Estrogen inhibits the apoptosis of spiral ganglion cells in C57BL/6J mice with presbycusis by downregulating AQP4.

Estrogen has a protective effect on age-related hearing loss (ARHL), but the mechanism has not been elucidated. Cochlear spiral ganglion neurons (SGN) are an important link between the cochlear hair cells and the auditory center. This study used C57BL/6J mice as models of elderly individuals to observe the protective effect of estrogen against cochlear spiral ganglion cell apoptosis. immunosorbent (HE) observe the morphological changes in the and was used to observe apoptosis of The expression of AQP4 was observed by immunouorescence, and the expression levels of AQP4, Caspase-3, Bax and Bcl-2 in the cochlea were determined by qRT-PCR. Cell experiments: Primary of a (SGC), DMSO Caspase-3, Bax and and and Bax mRNA in the spiral ganglion of elderly mice increased while the levels of the anti-apoptosis factor Bcl-2 decreased. These results indicated that the endogenous apoptosis pathway was involved in the loss of the spiral ganglion during cochlear aging. Moreover, in ovariectomized mice, the mRNA levels of the apoptosis proteins were signicantly upregulated, and the mRNA levels of the anti-apoptosis proteins was signicantly reduced. In the cell experiment, estrogen downregulated the increased mRNA levels of the apoptosis proteins caused by D-gal treatment. These results suggest that estrogen can inhibit the apoptosis of the spiral ganglion during aging.


Page 3/27
Background Presbycusis, also known as age-related hearing loss (ARHL), is the most common sensory dysfunction in the elderly population (Nienke C Homans et al., 2016). While ARHL seriously affects the quality of life of elderly people, it also threatens the physical and psychological health of this population (Ciorba A et al., 2015). Neurogenic ARHL is a kind of deafness characterized by progressive degeneration of spiral ganglion neurons in the cochlea and high-frequency hearing loss. Neurogenic ARHL is the most common form of deafness in the elderly population (Mcfadden Sl et al., 1999;Bao J et al., 2010). Spiral ganglion neurons are an important junction between cochlear hair cells and the auditory center. Therefore, the agerelated degeneration of SGN in the peripheral nervous system is the main reason for the occurrence and development of ARHL (Tang X et al., 2014).
Aquaporins (AQPs) are a family of membrane proteins that can transport water molecules (Ishibashi K et al., 2009). The imbalance of water homeostasis in the inner ear is closely related to inner ear disease, and AQP4 is widely distributed in the inner ear. Studies have shown that AQP4 can maintain endolymphatic homeostasis by regulating cellular osmotic pressure and K + concentrations (Li J et al.,2001). In addition, mice with AQP4 knocked out showed varying degrees of hearing loss (Eckhard A et al., 2012). Nathan Christensen et al. found that the content of AQP4 in the cochlea of CBA mice gradually decreased with age, and a study suggested that hearing loss may be related to the expression of AQP4 (Christensen N et al., 2009). These results suggest that the expression of AQP4 is closely related to hearing.
Estrogen is a steroid sex hormone, and studies have shown that it has a protective effect on hearing (Hill Ra et al., 2009). Studies have shown that the hearing threshold of women continues to decline after menopause, and the most signi cant decline occurs in the two years before and after menopause (K. Charitidi et al., 2010). Exogenous estrogen can reduce hearing loss in perimenopausal women (Zárate S et al., 2017). These results indicate that estrogen has a protective effect on ARHL, but how to exploit its protective effect is not yet clear. Estrogen can reduce nerve injury in the cerebral cortex of rats by regulating AQP4 (Shin Ja et al., 2011), and SGN apoptosis is an important reason for ARHL. Whether estrogen can inhibit the apoptosis of cochlear SGN by regulating AQP4 has not been reported.
In this study, C57BL/6 female mice were used as a model of ARHL. We performed ovarian removal and exogenous estrogen administration to observe the effect of estrogen on SGN cell apoptosis during ARHL.
We used D-gal to treat primary cultured spiral ganglion cells (SGCs) to establish a model of aging to further explore the protective effect of estrogen on ARHL at the cellular level.

