Hidden Hearing Loss: A Disorder with Multiple Etiologies and Mechanisms

Noise Exposure

Numerous studies in mice (Kujawa and Liberman 2009; Shi et al. 2015), rats (Lobarinas et al. 2017), and guinea pigs (Lin et al. 2011; Liu et al. 2012; Shi et al. 2013, 2016; Song et al. 2016) showed that moderate noise exposure induces HHL. These studies found that moderate noise exposures (e.g., 100 dB sound pressure level [SPL] for 2 hours in mice) produce acute but temporary shifts in auditory thresholds (ABR, CAP, and DPOAE, see Box 1) that recover within days or weeks (for review, see Hickox et al. 2017). These TTSs occur without hair cell loss. However, even after threshold recovery, cochlear responses to suprathreshold sound levels are significantly altered; that is, the amplitude of the first peak of the ABR waveform (ABR peak I) is reduced, consistent with a decreased number of AN fibers activated by the sound, and/or a decrease in their firing rate or synchrony. Although most early studies evaluated the effects of continuous noise, recent reports indicate that a single blast exposure may also result in HHL both in animals (Niwa et al. 2016; Hickman et al. 2018) and in humans (Bressler et al. 2017). The reduced neural responses associated with noise-induced HHL are expected to alter the coding of temporal and intensity features of suprathreshold sounds and to reduce the ability to perceive sounds in complex listening environments, such as those with background noise. The latter has been tested in rats with HHL following exposure to 109-dB SPL octave band noise, which resulted in TTS and permanent reductions in ABR peak I amplitudes (Lobarinas et al. 2017). Although thresholds recovered within 2 weeks, behavioral testing of exposed rats showed a poorer performance in a test of hearing in background noise.

In humans, the neural coding problems associated with HHL are expected to produce deficits in speech discrimination and intelligibility, especially in noisy environments (Kujawa and Liberman 2015; Wan and Corfas 2015). A number of studies are consistent with this prediction, suggesting that individuals who have experienced exposures to loud noise have greater difficulties in complex listening tasks despite near normal audiological thresholds. For example, subjects with higher reported noise exposures than controls with similar thresholds showed significant deficits in word recognition (Alvord 1983) and in accurate speech and sound detection in noisy background environments (Kujala et al. 2004; Kumar et al. 2012; Liberman et al. 2016). Furthermore, young adults with a history of exposure to loud recreational noise have problems in discrimination of bursts of narrowband sounds presented at low levels relative to those with less exposure (Stone et al. 2008). Effects of noise exposure have been reported for ABR and electrocochleogram peak I amplitudes in response to suprathreshold stimuli (Stamper and Johnson 2015; Liberman et al. 2016; Valderrama et al. 2018). Finally, a variety of physiological and perceptual measures were used by Bharadwaj et al. (2015) to investigate a group of young adult subjects (21–39 years old) with normal hearing sensitivity to determine potential relationships between coding of suprathreshold responses and perception of complex auditory stimuli. Although this study identified only “marginally significant” correlations between reported noise history and several measures of temporal coding, better temporal coding metrics were strongly associated with better perception of “competing” speech and to sounds that differ in their binaural timing characteristics. Together, these studies highlight the importance of accurate neural coding of suprathreshold sounds, independent of auditory sensitivity, for performance in complex listening environments, and suggest that accumulated noise exposures can degrade these abilities.

However, several recent studies have failed to find significant correlations among prior noise exposure, electrophysiological measures associated with HHL and perceptual hearing ability (Kluk et al. 2016; Fulbright et al. 2017; Grinn et al. 2017; Prendergast et al. 2017a,b; Guest et al. 2018). Additionally, studies in which small negative correlations were shown between reported noise exposure and ABR peak I amplitudes, often failed to identify a clear relationship with expected perceptual dysfunction (Grose et al. 2017; Valderrama et al. 2018). Many factors may underlie the discordant studies relating noise and HHL, including inaccuracies in self-reporting of lifetime noise exposure, potential confounding effects of age and central auditory system compensation mechanisms, and the effects of different underlying mechanisms on physiological and behavioral measures of hearing.

