Loss of central mineralocorticoid or glucocorticoid receptors impacts auditory nerve processing in the cochlea

Summary The key auditory signature that may associate peripheral hearing with central auditory cognitive defects remains elusive. Suggesting the involvement of stress receptors, we here deleted the mineralocorticoid and glucocorticoid receptors (MR and GR) using a CaMKIIα-based tamoxifen-inducible CreERT2/loxP approach to generate mice with single or double deletion of central but not cochlear MR and GR. Hearing thresholds of MRGRCaMKIIαCreERT2 conditional knockouts (cKO) were unchanged, whereas auditory nerve fiber (ANF) responses were larger and faster and auditory steady state responses were improved. Subsequent analysis of single MR or GR cKO revealed discrete roles for both, central MR and GR on cochlear functions. Limbic MR deletion reduced inner hair cell (IHC) ribbon numbers and ANF responses. In contrast, GR deletion shortened the latency and improved the synchronization to amplitude-modulated tones without affecting IHC ribbon numbers. These findings imply that stress hormone-dependent functions of central MR/GR contribute to “precognitive” sound processing in the cochlea.


INTRODUCTION
Hearing loss with age has recently been suggested to be an important modifying factor increasing the risk of dementia (Livingston et al., 2017;Montero-Odasso et al., 2020). The link between both pathologies is controversial. Until now, peripheral hearing and deficits in central cognitive processes, including cognitive decline and dementia, were predicted to be linked through limbic frontal brain dysfunction (Mudar and Husain, 2016), independent of cochlear functionality (Cope et al., 2015) and possibly disproportionate to peripheral hearing loss (Johnson et al., 2021). Among others, central hearing deficits include deficits in attention or executive function (Rutherford et al., 2018) that are also altered through chronic stress (Perez-Valenzuela et al., 2019;Panza et al., 2019;Canlon et al., 2013). Blockade of stress-hormone binding mineralocorticoid receptors (MR), for instance, impairs both memory tasks and selective attention (Wingenfeld and Otte, 2019;de Kloet et al., 2000de Kloet et al., , 2019. On the other hand, through neuronal atrophy and synaptic dysfunction, chronic stress can contribute to the degradation of synaptic plasticity and thereby influence cognitive functions (for review see Vyas et al., 2016). In the auditory system, numerous studies that analyzed stress-related hearing dysfunction in response to age, acoustic trauma (AT), or posttraumatic stress observed associated changes in cognitive functions (Jafari et al., 2019;Basner et al., 2014;Mazurek et al., 2019;Nadhimi and Llano, 2021;Johnson et al., 2021;Wang and Puel, 2020;Canlon et al., 2013;Meltser and Canlon, 2011). How stress-receptor activation links hearing and cognition is, however, currently elusive.
Glucocorticoid receptors (GR) and MR in the limbic system, including the prefrontal cortex and hippocampus, are suggested to mediate the top-down and bottom-up control of stress coping with environmental challenges through hypothalamic and extrahypothalamic prefrontal and hippocampal regions (de Kloet et al., 2000(de Kloet et al., , 2019. Previous studies implied that pharmacological or acoustic trauma-induced stress affects central auditory processing through sensorineural cochlear responses (Singer et al., 2013(Singer et al., , 2018. To examine whether the brain's ability to recognize and interpret sound depends on stress hormone receptors, we tested the effect of induced genetic disruption of central MR and GR on cochlear function. Employing the tissue-specific tamoxifen (TMX)-inducible Cre ERT2 /loxP system allowed for single and combined . After 4 weeks of recovery, the hearing was measured again and mice were sacrificed to collect brain and cochlea tissue. The TMX induction led to a tissue-specific deletion of MR and GR under the promoter of CaMKIIa, which is expressed in the forebrain, but not the cochlea. We conclude that our TMX-induction protocol promotes the activation of CaMKIIa-Cre ERT2 , which leads to an efficient deletion of MR and GR specifically in limbic brain regions (hippocampus) (Erdmann et al., 2007), but that TMX injection did not delete these receptors in the auditory periphery, as schematically depicted in Figure 1A.
Deletion of MR and GR in MRGR cKO mice exposed a negative impact of stress receptors on auditory temporal processing, independently of OHC We demonstrated that the hearing of Cre positive and Cre negative animals bearing floxed MRGR alleles was not different in the absence of TMX (data not shown). In addition, after TMX injection, ABR thresholds of control and MRGR cKO mice were not different, as shown for click-evoked and noise-evoked responses and for pure-tone stimuli between 2 and 32 kHz ( Figure 1B). When DPOAE thresholds were analyzed, providing information about electromechanical properties of OHC function, no difference between the genotypes was observed ( Figure 1C). Likewise, the population response thresholds to sound from ANF were not different between the two genotypes ( Figure 1D). In contrast, suprathreshold click-evoked ABR that are calculated as peak-to-peak amplitude growth functions of ABR wave I (originating from the auditory nerve) and ABR wave IV (originating from the inferior colliculus) (Melcher et al., 1996), were significantly increased in MRGR cKO mice ( Figures 1E and 1F, left and right panel). Thus the stimulus-level-dependent spreading of sound-response amplitudes was already elevated in the auditory periphery, where MR and GR levels were unaltered in MRGR cKO ( Figure S1B).
Together, these findings imply that the acute, combined deletion of MR and GR under the CaMKIIa promoter, while not influencing the electromechanical properties of OHC, leads to faster and more sensitive sensorineural cochlear processing, resulting in elevated brain responses to amplitude-modulated tones. This in turn suggests that the combined MR/GR activation during, e.g., elevated (chronic) stress might inhibit auditory neuronal synchronization and temporal auditory processing.
