Age-related changes in the biophysical and morphological characteristics of mouse cochlear outer hair cells

Outer hair cells (OHCs) are electromotile sensory receptors that provide sound ampliﬁcation within the mammalian cochlea. Although OHCs appear susceptible to ageing, the progression of the pathophysiological changes in these cells is still poorly understood. By using mouse strains with a different progression of hearing loss (C57BL/6J, C57BL/6NTac, C57BL/6NTac Cdh23 + , C3H/HeJ), we have identiﬁed morphological, physiological and molecular changes in ageing OHCs (9–12 kHz cochlear region). We show that by 6 months of age, OHCs from all strains underwent a reduction in surface area, which was not a sign of degeneration. Although the ageing OHCs retained a normal basolateral membrane protein proﬁle, they showed a reduction in the size of the K + current and non-linear capacitance, a readout of prestin-dependent electromotility. Despite these changes, OHCs have a normal V m and retain the ability to amplify sound, as distortion product otoacoustic emission thresholds were not affected in aged, good-hearing mice (C3H/HeJ, C57BL/6NTac Cdh23 + ). The loss of afferent synapses was present in all strains at 15 months. The number of efferent synapses per OHCs, deﬁned as postsynaptic SK2 puncta, was reduced in aged OHCs of all strains apart from C3H mice. Several of the identiﬁed changes occurred in aged OHCs from all mouse strains, thus representing a general trait in the pathophysiological progression of age-related hearing loss, possibly aimed at preserving functionality. We have also shown that the mechanoelectrical transduction (MET) current from OHCs of mice harbouring the Cdh23 ahl allele is reduced with age, highlighting the possibility that changes in the MET apparatus could play a role in cochlear ageing. interpretation of


Introduction
Age-related hearing loss (ARHL), also known as presbycusis, is the most common form of age-related sensory disability in humans, affecting ß60% of people by 70 years of age (Bowl & Dawson, 2019). ARHL is a complex disorder with both genetic predisposition and environmental factors (e.g. noise exposure) influencing its progression (Bowl & Dawson, 2019;Liberman, 2017), leading to a decrease in sensitivity threshold and speech discrimination (Gates & Mills, 2005;Gordon-Salant, 2005). Among the most common causes of ARHL are the loss of the sensory inner and outer hair cells (IHCs, OHCs) and their innervation within the cochlea (Liberman, 2017), with OHCs being particularly vulnerable to cochlear ageing (e.g. Kusunoki et al. 2004;Sergeyenko et al. 2013). Currently, very little is known about the functional changes that occur in ageing OHCs, or if any particular physiological or morphological characteristics are more vulnerable to change during the ageing process.
In mammals, normal hearing depends on the ability of the cochlea to amplify acoustic stimuli (Ashmore, 2019). This amplification is provided by the OHCs, which generate voltage-dependent axial forces driving electromotility (Brownell et al. 1985;Ashmore, 1987). Electromotility depends on prestin (encoded by Slc26a5: Zheng et al. 2000;Liberman et al. 2002), a protein expressed in the lateral membrane of mature OHCs (Belyantseva et al. 2000;Mahendrasingam et al. 2010) that changes conformation according to the membrane potential (Oliver et al. 2001). The voltage corresponding to the steepest voltage sensitivity of prestin has been estimated at around -40/-50 mV (Huang & Santos-Sacchi, 1993;Oliver & Fakler, 1999;Bai et al. 2019; see also : Ashmore, 2008). This relatively depolarized membrane potential of OHCs in vivo is likely to be set by the combined activity of the basolateral hyperpolarizing K + current I K,n (Mammano & Ashmore, 1996;Marcotti & Kros, 1999) and the depolarizing resting mechanoelectrical transducer current (Johnson et al. 2011). Another key feature of mature OHCs is that their activity is inhibited by the medial olivocochlear (MOC) efferent system (Liberman, 1980;Simmons et al. 1996;Maison et al. 2003). The release of acetylcholine (ACh) by the efferent neurons leads to OHC hyperpolarization (Oliver et al. 2000), thus reducing the mechanical amplification of the cochlear partition (Guinan, 1996;Fuchs & Lauer, 2019). Although mature IHCs are the primary target of the spiral ganglion afferent neurons (ß90-95% type I), OHCs receive the remaining 5-10% (type II: Ryugo, 1992), the role of which seems to be to transmit information about noxious sound stimulations (Fuchs, 2018).
The aim of this study was to identify the initial morphological and physiological changes that occur in ageing OHCs (in the 9-12 kHz cochlear region), using four different mouse strains (C57BL/6J, C57BL/6NTac, C3H/HeJ, C57BL/6NTac Cdh23+ ). We show that by 6 months of age all OHCs from the four mouse strains undergo a reduction in surface area, which was not a sign of apoptosis since even in 14-16-month-old mice OHC loss was minimal and did not affect distortion product otoacoustic emissions (DPOAEs) in the good-hearing mouse strain C3H/HeJ. Although OHC reduction in size was not linked to apoptosis, it seemed to be associated with a loss of basolateral membrane proteins, such as ion channels (KCNQ4) and the motor protein prestin, required for driving non-linear capacitance. Ageing OHCs also exhibited the progressive loss of afferent and efferent synapses. Efferent synapses onto OHCs from C3H/HeJ mice, which show the best hearing profile at the cochlear region used for the in vitro experiments (12 kHz), exhibited very little changes with age. Since most of the above changes occurred in OHCs from all mouse strains, we conclude that they represent a general trait in the normal pathophysiological progression in the ageing cochlea. Moreover, these changes could allow OHCs to preserve their functionality in the ageing cochlea. We have also shown that the mechanoelectrical transduction (MET) current of OHCs from mice harbouring the Cdh23 ahl allele is reduced with age. These data highlight the interesting possibility that alterations in the MET apparatus could contribute to the progression of ARHL as previously suggested (Johnson et al. 1997;Noben-Trauth et al. 2003).

Ethics statement
All animal work was performed at the University of Sheffield (UK) and licensed by the Home Office under the Animals (Scientific Procedures) Act 1986. Procedures were approved by the University of Sheffield Ethical Review Committee. Mice had free access to food and water and a 12 h light/dark cycle. Auditory functional measurements were conducted under anaesthesia using ketamine (100 mg kg -1 body weight, Fort Dodge Animal Health) and xylazine (10 mg kg -1 , Rompun 2%, Bayer), which were administered with intraperitoneal injection as previously described (Ingham et al. 2011). At the end of the procedure, mice were either killed in accordance with schedule 1 of the Act or recovered from anaesthesia with intraperitoneal injection of atipamezole (1 mg kg -1 ). Mice under recovery from anaesthesia were returned to their cage, placed on a thermal mat, and monitored for 2 to 4 h. Cages were returned to the holding racks once the mice were able to move well and respond to external stimuli. The wild-type C57BL/6NTac, C57BL/6J and C3H/HeJ, and the transgenic C57BL/6NTac Cdh23+ mice (Mianné et al. 2016) used for the experiments were bred at the MRC Harwell Institute.