Aging

The ability to understand speech in noisy environments decreases with age even in subjects with normal auditory thresholds in response to sounds up to at least 8 kHz (Dubno et al. 1984; Frisina and Frisina 1997; Pichora-Fuller and Souza 2003; Rajan and Cainer 2008). Likewise, aging subjects show declines in neural coding of the temporal features of sound that are likely important for speech perception in noise, and such deficits can also occur independently from increases in thresholds (Clinard and Tremblay 2013; Marmel et al. 2013; King et al. 2014). Decreases in the amplitudes of ABR peak I in aging human subjects are consistent with a cochlear neuropathy and could underlie these temporal processing defects (Konrad-Martin et al. 2012). Similar declines (>50%) across the life span in ABR peak I amplitudes have also been shown in mice (Sergeyenko et al. 2013; Muniak et al. 2018). Although these mice show a gradual increase in auditory thresholds over the course of their life span, this appears to be driven largely by OHC loss at later ages. The decreases in ABR peak I amplitudes in response to suprathreshold stimuli also progress over time but follow an earlier trajectory that parallels the pattern of IHC synapse loss (Sergeyenko et al. 2013), consistent with synaptopathy being a key contributor to age-related HHL in mice. Furthermore, Liberman and colleagues reported that a single noise exposure that causes TTS and HHL in young adult mice (4 months of age) significantly accelerates the rate of age-related overt hearing loss, including ABR threshold elevations and OHC loss. This data indicates that a single exposure to moderate noise levels (TTS type) early in life predisposes to an accelerated, progressive hearing loss across the life span (Kujawa and Liberman 2009; Fernandez et al. 2015).

Peripheral Neuropathy

A number of peripheral neuropathies can directly affect AN function and SGN survival (Rance and Starr 2015). The hearing loss associated with these disorders is often termed “auditory neuropathy,” as defined by intact OHC function together with altered ABRs. Although peripheral neuropathies often result in relatively large effects on AN function and thereby alter ABR thresholds, some patients show normal hearing sensitivity yet have significant perceptual difficulties, for example, in Guillain–Barré syndrome (GBS) and Charcot–Marie–Tooth (CMT) disease. GBS is caused by transient damage to peripheral myelinating Schwann cells (Kuwabara and Yuki 2013). Intriguingly, in GBS patients who suffered acute hearing loss, although hearing threshold gradually recovered, the latency of their ABR waveforms was persistently increased, indicative of HHL (Takazawa et al. 2012). CMT is an inherited, progressive peripheral neuropathy that affects both motor and sensory nerves and is genetically heterogenous (Rossor et al. 2013). Clinical evaluations of motor nerve conduction velocities have been used to classify CMT as either demyelinating (type 1) or axonal (type 2). Many individuals with CMT show classic auditory neuropathy characterized by normal OHC function but significant decreases in hearing sensitivity, often with ABR latency alterations (Rance 2005). Nonetheless, evaluation of a cohort of children genetically diagnosed with CMT1 or CMT2 showed that a majority of cases exhibited decreased speech understanding and altered temporal processing despite normal or near normal auditory thresholds (Rance et al. 2012). Similarly, a recent study of individuals diagnosed with CMT1A, the most common form of CMT that is associated with copy number variation of the gene encoding Peripheral Myelin Protein 22 (PMP22), provided additional evidence of HHL in this patient population (Choi et al. 2018). A cohort of 43 CMT1A patients ranging in age from 14 to 64 years showed normal pure tone threshold averages at a range of frequencies (250 Hz to 8 kHz), and their speech perception scores in quiet matched those of a group of age- and sex-matched controls. However, the mean score for speech perception in noise was reduced in the CMT1A patients, and their abilities to detect stimuli that require temporal and spectral acuity were also decreased. Early evaluation of auditory function in mouse models of CMT1 indicated a combination of elements of both HHL and auditory sensitivity defects associated with decreased AN myelination, including loss of AN proximal fibers and SGNs along with small increases in ABR thresholds and large decreases in peak I amplitudes in response to suprathreshold sounds (Zhou et al. 1995). Therefore, it is likely that HHL is an integral part of CMT pathology, whereas the degree of AN demyelination and SGN death dictates whether threshold shifts will be detected in audiometric tests.