Differential influences of MR or GR on discharge rate and synchrony of auditory nerve responses contribute to the overall phenotype of MRGR cKO on auditory processing As a possible rationale for the observed altered auditory responses in the MRGR cKO mice, we considered differences in the number of IHC ribbons, which influence auditory processing through defining an ANF's discharge rate (Kujawa and Liberman, 2009;Buran et al., 2010). To examine the ribbons, we used antibodies directed against the RIBEYE protein CtBP2 (Khimich et al., 2005). In MRGR cKO mice, IHC ribbon numbers were significantly reduced in high-frequency coding medial and midbasal cochlear turns ( Figure 2A). It is assumed that IHC ribbons can be subdivided into modiolar-sided vs. pillar-sided ribbon synapses, which contact postsynapses that differ functionally in spontaneous firing rates (SR) and thresholds for sound coding. The large modiolar ribbon synapses are known to have contact with small postsynapses of low-SR, high-threshold ANF, which are recruited at higher SPL and show little or no saturation. The small pillar ribbon synapses have contact with large postsynapses of high-SR, low-threshold ANF, which are activated in response to lower SPL and rapidly saturate (Liberman, 1982;Liberman and Oliver, 1984;Merchan-Perez and Liberman, 1996;Winter et al., 1990). In C57BL/6 mice the ratio between modiolar-sided and pillar-sided IHC ribbons is around 50:50 (Reijntjes et al., 2020). The same proportion was confirmed in MRGR control mice when analyzing the modiolar/pillar gradient of IHC ribbons in the midbasal turn ( Figure S2A). However, MRGR cKO had less ribbons located on the pillar side of the IHC in comparison to ribbons counted on the modiolar side ( Figure S2B; left panel). This was contrary to assumptions, as typically, a larger ABR wave I is expected to correlate with a larger, and not a reduced, number of IHC ribbons (Buran et al., 2010;Kujawa and Liberman, 2009). These inconsistencies might have their rational in an overlap of contrasting MR-influences or GR-influences on sensorineural responses (see below).
We hypothesized that a combined MRGR cKO may uncover possible differential effects of the individual MR or GR functions on auditory processing. Indeed, examining TMX-inducible single MR cKO and GR cKO mice revealed that IHC ribbons in MR cKO mice were like in MRGR cKO mice numerically reduced ( Figure 2B) and exhibited in high frequency cochlear regions a smaller number of IHC ribbons on the pillar  Figure S2B; middle panel). In contrast, in GR cKO mice, no changes of IHC ribbon numbers ( Figure 2C) between controls and GR cKO mice were seen, and the percentage of IHC ribbons numbers on pillar vs. modiolar sides in midbasal turns were not different in WT and GR cKO mice (Figure S2B;right panel). This may suggest that the observed reduction of IHC ribbon numbers, seen in MRGR cKO mice, is rather linked to the deletion of central MR.
When putative effects of the distinct limbic MR and GR ablation on ABR and DPOAE thresholds were analyzed, the ABR hearing thresholds in response to click-tone, noise-tone, or pure-tone stimuli ( Figures  S3A and S3C), as well as DPOAE thresholds ( Figures S3B and S3D), were not found to be different in MR cKO nor GR cKO compared to their respective controls. This supports the notion that neither the limbic MR nor GR affects basal hearing thresholds or electromechanical OHC response properties.
The reduction of IHC ribbons in MR cKO mice was linked to a reduction of the click-evoked ABR wave I amplitude when compared to controls ( Figure 3A, left panel). This was, however, centrally compensated, as evident from the normal-sized or even larger ABR wave IV (right panel). The smaller ABR wave I response in MR cKO mice was functionally reflected in slightly, but significantly, higher CAP thresholds compared to control mice ( Figure 3B, left panel), and in slightly prolonged CAP latencies (right panel). When ASSR were analyzed in MR cKO mice, we observed no difference to control mice in the modulation-depth function at 40 dB relative to threshold ( Figure 3C, left panel), in modulation growth functions (middle panel), or in the . This together suggested that the influence of MR deletion on auditory processing is restricted to IHC ribbon numbers and ABR wave I amplitudes. In contrast, the deletion of limbic GR, while leaving IHC synapse ribbon numbers unaffected, led to higher click-evoked ABR wave I ( Figure 4A, left panel) and wave IV amplitudes (right panel), as also observed in MRGR cKO mice ( Figure 1F). Consistently, compared with control mice, GR cKO mice exhibited a significantly lower CAP threshold ( Figure 4B, left panel), and shorter CAP wave I latency (right panel). The shorter CAP latency in GR cKO mice was coincident with stronger ASSR responses, evident in a larger signal than in controls when measuring responses to amplitude-modulated stimuli with variation in modulation depth ( Figure 4C, left panel), modulated in stimulus level ( Figure 4C, middle panel, or changes in modulation frequency (Figure 4C,right panel). Thus, GR deletion influences auditory processing through changes in spike timing and in the synchronization of neural auditory responses that were shown to be required for following amplitudemodulated stimuli (Johnson, 1980).
Overall, these findings in MRGR cKO mice suggested a negative influence of limbic MR deletion on auditory nerve responses, evidenced through reduced IHC ribbon numbers (Figure 2), reduced ABR wave amplitudes, elevated CAP thresholds, and prolonged CAP latencies in MR cKO mice ( Figure 3). This negative effect was counterbalanced after central GR loss that -as shown in GR cKO mice -results in elevated ABR amplitudes, lower CAP thresholds, and shortened CAP latencies, as well as facilitated iScience Article ASSR ( Figure 4). Both phenotypes together, reduction of the peripheral auditory processing after MR deletion and stimulation after GR deletion, contribute to the complex phenotype of MRGR cKO mice ( Figure 1).
Blood corticosterone levels do not account for the differential effects on peripheral hearing in MR cKO, GR cKO, and MRGR cKO mice Previously, the spatiotemporal deletion of GR in limbic forebrain regions was shown to enhance blood CORT levels through an unbalanced autoregulation of hippocampal GR on the hypothalamus-pituitary-adrenal (HPA)-axis (Erdmann et al., 2007). To investigate to what extent TMX-induced deletion of MR/GR might also exhibit its influence through an unbalanced HPA-axis, we analyzed the CORT level from blood plasma in anesthetized MR cKO, GR cKO, MRGR cKO, and control mice during hearing measurements. In MR cKO mice that exhibited smaller ABR wave amplitudes ( Figure 5A, left panel), the level of CORT was identical to that of controls ( Figure 5B, left panel), and the ABR wave I amplitude size, which was the mean of the three maximal amplitude values of the individual ears' suprathreshold amplitude growth function, was not correlated with an individual animal's stress-hormone levels ( Figure 5C, left iScience Article panel). In contrast, in GR cKO mice, characterized by an elevated ABR wave I ( Figure 5A, middle panel), a tendency toward higher CORT levels was observed ( Figure 5B, middle panel). Again, no direct correlation between ABR wave I amplitude size and an individual animal's stress levels was found ( Figure 5C, middle panel). In MRGR cKO both ABR wave I ( Figure 5A, right panel) and the CORT level were elevated ( Figure 5B, right panel). However, no correlation between the ABR wave I amplitude size and an individual animal's stress hormone levels was found ( Figure 5C, right panel). This may suggest that the GR-deletion-induced HPA-axis dysregulation is not the primary driver for the observed changes of MR/GR cKO on peripheral hearing.
In conclusion, GC activation works in a binary fashion on IHC synapses and auditory-nerve synchrony: by activation of either the stimulating limbic MR ( Figure 6, green plus) on the one hand and by activation of the inhibiting GR ( Figure 6, red minus) on the other hand. Signaling through these receptors thus provides a key auditory signature that may associate peripheral hearing with central auditory cognitive hearing defects ( Figure 6).

