Experimental drugs for the prevention or treatment of sensorineural hearing loss

ABSTRACT

Introduction

Sensorineural hearing loss results in irreversible loss of inner ear hair cells and spiral ganglion neurons. Reduced sound detection and speech discrimination can span all ages, and sensorineural hearing rehabilitation is limited to amplification with hearing aids or cochlear implants. Recent insights into experimental drug treatments for inner ear regeneration and otoprotection have paved the way for clinical trials in order to restore a more physiological hearing experience. Paired with the development of innovative minimally invasive approaches for drug delivery to the inner ear, new, emerging treatments for hearing protection and restoration are within reach.

Areas covered

This expert opinion provides an overview of the latest experimental drug therapies to protect from and to restore sensorineural hearing loss.

Expert opinion

The degree and type of cellular damage to the cochlea, the responsiveness of remaining, endogenous cells to regenerative treatments, and the duration of drug availability within cochlear fluids will determine the success of hearing protection or restoration.

KEYWORDS:

• inner ear therapeutics
• otoprotection
• regeneration
• small molecules
• synaptopathy
• sensorineural hearing loss

1. Introduction

Hearing loss affects more than 25% of the population over the age of 60. The World Health Organization estimates that one in ten people will suffer from significant hearing loss within 30 years [1,2]. Sensorineural hearing loss (SNHL) is irreversible and affects various cell types within the inner ear [3,4]. Cochlear damage leads to reduced detection of sounds and speech understanding and can affect all ages. Causes range from genetic hearing loss affecting children from birth, to environmental factors such as noise trauma or ototoxic drugs, to age-related changes in an ever-growing elderly population [5]. Current sensorineural hearing rehabilitation is limited to hearing aids for mild-to-moderate SNHL that provide sound amplification, or cochlear implants in cases of severe to profound SNHL with poor speech understanding that precludes benefit from amplification [6,7]. In either case, the result is subpar compared to normal hearing. Hearing aids do not provide significant high-frequency amplification and are unable to improve word understanding. Cochlear implants, while successfully improving speech perception, still fall short of ‘natural’ hearing, most notably with respect to music perception [6,7]. Current experimental drug treatments focus on protection or restoration of hearing after inner or outer hair cell damage, neuronal damage, or damage to the stria vascularis [2].

Recent advances in the fields of otoprotection and inner ear regeneration have led to the development of novel, minimally invasive approaches, which are geared toward protecting or restoring natural, biological hearing [8,9]. Here, we provide an overview of the latest experimental drug therapies for prevention and restoration of sensorineural hearing loss.

1.1. Anatomy

The cochlea, a portion of the inner ear, is a snail-shaped, fluid-filled compartment within the petrous portion of the temporal bone. It is divided into three chambers, two of which, the scala tympani, and scala vestibuli, are filled with perilymph (resembling cerebrospinal fluid and low in potassium chloride and high in sodium chloride), and the scala media, which contains endolymph (high in potassium chloride and low in sodium chloride). The scala media also harbors the cochlear duct, which is home to the organ of Corti with the sensory cells (hair cells) of the cochlea (Figure 1). Its unique structure allows frequency tuning along the cochlear axis – high frequencies are perceived at the base, while low frequencies are detected at the apex [10]. Along the organ of Corti, one row of inner hair cells and three rows of outer hair cells are flanked by nonsensory supporting cells and connect to the peripheral neurites of spiral ganglion neurons (SGNs). The cell bodies of the SGNs are located in Rosenthal’s canal, and their central axons project along the modiolus to the brainstem (Figure 1). Approximately 25 000 to 33 000 SGNs are found in the modiolus and connect with 3 500 inner and 12 000 outer hair cells [11].

Figure 1. Schematic depicting a cross section through the scala media of the cochlea with the organ of COrti housing the sensory epithelium with inner (IHC) and outer (OHC) hair cells, connecting to the primary sensory neurons, the spiral ganglion neurons (SGN) within the bony Rosenthal’s canal.

Inner hair cells have the ability to transduce mechanical movements into electrical signals via fiber-like structures at the apical end, the stereocilia [12]. Ribbon synapses at the basal end of hair cells release the neurotransmitter glutamate in response to sound-evoked stereocilia motion. Afferent spiral ganglion neurons are stimulated via their α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and N-methyl-D-aspartate (NMDA) glutamate receptors, and the electrical signal is then propagated along the auditory pathway. Outer hair cells act in response to auditory signals and function as modulators and amplifiers of sound [13]. The protein Prestin in the membrane of outer hair cells can undergo conformational changes in response to electrical potential, and this electromotility contributes to frequency selectivity and helps to amplify the auditory signal [13]. The electrical gradient between endolymph and perilymph is maintained by the stria vascularis, which flanks the scala media and organ of Corti, and which harbors various ion pumps to restore the resting potential [14].