Ototoxicity

High doses of aminoglycoside antibiotics such as gentamicin have long been known to cause auditory threshold shifts because of hair cell toxicity (Schacht et al. 2012). Interestingly, this type of exposure also has been shown to induce acute swelling of SGN terminal dendrites (Duan et al. 2000), similar to the morphology associated with excitotoxic damage of IHC synapses by noise exposure (Ruel et al. 2007). Most relevant to the focus of this review, doses of aminoglycosides that spare hair cells also have been linked to loss of IHC synapses, although the precise relationship to HHL and temporal processing in this case has not yet been determined. Liu et al. (2015) reported that systemic delivery of low dose gentamycin to mice for 2 weeks results in a temporary decrease in IHC synapse numbers, which partially recovered following cessation of the drug and suggested that synapses can be restored under these conditions. Oishi et al. (2015) found that a single middle ear administration of low amounts of gentamycin into mice results in small threshold increases associated with both synapse loss and SGN loss while sparing hair cells. This effect was observed although only in XBP-1 mutant mice with compromised stress responses but not in wild-type mice. Additional studies in mice have also implicated systemic gentamycin in acute losses of SGN dendrites (Ruan et al. 2014) or IHC synapses (Hong et al. 2018) at doses that did not cause substantial hair cell loss. In the latter study, synapse loss was accompanied by both small threshold shifts and large declines in ABR peak I amplitudes (Hong et al. 2018).

Cochlear Synaptopathy

The best documented mechanism for HHL is the degeneration of cochlear ribbon synapses without hair cell loss and SGNs (i.e., cochlear synaptopathy) (Fig. 1). These synapses, located at the basal end of IHCs, consist of a presynaptic specialization known as a “ribbon” that contains neurotransmitter vesicles and the release apparatus, as well as a postsynaptic zone in the nerve terminal with AMPA glutamate receptors (Reijntjes and Pyott 2016). The association between HHL and the loss of IHC-SGN synapses was first noted in studies of the effects of moderate noise in mice by Kujawa and Liberman (2009). These TTS exposures resulted in no loss of hair cells and full recovery of auditory thresholds but a permanent decline in AN responses to suprathreshold sounds as measured by the amplitudes of ABR peak I, or of CAPs measured at the cochlear round window. The suprathreshold changes were associated with permanent decreases of up to 50% of the number of presynaptic ribbon synapses as determined by immunostaining for RIBEYE/CtBP2, a component of the presynaptic ribbon. The decline in synapse numbers in this study was proportional to the relative decreases in suprathreshold responses and was found in basal, high-frequency regions of the cochlea that showed functional deficits. It has been speculated that the initiator of synapse damage and deafferentation may be glutamate-mediated excitotoxicity, which induces swelling of the SGN terminal fibers under IHC soon after damaging noise exposures (Ruel et al. 2007; Liberman and Kujawa 2017).

Similar characteristics of synaptopathy have also been shown in several mouse strains, and in rats, guinea pigs, chinchillas, and rhesus monkeys following TTS-type noise (Hickox et al. 2017; Valero et al. 2017). Not surprisingly, exposure to higher intensity noise that induces permanent threshold shifts (PTSs) associated with extensive OHC death and moderate IHC losses results in an even greater decrease in synapse numbers on surviving IHC relative to TTS noise exposure (Valero et al. 2017). Although evidence of TTS noise effects in mouse supports the permanence of the synaptic loss (Fernandez et al. 2015), most studies in other species have not examined longer-term effects of noise on synapse integrity. Several recent studies of guinea pigs exposed to less intense noise levels and evaluated at later time points suggest, however, that some synaptic repair may be possible following the initial damage (Liu et al. 2012; Shi et al. 2013, 2015, 2016; Song et al. 2016). As most studies used immunostainings of synaptic markers as indicators of synaptic integrity, it remains to be determined whether synaptic loss and regeneration involves structural alterations and/or dynamic changes in marker expression.