DISCUSSION
Physical and psychological stressors are manifested through the activation of the HPA-axis, and the production of GC exerts profound effects on neuronal networks and sensory gating. Acutely elevated and chronically elevated GC levels are assumed to influence sensory gating at the cortical or hippocampal level, independent of peripheral sensory function (Johnson et al., 2021;Basner et al., 2014). By investigating TMXinduced conditional single or combined deletion of MR/GR in forebrain regions, we report functional roles of limbic MR and GR as peripheral modulators of the IHC synapse and in ANF processing. This provides a concept for limbic MR and GR activities during acute or chronic stress, which should be reconsidered as modulators for subcortical processing and filtering elements during auditory perception. In this context, either limbic MR or GR function contribute to the precision of auditory processing, thereby possibly influencing speech comprehension and cognitive hearing function.

Specificity of the TMX-induced limbic deletion of MR and GR
Homozygous global GR KO (Cole et al., 1995), as well as Cre-mediated early embryonic ablation of GR by a constitutively active Cre in GR CaMKIIa KO mice (Erdmann et al., 2008), are lethal. As early as P8, animals with germline inactivation of MR, as shown in the homozygous global MR KO, suffer from hyperkalemia, hyponatremia, weight loss, and a strong increase in renin, angiotensin II, and aldosterone plasma concentrations. Consequently, a global lack of MR induces early postnatal death because of bodily dehydration (exsiccosis) as a consequence of massive renal sodium and water loss (Berger et al., 1998). To overcome the lethal phenotype of global GR KO and cell-specific (non-inducible) GR CaM-KIIa KO, and the severely diseased phenotype of young MR KO (Erdmann et al., 2007;Berger et al., 1998), we used TMX-inducible Cre driver lines (Berger et al., 2006;Erdmann et al., 2008) for the generation of adult MR cKO, GR cKO, and MRGR cKO mouse models. This enabled in parallel the extraction of specific roles of GR and MR in hearing in adulthood. As shown for MRGR cKO mice, both target genes were deleted in pyramidal cells of the forebrain (Figures S1C and S1D), but not in the cochlea (Figures S1E and S1F). The absence of CaMKIIa in the cochlea is in line with other studies (Meese et al., 2017). We observed high CORT levels in MRGR cKO mice, and partly also in the GR cKO, which corresponds well with an analogous GR CaMKIIa KO mouse line generated by Erdmann and coworkers, whereas no increase was seen in MR cKO mice ( Figure 5A) (Erdmann et al., 2007). This suggests that the TMX-inducible double deletion and single deletion of MR and GR mirrors previous mouse models with respect to HPAaxis dysfunction. However, GR cKO and MR cKO mice presented with distinct hearing phenotypes, which is in line with predictions that GR and MR serve nonredundant functions in neurons to control distinct transcriptional networks (Mifsud and Reul, 2016;Obradovic et al., 2004;de Kloet et al., 2000;McCann et al., 2021;Reul and de Kloet, 1985). As both nuclear receptors recognize the same specific DNA promoter sequences, called GC response elements (Rupprecht et al., 1993;Mifsud and Reul, 2016), the differential effects of MR and GR may stem from the different MR/GR expression profiles. GR is expressed in virtually all cell types in the rodent brain, whereas MR is expressed primarily in neurons of limbic regions such as the hippocampus, lateral septum, and amygdala (de Kloet et al., 2000;Chao et al., 1989;Reul and de Kloet, 1985). Besides their individual expression profile across the brain, differential MR and GR effects derive from different affinities of the individual receptors, with MR having a 10-fold higher affinity for CORT than GR, being occupied by ligands even under baseline, low-stress conditions ( iScience Article is stressed or during circadian periods when circulating CORT levels are naturally elevated (Mifsud and Reul, 2016;Joels et al., 2008;de Kloet et al., 1993).
Because MR has the same binding affinity for aldosterone, cortisol, and CORT (Gomez-Sanchez and Gomez-Sanchez, 2014), changes in presumptive aldosterone effects on hearing might contribute substantially to the MR cKO phenotype. Until now, only long-term effects of aldosterone on hearing have been described, acting through influences on the endocochlear potential (Bazard et al., 2020). The inactivation of cortisol and CORT by 11beta-hydroxysteroid dehydrogenase type 2 (11beta-HSD2) is, moreover, required to allow aldosterone to activate MR within aldosterone target cells (Gomez-Sanchez and Gomez-Sanchez, 2014). An immunoreactivity of 11beta-HSD2 has, however, not been detected in any inner ear tissues (Terakado et al., 2011).
A B Figure 6. Hypothesized effect of limbic MR/GR activation on auditory processing in the cochlea (A) Compound limbic deletion of MR and GR leads to reduced IHC ribbon numbers and enhanced auditory neural responses to modulated tones, indicating improved synchrony of neural responses. This phenotype is a mixture of MR cKO mice with reduced IHC ribbon numbers, reduced ABR wave I, but without effects on synchronized neural auditory responses, and GR cKO mice with normal IHC ribbon numbers and increased ANF synchrony. (B) Hypothesized physiological effect of the activation of limbic MR, possibly contributing to discharge rate, stabilized presynaptic IHC elements, and thereby to an improved ASSR, whereas the activation of limbic GR leads to the inhibition of ANF synchrony, thereby diminishing temporal auditory coding through degradation of synchronized neural auditory responses ll OPEN ACCESS iScience 25, 103981, March 18, 2022 iScience Article Finally, regarding the prominent circadian activity exhibited by the HPA-axis, functional hearing experiments were conducted during the same time period of the day. We were thereby able to avoid dynamic changes in the intrinsic properties of these MR-positive and/or GR-positive neural populations, which may alter the function of neural circuits affecting central or peripheral auditory pathways.
Together, the distinct central and cochlear MR/GR expression profiles, MR/GR-dependent changes in blood CORT levels, as well as the apparently contrasting auditory functions of limbic MR and GR, mirror, at least partly, the respective phenotypes of global and conditional MR/GR KO (Cole et al., 1995;Erdmann et al., 2008). Combined, these findings substantiate the specific MR and GR deletion in limbic forebrain regions of these mouse models.