Methods
Animal preparation and experimental groups C57BL/6J female mice (25-35 g) without ear disease were provided by the Beijing Weitonglihua Animal Center. The 50 mice were randomly divided into groups. The young group included 3-month-old mice (3 m, n=10). The elderly group included 12-month-old mice that were randomly assigned to the following four subgroups (n = 10 for each group): (1) the sham group (12 m); (2) the ovx group that underwent bilateral ovariectomy at the age of 9 months (12 m ovx); (3) the 12 months ovariectomy +1 month estrogen treatment group that underwent bilateral ovariectomy at the age of 9 months and, after 8 weeks, received estrogen (100 μg/kg/day) (Fabiola Olivieri et al., 2014) subcutaneously for 4 weeks until 12 months of age (E2 1 m); and (4) the 12 months ovariectomy +2 months estrogen treatment group that underwent bilateral ovariectomy at the age of 9 months, and after 4 weeks, received estrogen (100 μg/kg/day) subcutaneously for 8 weeks until 12 months of age (E2 2m). All the experimental procedures were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 80-23, revised 1996) and approved by the Committee of Animal Experimental Ethics of The First A liated Hospital of Medical College, Shihezi University (Shihezi, China) (No. A2018-160-01).

Auditory brainstem response (ABR) threshold test
The mice were anesthetized by intraperitoneal injection of 2% sodium pentobarbital and then placed in a sound screen room. A recording electrode was placed subcutaneously in the middle of the skull, a reference electrode was placed in the bilateral mastoid process, and a grounding electrode was placed at the tip of the nose. Both sides were given short acoustic stimuli with an intensity of 10~90 dB nHL and an interval of 20 db nHL. The scanning time was 10 s, the superposition was 1024 times, the stimulus interval was 11.10 times/s, and the V wave was used as the standard to determine the response threshold.
Measurement of estrogen levels by enzyme-linked immunosorbent assay (ELISA) Blood samples were taken from each group and coagulated in warm water for approximately 15 min. The samples were centrifuged for 20 min to separate and collect the serum, and then, the estrogen levels were measured following the instructions of the estrogen detection kit (Clone Cloud, Wuhan China).

Cochlear histology
The mice were anesthetized with 2% sodium pentobarbital, and then, a thoracotomy was performed. The mice were perfused with saline and 4% paraformaldehyde (PFA) for xation; then, the mouse necks were broken, and the bilateral cochlea were rapidly removed. The cochlea was xed with 4% PFA and incubated for 24 h at 4 ℃. Then, para n embedding was performed after decalci cation with 10% ethylenediaminetetraacetic acid (EDTA) for 1 week. Continuous sections were cut parallel to the cochlear axis with thicknesses of 4 μm. Then, the sections were stained with hematoxylin and eosin (HE) and observed under a microscope (Olympus BX71, Japan).
Terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) staining After dewaxing the para n sections, they were boiled in citrate solution for 15 min. For analysis with an in situ cell death assay kit (Roche Holding, Basel, Switzerland), the cochlear sections were incubated in the dark with a mixture of labeled nucleotides and terminal nucleotide transferases (37°C, 1 h). The sections were then observed with a Olympus laser confocal microscope (Zeiss LSM 510 META, Carl Zeiss AG), and the images were analyzed by Image-pro Plus 6.0 software.

Transmission electron microscopy (TEM)
The cochlea was exposed to 2.5% glutaraldehyde overnight at 4°C. The cells were decalci ed in 10% EDTA for 1 week and xed in 1% osmium tetroxide at room temperature for 2 h. The samples were immersed in an acetone/epoxy 618 mixture and then embedded in epoxy 618. The embedded material was sliced into ultrathin sections (50 nm) and then stained with uranium acetate and lead copper citrate. The ultrastructure of the stained sections was observed by transmission electron microscopy (FEI Tecnai G2 spirit, USA).