Tissue preparation
In vitro recordings were performed from apical-coil OHCs (ß9-12 kHz: Muller et al. 2005) of acutely dissected organs of Corti from mice of 1-18 months of age. Males and females were killed by cervical dislocation and the inner ear placed in a Petri dish containing ice-cold extracellular solution composed of (in mM): 135 NaCl, 5.8 KCl, 1.3 CaCl 2 , 0.9 MgCl 2 , 0.7 NaH 2 PO 4 , 5.6 D-glucose, 10 Hepes-NaOH. Sodium pyruvate (2 mM), amino acids and vitamins were added from concentrates (Thermo Fisher Scientific, UK). The pH was adjusted to 7.5 (osmolality ß308 mmol kg -1 ). The dissection procedure begins with the removal of the muscles and connective tissue surrounding the inner ear. The 'cleaned' inner ear is then moved to a new Petri dish containing fresh solution (i.e. free from debris). By holding the whole inner ear from the vestibular system with a pair of forceps, the bone covering the part of the apical-coil cochlear turn (around 8-20 kHz: see Ceriani et al. 2019) is removed using a second pair of forceps (Dumont #3 or #4, which are further polished using a sharpening stone). This initial step is aimed at exposing the middle bony-spiral structure containing the cochlear nerve, which is then gently cut with very fine forceps, a procedure that allows the removal of the entire bone covering the apical cochlear coil. When the apical coil is exposed, it is very important to keep the solution clean (i.e. adding fresh ice-cold solution) and avoid touching the delicate organ of Corti, which is where the hair cells are located. This is then followed by the removal of the bone covering the site of the apical coil of the organ of Corti in one single step. The final step requires the use of very fine forceps to cut the auditory nerve and gently remove the apical portion of the organ of Corti, which is then placed in the recoding chamber using a small stainless steel spoon (about 4 mm in diameter, custom made). The tip size of the forceps used for this procedure changes depending on the thickness of the bone, which changes with age and between strains; usually we have 3-4 sets of forceps to cover the full range of thicknesses. The same procedure was applied to all dissections performed in this study irrespective of the age of the animal. In the recording chamber, the isolated organ of Corti was immobilized using a nylon mesh fixed to a stainless steel ring and viewed using an upright microscope (Olympus BX51, Japan; Leica, DMLFS, Germany; Nikon, Germany). Hair cells were observed with Nomarski Differential Interface Contrast (DIC) optics (×63 or ×60 water immersion objective) and ×15 eyepieces.

Single-cell electrophysiology
For whole-cell basolateral K + current recordings, experiments were performed as previously described (Corns et al. 2018;Ceriani et al. 2019). The pipette intracellular solution contained (in mM): 131 KCl, 3 MgCl 2 , 1 EGTA-KOH, 5 Na 2 ATP, 5 Hepes-KOH, 10 Na-phosphocreatine (pH was adjusted with 1 M KOH to 7.28; osmolality was 294 mmol kg −1 ). Membrane currents and voltage responses were investigated at room temperature (20-24°C) using the Optopatch amplifier (Cairn Research Ltd, UK). Data acquisition was controlled by pClamp software using Digidata 1440A boards (Molecular Devices, USA). Recordings were low-pass filtered at 2.5 kHz (8-pole Bessel), sampled at 5 kHz and stored on data-drives for off-line analysis (Origin: OriginLab, USA). Membrane potentials in whole-cell recordings were corrected for the residual series resistance R s after compensation (usually 70-90%) and the liquid junction potential of -4 mV measured between electrode and bath solutions.
The presence of electromotile activity in OHCs was estimated by measuring nonlinear (voltage-dependent) capacitance using whole-cell patch clamp recordings. In order to block most of the ion channels in hair cells, the pipette intracellular solution contained (in mM): 125 CsCl, J Physiol 598.18 3 MgCl 2 , 1 EGTA-CsOH, 5 Na 2 ATP, 5 Hepes-CsOH, 5 tetraethylammonium (TEA), 5 4-aminopyridine (4-AP) (pH was adjusted with CsOH to 7.28; osmolality was 280 mmol kg −1 ). Real-time changes in nonlinear membrane capacitance (C N-L ) were investigated using the capacitance track-in-mode of the Optopatch amplifier (Cairn Research Ltd, UK) during the application of a 4 kHz sine wave of 13 mV RMS. From the holding potential of −84 mV, hair cells were subjected to a voltage ramp from −154 mV to +96 mV over 2 s. The capacitance signal from the Optopatch amplifier was filtered at 250 Hz and sampled at 5 kHz.
MET currents in young adult and aged mice were recorded from OHCs of Tecta EGFP/EGFP ; Tectb −/− double knockout mice (Tecta/Tectb −/− ), in which the tectorial membrane is detached from the OHC hair bundles. MET currents were elicited using a fluid jet from a pipette driven by a 25 mm diameter piezoelectric disc (Corns et al. 2014;Corns et al. 2018). The fluid jet pipette tip had a diameter of 8−10 µm and was positioned at about 8 µm from the hair bundles to elicit a maximal MET current. Mechanical stimuli were applied as steps or 50 Hz sinusoids. The resting MET current becomes evident when the MET channels shut off in the inhibitory phase of the stimulus: the holding current minus the current present during inhibitory bundle deflection (Corns et al. 2014).

Auditory brainstem responses
Auditory brainstem responses (ABRs) were recorded from male and female mice between 1 and 18 months of age. Recordings were performed in a soundproof chamber (MAC-3 Acoustic Chamber, IAC Acoustic, UK) as previously described (Ingham et al. 2011). Briefly, stimuli were delivered to the ear by calibrated loudspeakers (MF1-S, Multi Field Speaker, Tucker-Davis Technologies, USA) placed 10 cm from the animal's pinna. Sound pressure was calibrated with a low-noise microphone probe system (ER10B+, Etymotic, USA). Experiments were performed using BioSigRZ software (Ingham et al. 2011) driving an RZ6 auditory processor (Tucker-Davis Technologies). Response thresholds were estimated from the resulting ABR waveforms and defined as the lowest sound level where any recognisable feature of the waveform was visible. Each set of ABR data was independently assessed by at least three members of the laboratory without prior knowledge of the mouse strain, age or sex. Responses were measured for click and stimulus pure tones of frequencies between 3 and 42 kHz. Stimulus were typically 0-95 dB sound pressure level (SPL) and presented in steps of 5 dB. The brainstem response signal was averaged over 256 repetitions. Tone bursts were 5 ms in duration with a 1 ms on/off ramp time, and were presented at a rate of 42.6/s.