Differential impact of limbic MR and GR activation on auditory-nerve processing is independent of OHC Activation of both MR and GR in frontal brain regions and the hippocampus has a direct and differential impact on peripheral auditory processing. This can be concluded from the observation that the induced, combined limbic deletion of MR and GR in adult animals significantly enhanced amplitudes and reduced delays in auditory-nerve responses, independently of OHC function, as no difference was found in the DPOAE measurements of control and MRGR cKO mice ( Figure 1). The amplitude of the suprathreshold evoked ABR wave I (Kujawa and Liberman, 2009), the peripheral neural response that is affected in MR and GR cKO mice, is determined through discharge rate and synchrony changes of the auditory nerve (Johnson and Kiang, 1976;Kujawa and Liberman, 2009). ABR wave I spreads centrally to generate ABR wave IV at the level of the inferior colliculus (Melcher et al., 1996). At the cochlear level the effect of MR/ GR deletion was quantified through the altered threshold and latency of acoustically-evoked CAP responses reflecting summed single-fiber action potentials of ANF (Earl and Chertoff, 2010; Rü ttiger et al., 2017). These CAP measurements allow, moreover, a first estimate of possible distinct contributions of ANF types, which exhibit either high-SR, low-threshold ANF, which are activated in response to lower SPL and rapidly saturate, or low-SR, high-threshold ANF, which are recruited at higher SPL and show little or no saturation (Winter et al., 1990). CAP thresholds are insensitive to changes in low-SR ANF contributions, and only shift when high-SR ANF are affected (Bourien et al., 2014). As high-SR ANF are responsible for the shortest latencies of auditory responses at any given characteristic frequency and are suggested to define perception thresholds (Meddis, 2006;Heil et al., 2008), changes in the thresholds or latencies of an evoked CAP response might point to a contribution of such high-SR ANF.
The deletion of both MR and GR in MRGR cKO mice resulted in larger and faster ABR wave I and IV waves, with both changes occurring independently of any changes in DPOAE function. This improved neural auditory response in MRGR cKO mice was confirmed through shorter CAP latencies and improved ASSR resolution (Figure 1), indicating that compound limbic MR and GR activation might exhibit an overall negative impact on auditory temporal processing.
Interestingly, an analysis of induced, tissue-specific single MR or GR deletions revealed that the response pattern of MRGR cKO mice results from the contrasting effects of MR and GR on either auditory nerve activity and discharge rates (MR cKO) or on spike timing and synchrony (GR cKO). Thus, the lack of MR in the limbic system results in reduced IHC ribbon numbers, and specifically to a reduction of pillar-sided IHC ribbons, and reduced amplitudes of ABR wave I, while leaving ASSR responses intact (Figures 2B and 3;Figure S2B). IHC ribbons contribute through their influence on the readily-releasable vesicle pool at IHC synapses to spontaneous and evoked discharge rates of ANF, and when absent, lead to a severe reduction in ABR wave I amplitude and to deficits in onset responses and first-spike latencies (Buran et al., 2010). In the absence of IHC ribbons, the hearing threshold, dynamic range, and response precision to amplitudemodulated tones remain intact, and spreading auditory response amplitudes of ABR wave II, generated in the superior olivocochlear complex (Melcher et al., 1996), are centrally compensated (Buran et al., 2010). In MR cKO mice, the IHC ribbon numbers and ABR wave I amplitude are reduced, the ABR wave IV is enhanced, and ASSR are normal; from this we may conclude that the limbic MR affects the temporal resolving power of the auditory nerve responses through its influences on IHC ribbon integrity ( Figure 6). The spatial gradient of IHC ribbon synapses provides information about postsynaptic ANF that functionally differ in SR and thresholds, if localized either on the pillar or modiolar side (Liberman, 1982;Liberman and Oliver, 1984;Merchan-Perez and Liberman, 1996) iScience Article contact stabilization to ANF with high-SR and low thresholds that dominate more on the pillar side. It is challenging to consider for future studies if retrocochlear influences of MR activities may be realized through stress-MR-mediated effects on efferent cochlear feedback control.
In contrast, GR cKO mice showed higher-amplitude ABR waves I and IV, a shorter CAP wave I latency, reduced CAP threshold, and a better ability to detect temporally challenging, frequency-modulated tones (Figure 4), while leaving IHC ribbon numbers intact ( Figure 2C). Intact IHC ribbon numbers, but altered ASSR responses, suggest that limbic GR activation exerts a ''negative'' top-down effect on temporal auditory processing through influencing the synchronization of spike times. Precise spike synchronization is influenced through, e.g., the fast kinetics of voltage-gated calcium channels (Ca v 1.3) in IHC (Zidanic and Fuchs, 1995;Neef et al., 2009). These are required to ensure Ca 2+ -binding during the docking of release-ready vesicles and exocytosis (Wong et al., 2013;Rutherford et al., 2021;Johnson et al., 2019). This process critically depends on the metabolically demanding and timely retrieval of vesicles through endocytosis (Wu et al., 2014). Although the underlying details need to be determined further, limbic GR activation may provide the proper metabolic supply for precise spike synchronization. This concept is strengthened through the profound influences of CAP threshold and latency shifts in GR cKO-and MRGR cKO mice (Figures 1 and 4), that may indicate a specific vulnerability of high-SR ANF responses to limbic GR activation, as only high-SR ANF alter CAP thresholds and latencies (Bourien et al., 2014).
In addition, previous findings indirectly pointed to a GR contribution to synchronization of ANF responses. Thus, AT-induced changes in the dynamic range of auditory nerve response were weakened by GR, but not MR antagonists (Singer et al., 2018). As described in Buran et al. (2010) the dynamic range of ANF, which is defined as the range of stimulus levels over which discharge rate increases, is unaffected upon a loss of synaptic ribbons, following Bassoon deletion. This would mean that the dynamic range of the auditory nerve response is strongly affected by synchronized spike responses, but less through altered discharge rates (Buran et al., 2010). In total, the findings in GR cKO and MRGR cKO mice suggest that limbic GR activation may influence temporal auditory processing independently of discharge-rate changes of ANF, but rather through its negative impact on spike timing and synchronization of neural responses by affecting high-SR ANF processing ( Figure 6). Although further studies are required to confirm the differential effect of MR and GR on discharge rate and synchronization of ANF responses, it is a striking observation that the GR function in auditory nerve processing did not correlate with blood CORT level changes. Previous studies observed limbic MR/GR functions to be the result of locally synthesized CORT on MR and GR in hippocampal cells, independently of GC produced by adrenal glands (Croft et al., 2008;Taves et al., 2011;McCann et al., 2021). Such local MR effects have previously been shown to alter e.g., social behavior or behavioral responses to novelty (McCann et al., 2021). Further studies are required to elucidate the signaling cascade by which limbic MR and GR improve or degrade auditory neural responses.