Immuno uorescence
The para n sections were dewaxed and boiled in a citrate solution (Sigma-Aldrich) to repair the antigens.
The sections were permeabilized in 1% Triton-X100 and sealed with a 5% BSA (Sigma-Aldrich) solution for 1 h. The primary antibodies were diluted with a 5% BSA solution, and the sections were incubated with these antibodies at 4°C for 12 h. The antibodies included rabbit anti-AQP4 (1:200 Abcam, ab46182) and mouse anti-TUJ1 (1:200, Abcam, ab78078). After washing with PBS (Sigma-Aldrich, USA) three times, the sections were incubated with the secondary antibodies (diluted with 5% BSA solution) at room temperature without light for 1 h. The secondary antibodies included goat anti-rabbit or anti-mouse secondary antibodies (1:100; Beijing Fir Jinqiao Biotechnology Co. Beijing, China). After staining with 4,6diamino-2-phenyl indole (DAPI; Solarbio), the sections were sealed with an anti-uorescence quenching agent, and the uorescence intensity was observed by the LCSM method. The average uorescence intensity (MFI) was used to present the protein expression level. The reaction conditions of the qRT-PCR were as follows: predenaturation at 95°C for 2 min, denaturation at 95°C for 5 s, and annealing at 60°C for 10 s, and this cycle was repeated 40 times.

SGC primary culture and identi cation
Newborn C57BL/6J suckling mice (0-3 days old) were selected, and 75% alcohol immersion disinfection was performed for 3~5 min. Anesthesia was followed by decapitation, and then the bilateral temporal bone was separated and immersed in prechilled D-Hanks solution (containing 100 U/ml penicillin +100 U/ml streptomycin) at 4°C. The auditory vesicles were fully exposed under an anatomical microscope.
The membranous labyrinth, helical ligament and stripped basement membrane were removed. The tissue of the cochlear axis spiral tube containing the spiral ganglion neurons was rapidly transferred into a 1-ml centrifuge tube containing 0.125% type collagenase and 0.125% trypsin. The centrifuge tube was incubated in a 37°C incubator for 20 min, and cell culture medium (90% DMEM/F12 (No. C11330500, Gibco, Thermo Fisher Scienti c, China)+10% FBS (No. 04-001-A; Biological Industries, Cornwell, CT, USA)+100 u/ml penicillin + streptomycin) was added to stop the digestion. The cell suspension was centrifuged, and the supernatant was discarded. The cell culture medium was added to resuspend the SGCs. Cell counts were performed under an optical microscope, and the cells were plated at a density of 1×10 6 cells/ml on a cover glass or 6-well plates precoated with 0.05% polylysine. The culture medium was changed to B-27 (Gibco, Grand Island, NY, USA) neural base medium (90% DMEM/F12+10% b-27+100 u/ml penicillin + streptomycin) after the cells attached to the well. Approximately 50% of the medium was changed every 3 days. After 7 days, the neurons were treated with D-gal (V900922-100G; Sigma-Aldrich), which was dissolved in serum-free 10% B-27-supplemented neurobasal medium. The cells were cultured for 7 days, and 50% of the medium was changed every 3 days during this period. The medium was replaced after adding the AQP4 inhibitor TGN020 (100 nM) (Olivieri F et al., 2014) (Sigma-Aldrich, SML0136, St. Louis, MO, USA) to the pretreated neurons with D-gal. Estrogen (10 nM) (Wu Sy et al., 2016) (Duenas M et al., 1996) was added to the cells in petri dishes and incubated for 24 h. All the neuron cells were cultured in an incubator with 5% carbon dioxide at 37°C.
Cell identi cation: After the primary SGCs were cultured, the medium was removed from the petri dish, and 4% paraformaldehyde was added for xation. The cells were permeabilized with 0.3% Triton X-100 for 15 min and then blocked with 5% BSA. The primary antibody against the TUJ-1 protein (1:100) was added, and the cells were placed in a wet box and incubated at 4°C overnight. The cells were then incubated with a FITC-conjugated secondary antibody (1:100) for 1 h at room temperature in the dark. The nuclei were stained with DAPI, and an anti-uorescence quenching agent was added. Then, the cells were examined under a uorescence microscope to collect the images.