Distortion product otoacoustic emissions
OHC function was assessed in vivo by measuring the distortion product otoacustic emissions (DPOAEs). Recordings were performed in a soundproof chamber (MAC-3 Acoustic Chamber, IAC Acoustic, UK). DPOAEs were recorded at 2f1-f2 in response to primary tones f1 and f2, where f2/f1 = 1.2. The f2 level (L2) was set from 20 to 80 dB in 10 dB increments, and the f1 level (L1) was set equal to L2. Frequency pairs of tones between f2 = 6.5 kHz and f2 = 39.2 kHz were presented directly into the ear canal by means of a metal coupler connected to two calibrated loudspeakers (MF1-S, Multi Field Speaker, Tucker-Davis Technologies, USA). The emission signals were recorded by a low-noise microphone (ER10B+: Etymotic Research Inc, USA) connected to the coupler. Experiments were performed using BioSigRZ software driving an RZ6 auditory processor (Tucker-Davis Technologies). The DPOAE thresholds were defined by the minimal sound level where the DPOAEs were above two standard deviation of the noise. The determined DPOAE thresholds were plotted against geometric mean frequency of f1 and f2. Stimulus sound pressure levels were typically 0-80 dB SPL, presented in steps of 10 dB. The response signal was averaged over 500 repetitions.

Immunofluorescence microscopy
The dissected inner ears from the mice were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS, pH 7.4) for 20 min at room temperature. Cochleae were microdissected, rinsed three times for 10 min in PBS, and incubated for 1 h at room temperature in PBS supplemented with 5% normal goat or horse serum and 0.5% Triton X-100. The samples were then incubated overnight at 37˚C with the primary antibody in PBS supplemented with 1% of serum. Primary antibodies were: mouse anti-myosin 7a (1:1000, Developmental Studies Hybridoma Bank, #138-1C), rabbit anti-myosin 7a (1:200, Proteus Biosciences, #25-6790), rabbit anti-prestin (1:5000, kindly provided by Robert Fettiplace), rabbit anti-SK2 (1:500, Sigma-Aldrich, P0483) goat anti-choline acetyltransferase (ChAT, 1:500, Millipore, AB144P), and mouse anti-CtBP2 (1:200, Biosciences, #612044). All primary antibodies were labelled with species-appropriate Alexa Fluor or NorthernLights secondary antibody for 1 h at 37˚C. Samples were then mounted in VECTASHIELD. The z-stack images were captured with a Nikon A1 confocal microscope equipped with Nikon CFI Plan Apo 60X Oil objective, which is part of the Light Microscope Facility at the University of Sheffield. Image stacks were processed with Fiji Image Analysis software. The number of synaptic ribbons (CtBP2) and SK2 puncta was estimated from the z-stack images of the immunolabelled proteins using Fiji ImageJ. For the SK2 puncta, the threshold for staining identification was manually set on a per sample basis at a level at which puncta of the expected size, shape and intercellular location were clearly distinct from the background. Within each mouse strain, 1-month and 15-month-old mice were fixed and processed for imaging simultaneously, while 6-month-old mice were done separately. Images were taken within a few days of fixing the samples. To identify the presence of colocalized SK2 and ChAT puncta we used the blob finder function within Arivis Vision4D 3.1 software on Z-projected images. Briefly, images were processed with a Gaussian filter, followed by an automatic seed finding and a watershed algorithm. A threshold at <1 µm was set to define the juxtaposition of SK2 and ChAT puncta and the results were confirmed by visual inspection. Each set of experiments was performed on at least three mice.
In order to investigate whether Tecta/Tectb double knockout mice (Tecta/Tectb −/− ) harbour the Cdh23 ahl allele (Johnson et al. 1997;Noben-Trauth et al. 2003), we performed DNA sequencing. Genomic DNA from the animals was extracted using NaOH. The targeted region was PCR amplified using GoTaq Master Mixes (M7123, Promega) and primers 5 -TCAGGCTCCCCTGC TTCTAT-3 and 5 -CCAAGCTGTTCCTGGTAGCT-3 . PCR products were purified and sequenced by Sanger sequencing using the DNA sequencing service from Core Genomic Facility, University of Sheffield.

Scanning electron microscopy (SEM)
For SEM, cochleae were excised from aged C57BL/6N mice (15 months). They were fixed by intralabyrinthine perfusion using a fine hypodermic needle through the round window with 2.5% vol/vol glutaraldehyde in 0.1 M sodium cacodylate buffer containing 2 mM calcium chloride (pH 7.4) and then immersed in this fixative for 2 h. They were stored in fixative diluted 1/10th in buffer and subsequently dissected by removing the bone from the apical coil to expose the organ of Corti and then immersed in 1% osmium tetroxide in the cacodylate buffer for 1 h. For osmium impregnation, which avoids gold coating, cochleae were incubated in solutions of saturated aqueous thiocarbohydrazide (20 min) alternating with 1% osmium tetroxide in buffer (2 h) twice (the OTOTO technique) (Furness & Hackney, 1986). They were dehydrated through an ethanol series and critical point dried using CO 2 as the transitional fluid, then mounted on specimen stubs using conductive silver paint (Agar Scientific, Stansted, UK) and examined in a Hitachi S4500 field emission SEM operated at 5 kV accelerating voltage.

Toluidine blue staining
After glutaraldehyde fixation, cochleae were washed three times in 0.1 M sodium cacodylate buffer pH 7.2 and post-fixed in 1% osmium tetroxide in 0.1 M sodium cacodylate buffer for 3 h at room temperature. Samples were then washed three times in sodium cacodylate buffer and decalcified in 0.5 M EDTA pH 8.0 containing 0.1% glutaraldehyde for 3 days at 4°C. Samples were then washed briefly in water, dehydrated through an ascending ethanol series, equilibrated in propylene oxide and embedded in epoxy (TAAB 812) resin. Blocks were cured at 60°C for 24 h and trimmed with a glass knife after which semi-thin 1 micron thick sections were cut on a Reichert Ultracut E ultramicrotome using a histo-grade Diatome diamond knife. Sections were dried onto glass slides and stained briefly with toluidine blue before viewing on a Zeiss Axioplan 2 wide-field microscope. Images were captured with a Spot RT digital camera.

Statistical analysis
Statistical comparisons of means were made by Student's two-tailed t test or for multiple comparisons, analysis of variance (one-way and two-way ANOVA followed by Tukey's or Sidak's post hoc test). P < 0.05 was selected as the criterion for statistical significance. Only mean J Physiol 598.18 values with a similar variance between groups were compared. Mean values are quoted in text and figures as means ± SD For DPOAE experiments, due to the presence of 'not found' values (i.e. above the upper threshold limit of our equipment), the non-parametric aligned ranks transformation two-way ANOVA was used. Mann-Whitney U-test was used for pairwise comparisons with Bonferroni-adjusted P values) statistical test. Data are quoted as median, and first and third quartiles.