Relation of the finding to corticosterone blood level changes
It was surprising that -despite obvious differences in the blood CORT level between GR cKO, MRGR cKO, and their littermate controls -we could not show a correlation between auditory nerve processing and the individual blood CORT level ( Figure 5). As blood CORT is a central determinant of MR and GR function, we discuss reasons for this apparent disparity in depth as follows: (i) Blood samples were taken under exactly the same conditions to avoid CORT changes. Thus, differences in GR and MR cKO CORT blood levels were not because of influences of, e.g., anesthesia level, which can dramatically change CORT (see Arnold and Langhans, 2010). We further made sure that samples were always collected by the same experimenter during the same time of the day, to minimize circadian variation of blood CORT (see Atkinson and Waddell, 1997).
(ii) Although effect sizes in male and female MR cKO, GR cKO, and MRGR cKO mice and their respective control groups were almost identical, we cannot exclude gender-related effects of blood CORT on auditory nerve function.
(iii) As differences in the blood CORT level were only observed in GR cKO and MRGR cKO, but not MR cKO mice, significant effects of central MR on auditory nerve responses are unlikely to be controlled by plasma CORT levels.
(iv) High levels of basal CORT reportedly impaired auditory nerve responses (Singer et al., 2018). In the present study, GR cKO mice had the largest auditory nerve responses, but at the same time the iScience Article highest levels of basal blood CORT. In this group of animals, we saw an opposite relation between mean CORT and ABR wave I amplitude, which is another hint that CORT itself is not responsible for the reduced inhibition of auditory nerve sensitivity.
However, in many disorders, HPA-axis activity abnormalities are not evident in baseline blood plasma CORT levels, but are observed in the presence of an acute stressor. Further studies are required to test the extent to which stress levels in MR, GR, or MRGR cKO are different under acute stress and if such changes affect the processing of sound. So far, our data allow us to conclude that the changes in peripheral auditory processing and IHC ribbon synapses in MR cKO, GR cKO, and MRGR cKO are not directly related to changes in the blood CORT level. Mechanisms underlying this top-down signaling from central MR/GR to the auditory periphery remain, however, unclear at present.
Besides CORT itself, other hormones influence the body's responses after exposure to stressful events. Corticotropin releasing hormone (CRH) is highly expressed and widely distributed in neurons of the CNS. A large amount of CRH is synthesized in the paraventricular nucleus, and its release stimulates the pituitary gland to produce and secrete adrenocorticotropin (Richard and Lopez, 2013). Beside its modulation of the HPA-axis, CRH is responsible for food intake and energy expenditure (Richard and Lopez, 2013), and in several species CRH also controls the HPA-axis by inducing the secretion of thyroid-stimulating hormone (De Groef et al., 2006) and thereby possibly plays a key role in the endocrine regulation of life-stage transitions. Effects of CRH on ABR wave I are unclear, although it has previously been shown that CRH plays a role in cochlear hearing sensitivity and noise vulnerability (Vetter, 2015;Graham et al., 2010Graham et al., , 2011 and is expressed during the development of hair-cell innervation (Graham and Vetter, 2011). Further quantification of CRH blood level and its local expression in the cochlea should help clarify the role of CRH in ANF signaling in MR cKO, GR cKO, and MRGR cKO mice.
Previous studies observed limbic MR/GR activation in response to locally synthesized CORT in hippocampal cells, independently of GC produced by adrenal glands (Croft et al., 2008;Taves et al., 2011;McCann et al., 2021). Moreover, local MR activation has previously been shown to alter, e.g., social behavior or behavioral responses to novelty (McCann et al., 2021). CA2 pyramidal cells are further distinguished from neighboring CA1 and CA3 pyramidal cells in that they exhibit a unique pattern of gene expression that permits tight regulation of synaptic plasticity of Schaffer's collateral synapses (McCann et al., 2021), and confers sensitivity to the social neuropeptides oxytocin and vasopressin (Pagani et al., 2015;Raam et al., 2017). Vasopressin and oxytocin are expressed not only in the limbic forebrain (Cilz et al., 2019) but also in the cochlea (Reuss et al., 2009). Oxytocin also stimulates soluble guanylate cyclase and increases intracellular cGMP (Porzionato et al., 2010;Conrad et al., 1993), the signaling cascade that was shown to act on auditory nerve processing (Chumak et al., 2016;Marchetta et al., 2020Marchetta et al., , 2021. Alternatively, MR and GR forebrain activities may exhibit retrocochlear top-down effects to the periphery through e.g., the olivocochlear efferent feedback system, which has been shown to be activated by selective attention (for review see (Lopez-Poveda, 2018)). Here, one very important modulator is the lateral olivocochlear system (Guinan, 2006), where among other neurotransmitters dopamine plays a crucial role (for review see (Lopez-Poveda, 2018)). In the cochlea, dopamine acts at a postsynaptic level through axodendritic auditory nerve terminal, where it tonically decreased CAP wave I amplitude and prolonged latency when the cochlea was perfused with dopamine (Ruel et al., 2001). It is thus challenging to consider a contrasting modification of dopamine mediated tonic inhibition as a target of either GR or MR triggered retrocochlear feedback.