Cell viability and cell senescence tests
The cells were cultured in a 96-well microwell plate for 7 days, and D-gal was added at different concentrations for 7 days of intervention. The cell viability was then measured by the cell counting kit-8 (CCK-8, Dojindo Laboratories, Kumamoto, Japan). The cellular senescence was evaluated by means of a β-galactosidase staining kit (No. C0602; Beyotime Biotechnology, Haimen, China).

Flow cytometry
An Annexin V-FITC/PI double-staining apoptosis kit was used to detect the apoptosis of cultured neurons. The cells were rinsed twice with cold PBS and incubated with EDTA-free 0.25% trypsin at 37°C until 80% of the cells were detached. The cell suspensions were collected, centrifuged at 1000 rpm for 5 min. Then, the supernatant was discarded, and the cells were washed twice with PBS. The cells were resuspended in 500 µl of 1× binding buffer containing 10 µl of annexin V-FITC and 5 µl PI. The cells were incubated at room temperature in the dark for 15 min and then analyzed by ow cytometry.
Statistical analysis SPSS 19.0 software was used for the statistical analyses, and GraphPad Prism 6.01 software was used to generate the graphs. All the data are expressed as the means ± standard deviations. One-way ANOVA was used for comparisons between groups, LSD-t tests were used for pairwise comparisons between groups, multivariate analysis of variance was used to compare multiple groups at different time points. A p-value of less than 0.05 was considered to be statistically signi cant.

Results
Effects of estrogen on hearing in mice ( Fig. 1) Estrogen levels in the mice in each group Figure 1A shows a schematic diagram of estrogen administration. The ELISA results showed that there was no signi cant difference in the estrogen levels between the 3 m group and the 12 m group. Compared with those in the 12 m group, the estrogen levels in the 12 m ovx group decreased signi cantly (P<0.001). Compared with those in the 12 m ovx group, the estrogen levels in the E2 1 m and E2 2 m groups were signi cantly increased (P<0.001). ( Figure 1B).
ABR detected changes in the hearing threshold in each group The auditory brain stem response (ABR) was used to assess the hearing of the C57BL/6 J mice. Within an interval of 10 dB nHL, the sound stimulus intensity gradually decreased from 90 dB nHL to 10 dB nHL, and then, we recorded the changes in the V waves in the acoustic stimulus threshold of each group. The results showed that the ABR threshold increased with age, and compared with that of the 3 m group, the threshold of the 12 m group was signi cantly increased (P<0.001). After ovariectomy, the threshold decreased compared with that of normal mice. Compared with that in the 12 m group, the threshold in the 12 m OVX group was increased (P<0.05). After estrogen injection, the thresholds of the E2 1 m group and the E2 2 m group were signi cantly lower than those of the 12 m OVX group (P<0.05) ( Figure 1C).
Protective effect of estrogen on spiral ganglion neurons in mouse cochlea (Fig. 2) Histopathological changes of the cochlea SGN The left cochlea was harvested from each group, and the histopathological changes in the SGN were observed by HE staining. As shown in Figure 2A, the SGN structure in the 3 m group was clear, the abundant nerve cells were regularly arranged, and the ratio of nuclear pulp was normal ( Figure 2Aa). The number of nerve cells in the 12 m group decreased, and the ratio of the nuclear pulp was altered ( Figure  2Ab). The SGN cells in the 12 m ovx group were loose in structure, irregular in shape and disordered in arrangement, nuclear condensation, and chromatin edging (Figure 2Ac). In contrast, the E2 1 m group ( Figure 2Ad) and E2 2 m group (Figure 2Ae) exhibited different degrees of reduced nerve cell injury, and the number of injured nerve cells with relatively complete cell structure was reduced.