Results
Mouse mutants have been instrumental in the identification and characterization of genes underlying congenital and early-onset hearing loss (Mianné et al. 2016;Bowl et al. 2017). In addition, the mouse has been utilized for the study of ARHL, taking advantage of the fact that different inbred strains, while having very similar hearing thresholds when young, exhibit differing rates of hearing loss with age (Ohlemiller et al. 2016). Some strains, such as C3H/HeJ and CBA, are labelled 'good' hearing strains and show a very slow decline in their hearing thresholds with age (Trune et al. 1996;Spongr et al. 1997;Sha et al. 2008). However, other strains, such as C57BL/6NTac and C57BL/6J, exhibit an early progressive hearing loss (starting from 3 months of age at high frequency), which is due to the presence of a fixed hypomorphic allele in Cadherin 23 (Cdh23 ahl : Johnson et al. 1997;Noben-Trauth et al. 2003). Cdh23 encodes cadherin-23 that, together with protocadherin-15, forms the stereocilia tip links required for gating the mechanoelectrical transducer channels (Kazmierczak et al., 2007;Richardson et al. 2011). A recent study has shown that when the Cdh23 ahl allele in C57BL/6NTac mice is 'corrected' , using CRISPR/Cas9 homology-directed repair (C57BL/6NTac Cdh23+ ), the mice exhibit improved age-related hearing with normal thresholds observed at 8 months of age (Mianné et al. 2016). Because of the complexity and heterogeneous nature of ARHL, we investigated the function of OHCs in mouse strains that show early-onset hearing loss (C57BL/6NTac: 6N; C57BL/6J: 6J) and strains that have good age-related hearing (C57BL/6NTac Cdh23+ : 6N-Repaired; C3H/HeJ: C3H), with the aim of identifying early signs of any morphological and physiological change in OHCs associated with ARHL.

Age-related changes in hearing function
We initially determined the hearing loss in the four mouse strains, which were kept under the same environmental conditions, in order to provide a more comprehensive understanding of the progression of the pathology. ABRs, which indicate the activity of the afferent auditory pathway downstream of IHCs, were made using the same experimental set-up. At 12-14 months of age, good-hearing C3H mice had ABR thresholds for click and pure-tone evoked ABRs (Fig. 1A,B) comparable to those published previously (Trune et al. 1996). ABR thresholds to clicks were significantly different among strains (P < 0.0001, one-way ANOVA). Compared with C3H mice, aged-matched 6J and 6N strains showed significantly increased ABR thresholds for clicks (6J, P < 0.001; 6N, P < 0.05: Tukey's post hoc test, one-way ANOVA: Fig. 1A). 6N-Repaired mice had ABR click stimulus thresholds comparable to C3H mice (P > 0.05), but significantly lower values than those measured in the co-isogenic 6N strain (P < 0.01) and 6J mice (P < 0.001:

Figure 1. ABR thresholds evoked by frequency-specific stimuli
A and B, mean ABR thresholds for clicks (A) and frequency-specific pure-tone stimulation from 3 kHz to 42 kHz (B) obtained from C3H, 6N, 6N-Repaired and 6J mice at 12-14 months of age. The arrows in B represent values above the upper threshold limit of our equipment (95 dB SPL). ABR thresholds were significantly elevated in 6J and 6N mice compared with both C3H and 6N-Repaired mice. For clicks: one-way ANOVA from single data points was: P < 0.0001. For frequency-specific stimulations, two-way ANOVA was P < 0.0001. For post hoc test comparisons, see Results. * indicates statistical significance when compared to C3H mice. Number of animals tested (males and females) is indicated by 'n' and shown next to the different mouse strains. Values are means ± SD.
In order to investigate the specific contribution of OHCs to these hearing phenotypes, we performed DPOAEs (Figs 2 and 3), which are readouts of cochlear amplification caused by the displacement of OHC stereociliary bundles during sound-induced motion of the cochlear partition. We initially compared the DPOAE thresholds between males and females at 13-14 months of age, and we found no significant differences in all four strains (C3H, P = 0.3853; 6N, P = 0.5136; 6N-Repaired, P = 0.1988; 6J, P = 0.1956, aligned ranks transformation two-way ANOVA). Because no statistical significance was detected between strain-matched males and females, data from males and females were pooled together to investigate possible inter-strain differences in DPOAE thresholds (Fig. 2). We found no significant difference between C3H and 6N-Repaired mice (P = 0.5737) or between 6N and 6J mice (P = 0.1800, aligned ranks transformation two-way ANOVA). However, DPOAE thresholds were significantly raised in both 6N and 6J mice compared with those recorded from C3H and 6N-Repaired mice (P < 0.0001 for all four combinations, aligned ranks transformation two-way ANOVA). These results also indicate that OHC defects are partially responsible for the different ABR thresholds between early-onset hearing loss and good-hearing mice (Fig. 1). We then used male mice to investigate age-related changes in DPOAE thresholds (6-24 kHz range: Fig. 3A-F). As expected, the C3H and 6N-Repaired mice showed no significant change in DPOAE thresholds between 1 and 17 months of age (C3H, P = 0.4791, Fig. 3A, D; 6N-Repaired, P = 0.8290, Fig. 3B, E), while early-onset hearing loss mice exhibited highly increased DPOAE thresholds (6N, P < 0.0001, Fig. 3C, F, aligned ranks transformation two-way ANOVA). At the 12 kHz region, which was used for the in vitro measurements described below, we found that DPOAE values were significantly different between strains (P < 0.0001, aligned ranks transformation two-way ANOVA). At 14-17 months, DPOAE thresholds were highly elevated in 6N compared with both C3H and 6N-Repaired strains (P = 0.0108; P = 0.0002, respectively), but not significantly different between C3H and 6N-Repaired mice (P = 0.5808, Tukey's post hoc test from aligned ranks transformation two-way ANOVA: see Methods). These findings are in agreement with the extensive literature describing the progressive changes in DPOAE threshold in early-onset hearing loss and good-hearing mice (e.g. Spongr et al. 1997;Jimenez et al. 1999;Zheng et al. 1999).