Role of central/limbic MR in perceptual auditory object formation
In conclusion, we suggest that MR in the limbic brain or GR in central brain regions, both shown to optimize stress-coping (de Kloet et al., 2019), control top-down signaling to the auditory system by improving or weakening IHC synapse processing.
Numerous studies that describe positive effects of acute or low CORT levels on hearing (Meltser and Canlon, 2011;Canlon et al., 2013) should be reconsidered in the light of our present findings on limbic MR activities. This includes the acute and subchronic administration of hydrocortisone, which was shown to transiently enhance the amplitude of auditory evoked potentials in normal subjects ( iScience Article includes acute restraint stress, which was observed to improve sound-induced responses in the auditory cortex (Ma et al., 2015), and finally low-dose CORT effects that enhanced the amplitude of auditory evoked potentials recorded from electrodes placed in the CA3 region of the hippocampus (Maxwell et al., 2006). In all cases, limbic MR activation explains these findings because of its direct impact on peripheral auditory processing. Limbic MR activities might be a most attractive key top-down signature linking limbic and auditory neural circuits that are causatively involved in improved hearing, speech discrimination, or communications skills (Jauset-Berrocal and Soria-Urios, 2018; Sihvonen et al., 2017;Sinha et al., 2011;Schaffert et al., 2019;Micheyl et al., 2006). A reduced expression of MR could be linked to neurodevelopmental disorders, such as autism spectrum disorder (Patel et al., 2016) and a human genome mutation leading to a stop-gain alteration of MR protein was found in three brothers with autism (Cukier et al., 2020). Autism spectrum disorder has been hypothesized to be linked with auditory processing deficits in humans ( , and the increasing evidence suggesting a role of high-SR ANF processing deficits in tinnitus (Knipper et al., 2013), must be adapted to reflect the herein described limbic GR effects on cochlear function. Interestingly it was shown that patients suffering from a hyperfunctioning pituitary tumor (Cushing Disease) have an increased risk for hearing impairments as comorbidities (Kuan et al., 2017). On the other hand, adrenal cortical insufficiency led to lower hearing thresholds and higher hearing sensitivity in patients, such as in our GR cKO mice (Henkin et al., 1967).
Within this context, the present findings should guide consideration of limbic MR and GR effects on auditory processing, possibly being one of the enigmatic contributors to perceptual auditory object recognition. Although all acoustic input in the environment is detected in multiple ''streams,'' attention can be laid on each of the streams selectively and is used when a person follows e.g., a musical instrument in the middle of an orchestra (Pressnitzer et al., 2008). Such streaming allows the suppression of responses to unimportant auditory cues in a sound mixture, and thereby influences auditory perception, understanding, and behavioral responses in hearing function in an everyday setting. For object recognition, natural auditory environments or ''scenes'' require listening to sounds at different time points and frequencies to match the incoming auditory input with stored central information to, e.g., isolate and match voices in a crowded environment to previously memorized information and thus recognize the person (Cope et al., 2017). This process of auditory perception has been found to use ''precognitive'' subcortical processing information (Michie et al., 2016; Perez-Gonzalez and Malmierca, 2014; Antunes and Malmierca, 2021), possible tuned as low as the cochlea (Pressnitzer et al., 2008). The key signature that bridges the ''central hearing'' and the ''peripheral hearing'' to extract the information in ''scenes'' is currently missing (Johnson et al., 2021). As MR plays a role in selective attention (Cornelisse et al., 2011), we may hypothesize that recruitment of limbic MR activities could be a possible candidate mediator to bridge central processing with peripheral processing during the streaming process.
Overall, the roles that MR activation is predicted to play in attention, decision-making, and empathy (Wingenfeld and Otte, 2019; Joels, 2018;Chumak et al., 2016), and that GR activation is predicted to play in memory deficits, cognitive decline, and psychopathologies including Alzheimer's disease (Finsterwald and Alberini, 2014;Ouanes and Popp, 2019;Johnson et al., 2021), make MR and GR most attractive candidates for positive and negative ''precognitive'' cochlear processing during auditory perception and auditory cognitive dysfunction (Johnson et al., 2021).

Limitations of the study
The present study demonstrates that central/limbic deletion of MR and/or GR leads to changes in peripheral auditory processing and IHC ribbon synapses.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

DECLARATION OF INTERESTS
The authors declare no competing interests.

Lead contact
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Marlies Knipper (marlies.knipper@uni-tuebingen.de).

Materials availability
This study did not generate new, unique reagents.

Data and code availability
All data reported in this paper will be shared by the lead contact upon request.
This paper does not report original code.
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