Cochlea SGN injury in each group
The SGN ultrastructure was observed by transmission electron microscopy to evaluate neuronal damage ( Figure 2B). The results showed that in the 3 m group, the cell structure was basically intact, the cytoplasm was slightly dissolved, the mitochondrial arrangement was orderly, and the chromatin was basically normal (Figure 2Ba). In the 12 m group, the cytoplasmic vacuoles were obvious, the mitochondrial cristae were dissolved, and the euchromatin was dissolved in the nucleus (Figure 2Bb). In the 12 m ovx group, the cytoplasm was obviously vacuolized, the numbers of mitochondria were reduced, the nuclear membrane was pitted and dissoluted, and the euchromatin and heterochromatin were absent ( Figure 2Bc). In the E2 1 m group, cytoplasmic vacuoles, mitochondrial cristolysis, organelle depletion, nuclear chromatin dissolution, and a relatively intact nuclear membrane were observed (Figure 2Bd). In the E2 2 m group, cytoplasmic vacuoles were formed, mitochondrial cristae were dissolved, euchromatin was absent in the nucleus, and the nuclear membrane was complete (Figure 2Be).
Injury to SGN neurons in the cochlea and apoptosis in each group TUNEL staining was used to observe the apoptosis of the helical ganglion cells in each group. Apoptosis was more serious in the 12 m group than in the 3 m group, and apoptosis was increased in the 12 m ovx group compared with the 12 m group (P<0.001). The degree of apoptosis in the E2 1 m group was decreased after estrogen administration compared with that in the 12 m ovx group (P<0.001). Compared with that of the 12 m ovx group, the apoptosis of the E2 2 m group was signi cantly reduced (P<0.001), and the apoptosis of the E2 1 m group was signi cantly improved (P<0.001) ( Figure 2C).
Estrogen downregulated AQP4 and apoptotic proteins (Fig. 3) Effect of estrogen on AQP4 expression The expression of AQP4 in the helical ganglia was observed by immuno uorescence. As shown in Figure  5A, AQP4 was labeled with green uorescence, and TUJ1 was labeled with red uorescence. As shown in Figure 5B, the semiquantitative statistical analysis of AQP4 indicated that compared with that in the 12 m group, AQP4 expression was enhanced in the 12 m ovx group (P<0.001). In addition, AQP4 expression decreased in the E2 1 m group and E2 2 m group (P<0.001) ( Figure 3A).
Expression of AQP4, Caspase3, Bax and bcl-2 mRNA in each group qRT-PCR was used to detect the mRNA levels of Bcl-2 in the spiral ganglion. Compared to those in the 3 m group, the mRNA levels of the apoptotic proteins Caspase3 and Bax in the 12 m group were increased (P<0.05), and the mRNA level of the anti-apoptotic protein Bcl-2 was signi cantly decreased (P<0.01). Compared with those in the 12 m group, the mRNA levels of AQP4, Caspase3 and Bax were signi cantly increased, while the level of Bcl-2 was increased, in the 12 m ovx group (P<0.01). Compared with that in the 12 m ovx group, the apoptotic protein caspase-3 mRNA level was decreased (P<0.01), the Bax level was decreased (P<0.001), and the anti-apoptotic protein Bcl-2 mRNA level was increased (P<0.05).
Compared with those of the 12 m ovx group, the caspase-3 and Bax mRNA levels of the E2 2 m group were decreased (P<0.01), and the mRNA level of the anti-apoptotic protein Bcl-2 was increased (P<0.01). All the above data are statistically signi cant ( Figure 3B, Figure 3C).
SGC primary culture and senescence model preparation (Fig. 4) SGC primary culture and identi cation Primary spiral ganglion cells were cultured and puri ed ( Figure 4A). The spiral ganglionic cells were identi ed with the nerve cell marker TUJ1 (green uorescence), and prokaryotes were labeled with DAPI (blue uorescence) ( Figure 4B).
D-gal treatment of SGCs to establish a senescence model D-gal decreased neuronal activity in a dose-dependent manner. The results were statistically signi cant at a concentration of 10 mg/ml (P < 0.05, Figure 4E). When the concentration of D-gal was greater than 15 mg/ml, the cellular activity was signi cantly decreased, and the D-gal staining results are shown in Figure  4C. In addition, the number of positive cells decreased (P<0.001, Figure 4D). These results suggested that we successfully established a senescent cell model, and thus, we chose 10 mg/ml D-gal to induce neuronal senescence in the subsequent experiments.