Most of the synapse loss following TTS noise exposure occurs within hours of the noise and appears to preferentially occur in synapses associated with AN fibers that have low spontaneous rate firing activity (low SR) and high threshold response properties (Furman et al. 2013; Liberman and Liberman 2015). These low SR fibers also show larger dynamic ranges (Schalk and Sachs 1980), preserve information concerning stimulus timing and amplitude modulation (AM) more efficiently (Joris et al. 1994; Rhode and Greenberg 1994; Frisina et al. 1996), and are less vulnerable to masking in background noise than high SR fibers (Costalupes et al. 1984). Preferential loss of low SR fibers would therefore be consistent with the decreased responses to suprathreshold sounds and auditory acuity in complex listening environments that have been associated with the synaptic damage from TTS noise exposures. The synapses associated with high SR fibers that remain following TTS noise retain normal threshold and frequency tuning characteristics and presumably explain the recovery of normal auditory thresholds (Furman et al. 2013).

A strong correlation between synapse loss and reduced suprathreshold responses was found in mice during aging, supporting the notion that synaptopathy is a key contributor to progressive auditory neuropathy and HHL in the absence of noise overexposure (Sergeyenko et al. 2013). In addition, SGN loss in aging mice occurred at approximately the same rate as the synaptopathy but with several months delay, suggesting that synapse loss predisposes neurons to degeneration (Sergeyenko et al. 2013). Indeed, mice with noise-induced synaptopathy in the cochlear base at early ages show higher degrees of SGN loss later in life than nonexposed mice, and this occurs in the same cochlear regions affected by the exposure (Fernandez et al. 2015). Electrophysiological studies in aging gerbils indicate that, as in TTS noise-exposed mice, low SR AN fibers are preferentially lost, and supports the idea that loss of such fibers could give rise to the suprathreshold response and temporal coding deficits associated with HHL during aging (Schmiedt et al. 1996).

SGN counts of a large set of human cochleae from a range of subject ages with relatively intact hair cell numbers and no evidence of otologic disease indicated a similar rate of SGN loss across the life span (Makary et al. 2011). In addition, immunostaining for IHC ribbon synapses and peripheral AN fibers in separate sets of human cochleae with relatively intact IHC populations suggested a progressive deafferentation that proceeds faster than the rate of SGN loss in the Makary et al. study (Viana et al. 2015; Wu et al. 2018). These findings are similar to those obtained in aging mice, suggesting that synaptopathy in humans may also predispose SGN to degeneration.

Demyelination

Proper development and maintenance of AN myelination by Schwann cells is critical for auditory processing fidelity and defects in myelination have been implicated in hearing loss in both humans and animal model systems (Long et al. 2018). Schwann cells are critical for the formation of nodes of Ranvier, the specialized regions along myelinated fibers at which voltage-gated sodium and potassium channels are clustered for regeneration of action potentials and the fast, synchronized transmission of electrical signals (Rasband and Peles 2015). In addition to demyelinating disorders such as CMT1 that clearly affect temporal processing in AN, other risk factors for HHL such as noise and aging have also been associated with myelin dysfunction. Exposure of rats to loud noise has been associated with thinning of myelin surrounding AN and also changes to paranodal myelin regions that are important for function of the nodes of Ranvier (Tagoe et al. 2014). Although this noise level produced increases in threshold shifts, reductions in ABR peak I amplitudes and increases in latency were also observed. Similar myelin defects and changes to suprathreshold stimuli have also been observed in mice exposed to noises that produce a combined TTS-PTS (Panganiban et al. 2018). Other studies based on electron microscopy reported thinning and degeneration of myelin sheaths in the AN of aging mice, which correlated with decreases in myelin basic protein (MBP) levels and declines in ABR peak I amplitudes (Xing et al. 2012). Age-associated decreases in MBP levels were also noted in AN from a set of human temporal bone specimens (Xing et al. 2012), suggesting myelin loss could contribute to the temporal processing abnormalities described in aging humans, including increases in response latency and dyssynchrony (Plack et al. 2014; Harris and Dubno 2017).