EXPERIMENTAL MODEL AND SUBJECT DETAILS
In the present study, three tamoxifen-inducible conditional knock-out mouse lines were studied, in which MR and GR or either MR or GR are deleted, mainly in the forebrain. To generate the MRGR cKO, MR cKO and GR cKO mice and corresponding control animals, three different mouse lines were used. We received from Prof. Gü nther Schü tz (DKFZ, Molecular Biology of the Cell I, Heidelberg, Germany) homozygous floxed MR and GR mouse lines (Berger et al., 2006;Erdmann et al., 2007), in which the exon 3 of either Mr or Gr is flanked by loxP sites. These lines were crossed in our laboratory to obtain a homozygous MRGRflox line. These three lines were than bred with a CaMKIIa Cre ERT2 line (Erdmann et al., 2007) (kindly provided by Prof. Gü nther Schü tz) in which the Cre-Recombinase is expressed under the CaMKIIa promoter after Tamoxifen (TMX) injection. In brief, after confirmation of a normal hearing function, mice received an intraperitoneal injection of 1 mg TMX in 100 ml TMX-solution (Sigma-Aldrich, T-5648, Munich) twice a day on 5 consecutive days at the age of approximately 8 weeks. For the solution, 50 mg TMX was dissolved in 500 ml Ethanol abs. (Merck, Darmstadt) and mixed with 4.5 ml sunflower oil (Sigma-Aldrich, S-5007). After the last injection, the animals were allowed to recover in their home cages for four weeks before the experiments started. The Cre-recombinase leads to the excision of exon 3 of Mr and Gr or either Mr or Gr in cells where CaMKIIa is expressed. To verify the deletion pattern of MR and GR, CaM-KIIaCre ERT2 transgenic mice (Erdmann et al., 2007), were crossed with a Rosa tdTomato reporter mouse line (Madisen et al., 2010), kindly provided by Prof. Hubert Lö wenheim from the Tü bingen Hearing Research Centre (THRC). For all transgenic mouse lines, homozygous floxed Cre-negative littermates that also received TMX were used as controls. For all lines, mice of both sexes aged between 1.8 and 8.4 months were used. The genetic status of all mouse lines was confirmed by genotyping using gene-specific PCR protocols.
Mice were housed in the animal facility of the ENT University Hospital of Tü bingen and had access to water and food pellets ad libitum. They were housed alone or in groups of 2 to 5. Females and single males had a wooden house or tunnel in their cages. The dark-light cycle was 12-12 h, with a light period from 6 am to 6 pm summer time. Humidity was 55 (G5) % and temperature 21.5 (G1) C. iScience Article controlled on every experimental day. The average noise level in the animal facility was below 50 -60 dB SPL.
Animal care, procedures, and experimental protocols corresponded to national and institutional guidelines and were reviewed and approved by the University of Tü bingen Veterinary Care Unit and the Animal Care and Ethics Committee of the regional board of the Federal State Government of Baden-Wü rttemberg, Germany. All experiments were performed according to the European Union Directive for the protection of animals used for experimental and other scientific purposes (2010/63/EU). Mice were kept according to the national guidelines for animal care in a specifically pathogen-free animal facility.

Hearing measurements
Hearing function was studied before and after TMX-induction by measuring

ABR
The anaesthetized mice lay on a pre-warmed resting pad (37 C) in the soundproof chamber. ABR in anesthetized mice were evoked by short-duration sound stimuli with the same stimulus parameters for all groups of KO and control animals. They represent the summed activity of neurons in distinct anatomical structures along the ascending auditory pathway recorded from subcutaneous cranial electrodes. A microphone (Bruel & Kjaer 4191, Naerum, Denmark) was used to calibrate and record the acoustic stimuli. ABR thresholds were elicited with click (100 microsecond duration with an FFT mean of 5.4 kHz), noise-burst (1 ms duration, FFT mean of 7.9 kHz), or pure-tone stimuli (3 ms duration, including 1 ms cosine squared rise and fall envelope, 2-32 kHz). The stimulus level was increased stepwise from 10 to 100 dB SPL in 5 dB steps. Stimuli were generated with an I-O-card (PCI-6052E, PCI-6251, or PCIe-6259, National Instruments, Austin, Texas, USA) in an IBM compatible computer. The SPL of the stimuli was modulated by custom-made amplifier and attenuator systems (Wulf Elektronik, Frankfurt). The measured signals were band-pass filtered from 200 Hz to 5 kHz (F1, 6-pole Butterworth hardware Filter, Wulf Elektronik, Frankfurt) and amplified by 100,000. The analog/digital (A/D) rate was 20 kHz. Each stimulus had a recording interval of 16 ms and was directly repeated and averaged up to 512 times (256 for pure-tone stimuli).

Auditory steady-state responses
The response on amplitude modulation was tested on the ear with the lower click-and noise-evoked threshold directly after finishing the standard ABR protocol, with similar electrode positions. iScience Article depth varying from 0% (unmodulated) to 100% (maximal modulation) and 512 Hz modulation frequency, at 40 dB relative to threshold. For the growth function, the stimuli, modulated 100% with 512 Hz, were presented between 0 and 60 dB relative to threshold in 5 dB steps. For the transfer function, stimuli were modulated with a frequency between 64 and 2048 Hz at 100% modulation depth and 40 dB relative to threshold.

Electrocochleographic recordings
We studied electrical potentials of auditory nerve fibers (ANF) by electrocochleography in living anaesthetized mice. The mice were anesthetized as described above, 20-40 ml Xylocain 2% (AstraZeneca, Wedel, Germany) was applied subcutaneously at sites of surgical incisions and the mice were laid on a pre-warmed resting pad (37 C). The bony auditory bulla was exposed by cutting the skin behind the ear and carefully moving muscles, nerves, and connective tissues beside. A small hole (0.6 mm diameter) was drilled into the bulla, and the round-window niche of the cochlea visualized. A silver wire electrode insulated by varnish and silicone and ending in a small silver bead was placed within the niche. The skin above the ear was closed and the mouse placed in the sound-attenuating booth in front of a loudspeaker for recording. Compound action potential (CAP) threshold responses from the auditory nerves were measured by stimulation with short tone pips (3 ms duration including 1 ms on-and off-ramp cos-square shaped, 32-96 repetitions with stimulus interval 16 ms and alternating polarity) presented with 5 dB 12 incremental steps from 0-100 between 2 and 34 kHz. Electrical potentials were amplified (80 dB) and filtered between 0.2 and 5 kHz before being sampled at 20 kHz A/D rate, averaged, and saved to file. Thresholds were determined from individual ears from averaged waveform responses as the lowest SPL, resulting in a signal visually distinguishable from noise.
For the CAP latency, electrical responses were recorded for 100 ms click stimuli of 0 to 100 dB SPL. Responses were amplified, filtered (DC, 50 kHz low pass), sampled at 100 kHz A/D rate, and averaged for 64 repetitions (ISI 50 ms). For CAP input-output analysis, the averaged waveform was manually inspected for the first negative amplitude deflection after stimulus onset. The latency of the CAP is registered for each stimulus intensity for each individual ear and the resulting growth function averaged and presented as the mean and SEM.