Effect of estrogen on AQP4 expression in SGCs
Immuno uorescence was used to observe the expression of AQP4 in each group. AQP4 was labeled green, TUJ1 nerve cells were labeled red, and DAPI nuclei were labeled blue. The AQP4 expression increased after D-gal treatment (P<0.001). After E2 was administered, the AQP4 expression in the D-gal+E2group was decreased in the D-gal group (P<0.001). In addition, the results showed that there was no signi cant difference in AQP4 expression between the DMSO group and SGC group ( Figure 5A).

Apoptosis of SGCs in each group
Cell apoptosis was detected by ow cytometry, and the results showed that there was no signi cant difference in the apoptosis rate between the DMSO group and SGC group. After D-gal treatment, the number of apoptotic cells was increased (P<0,001). Compared with the D-gal group, the D-gal+E2 group showed decreased apoptosis (P<0,001). The D-gal+TGN020 group showed decreased apoptosis compared with the D-gal group (P<0,001). Compared with the D-gal group, the D-gal+E2+TGN020group showed decreased apoptosis (P<0,001) ( Figure 5B).

Expression of AQP4, Caspase3, Bax and bcl-2 mRNA in SGCs of each group
The mRNA levels of AQP4 and the apoptotic proteins Caspase3, Bax, and Bcl-2 in each group were determined by qRT-PCR. The results showed that there was no signi cant difference in the mRNA levels between the DMSO group and the SGC group. Compared with those in the SGC group, the mRNA level of AQP4 was increased (P<0.05), the mRNA level of the apoptotic proteins Caspase3 and Bax were increased (P<0.001), and the mRNA level of the anti-apoptotic protein Bcl-2 was decreased (P<0.001) in the D-gal group. Compared with those in the D-gal group, the mRNA level of AQP4 was decreased (P<0.05), the mRNA levels of Caspase3 and Bax were decreased (P<0.01), and the mRNA level of Bcl-2 was increased (P<0.001) in the D-gal+E2group. Compared with those in the D-gal group, the mRNA level of AQP4 was reduced (P<0.001), the mRNA levels of Caspase3 and Bax were reduced (P<0.05), and the mRNA level of Bcl-2 was increased (P<0.01) in the D-gal+TGN020 group. Compared with those in the Dgal group, the mRNA level of AQP4 was reduced (P<0.001), the mRNA levels of Caspase3 and Bax were decreased (P<0.001), and the mRNA level of Bcl-2 was increased (P<0.001) in the D-gal+E2+TGN020group ( Figure 5C, Figure 5D).

