricting decisions must be used to direct the differentiation of the stem cell into a specific functional cell. Transplanted stem cells can survive in the cochlea to varying degrees but the cell type it differentiates into depends on the type of stem cell it originated from and the molecular signals it receives [81]. Directing the cell fate of stem cells requires knowledge of regulation of the cell cycle and the genes needed for specification of the desired tissue [1,39]. For example, a pluripotent stem cell (such as an iPSC) must first be stepwise restricted to an ectodermal fate, then directed toward an inner ear neurosensory cell fate, and finally to a hair cell fate. In contrast, an ectodermal multipotent stem cell (such as EPI-NCSC) need only be restricted to a neurosensory cell fate and then to a hair cell fate (Figure 1). Obviously, fewer local regulations (as seems to be the case in postnatal ears) [82] driving the direction of differentiation must be substituted by more specification prior to injection of cells to ensure that hair cells differentiate in the right type at the right place.

If transcription factors are the hands on a clock, miRs may well be the gears that really make the clock tick. MiRs are small non-coding 20-25 nucleotide single stranded RNAs that have a two to eight nucleotide “seed” region that interacts with the 3′ end of mature mRNAs. MiRs repress translation of mRNA or degrade mRNA depending on degree of complementarity of the seed region to the target mRNA. MiRs are not fully functional until processed from their immature strands. Abnormalities of processing enzymes, specifically Dicer knockouts, have revealed a large impact of miRs (and other small RNAs) on the proper development of the embryo. Dicer cleaves the hairpin loop of the pre-miRNA to form the mature miRNA and controls regulatory activity [169]. Pax2-Cre Dicer conditional knockouts (CKOs) lose the ability to differentiate into the three germ layers and show defective differentiation [171] and conditional deletion of Dicer with Foxg1-Cre results in loss of almost all cochlear neurosensory elements [174]. However, Dicer also plays other roles making the correlation with miR loss somewhat tenuous. The DICER enzyme is also necessary for siRNA production, an area not well understood in its rapidly growing significance that will not be dealt with here.

Gene therapy has been achieved in principle in other systems [233-236] but the stereotypic order of cells combined with the complexity of the organ of Corti cellular distribution remains a roadblock in the ear. We have previously discussed the importance of understanding stem cell regulation and what genes are necessary for stem cell differentiation to form functional hair cells, but all this falls on deaf ears unless the stem cells can target the appropriate tissue in vivo and the needed genes (i.e. Atoh1) can be regulated within the stem cells. Therefore, technical consideration must be given to both the targeting of stem cells to damaged tissue (and their survival) and the regulated gene expression within the stem cells (or in vivo newly proliferated cells).

The endolymph can safely accept up to 0.5 μL of stem cells. Stem cells are capable of efficient migration from the point of injection through the entire scala media [40,96,237]. A number of attempts have been made to target damaged tissue in the ear using various stem cells including neural stem cells [96], embryonic stem cells [128,238], and bone-marrow derived stem cells [239,240] but many of these cells were immunorejected or otherwise had unremarkable results [40]. To date, targeting and survival of stem cells remains the limiting factor in applying stem cell therapy to the ear.

Once established in damaged tissue, regulation of gene expression in the stem cell is necessary for tissue restoration. Regulation of stem cells in vitro has successfully established hair cell like cells, but manipulation of stem cells in vivo has not been accomplished. However, the ear is an ideal playground for viral transduction due to the encasement by the bony mastoid [241]. Three viral vectors are commonly used to target the inner ear tissue: Adenovirus (AV), Adeno-Associated Virus (AAV) and Lentivirus (LV). Each of these three viruses target cochlear cell types. AV targets hair cells, supporting cells, spiral ganglion neurons, fibrocytes, stria vascularis, and Reissner's membrane [40]. AAV targets hair cells, supporting cells, spiral limbus, spiral ganglion neurons, Reissner's membrane, and the spiral ligament [40]. LV targets spiral ganglion neurons, fibrocytes and Reissner's membrane [40]. Each of these vectors have been used to successfully transfect target cells with different genes: GDNF [242], HGF [243], Catalase [244], BDNF [245-247], NTF-3 [248], Atoh1 [249], EGFP [250, 251], and GJB2 [252] in the inner ear. Unfortunately, each of these vectors have shown off-target effects, notably in vivo injections of Atoh1 in the mammalian cochlea have resulted in both normal hair cell formation [158,249,253,254] and Atoh1 positive hair cells outside the organ of Corti [255]. While viruses do have some tropic specificity, they can target ectopic tissue. Furthermore, viruses themselves may pose a health risk if injecting DNA. The ability to stably and specifically target damaged tissue safely with both exogenous stem cells and the necessary genes is the final step necessary for hearing loss restoration. Multiple laboratories are working on other stem cell targeting and gene regulation expression methods including the use of nanoparticles [256] and other non-viral vectors [257-260] that will be safer, more efficient [261] and will reduce auditory trauma [237,262-264]. For example, regulated expression of target genes could be accomplished through the conditional activation of non-mammalian promoters that drive hair cell specific genes in target tissue, not ectopic sites [265]. Regardless of method, implanted stem cells must be cell fate restricted through regulated gene expression to achieve gene therapy in the ear.

With the dawn of gene therapy and technological booms in stem cell technology, regenerative medicine appears to be the wave of the future, promising cures to a wide array of human diseases. While regenerative medicine is still in its infancy and years away from clinical adaptation, it represents new interactions between basic scientists and clinicians. Unprecedented in applicability, regeneration of any tissue requires a solid foundational knowledge of the molecular basis of developmental processes and stem cell regulation. As basic scientists dredge through thousands of genes in the hopes of elucidating the minimal essential gene networks required for induction of directed differentiation and clinicians have increasing numbers of patients asking about permanent cures, the race from bench to bedside has never been more important as well as challenging. Progress in network analysis highlights the controllability of a cell network and the minimal nodes needed to control that network [266].

Millions of patients suffer from hearing loss resulting from lost hair cells in the organ of Corti. Gene therapy and stem cells may offer a permanent cure for afflicted individuals, but the complexity of the ear compared to other systems, makes regeneration of hair cells a daunting task. Yet, the past ten years have shown an incredible advancement in the molecular understanding of ear development [15,267]. A continued improvement in understanding of ear development amalgamated with the nonstop advancements of gene therapy and stem cell biology may someday be fine-tuned music to the ears of many.