Each SGN only contacts one hair cell, but multiple SGNs can be in contact with a single hair cell. This allows for frequency selectivity along the cochlear duct [15]. These primary afferent (Type I) neurons comprise roughly 95% of the cochlear SGN population. A small subgroup of neurons (Type II) are efferent neurons reaching the outer hair cells [16,17]. Several subtypes of Type I afferent neurons with different thresholds and response rates contribute to tuning at the same frequency [18]: Low threshold, high-spontaneous rate (SR) neurons; high threshold, low SR neurons; and neurons with intermediate response patterns all contribute to hearing at a given frequency. High threshold, low SR fibers may be particularly relevant for sound perception in background noise, while low threshold, high SR fibers may be important for sensitivity of sound detection [19]. All the sensory and non-sensory cells within the membranous labyrinth of the inner ear are susceptible to damage and aging to varying degrees. From a clinicopathological standpoint, loss of OHCs has been associated with decreases in threshold detection, while IHC and SGN loss has been associated with decreases in clarity of sound detection, as measured by word recognition [20].

Most mammals have highly similar cochlear structures to the human inner ear, making them valuable tools to study damage, protection, and restoration of hearing [10].

1.2. Functional hearing measurements

Various subjective and objective tests can be utilized to measure the degree and type of hearing loss in the human. The workhorse of audiologic testing is the pure tone threshold audiogram, which detects tones of various frequencies (usually recorded between 125 and 8000 Hz) at their lowest detection level (sound pressure level). Air conduction and bone conduction thresholds can be determined, but the test is dependent on the patient and therefore subjective [21]. Word recognition or speech discrimination testing refers to a suprathreshold level test above the minimum hearing level for speech and measures the ability to recognize phonemes or monosyllabic words [22], whereas speech in noise (SIN) testing comprises various different tests to evaluate sound, word, and sentence understanding in background noise, and simulates a more ‘real life’ hearing situation [23].

Objective tests of auditory function include auditory evoked responses such as otoacoustic emissions (OAEs) and auditory brainstem responses (ABRs). OAEs, specifically distortion product OAEs (DPOAEs), offer a way to test cochlear (outer hair cell) function independent of cochlear nerve function [24]. The threshold response, as well as suprathreshold amplitudes, are helpful to determine inner ear health. Present DPOAEs in the setting of an abnormal ABR response can indicate isolated auditory neuropathy.

The ABR records an electrical response consisting of synchronized action along the auditory circuit in response to either click or tone stimuli. The evoked, summed signal consists of five waves (I-V), the first being the activity of the spiral ganglion nerve fibers of the cochlea, and the latter ones representing the subsequent relay stations of the auditory pathway all the way to the auditory cortex [25]. ABR testing can determine a hearing threshold, as well as the overall wave latency, interpeak latencies, and suprathreshold amplitudes.

In animal models without prior noise exposure and under highly controlled conditions, a permanent reduction of the Wave I suprathreshold amplitude can be used to determine the degree of synaptic loss after noise exposure with only temporary threshold shifts [26–28]. However, in humans, the interpretation of changes in the Wave I ABR amplitude may be less strictly indicative of synapse loss, particularly in elderly patients [29,30]. Envelope following responses (EFR) and the middle ear muscle reflex (MEMR) have been proposed as diagnostic tools to measure synaptic damage in animals, but the interpretation of these tests may also be less straightforward in humans as the outcome measure (in this case word recognition scores) varied widely and could not reliably serve as diagnostic tool for each individual patient with difficulty hearing in noise [31–35].

2. Otoprotection

A variety of drugs have ototoxic potential and can lead to manifold types of cellular damage within the cochlea. We will exclude vestibulotoxic drugs for the purpose of this review and focus on two major ototoxic drug groups that can cause SNHL. The most common ototoxic effects are seen after treatment with platinum-containing chemotherapeutic drugs, such as cisplatin and aminoglycoside antibiotics. Ototoxicity can only be proven if pre- and post-treatment audiologic testing is available, but often the diagnosis is made clinically, when a new subjective hearing loss is confirmed by audiometry, and a history of recent treatment with ototoxic medication is obtained [36,37].