Discussion
ARHL is a common sensory dysfunction in the elderly population. It has been reported that estrogen has a certain protective effect on ARHL, but ways by which to exploit this effect remain unclear. In this experiment, C57BL/6J mice were selected to establish a model of ARHL. By observing the changes in spiral ganglion cellular apoptosis and AQP4 expression in the elderly mice, as well as the effect of estrogen on this change, the protective effect of estrogen against ARHL, which is achieved by regulating AQP4 to inhibit spiral ganglion cell apoptosis, was observed.
C57BL/6J mice are inbred mice with Cdh23 gene mutations, are characterized by progressive hearing loss and precocious puberty, and are the most commonly used model mice for ARHL research (Keithley Em et al., 2004;Someya S et al., 2010). In this experiment, 3-month-old mice were included in the young group, and 12-month-old mice were included in the elderly group. The ABR test showed that the hearing threshold of the 12-month-old mice was signi cantly increased. Some studies have reported that in normal rats, ovariectomy leads to an increase in the ABR threshold, and estrogen replacement can reverse these changes (Coleman Jr et al., 1994). Moreover, estrogen replacement therapy can improve the amplitude and latency of DPOAEs and ABR in mice (Williamson Tt et al., 2019). In our study, after ovariectomy, the in vivo estrogen level was lower and the ABR threshold was higher than it was in normal elderly female mice; however, after exogenous estrogen administration, the in vivo estrogen level was maintained within the normal range, and the ABR threshold was lower than that of the ovariectomized mice. The results are consistent with the experiments mentioned above, suggesting that stable in vivo estrogen levels have a certain protective effect against hearing loss in elderly mice. Sensorineural ARHL is the most common type of ARHL, and it is mainly characterized by the gradual degeneration of cochlear hair cells and spiral ganglion neurons. The necrosis or apoptosis of spiral ganglion cells is one of the important mechanisms of ARHL (Sang L et al., 2017). At present, studies have preliminarily shown that the protective effect of estrogen against hearing loss may be related to the survival of neurons. Estrogen may play a protective role in nerve cells by regulating neurotrophic factors (Coleman Jr et al., 1994) and inhibiting apoptosis (Hill Ra et al., 2004;Bean La et al., 2015). In this experiment, histopathological changes in the spiral ganglion of the cochlea and ultrastructural neuronal damage were observed by HE staining and transmission electron microscopy. The results showed that the spiral ganglion of elderly C57BL/6J mice was obviously missing and that the neurons were damaged. After ovariectomy, the damage to the spiral ganglion in these mice was exacerbated. After estrogen was administered, the loss of the spiral ganglion and the degree of neuronal damage were signi cantly improved. The results suggest that estrogen can reduce the loss of the spiral ganglion and improve damage to neurons during ARHL. In the auditory system, damage to the spiral ganglion and hair cells is irreversible, which leads to permanent sensorineural deafness. At present, there is no effective drug for the prevention or treatment of ARHL, and hearing can only be improved through sound reinforcement by hearing aids or through electrical stimulation by cochlear implants. Cochlear implants can effectively replace the mechanical sensory transduction function of the lost hair cells through direct electrical stimulation of the spiral ganglion, but these devices only work when su cient spiral ganglion is retained (Roehm Pc et al., 2005). Therefore, the protection of the spiral ganglion is very important for the prevention and treatment of ARHL. Studies show that after ERβ is knocked out, the number of cells spiral ganglion cells in 12-monthold mice decreases, and stable estrogen levels may help to maintain the integrity of ERs. The estrogen ligand and ER can be mutually regulated (Hill Ra et al., 2004), thus delaying the degeneration of auditory neurons. This observation is consistent with our results.