The biophysical properties of OHCs in ageing mice
In order to investigate possible changes in the biophysical properties of OHCs in ageing mice, we performed electrophysiological recordings from apical (ß9-12 kHz) OHCs  6N (C) at different ages for frequencies between 6 and 24 kHz. The number of mice with 'found' and 'not found' values at each frequency is shown below and above the median, respectively. D-F, comparison of age-related changes in the median DPOAE threshold values (from panels A-C) for the different mouse strains. Number of mice used for each strain/age is also shown.
of 12-13-month-old male and female mice. Mature mouse OHCs express a negatively activating K + current called I K,n , which is carried by KCNQ4 channels (Kcnq4 gene: Kubisch et al. 1999) and represents the major K + current present in these cells (Marcotti & Kros, 1999). Using whole-cell patch clamp, we recorded the K + currents from OHCs by applying a series of depolarizing voltage steps in 10 mV increments from −124 mV (holding potential was −84 mV). We found that a large I K,n was present in the OHCs of females and males from all four mouse strains (Fig. 4A-D shows female OHCs). The peak current-voltage (I-V) relationship of the total K + current (I K : Fig. 4E) in OHCs was not significantly different in females and males of all four mouse strains (C3H: P = 0.6558; 6N-Repaired: P = 0.3563; 6N: P = 0.7571; 6J: P = 0.2578, two-way ANOVA). The size of the total K + current (Fig. 4F), which was measured at 0 mV for consistency with previously published work in adult  6N-Repaired (C) and 6J (D) using 10 mV depolarizing voltage steps from −124 mV to the various test potentials shown by some of the traces. The activation of the adult-type current I K,n (indicated by the arrows) was present in all OHCs. E, average current-voltage (I-V) curves obtained from 12-13-month-old OHCs from females and males of the above four mouse strains. F, average size of the peak K + current measured at the membrane potential of 0 mV, as previously described (Marcotti & Kros, 1999) OHCs (Marcotti & Kros, 1999), was similar across all four mouse strains and sex (P = 0.9218, two-way ANOVA). We then investigated two crucial biophysical properties of OHCs, their membrane capacitance (C m ), which provides an indication of the cell surface area, and their resting membrane potential (V m ). Both were comparable between strains and sex (C m : P = 0.1990, Fig. 4G; V m : P = 0.4705, Fig. 4H, two-way ANOVA). The resting V m of OHCs was measured under whole-cell current clamp conditions and using 1.3 mM extracellular Ca 2+ , which also includes the effect of the very small resting MET current. MET channel blockers were not used because they normally affect the basolateral membrane conductances (e.g. Kenyon et al. 2017).
Following this initial characterization, we sought to investigate whether the biophysical characteristics of OHCs changed with age. Considering that we did not find any specific sex differences in aged mice in terms of hearing ability (Fig. 2) or OHC biophysics (Fig. 4), the following experiments are reported by combining the data from males and females. The size of I K decreased significantly with age in each mouse strain (C3H: P = 0.0096; 6N-Repaired: P = 0.0018; 6N: P = 0.0292; 6J: P = 0.0007, one-way ANOVA: Fig. 5A-C), but was not significantly  6N (B) mice at 1 month (left panels), 12 months (middle panels) and 16-17 months (right panels) of age. Currents were recorded using the voltage protocol described in Fig. 4. C, peak K + current (I K ) measured at the step potential of 0 mV from all four mouse strains at different ages. D, membrane capacitance (C m ) of OHCs as a function of age in the four mouse strains. E, normalized peak I K (panel C) to the corresponding C m (panel D) for each OHC tested. F, resting membrane potential (V m ) as a function of age in all the above mouse strains. In panels C-F, single cell values are also shown behind the average data. Number of OHCs investigated is shown above the average data points. Number of mice used: C3H (1 month: 6; 6 months: 6; 12-13 months: 12), 6N (1 month: 5; 6 months: 8; 12-13 months: 15; 15-17 months: 4), 6N-Repaired (1 month: 7; 6 months: 5; 12-13 months: 12; 15-17 months: 4) and 6J (1 month: 19; 6 months: 7; 12-13 months: 9; 15-17 months: 2). For statistical analysis, see Results.
Values are means ± SD. different between strains (P = 0.1267, all strains in the 1-13 month range; P = 0.1484, three strains in the 1-17 month range, two-way ANOVA). Within each strain and over the full age range investigated, the C m of OHCs was also significantly reduced with age (P < 0.0001 in each of the four strains, one-way ANOVA: Fig. 5D), with OHC surface being reduced by ß40% in 6J, 6N and 6N-Repaired and ß25% in C3H mice by 15-17 months. This reduction was not significantly different between strains (P = 0.1243, all strains in the 1-13 month range; P = 0.4959, three strains in the 1-17 month range, two-way ANOVA). We have also found that C m at 6 months was already significantly smaller than that at 1 month in 6J, 6N and C3H mice (P < 0.0001), but not in 6N-Repaired mice (P = 0.0727, Tukey's post hoc test, one-way ANOVA). When the size of I K in OHCs was normalized to their respective C m (i.e. current density), it was no longer significantly different for any of the four mouse strains (C3H: P = 0.1125; 6N-Repaired: P = 0.7086; 6N: P = 0.4292; 6J: P = 0.1577, one-way ANOVA, Fig. 5E), and also between strains (four strains in the 1-13 month age range: P = 0.0537; three strains in the 1-17 month range P = 0.3055, two-way ANOVA). Despite the reduced size of I K , the V m of OHCs remained stable until at least 16-17 months of age in all mouse strains investigated (C3H: P = 0.0622; 6N-Repaired: P = 0.9826; 6N: P = 0.5555; 6J: P = 0.1125, one-way ANOVA, Fig. 5F), and also between strains (four strains in the 1-13 month age range: P = 0.2861; three strains in the 1-17 month range P = 0.7763, two-way ANOVA).