2.1. Aminoglycoside antibiotics

Aminoglycoside antibiotics are broad-spectrum antibiotics often used for the treatment of severe gram-negative infections, e.g. severe mycobacterial infections or meningitis [36]. Up to 3% of neonates have been found to have hearing loss after treatment with aminoglycosides [38], and between 20% and 60% of patients have demonstrated SNHL after multiday treatment with aminoglycosides [37,39,40]. Nonetheless, this group of antibiotics is frequently used, especially in developing countries, due to their low cost and good effectiveness in treating severe infections [38]. Aminoglycoside ototoxicity is dependent on the cumulative dose, frequency of treatment courses and times, and serum levels [41].

2.2. Cisplatin

Cisplatin is a platinum-containing chemotherapeutic drug that is in wide use for the treatment of various pediatric and adult malignancies. It is the work horse of multimodality chemotherapies for lung, colon, testicular, ovarian, bladder, and head and neck cancers. As with aminoglycosides, the ototoxic effect is dependent on the cumulative dose and treatment frequency [37,42].

2.3. Ototoxic effects in the cochlea

Cisplatin and aminoglycosides can enter the perilymph through the blood labyrinth barrier and the endolymph through the stria vascularis, presumably through a combination of ion channels, transporters, and transcytosis [43]. From there, they are taken up via stereocilia at the apical portion of the hair cells [44,45]. The nonselective cation channel TMC1, as well as various transient receptor potential (TRP) channels such as TRPV1 and TRPV4 have been implicated as entry points for aminoglycosides [46–48]. Lysosomal trafficking and nonreceptor-mediated endocytosis have also been reported as entry mechanisms into hair cells for aminoglycosides, and passive diffusion allows cisplatin to enter across the hair cell membrane [49,50]. Outer hair cells of the high frequencies are most susceptible to these ototoxic drugs. Outer hair cells at mid and lower frequencies and inner hair cells are subsequently affected in a dose-dependent manner, leading to permanent SNHL [51]. While cisplatin appears to prefer cochlear hair cells, aminoglycosides have demonstrated equally damaging effects in cochlear and vestibular hair cells [50]. The wide-ranging mechanisms that lead to hair cell death have not been fully elucidated, but likely include generation of reactive oxygen species by both aminoglycosides and cisplatin, leading to mitochondrial dysfunction and, ultimately, apoptosis [36]. Patients with concurrent inflammatory processes, renal failure, or various genetic polymorphisms are at higher risk for SNHL when treated with aminoglycosides or cisplatin [37,52,53]. Interestingly, a genetic susceptibility to aminoglycosides has been reported. Patients carrying mitochondrial mutations for 12S rRNA, A1555G, and C1494T are most commonly at risk of developing ototoxicity [54].

2.4. Otoprotective drugs

As both aminoglycoside antibiotics and cisplatin are important medications, multiple in vitro and in vivo studies have investigated novel compounds with a goal to protect the inner ear from the ototoxic effects of these drugs, while maintaining their effectiveness elsewhere in the body. Although the majority of these studies are at the preclinical stage, a few substances are already being tested in clinical trials [2,36]. Prevention of reactive oxygen and nitrogen metabolites and apoptotic cascades are at the center of current drug development strategies, and various antioxidants (including D-methionine and N-acetylcysteine, NAC; alpha-tocopherol, ebselen and sodium thiosulfate) have shown otoprotective effects in vitro and in animal studies [55–57]. BRAF inhibitors are in preclinical trials with a goal to reduce cisplatin toxicity without compromising its chemotherapeutic effectiveness. Various substances of this group also act as chemotherapies, and it may be possible to combine them with cisplatin and reduce overall dosages of both drugs to reduce side effects [58].

Additional otoprotective treatments have focused on blockage of ototoxic drugs at the blood labyrinth barrier, as well as direct modification of ototoxic drugs to ameliorate hair cell damage [43,55,59,60]. Table 1 summarizes experimental drugs that are currently being tested in ongoing clinical trials.

Table 1. Otoprotective drugs in clinical trials.