Tissue preparation
The mice were euthanized by exposure to CO 2 . After decapitation the skull was opened and the complete brain was removed. The residual skull was cut in half and the cochleae were removed from the temporal bone under the microscope.
For cochlear whole-mounts the cochleae were isolated, fixed by immersion in 4% paraformaldehyde, 125 mM sucrose in 100 mM phosphate buffered saline (pH 7.4) for 15 min on ice and then dissected for whole-mounts. For cochlear cross-sections, the cochleae were fixated in 2% paraformaldehyde for 2h at 4 C on a rotating wheel, decalcified until the bone was soft and stored in Sucrose-Hank's solution (4 C) over night. Afterwards the cochleae were embedded in Tissue-Tek, frozen at -80 C and sliced in 10 mm sections, using a Cryostat (Leica Cryostat 1720 Digital Leica, Wetzlar, Germany).
Brains were fixed by immersion for 48 h in 2 % paraformaldehyde (exchange of fixative solution after 24 h) and then stored in 0.4 % paraformaldehyde until embedded in 4% agarose. Brains were cut in 60 mm slices with a vibratome (Leica VT 1000S) and stored at -20 C in cryoprotectant (mix 150 g of sucrose in 200 ml 13 phosphate buffer saline (PBS) and 150 ml ethylene glycol) until used for immunohistochemistry.

Immunohistochemistry and ribbon counting
Object slides with cochlear sections or whole-mount preparations were thaw at room temperature for 30 min. After permeabilization for 10 min, the tissue was rinsed with PBS and blocking solution was given to each slice for 30 min. The primary antibody was diluted in reaction buffer and applied to the object slides. After incubation overnight at 4 C and 3x washing in PBS, the secondary antibody, which was diluted in reaction buffer was pipet and incubated for 1h at room temperature. After 3x rinsing in PBS, the cochlear tissue was covered using Vectashield mounting medium with DAPI.
For free-floating brain immunohistochemistry, brain slices were taken out of the cryosolution and transferred into 13 PBS. Corticosterone (CORT) concentration was measured in venous blood samples that were collected into heparin-coatedmicrovettes (Sarsted) directly after the onset of anesthesia for hearing measurements (i.e. < 5 min after handling and injection) in the time window between 9 and 12 am. Each blood sample was around 50 -70ml. After collection, the sample was centrifuged at 1800 x g for 5 min and the plasma was pipetted to 1.5 ml Eppendorf tubes and stored at -80 C. The analysis was assessed by using CORT ELISA kit (Catalog Nr. ADI-901-097) from Enzo Life Sciences Inc. (Farmingdale, NY, USA), following the manufacturers protocol. The plates were read with an optima FLUOstar microplate reader at 405 nm.

QUANTIFICATION AND STATISTICAL ANALYSIS
All statistical information and n numbers can be found in the results section and in Table 1. In figures, significance and a trend for significance is indicated by asterisks ((*) p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001). n.s. denotes non-significant results (p R 0.05). The p-values of the 2-way ANOVAs refer to the main effect of the genotype.

Hearing measurements DPOAE
For analysis, the respective thresholds for the single frequencies were identified. The criteria for reaching the threshold, were a signal level above -15 dB SPL, 5 dB above noise level, and as part of an increasing function. Data were presented as the mean and SEM and analyzed by 2-way ANOVA with a = 0.05, followed by Bonferroni post hoc test (GraphPad Prism).

ABR
The threshold for all click, noise, and pure-tone ABR measurements were manually defined as the lowest sound level at which a clear signal could be discriminated from the baseline. Data are shown as the mean G SEM. Click-and noise-evoked ABR measurements were compared between genotypes by an iScience Article unpaired Student's t-test with a = 0.05. F-ABR measurements were group analyzed by 2-way ANOVA with a = 0.05 followed by Bonferroni post hoc test (GraphPad Prism).
Supra-threshold click-evoked ABR waveforms were analyzed for consecutive amplitude deflections (waves), with each wave consisting of a starting negative (n) peak and the following positive (p) peak. Two peak classes were selected: (1) early peaks (wave I: In À Ip at 1.2-1.8 ms) interpreted as the sum of the first stimulus-related action potential within the auditory nerve, and (2) late peaks (wave IV: IVn À IVp at 4.1-4.9 ms), the response from the auditory midbrain. All ABR wave supra-threshold amplitude growth functions were calculated for increasing stimulus levels with reference to the ABR thresholds (from 0 to a maximum of 80 dB above threshold). For Figure 5 the ABR response amplitude size was calculated by averaging the three maximal amplitude values of the individual ears' supra-threshold amplitude growth  iScience Article function. In Figure 5 for each animal the mean of both ears' ABR wave I size was correlated with the CORT level of the animal. Data were shown as the mean G SEM and analyzed by 2-way ANOVA with a = 0.05 followed by Bonferroni post hoc test (GraphPad Prism).

ASSR
For analysis, a fast Fourier transform (FFT) of the response was calculated. From the FFT, the spectral amplitude at the modulation frequency was extracted, along with the first five harmonics. Additionally, from the FFT, the noise level from the neighboring frequency spectral amplitude (G 4 Hz) was extracted. From this, the signal above noise (mV) was calculated. Data were shown as the mean G SEM and analyzed by 2-way ANOVA with a = 0.05 followed by Bonferroni post hoc test (GraphPad Prism).

Electrocochleographic recordings
Thresholds were determined from individual ears from averaged waveform responses as the lowest SPL resulting in a signal visually distinguishable from noise. Data were analyzed by 2-way ANOVA with a = 0.05 (GraphPad Prism). For CAP input-output analysis, the averaged waveform was manually inspected for the first negative amplitude deflection after stimulus onset. The latency of the CAP was registered for each stimulus intensity for each individual ear, and the resulting growth function averaged and presented as the mean and SEM. Data were analyzed by 2-way ANOVA with a = 0.05 followed by Bonferroni post hoc test (GraphPad Prism).

Ribbon counting
Ribbons are shown as average ribbon number per IHC G SEM. Statistical analysis was performed using 2way ANOVA with a = 0.05 followed by Bonferroni post hoc test (GraphPad Prism). The amount of pillar/ modiolar ribbons was presented as the mean and SD and analyzed by unpaired Student's t-test with a = 0.05 (GraphPad Prism).

Corticosterone analysis
Calculation of the CORT levels were performed using an online data analysis tool (myassays.com). Finally, CORT levels were averaged per genotype and presented as the mean G SEM. Statistical analysis was performed by unpaired Student's t-test with a = 0.05 (GraphPad Prism).