Electromotility is maintained with ageing
One of the characteristic signatures of mature OHCs is the expression of the motor protein prestin (gene name: Slc26a5), which drives their somatic motility and thus generates DPOAEs. Therefore, we investigated whether prestin was maintained in ageing OHCs. We found that prestin was expressed in ageing OHCs from all four mouse strains until at least 16 months of age (6N and 6N-repaired: Fig. 6A-D). Since prestin is also used as an OHC marker, it allowed us to investigate any change in the number of OHCs present in the equivalent apical coil region used to perform the above electrophysiological experiments. Myosin 7a antibodies (e.g. Fig. 8) were also used as an additional hair cell marker to avoid underestimating the loss of OHC number with age due to the reduced expression of prestin. We found that the number of OHCs, which was measured over a distance of 140 µm in the ß9-12 kHz region, remained unchanged between 1 and 6 months of age in C3H (P = 0.9999), 6N (P = 0.2510) and 6N-Repaired mice (P = 0.9981, Tukey's post hoc test, one-way ANOVA). However, by 14-16 months of age the number of OHCs significantly decreased compared with one month in all four strains (C3H: P = 0.0019, 6N-Repaired: P = 0.0018, 6N: P < 0.0001, 6J: P < 0.0001, one-way ANOVA; Fig. 6E-H).
Although aged OHCs expressed the motor protein prestin (Fig. 6), we sought to test whether it was able to induce voltage-dependent changes in OHC motility (Zheng et al. 2000;Liberman et al. 2002). OHC electromotile activity can be assessed by either measuring OHC movement visually, or by using the non-linear (voltage-dependent) capacitance change (C N-L ) under whole-cell patch clamp, which is an electrical signature of electromotility (e.g. Abe et al. 2007;Bai et al. 2019;Zhai et al. 2020;Jeng et al. 2020). In the present study we preferred to use C N-L because it is particularly difficult to measure visual OHC motility on a precise horizontal plane and because displacement can be misleading through its dependence on intracellular turgor pressure, which is variable. We found that the maximum size of C N-L recorded in OHCs decreased significantly between 1 month and aged 6N (P = 0.0013), 6J (P = 0.0160) and 6N-Repaired mice (P = 0.0017, Tukey's post hoc test, one-way ANOVA), but was similar among strains (P = 0.6120, two-way ANOVA) (Fig. 7A-C). When the C N-L was normalized to the OHC membrane capacitance, there was no significant difference among all data groups (P = 0.0589, one-way ANOVA) or between strains (P = 0.2306, two-way ANOVA) (Fig. 7D). These data indicate that the electromotile activity is maintained in aged OHCs, but reduced in amplitude most likely due to the cells becoming smaller and losing some prestin in their lateral membrane (Fig. 7D). We have also investigated whether the expression level of prestin was affected in the aged cochlea by performing quantitative PCR experiments at 1 and 15-17-month-old 6N and 6N-Repaired mice. We found that the expression of prestin (Slc26a5) was significantly reduced in both aged mouse strains (6N, P = 0.0117; 6N-Repaired, P = 0.0248, Student's t test) (Fig. 7E). We also found that Ocm, which encodes for the main Ca 2+ binding protein oncomodulin expressed in OHCs (Simmons et al. 2010), was significantly reduced with age (6N, P = 0.0110; 6N-Repaired, P = 0.0028, Student's t test; Fig. 7F) in both strains. Although these qPCR experiments are intrinsically biased by the fact that OHC loss along the length of the cochlea is likely to be different among mouse strains (e.g. Hequembourg Note that the cell membrane capacitance (C m ) was added to the measured C N-L . C N-L was absent in the non-electromotile inner hair cells, which was used as a control for our experiments. C and D, average C N-L was reduced in OHCs from both mouse strains with age (C), but not after normalization to the OHC membrane capacitance C m (D). C N-L was calculated as the difference between the peak of the recording near -40 mV and the lowest value at positive membrane potentials. For statistical analysis, see Results. Number of cells investigated is shown above the data. Individual measurements are plotted behind the average data points and were collected from both males and females. Values are means ± SD. E, F, quantitative real-time PCR (qPCR) from 6N and 6N-Repaired mice at 1 and 15 months of age. Genes investigated: Slc26a5 (E, prestin) and Ocm (F, oncomodulin). Number of mice tested is shown above the data. Values are means ± SD. G and H, toluidine blue stained semi-thin plastic sections from the 8 kHz and 20 kHz region of the cochlea of 17-month-old 6N (G) and 6N-Repaired mice (H) mice. The tectorial membrane (TM) is attached to the spiral limbus (arrows) and lying on top of the OHCs (arrowheads). & Liberman, 2001;Francis et al. 2003;Zachary & Fuchs, 2015;Sha et al. 2008;Ohlemiller et al. 2010), the results show a similar reduction in prestin level between the early-onset (6N) and late-onset (6N-Repaired) hearing loss mice, suggesting that its expression level is likely to be similarly affected by ageing. OHC function also depends upon the tectorial membrane (TM), which is an acellular structure attached to the stereociliary bundles of the OHCs that gradually degrades with age especially in C57BL/6J mice . The presence and structure of the TM was investigated using light microscopy and toluidine blue staining. In both 6N and 6N-Repaired mice, the TM was attached medially to the spiral limbus, and laterally sitting atop the OHC stereociliary bundles (Fig. 7G,  H), indicating a similar gross structural organization in 17-month-old mice. However, a recent ultrastructure study has shown that the TM progressively detaches from the spiral limbus in very old CBA mice (Bullen et al. 2019), indicating some possible strain differences.

Afferent and efferent synapses are reduced in aged OHCs
We evaluated the number of afferent synapses on OHCs from 6N, 6J, 6N-Repaired and C3H mice using an anti-CtBP2 antibody to label the presynaptic ribbon protein RIBEYE. We found that CtBP2 puncta were present at the OHC presynaptic site in the different mouse strains (Fig. 8A, B for 6N and 6N-Repaired mice). The number of CtBP2 puncta in OHCs, over a distance of 140 µm in the ß9-12 kHz region, was significantly different when comparing strains tested at 1, 6 and 15 months (C3H, 6N and 6N-Repaired: P = 0.0025) or all four strains tested at 1 and 15 months (P = 0.0054, two-way ANOVA). Within each strain, the number of CtBP2 puncta decreased significantly in all strains (C3H and 6N-Repaired: P < 0.0001, 6N: P = 0.0090, one-way ANOVA; 6J: P < 0.0001, Student's t test, Fig. 8C-F).
Mature OHCs are mainly innervated by the cholinergic MOC neurons (Liberman, 1980;Maison et al. 2003), which release the neurotransmitter acetylcholine (ACh) (Simmons et al. 1996). The role of the MOC system is to modulate OHC electromotility and thus mechanical amplification in the cochlear partition (Guinan, 1996). The effects of efferent activity on OHCs is inhibitory because the release of ACh causes Ca 2+ influx through α9α10-nAChRs, which leads to the activation of a hyperpolarizing current through small-conductance Ca 2+ -activated K + channels (SK2: Oliver et al. 2000;Katz et al. 2004;Lioudyno et al. 2004;Marcotti et al. 2004). We investigated whether the efferent innervation was retained in the aged cochlea of all four mouse strains by immunolabelling with a choline acetyl transferase (ChAT) antibody, which labels the presynaptic efferent terminals, and with the postsynaptic SK2 antibody. We found that juxtaposed SK2 channels (in OHCs) and ChAT-immunoreactivity (efferent terminals) were still present in most of the remaining OHCs of 15-month-old mice (Fig. 9A, B for 6N and 6N-Repaired mice). The number of SK2 puncta per OHC, within a cochlear J Physiol 598.18 region of 140 µm of the apical coil (9-12 kHz), was not significantly reduced between 1 and 6 months (C3H: P = 0.4595; 6N: P = 0.5089; 6N-Repaired: P = 0.2419, Tukey's post hoc test, two-way ANOVA). However, between 1 and 15 months, the number of SK2 puncta was significantly reduced in 6N, 6N-Repaired and 6J (P = 0.0005, P = 0.0030, P < 0.0001, respectively, Tukey's post hoc test, one-way ANOVA) but not in C3H mice (P = 0.6580) (Fig. 9C). We also found that the percentage of most of the SK2 channels in OHCs were juxtaposed with ChAT staining (Fig. 9D), with no significant age-related changes in all mouse strains (P = 0.7668, two-way ANOVA). However, some OHCs showed only efferent terminals (Fig. 9A,B, arrowheads) or SK2 channels (Fig. 9A,B, arrows).