Recently, Wan and Corfas (2017) reported that transient demyelination also results in HHL (Fig. 2). This study showed that ablation of Schwann cells via genetic means in adult animals causes a near total loss of AN myelin within 1 week. Remarkably, this loss does not alter auditory thresholds, yet induces a permanent HHL, that is, suprathreshold ABR peak I amplitudes are significantly decreased. In contrast to the HHL produced by noise exposure, ABR peak I latency was also increased by demyelination. These alterations persist even after complete AN remyelination through proliferation and differentiation of new Schwann cell precursors. Although no changes in IHC-SGN synapses were observed following the acute demyelination, immunostaining studies showed a permanent defect in the nodal structures closest to the IHCs, known as heminodes. Electrophysiological and immunostaining studies showed that the heminode is the action potential generator in AN (Hossain et al. 2005; Rutherford et al. 2012; Kim and Rutherford 2016). Together, these observations indicate that heminode disruption is likely to be the basis for the HHL observed following transient demyelination (Wan and Corfas 2017). The temporary hearing sensitivity deficits, together with persistently increased ABR latencies observed in some humans with transient demyelination caused by GBS, appear to match the phenotype of this mouse model (Takazawa et al. 2012).

Hair Cell Dysfunction

Two recent studies suggested that noise-induced hair cell dysfunction could also contribute to HHL. Mulders et al. (2018) reported that guinea pigs with noise-induced HHL also have persistent reductions in summating potential (SP) amplitudes. As the SP reflects the sound-induced IHC receptor current (Zheng et al. 1997), the investigators suggested that long-lasting changes in IHC electrophysiological function might contribute to HHL. However, this finding contradicts other studies in which the SP was reported to be stable after noise exposure in HHL mouse models (Kujawa and Liberman 2009) and during aging (Sergeyenko et al. 2013). The second study involved human listeners and suggested OHC involvement in HHL (Hoben et al. 2017). This group reported correlations among very mild threshold elevations, altered DPOAEs, and difficulties in speech in noise. Importantly, they found no correlation of those variables with CAP amplitudes. Based on these observations, the investigators hypothesized that loss of OHCs results in sound-induced deflection of a broader region of the basilar membrane, thereby recruiting a larger population of AN fibers and compensating for decreased cochlear amplifier function at low sound levels (i.e., normal CAP thresholds), but that this process leads to altered temporal acuity and compromises speech understanding in noisy backgrounds (Hoben et al. 2017). Further exploration of the contribution of hair cell dysfunction to HHL is necessary.

ABR Measurements

The most frequently applied tests for HHL are based on ABR recordings and electrocochleograms, with analysis of specific features of the waveforms. It has been well documented in animal models that HHL can be diagnosed by a characteristic reduction in ABR peak I amplitudes in the absence of ABR threshold or latency changes (for review, see Hickox et al. 2017). Importantly, the degree of peak I amplitude reduction correlates well with the degree of cochlear synaptopathy (Kujawa and Liberman 2009; Sergeyenko et al. 2013; Wan et al. 2014). Interestingly, mouse studies indicate that demyelination-related HHL has the additional phenotype of increased ABR peak I latency (Wan and Corfas 2017). Similarly, in humans, both aging and noise exposure have been associated with reduction of ABR peak I suprathreshold amplitudes in the absence of threshold changes (Konrad-Martin et al. 2012; Stamper and Johnson 2015). However, unlike in animal models, ABR peak I amplitudes recorded from human subjects are much smaller and more variable, limiting their clinical application for routine HHL diagnosis in individual patients (Beattie 1988; Lauter and Loomis 1988; Trune et al. 1988).