AGENT	MECHANISM	TARGET	CLINICAL TRIAL	ADMINISTRATION
N-acetylcysteine, NAC	Glutathione precursor, antioxidant	C	NCT04291209 NCT04226456	IT, IV
D- Methionine	Antioxidant, ROS scavenger	N	NCT02903355	Oral, (IT)
Sodium thiosulfate, STS	Antioxidant	C	NCT05382338 NCT00716976	IV
Ebselen, SPI-1005	Gluthatione mimic, antioxidant	C, AG	NCT01451853 NCT02819856	Oral
SENS-401, Azasetron	Selective 5-HT3 receptor antagonist	C	NCT03603314	Oral
Steroids	Anti-ROS	C, N	NCT04766853	IT
Statins	Antioxidant	C	NCT04915183 NCT04817904	Oral
Zonisamide	Anticonvulsant	N	NCT04768569	Oral
DB-020 (STS formulation)	Antioxidant	C	NCT04262336	IT
ACOU085	KCNQ4 agonist	C, SNHL	Germany

Note: AG, Aminoglycosides; C, Cisplatin; CA, Carboplatin; N, Noise; SNHL, sensorineural hearing loss; IT, intratympanic injection; ROS, reactive oxygen species.

With respect to cisplatin, such antioxidant approaches need to be evaluated in the context of the patient recipient’s overall cancer treatment. Some studies have highlighted reduced efficacy of cancer drugs with concurrent, systemic treatment of antioxidants, and the safety profile as well as potential drug interactions with the otoprotective drug need to be assessed [36]. To limit systemic exposure to the antioxidant activity, ongoing clinical trials have increasingly utilized intratympanic (IT) injections for delivery in close proximity to the inner ear.

However, despite the wealth of preclinical studies on otoprotection for aminoglycosides and cisplatin, there is no clear clinical consensus regarding the best approach to pursue. This may in part be related to variable clinical trial outcomes. Completed clinical trials for otoprotection over the last 10 years have demonstrated mixed efficacy, and clinical studies frequently failed in later clinical trial phases [61–63]. Ongoing drug development and pre-clinical studies, however, may hold promise [60].

3. Drug delivery

The inner ear is protected by a blood – labyrinth barrier similar to the blood-brain barrier, which greatly limits drug entry after systemic administration [64]. Therefore, local transcanal drug delivery via the middle ear has gained significant interest, as it increases drug concentration at the target site and reduces unwanted systemic pharmacological side effects [8,65,66]. Various established and experimental access routes to the inner ear are available in animal models and humans for direct and indirect delivery, with the caveat that the inner ear is encased in the temporal bone and the only membranous communication with the middle ear exists at the round window membrane (RWM) and the oval window [8,65,67,68]. Transtympanic drug delivery, a relatively safe and minimally invasive approach, is already in clinical use for the delivery of steroids and gentamicin [69]. This indirect approach involves injection of the drug through the tympanic membrane into the middle ear cavity. Its advantage lies in preservation of cochlear structures and residual hearing, but drug diffusion across the RWM and OW is limited by the intrinsic characteristics of a given drug, anatomical features of the RWM such as thickening of the RWM or the presence of a RWM pseudomembrane, as well as the persistence of the fluid depot at the RWM and in the middle ear [69,70]. More sustained drug exposure at the RWM and increased inner ear penetration has been attempted and variably successful with beads, various hydrogels, lipid formulations, thermosensitive polymers, and nanoparticles [70–73].

Direct access to the RWM for drug deposition or injection of compounds through the RWM into the inner ear is possible, but requires a surgical approach, with elevation of a tympanomeatal flap [74]. This approach potentially widens the scope of drugs deliverable to the cochlea and may speed the development of other novel approaches, including that of virally mediated gene therapy. The major caveat of such an attempt lies in the higher risk of iatrogenic SNHL, as well as the possibility of meningitis and perilymph fistula [75,76].

4. Hearing restoration

Sensorineural hearing loss in humans and mammals has long been considered irreversible, as damaged cellular structures do not undergo spontaneous repair or renewal. Loss of hair cells, synapses, and neurons, as well as defects within the nonsensory cells of scala media, have been well-characterized in preclinical models and in postmortem human temporal bone specimens. Twenty years ago, the discovery of progenitor-like cells within the cochlea initiated a new era for hearing restoration. While these ‘dormant’ cells failed to induce regeneration in vivo after damage, they demonstrated regenerative potential when stimulated in vitro, albeit at a low yield [77,78]. In the two decades since, the inner ear regeneration field has been on a quest to 1) identify and characterize inner ear progenitors, to 2) identify upstream manipulators to initiate cell fate switches, and to 3) increase the yield of newly generated hair cells.