Mechanoelectrical transduction is reduced in OHCs from ageing mice
Another key biophysical property of hair cells is their ability to convert sound into an electrical signal, a process performed by the MET apparatus. The gross morphology of the OHC stereociliary bundles appeared normal in the apical coil of aged 6N mice ( Fig. 10A; see also Bullen et al. 2019), despite their elevated ABR and DPOAE thresholds (Figs 1-3). We therefore investigated whether other properties of the MET apparatus had changed in ageing mice. Quantitative PCR revealed that the expression of Cdh23 was significantly reduced at 15 months of age in both 6N and 6J mice (P = 0.0119 and P < 0.0001, respectively, Student's t test, Fig. 10B). A reduction in Cdh23 expression also occurred in the 6N-Repaired strain (P = 0.0073), which have overall better hearing than both 6N and 6J mice, and like CBA (Sha et al. 2008;Sergeyenko et al. 2013) are unlikely to lose many OHCs along the cochlea at old ages. On the other hand, Pcdh15 (protocadherin-15) was significantly increased in 6N and 6J (P = 0.0375 and P = 0.0057, respectively) but not in 6N-Repaired mice (P = 0.0891, Student's t test, Fig. 10B). These data indicate that despite the normal morphology of the stereociliary bundles, the function of the MET apparatus is likely to change in the ageing cochlea, which could be investigated by recording the MET current. However, recording the MET current in OHCs is usually limited to pre-hearing ages because of the difficulties associated with making measurements from adult hair cells. Furthermore, the stereociliary bundles of OHCs are easily damaged during the physical removal of the attached TM. In order to provide some information on possible changes in the MET current with age, we attempted a few recordings from mice (Tecta/b −/− : see Methods), in which the TM is not attached to the OHC stereociliary bundles and as such its removal does not disrupt the classical V-shaped bundle structure (Fig. 10C, D). As for the 6J and 6N strains, we found that these Tecta/b −/− mice, which are primarily on the 6N background, harbour the Cdh23 ahl allele. The OHCs of 11-month-old Tecta/b −/− mice were all present in the 9-12 kHz cochlear region, and their basolateral current profile was comparable to that of 1-month-old OHCs (Fig. 10E, F), but with a significantly reduced K + current (P = 0.0373, Student's t test, Fig. 10G), as also shown for the other mouse strains (Fig. 5C). Using these Tecta/b −/− mice, we showed that a large MET current was present in OHCs from both 1-month and 11-month-old mice (Fig. 10H-I), although in aged cells it was significantly reduced compared with 1 month (P = 0.0003, Student's t test, Fig. 10J). Despite the different MET current size, the open probability of MET channels at rest and in 1.3 mM Ca 2+ was comparable between the two ages (P = 0.9403, Fig. 10K). These data  Fig. 4. G, peak K + current (I K ) measured from Tecta/b −/− mouse OHCs at the step potential of 0 mV, from the holding of -84 mV, at 1 and 11 months of age. H and I, saturating MET currents recorded from OHCs of 1 month (H) and 11 month (I) Tecta/b −/− mice by applying sinusoidal force stimuli of 50 Hz to the hair bundles from a holding potential of -84 mV. The driver voltage (V Piezo ) signal of ± 40 V to the fluid jet is shown above the traces (positive half-cycles of the V Piezo o are inhibitory). The extracellular Ca 2+ concentration was 1.3 mM. The arrows indicate the complete closure of the transducer currents elicited during inhibitory bundle displacements (note that the difference in current between this level and the dashed lines represents the size of the resting MET current). Dashed lines indicate the holding current, which is the current at the holding membrane potential. J and K, maximal size (J) and resting open probability (K) of the MET current recorded at the two different ages from OHCs of Tecta/b −/− mice. The number of OHCs tested is shown above the data points.
suggest the intriguing possibility that changes in the MET current with age could be a mechanism involved in the progression of ARHL in ageing mice, at least in those harbouring the Cdh23 ahl allele as previously suggested (Johnson et al. 1997;Noben-Trauth et al. 2003;Johnson et al. 2017).

Discussion
In this study, we have provided evidence that several of the biophysical properties of OHCs remained unaffected by ageing. Despite the highly elevated DPOAE thresholds in the ß12 kHz cochlear region of aged early-onset hearing loss (6N and 6J) compared with C3H and 6N-Repaired mice, OHCs exhibited similar basolateral membrane characteristics irrespective of mouse strain. The first morphological alteration we observed in ageing OHCs (9-12 kHz region) was the reduction in their surface area, which was already evident in 6N, 6J and C3H mice at 6 months of age. The reduction in OHC surface area was not a sign of apoptosis in this case, since OHC loss was minimal even in 14-16-month-old mice, and did not significantly affect DPOAEs in C3H and C57BL/6NTac Cdh23+ mice. In addition, OHCs showed a reduction in the basolateral K + current size and electromotile activity, which was proportional to their smaller membrane surface area, but remained viable with normal resting membrane potentials. The size, but not the open probability of the MET current was also reduced in aged OHCs compared with young adult cells. At the synaptic level, the loss of afferent and efferent synapses was evident at 15 months. OHCs from C3H mice, which showed the best hearing profile at 12 kHz, exhibited very little change to their efferent synapse architecture in old mice, highlighting specific genetic background differences. We propose that with ageing, the basolateral membrane of OHCs undergoes several biophysical and morphological changes (Fig. 11). We have also provided evidence that the MET current of OHCs from mice harbouring the Cdh23 ahl allele is reduced with age, highlighting the possibility that it could be playing a crucial role in ARHL.