Several studies have used alternative measurements of ABRs or electrocochleography in an attempt to more accurately evaluate AN activity, particularly in humans, to relate changes in these measurements with performance on auditory tasks that rely on precise neural coding of sound. For example, the ratio between ABR peak I (AP) and SP amplitudes (AP/SP ratio) correlates with performance in a test for word comprehension in noise and self-reporting of noise exposure (Liberman et al. 2016). Similarly, latencies of peak V, which is generated by processing in the auditory midbrain, show lower degrees of change in background noise in subjects with poorer abilities in binaural discrimination tasks (Mehraei et al. 2016), consistent with modeling predictions for loss of a subset of AN fibers (Verhulst et al. 2013). Comparable changes were also identified in mice exposed to TTS-level noise that induced ABR peak I amplitude changes and loss of IHC synapses, suggesting that ABR peak V latencies in a masking noise may be useful in assessing HHL in humans (Mehraei et al. 2016). Finally, a number of modeling studies have been used to simulate various cochlear hearing defects, including HHL, to better predict how various aspects of auditory waveforms will behave in response to different stimulus conditions, including pure tone thresholds in noisy backgrounds (Ridley et al. 2018) and by suprathreshold sounds (Verhulst et al. 2016). Further model validation in animals with experimentally induced HHL may lead to more precise test approaches in human, including the ability to isolate synaptopathic effects from more overt hearing loss.

Envelope Following Responses

Envelope following responses (EFRs), also called frequency following responses (FFRs), are far-field responses to AM tones that reflect neural activity at multiple locations along the auditory pathway, with relative responsiveness dependent on the AM frequency (Kuwada et al. 2002; Krishnan 2006). The AM tones are sinusoidal and continuous rather than transient as those used in ABR evaluations, and deficits in EFRs have been correlated with reduced signal detection in noise (Dimitrijevic et al. 2004) and in other listening tasks that require detection of timing cues (Ruggles et al. 2011; Bharadwaj et al. 2015). Recent animal studies indicated that declines in EFR responses to modulation frequencies near 1 kHz correlate well with synaptic loss and HHL induced by TTS noise and by aging (Shaheen et al. 2015; Parthasarathy and Kujawa 2018). Similarly, in a recent study of humans with normal hearing thresholds, control subjects had better EFRs to a 5 kHz, 85 Hz AM tone in background noise than subjects with suspected HHL, whose EFRs were reduced by the noise (Paul et al. 2017). Furthermore, subjects with prior recreational noise exposure appear to have reduced EFR activity compared with control subjects without noise exposure but similar hearing thresholds (Plack et al. 2014). These comparative human and animal studies support the use of EFR as an assay for cochlear synaptopathy and HHL.

Middle Ear Muscle Reflex

Low SR ANs, which are preferentially affected during cochlear synaptopathy, are also required for the middle ear muscle reflex (MEMR) (Liberman and Kiang 1984; Kobler et al. 1992). The MEMR can reduce the sound-evoked excitation of IHCs by increasing the impedance of the middle ear, thus acting as a signal attenuator to protect the cochlea from damaging sounds. MEMR is measured by monitoring the changes in sound pressure in the ear canal ipsilateral to a probe tone while eliciting the MEMR with a sound in either the ipsilateral or contralateral ear. In mouse models of HHL, elevation of MEMR threshold and reduction in suprathreshold MEMR reflex strength are correlated well with synaptopathy (Valero et al. 2016, 2018). In human subjects with tinnitus and normal audiological thresholds, the presence of HHL was suspected based on their significantly weaker MEMR responses relative to control subjects without tinnitus (Wojtczak et al. 2017). Emerging evidence suggests that MEMR may be more sensitive than the suprathreshold amplitude of ABR peak I as an indicator of moderate synaptopathy (Valero et al. 2018).