4.1. Hair cell regeneration

Hair cell loss has been implicated as a major factor in age- and noise-related SNHL, as well as ototoxic and genetic hearing loss. Non-mammalian species such as amphibians and birds have been known to spontaneously regenerate hair cells during the aging process and after damage. In these species, hair cell pools are replenished from continuously dividing and transdifferentiating supporting cells [79–81].

Researchers have therefore sought to understand the underlying molecular mechanisms that enable hair cell regeneration in birds, amphibians, and fish, in the hopes that their findings might one day be extrapolated to the mammalian and human inner ear. The discovery of the transcription factor Atoh1 (Math1) was the first major milestone toward inner ear regeneration in mammals [82–85]. Atoh1 activation proved to be necessary and sufficient to drive inner ear progenitors or nonsensory cells within the organ of Corti toward a hair cell lineage and initiate the inner ear hair cell phenotype [86,87]. The upstream governing pathways and transcription factors leading to Atoh1 activation provided ample new targets for drug development [88,89]. Specifically, the interplay between the Notch and Wnt pathways for upstream stimulation of Atoh1 has taken center stage in the field of hair cell regeneration [90–94].

Two main mechanisms of hair cell regeneration have been observed after manipulation of the Notch and Wnt pathways. Inhibition of the Notch pathway alone induced direct transdifferentiation of supporting cells into hair cells in mammalians in vitro and in vivo, while concomitant activation of the Wnt and modulation of the Notch pathway in a subset of Wnt-responsive supporting cells induced mitotic, asymmetric division of supporting cells with subsequent differentiation into hair cells and supporting cells [95–97] (Figure 2). These findings have since become the basic modules and tools of ongoing preclinical and early clinical hair cell regeneration attempts.

Figure 2. Hair cell regeneration principles. Direct transdifferentiation or mitotic, asymmetric division of supporting cells.

Active lateral inhibition determines the fate of hair cells or supporting cells from common progenitors during development, and active Notch signaling in mature supporting cells prohibits them from transdifferentiating into hair cells [98]. Signal transduction of the transmembrane protein Notch is induced by cleavage of its intracellular domain via the enzyme γ-secretase. Inhibition of γ-secretases by small molecule drugs thereby renders the Notch pathway inactive and enables hair cell differentiation from surrounding supporting cells [99]. More than 100 different γ-secretase inhibitors of two different subtypes have been synthesized, originally for applications in the central nervous system (CNS) [100]. Previous drug screens initially identified LY411575 as one of the most effective inhibitors in the inner ear [99,101], but since then, additional γ-secretase inhibitors with inner ear activity have been discovered, some of which are currently being tested in clinical trials for SNHL (Table 2). There is ongoing debate about the utility of a transdifferentiation approach, which in principle would deplete the supporting cell pool, thereby potentially removing necessary trophic support for hair cells [95].

Table 2. Hair cell and synapse regeneration.

AGENT	MECHANISM	TARGET	CLINICAL TRIAL	ADMINISTRATION
LY3056480	Notch inhibition	SC, SNHL	NCT05061758	IT (4 doses)
OTO-413 (BDNF formulation)	Neurotrophin	IHC synapse	NCT04129775	IT (single dose)
PIPE-505	Notch inhibition, Netrin modulation	SC, IHC synapse, SNHL+SIN	NCT04462198	IT (single dose)
LY3056480	Notch inhibition	SC, SNHL	REGAIN, Europe	IT (3 doses)
FX-322 (CHIR99021, VPA)	Notch, Wnt modulation	SC, SNHL, SSNHL	NCT04120116 NCT04601909 NCT03616223 NCT05086276	IT (1 dose, 2 doses, 4 doses)

Note: SC, supporting cells; IHC, inner hair cells; SNHL, sensorineural hearing loss; SSNHL, Sudden sensorineural hearing loss; IT, intratympanic injection.

Nonetheless, γ-secretase inhibitors represented one of the first preclinically effective therapies to initiate cell fate switch and increase the yield of new hair cells from supporting cells. This success culminated in the start of several clinical trials for hearing restoration with γ-secretase inhibitors. A phase I/IIa clinical study using LY3056480 (AUD1001) confirmed safety and good tolerability of drug in patients with mild-to-moderate SNHL. According to the company, the (unpublished) results suggest an improvement in hearing speech in noise.

It remains to be seen if transdifferentiation of supporting cells into hair cells will be effective, especially given that response to Notch modulation changes in early postnatal supporting cells [102]. Pipeline Therapeutics similarly focused on the local treatment of patients with γ-secretase inhibitors, specifically in subjects with difficulty hearing in background noise, with a goal of inducing Netrin-dependent synapse regeneration between inner hair cells and auditory neurons in addition to driving hair cell regeneration [103,104]. No results have been published thus far regarding the outcomes of the trial.