Age-related hearing loss in different mouse strains
The mouse has become the primary animal model to study ageing in the auditory system (Ohlemiller et al. 2016;Bowl & Dawson, 2019). The commonly used 6J (C57BL/6J) and 6N (C57BL/6NTac) mice have a single-nucleotide polymorphism in exon7 of the cadherin23 (Cdh23 ahl or Cdh23 753A : Johnson et al. 1997;Noben-Trauth et al. 2003;Johnson et al. 2017), which causes them to already exhibit hearing loss in the high-frequency region by 3-6 months and become almost completely deaf by 15 months of age (Hequembourg & Liberman, 2001;Kane et al. 2012). The Cdh23 ahl allele is present in several of the early-onset, but not in the late-onset hearing loss mouse strains, such as C3H mice (Trune et al. 1996). To investigate the possible involvement of strain-specific effects, we have also used the 6N co-isogenic C57BL/6NTac Cdh23+ mice (6N-Repaired), in which the Cdh23 ahl allele was repaired with targeted CRISPR/Cas9 gene editing (Mianné et al. 2016). We found that, compared with the C3H and 6N-Repaired strains, aged mice from both 'early-onset hearing loss' strains (6N and 6J) had significantly raised ABR and DPOAE thresholds as previously shown (e.g. Parham, 1997;Spongr et al. 1997;Jimenez et al. 1999;Zheng et al. 1999;Hequembourg & Liberman, 2001;Sha et al. 2008). Although sex differences in DPOAE thresholds have been described in CBA mice, they normally occur starting from 14-16 months (Guimaraes et al. 2004), which is just after the age at which we performed the recordings in the four mouse strains. In the frequency region used to investigate the morphological and physiological properties of OHCs (9-12 kHz), DPOAE thresholds were similar in all four mouse strains until about 9 months of age, but became significantly raised thereafter in both 6N and 6J mice.

The reduced surface area of OHCs is an early sign of cochlear ageing
We have shown that the surface area of aged OHCs is smaller than that measured in young adult OHCs. Although cellular shrinkage is generally associated with apoptotic mechanisms, this is not the case for the OHCs. At a time when the reduction in surface area was already present in all OHCs from 6-month-old mice (9-12 kHz cochlea region), we found no OHC loss in any of the four mouse strains. Moreover, in 14-16-month-old mice only about 20% of the apical OHCs were missing, which is consistent with previous reports for both C57BL/6 (Hequembourg & Liberman, 2001;Francis et al. 2003;Zachary & Fuchs, 2015) and CBA mice (Sha et al. 2008;Ohlemiller et al. 2010;Sergeyenko et al. 2013). The normal DPOAE thresholds in aged C3H and 6N-Repaired mice, despite having fewer OHCs present in the apical coil, are likely to be due to the smaller contribution of cochlear amplification for low-frequency sounds compared with higher-frequency sounds (Liberman et al. 2002).
The presence of healthy OHCs despite a reduced surface area is also supported by a previous study showing that cell shrinkage is a morphological change that can be separate from apoptosis (Bortner & Cidlowski, 2003). There are several lines of evidence indicating that a cell's size can be adaptively regulated to an optimal volume in response to a variety of physiological stimuli (Ginzberg et al. 2015;Miettinen & Björklund, 2016), with larger cells tending to have higher metabolic activities (Giordano et al. 1993;Ginzberg et al. 2015;Miettinen & Björklund, 2016). The cochlea has been shown to be highly dependent on energy provided by mitochondria, and their dysfunction exacerbates the progression of ARHL (Pickles, 2004;Crawley & Keithley, 2011). Therefore, it is possible that the reduction in OHC size reduces energy consumption, thus maintaining functionality. Since a similar temporal progression in the reduction in size and loss of OHCs is present in both 'early-onset hearing loss' (6J and 6N) and C3H and 6N-Repaired mice, they are unlikely to contribute directly to their different DPOAE thresholds, implicating OHC reduction in size as a general trait in the normal pathophysiological changes in the cochlea with age.

Changes in the physiology of OHCs with age
In addition to the reduction in OHC surface area, we observed a decrease in the size of basolateral K + current I K,n and non-linear capacitance (NLC). However, when the size of I K,n and NLC were normalized to OHC surface area they were comparable with those in younger cells, suggesting that the loss of OHC membrane is responsible for the loss of proteins. A reduced OHC surface area and loss of prestin has also recently been shown in OHCs from 8 month old C57BL/6 mice (Bai et al. 2019). In addition to the loss of basolateral membrane channels and prestin, aged OHCs showed a similar progressive loss of afferent synapses irrespective of the mouse strain. This seemed to also correlate with the reduction in OHC surface area since their number was still normal at 6 months of age. The efferent synapses were also largely affected with age. However, these age-related changes were minimal or absent in C3H mice, which have the better hearing profile. Despite the progressive reduction in basolateral membrane currents, the in vitro resting membrane potential (V m ) remained similar between young and aged OHCs and to that previously reported (Marcotti & Kros, 1999). The consistency in resting V m with age indicates that OHCs are likely to be viable and functional, which is also suggested by the normal DPOAE thresholds in aged C3H and 6N-Repaired mice. Therefore, similar to the reduction in OHC size, the above biophysical and synaptic changes appear to be a general mechanism that maintains cell function in the ageing cochlea. One change that seems to be independent from the smaller basolateral membrane surface area was the reduction of the MET current. These data provide the intriguing hypothesis that the early-onset hearing loss in mice harbouring the Cdh23 ahl allele could be due to defects in the MET apparatus. Being able to directly address this hypothesis will require the generation of the Tecta/b −/− mouse on the 6N-Repaired background. However, we recently found that in IHCs, whose stereocilia are not attached to the TM and, therefore, less damaged by its removal, the MET current was found significantly smaller in mice harbouring the hypomorphic Cdh23 allele Cdh23 ahl compared with that measured in IHCs from 6N-Repaired mice (data not shown).

Mechanisms leading to progressive hearing loss with age
Our results have shown that in the apical coil (9-12 kHz) of the aged cochlea from all four mouse strains, OHCs retained similar resting membrane potentials to those measured in cells from young adult mice. Although the basolateral membrane protein profile of OHCs from all mouse strains was retained at older ages, the expression level of different proteins was significantly reduced compared with that found in young adult cells, most likely due to their reduced surface area (Fig. 11). The extent of OHC and afferent loss with age was comparable between all four mouse strains, indicating that these age-related changes are a general characteristic of the ageing cochlea. The loss of efferent synapses appeared strain-specific because it was very minimal or absent in C3H mice, which is the best-hearing strain.
Our finding that MET current size is reduced in ageing mice carrying the Cdh23 ahl allele is consistent with previous work showing that this allele is likely to be J Physiol 598.18 responsible for the early-onset hearing loss phenotype in 6J and 6N mice (Johnson et al. 1997;Noben-Trauth et al. 2003), which agrees with our finding of a reduction in MET current size with ageing in mice with the Cdh23 ahl allele. The contribution of the MET apparatus to ARHL is also supported by the progressive alteration in the structural and functional integrity of the stereociliary bundles observed with age in the chinchilla, mouse and human (Bohne et al. 1990;Bullen et al. 2019). Additional functional alterations in hair cells that could contribute to the progression of hearing loss with age include the biophysical properties of IHCs and their synapses, with the latter also being correlated with the degree of hearing loss in mice (data not shown).