Neurotrophins as Potential Therapies for Noise-Induced HHL

There is increasing evidence that the neurotrophins, in particular neurotrophin 3 (NT-3), might have the ability to induce IHC synapse regeneration after HHL-inducing noise exposure (Wan et al. 2014; Cunningham and Tucci 2015). We have shown that the levels of NT-3 expression by cochlear supporting cells regulate the density of IHC synapses during cochlear maturation (Wan et al. 2014). Specifically, NT-3 overexpression in supporting cells results in higher synapse density in the cochlear base and higher auditory sensitivity in the corresponding high frequencies. Correspondingly, reductions in NT-3 expression by supporting cells at early postnatal ages results in decreased IHC synapse density in the cochlear base and a mild high-frequency hearing loss. Most relevant to HHL therapeutics, in this study we showed that NT-3 overexpression in supporting cells does not prevent the synaptopathy and TTS elicited by an HHL-inducing exposure, but IHC synapses and ABR peak I amplitudes recover within 2 weeks after the noise, consistent with NT-3 inducing synaptic regeneration. More recent studies have shown that round window delivery of either NT-3 (Suzuki et al. 2016) or NT-3 plus BDNF (Sly et al. 2016) within 24 hours following noise exposure have the same effects as transgenic NT-3 overexpression. The observation that the neurotrophin treatment worked both in mice (Suzuki et al. 2016) and guinea pigs (Sly et al. 2016) strongly supports the potential efficacy of NT-3 as a therapeutic for HHL (Fig. 4).

Modulation of Efferent Feedback

Olivocochlear efferent fibers originating in the auditory brainstem innervate the cochlea and provide feedback control of cochlear activity by the central nervous system (CNS) (Fuchs and Lauer 2018). Lateral olivocochlear (LOC) efferent fibers synapse with the SGN terminals near IHCs, whereas medial olivocochlear (MOC) efferent fibers target OHCs (Frank and Goodrich 2018). Acetylcholine released by MOC neurons bind and activate OHC nicotinic acetylcholine receptors (nAChRs) comprised of the α9/α10 subunits. Activation of these receptors decrease cochlear amplification. Signaling through LOC alters AN response properties and works in part through dopaminergic signaling (Guinan 2018). A number of studies have implicated both MOC and LOC efferents in cochlear protection from noise damage (for review, see Le Prell et al. 2003; Fuente 2015). Although the majority of these studies evaluated protection from intense noise exposures that result in PTS, several recent studies have evaluated the involvement of efferent responses during aging and in response to moderate noise exposures relevant to HHL.

The Liberman laboratory used a surgical ablation approach in mice to show the importance of MOC efferents for hearing protection from an extended exposure to moderate noise (84 dB for 1 week) (Maison et al. 2013). Specifically, they determined that mice with the selective MOC lesion show larger ABR peak I amplitude declines along with greater decreases in synapses than those with intact MOC efferents. A separate study showed that a lesion that removes ∼50% of both MOC and LOC fibers results in accelerated age-related HHL and IHC synapse loss in the absence of overt noise exposure (Liberman et al. 2014). Evaluation of innervation in the surgically treated mice suggested that MOC efferents are most important for synaptic protection in the apical half (lower frequencies) of the cochlea, whereas LOC efferent affects predominately the basal half (higher frequencies). More recently, the Gomez-Casati laboratory explored the role of efferents in noise protection by modulating MOC responses in OHCs using mice carrying either a knockout (KO) allele of the OHC α9 nAChR or a knockin (KI) allele of the receptor that increases its activity (Boero et al. 2018). They reported that, as expected, wild-type mice show HHL and synaptopathy 1 week after exposure to a TTS-type noise (100 dB for 1 hour). In contrast, KO mice with loss of OHC cholinergic function develop PTS under these conditions, whereas KI mice with enhanced nAChR function do not show signs of HHL. Together, these studies suggest that augmentation of efferent signaling may be a potential strategy for minimizing both HHL resulting from moderate noise exposures, as well as PTS in response to more damaging noise (Maison et al. 2002; Taranda et al. 2009). However, the possibility that the transient MOC connections with IHC that occur during development (Glowatzki and Fuchs 2000; Katz et al. 2004) contribute to these observations cannot be ruled out. Because LOC effects on IHC-SGN synaptic connections may operate via dopaminergic-mediated mechanisms (Ruel et al. 2001, 2007), it has been speculated that dopaminergic agonists might impact HHL. Along similar lines, NMDA glutamate receptors are expressed by AN fibers and NMDA receptor antagonists have been shown to block AN excitotoxicity (Pujol et al. 1993). Accordingly, a recent report indicated that delivery of the NMDA receptor antagonist esketamine (AM-101) following noise exposure reduces HHL, although postsynaptic integrity was not evaluated in this study (Bing et al. 2015).