Recently, the search for the elusive ‘dormant’ stem cells of the inner ear has revealed a subset of Wnt-responsive supporting cells as the main progenitor source for auditory hair cells. The canonical Wnt/β-catenin pathway is evolutionarily conserved, and, among multiple other functions, regulates organogenesis and stem cell renewal. When Wnt signaling is active, cytoplasmic β-catenin is stabilized and can translocate into the nucleus. There, it acts as transcriptional co-activator, complexing with LEF/TCF transcription factors [105]. Activation of the Wnt pathway in cochlear supporting cells in vitro or in transgenic mice in vivo drove differentiation of these supporting cells into hair cells. GSK3-β inhibitor-mediated activation of Wnt/β-catenin signaling has been a major target of studies aimed at modulation of the Wnt pathway with small-molecule drugs [95,106]. Active Wnt pathway signaling leads to the reduction of glycogen synthase kinase 3β (GSK3β), an enzyme involved in destabilizing β-catenin. When in its stable form, β-catenin can translocate into the nucleus and activate transcription of hair cell genes, most prominently Atoh1 [106]. GSK3β inhibitors such as CHIR99021 therefore mimic active Wnt signaling and promote hair cell regeneration from Wnt-responsive supporting cells [92].

The combination of CHIR99021 with HDAC inhibitor valproic acid (VPA), which is thought to modulate the Notch pathway, has been successful in increasing the yield of newly differentiated hair cells from Wnt responsive supporting cells [95]. CHIR990121 and VPA together initiate and increase mitotic division of supporting cells, which then give rise to a new hair cell and a new supporting cell. This approach, in theory, can replenish the supporting cell pool during hair cell regeneration [95,97,107].

This concept has been the basis of clinical trials with FX-322, a combinatorial preparation of CHIR99021 and VPA aiming to induce hair cell regeneration from supporting cells after local intratympanic application. The initial data from a Phase Ib clinical trial in subjects with moderate-to-severe SNHL was recently published [108]. Stable improvement was documented for speech discrimination scores in a subpopulation of subjects with preexisting reduced word recognition scores after one injection of FX-322, although pure tone thresholds did not change significantly in this small group of 15 subjects. A positive trend at the extended high frequencies (at and above 8 kHz) was noted, however. Additional follow-up trials, which included single and multiple applications of FX-322, demonstrated a trend toward improved speech discrimination, but no significant difference in threshold audiograms. The lack of improvement of pure tone thresholds begs the question as to whether the claimed hair cell regeneration is in fact taking place in humans after FX-322 treatment. Based on the published PK data, the drug may only be present in therapeutic concentrations at the extended high frequency (EHF) range of around 8 kHz but may not reach lower frequencies at sufficient levels. This may be relevant, as others have theorized that EHFs may be responsible for impaired hearing in noise in patients with normal threshold audiograms [109]. Additional studies are needed to determine if the improvement in word recognition could be related to ongoing synaptic regenerative processes. Table 2 summarizes clinical trials for inner ear regeneration.

Other small molecule approaches such siRNA-based technologies remain at the preclinical stage [110], and development of novel potential drug targets for hair cell regeneration is ongoing. Promising candidates include the cell cycle regulator genes MYC and P27kip1, manipulators of the Hippo/YAP and Lin28/Let-7 pathways, as well as epigenetic modifiers that have been shown to influence DNA methylation and histone acetylation and deacetylation in the cochlea [111–117]. The goal remains stimulation of supporting cells to undergo either direct transdifferentiation or mitotic division in an attempt to regenerate the hair cell pool [112,118].

4.2. Synapse regeneration

The traditional view of SNHL revolved around the concept that in the vast majority of cases- apart from isolated, rare cases of primary auditory neuropathy- cochlear insults led to direct damage and loss of inner and outer hair cells, followed by secondary degeneration of primary afferent auditory neurons [4,119]. This was supported by early animal studies and human temporal bone histopathology, where loss of hair cells was apparent quickly after cochlear injury, while spiral ganglion nerve cell loss was not observed until later timepoints [120]. However, the discovery of ‘hidden’ hearing loss and auditory synaptopathy has fundamentally altered this early conceptualization of SNHL and provided new concepts regarding hearing loss and regenerative strategies [28,121,122] (Figure 3).

Figure 3. Principle of synaptopathy and synapse regeneration following noise damage. Synaptic connection between peripheral spiral ganglion axons and inner hair cells is lost and leads to synaptopathy with ‘hidden’ hearing loss. Neurotrophic compounds can regenerate ribbon synapses.

Cochlear synaptopathy was initially observed in animal studies studying reversible and irreversible noise damage. After exposure to a noise band below the threshold for irreversible SNHL, animals exhibited a temporary threshold shift in ABR and DPOAE testing, which recovered within 1–2 weeks. Hair cells and neurons survived noise injury at this level. However, synapses between the inner hair cells and afferent auditory neurons at frequencies above the noise exposure band were permanently lost [28]. Loss of these synapses correlated with a permanent decline in suprathreshold Wave I ABR amplitudes [28]. Further studies in guinea pigs revealed that the high-threshold low SR auditory nerve fibers may be the most vulnerable to noise, and might account for loss of the suprathreshold Wave I amplitude without affecting threshold detection [123]. The phenomenon of threshold maintenance in the setting of decreased suprathreshold responses has led to the term ‘hidden hearing loss’ [28]. The loss of inner hair cell synapses may be a result of excessive, glutamatergic transmitter release in response to loud noise, leading to excitotoxic overstimulation of AMPA and NMDA receptors on spiral ganglion neurites [124,125].

Electron microscopy studies have revealed loss of synapses and afferent spiral ganglion nerve fibers prior to the loss of hair cells in age- and noise-related hearing loss [126–128]. Early postmortem histopathology of human temporal bones has similarly identified a decrease in synapses associated with aging. These studies laid the basis for our growing understanding of how synaptopathic changes influence speech understanding in background noise even when auditory thresholds are seemingly within normal range [28,128,129].

Synaptic regeneration could therefore be a treatment target for millions of patients with noise- and age-related hearing changes. Neurotrophins are soluble growth factors that function to provide trophic support to neurons during development and adulthood and promote survival of neurons and synaptic connections. Two neurotrophins, brain-derived neurotrophic factor (BDNF) and neurotrophin-3 (NT-3), and their respective receptors, TrkB and TrkC, govern establishment and maintenance of inner ear ribbon synapses and support spiral ganglion neuron survival [130,131]. Multiple in vitro culture models have confirmed the beneficial effects of BDNF and NT-3 on neurite outgrowth, SGN survival, and synapse regeneration [132,133].

In animals with auditory synaptopathy, local treatment with neurotrophic proteins or overexpression of neurotrophic genes resulted in improved axonal regrowth, SGN survival, and restoration of ribbon synapses [134–140]. A search for neurotrophic analogues and TrkB and TrkC agonists led to various novel compounds that possess neurotrophic activity and promote synapse regeneration in experimental studies. BDNF and NT-3 themselves, as well as monoclonal antibodies for selective activation of TrkB or TrkC and engineered neurotrophins have been tested, in addition to small molecule neurotrophic analogues and receptor agonists [141–144]. Axonal guidance molecules and Notch modulators are also being actively investigated in synapse regeneration [145,146].

Preclinical studies have attempted to address the difficulties associated with local exposure and RWM penetration of various neurotrophic compounds, and are tackling the challenge of sustained delivery and presence of these molecules for patients with more severe or long-term SNHL and significant neurite retraction [147,148]. Promising results emerged from novel hybrid molecules with neurotrophic activity. Small-molecule BDNF and NT3 analogues were linked to bisphosphonates, bone-binding drugs commonly used for osteoporosis treatment. These compounds promoted neurite outgrowth and synapse regeneration in in vitro neurite outgrowth studies and a cochlear synaptopathy model, and may allow for long-term inner ear presence within cochlear bone [141,142,149,150].

A first Phase I/IIa clinical trial has just recently been completed to address patients with potential cochlear synaptopathy and difficulty with speech perception in noise (Table 2). A transtympanic injection was chosen to deliver a BDNF-Poloxamer 407 hydrogel formulation for longer-term local exposure at the RWM [151].

5. Expert opinion

In spite of intense effort, there remains a tremendous unmet need for novel therapies aimed at the prevention and treatment of hearing loss. However, several major barriers remain to be overcome. One major barrier is the inability to predict the histologic correlate of a patient’s phenotypic hearing loss. For example, our data rely mostly on the knowledge we have gained from studies on noise and drug-related damage to the inner ear, while we know very little about the pathophysiology underlying viral etiologies of hearing loss. The closed, delicately balanced system of the inner ear precludes invasive biopsy in patients, due to an unacceptably high risk of hearing loss. Specialized imaging techniques may eventually provide cellular-level detail in a noninvasive manner, but current techniques remain far short of this goal [152]. Most current approaches to the treatment of hearing loss are focused on addressing one cell type or connection between cells. However, once sensorineural hearing loss is manifested in deficits both in threshold detection and word recognition, multiple cell types have already been damaged or destroyed [128]. It is therefore likely that combinations of drugs will eventually be required to restore the full range of hearing for patients. With respect to an initial focus, however, it seems plausible that regeneration of the inner hair cell – ribbon synapse – neuron axis might provide more immediate benefit for patients than regeneration of outer hair cells. Since the connection between inner hair cells and neurons appears to be more critical for word recognition, such improvements would be expected to provide significant hearing benefits, even if amplification is required secondary to outer hair cell loss. On the other hand, outer hair cell regeneration in the setting of persistent inner hair cell, ribbon synapse, and neuronal loss would be predicted to provide greater ability to detect sound with minimal improvement in clarity or speech recognition, which we assess may provide less clinical benefit.

The relative inaccessibility of inner ear structures also holds significant implications for delivery of drugs into the inner ear. Local delivery holds intrinsic appeal, as side effects can be minimized and the amount of drug delivered to the inner ear maximized. Indeed, intratympanic delivery of steroids for sudden hearing loss and gentamicin for Meniere’s disease is commonly performed in the office by otolaryngologists. However, two major potential barriers remain in this regard. The first is retention of drug within the middle ear space for prolonged delivery to the inner ear. Thermosensitive gel compounds that polymerize from a liquid to gel upon increase in temperature when contacting the middle ear have been tested in human clinical trials. Although there is a concern for transient conductive hearing loss from the gel, at least in animal studies [153], prior clinical trials have demonstrated the safety of such hydrogels [154,155]. Multiple different hydrogels are currently being studied for intratympanic injection [156], but ultimately, the compound that is being applied still needs to have the ability to cross the RWM. Nonetheless, it is likely that prolonged exposure to the RWM will increase the relative availability of drug in the inner ear, and with it the chance of treatment success. The need for persistent treatment beyond the life of the hydrogel, however, may still demand repeated treatments.

A second potential barrier to the efficacy of local delivery is the ability of a compound to transit across the RWM into the labyrinth. In this regard, properties that may improve the ability of drugs to cross the RWM have been described [75]. Novel approaches, such as conjugation of drugs to cochlear bone-binding bisphosphonates, could allow both for prolonged delivery and transit across the RWM [141,142]. Direct intracochlear injections provide a tempting alternative, in that much higher local concentrations of a drug can be achieved. This approach does, however, significantly increase the risk of iatrogenic sensorineural hearing loss. Several intracochlear drug delivery pumps have been developed in the past, designed to introduce drug through the scala tympani, the oval window, or the posterior semicircular canal [157]. These pumps can overcome the need for repeated intracochlear injections, but may bear a similar risk of sensorineural hearing loss and spread of infection into the labyrinth and the brain. Microneedle fenestration of the RWM may hold promise for safer intracochlear delivery [158].

The timing and the frequency of delivery for either hair cell or synapse regeneration remains a topic of speculation. We can only guess the duration required for regeneration of either element. We therefore have yet to better understand the risk-benefit of repeated injections for higher local inner ear drug concentrations, or longer treatment window, versus the increased chance of persistent damage to the tympanic membrane and middle ear from repeated procedures. In this regard, the previously conducted clinical trials had initially chosen a single delivery injection, but the FX-322 follow-up trials included several injections, likely based on prior animal data.

Ultimately, the degree and type of cellular damage in the cochlea, the responsiveness of remaining cells to regenerative treatments, and the duration of drug availability within cochlear fluids will determine the success of hearing protection or restoration.

Article highlights

• Introduction into anatomy and physiology of the ear

• Discussion of novel, minimally invasive approaches geared toward protecting or restoring natural, biological hearing

• Overview over clinical trials for otoprotection and ongoing drug development and pre-clinical studies

• Summary of current knowledge on hair cell and synapse regeneration in the inner ear.

Declaration of interest

DH Jung acts as a consultant for Akouos and is an expert witness in the ongoing 3 M/Aero earplug litigation (both unrelated to the topics in this review). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This paper was not funded.