Studies have found that activation the apoptosis pathway in the mouse cochlea increases with age, and the age-related absence of spiral ganglion cells may be related to apoptosis. However, the mechanism by which spiral ganglion cell apoptosis participates in the occurrence and development of ARHL is still unclear. Apoptosis is mainly mediated by the exogenous pathway, which is triggered by the death receptor, and the endogenous pathway, which is initiated by mitochondria-and endoplasmic reticulumdependent pathways (Someya S et al., 2010). Estrogen is known to inhibit neuronal apoptosis and increase neuronal survival. We observed apoptosis of the spiral ganglion in the mouse cochlea by TUNEL staining. The apoptosis in the elderly mice was obvious. The ovariectomized mice exhibited more severe apoptosis than the normal elderly female mice. However, after estrogen was administered, the apoptosis decreased. In addition, we further con rmed apoptosis at the cellular level in cell model of aging. D-gal is widely used in experimental research to establish models that simulate aging (Cui X et al., 2006).
Research shows that D-gal can induce aging in neuron cells (Hattori F et al., 2007;Arnér Es et al., 2009) and even lead to apoptosis (Zeng L et al., 2014) by promoting the generation of ROS in cells. We used Dgal to treat spiral ganglion cells to establish a model of aging, and ow cytometry was used to assess apoptosis. The results showed that the apoptosis rate increased after D-gal treatment but decreased after estrogen administration.
To further explore the apoptotic pathway in the spiral ganglion during ARHL, we used qRT-PCR to detect changes in the Caspase-3, Bax, and Bcl-2 mRNA levels related to the endogenous pathway of apoptosis.
Caspase-3 is considered the nal effector protein of apoptosis, and its expression and activation levels will cause cell apoptosis to enter an irreversible stage (Fujimura M et al., 1999). We observed that the levels of Caspase-3 and Bax mRNA in the spiral ganglion of elderly mice increased while the levels of the anti-apoptosis factor Bcl-2 decreased. These results indicated that the endogenous apoptosis pathway was involved in the loss of the spiral ganglion during cochlear aging. Moreover, in ovariectomized mice, the mRNA levels of the apoptosis proteins were signi cantly upregulated, and the mRNA levels of the anti-apoptosis proteins was signi cantly reduced. In the cell experiment, estrogen downregulated the increased mRNA levels of the apoptosis proteins caused by D-gal treatment. These results suggest that estrogen can inhibit the apoptosis of the spiral ganglion during aging.
In the aging cochlea, glutamate excitotoxicity may lead to apoptosis of the spiral ganglion, while estrogen has been proven to inhibit the neuronal damage and apoptosis induced by glutamate excitotoxicity in the central nervous system (Zhao L et al., 2004). In addition, estrogen can reduce gentamicin-induced HC apoptosis by inhibiting the JNK-mediated apoptosis pathway (Nakamagoe M et al., 2010). Estrogen can also inhibit ROS-induced cochlear HC apoptosis by upregulating anti-apoptosis factors. Among them these factors, E2 is known to upregulate Bcl-2 and Bcl-XL in neurons in the central nervous system (Garcia-Segura Lm et al., 1998;Christian, J et al., 1999), and the presence of ERE sequences in these genes suggests that they may be direct targets of ERs (Perillo B et al., 2000). In addition, estrogen can induce Bcl-2 expression by rapidly activating the PI3K/Akt signaling pathway (Yune Ty et al., 2008). These pathways may be potential targets for the estrogen-mediated inhibition of apoptosis in spiral ganglion cells.
AQP4 is one of the most important water-selective channel proteins in the central nervous system. AQP4 is widely distributed in the astrocytes and retinal optic ganglia in the brain (Previch Le et al., 2016), but its expression and function in the spiral ganglion of the cochlea have been rarely studied. The expression of AQP4 in the spiral ganglion of the cochlea of C57BL/6J mice was determined by immuno uorescence using an AQP4 antibody and the nerve cell marker TUJ1. We found that the expression of AQP4 in spiral ganglion cells of the ovariectomized mice was increased. After estrogen was administered, the expression of AQP4 was decreased. In the cell experiments, after D-gal treatment, the expression of AQP4 in the spiral ganglion cells was increased, and after estrogen was administered, the expression was decreased. It is suggested that estrogen can downregulate the expression of AQP4 in the spiral ganglion.
Research shows that estrogen can upregulate AQP4 in the cerebral cortex of rats in a rat model of stroke (Szabo C et al., 2014), while estrogen can also downregulate AQP4 in the brain of rats in a model of traumatic brain injury to achieve protection of the brain . Therefore, the evidence regarding the regulation of aquaporins by estrogen is contradictory, and the differences may depend on the type of injury. The regulation of AQP4 by estrogen can be inhibited by ICI182780, which is a common antagonist of the ERα or ERβ receptors, suggesting that estrogen may regulate AQP4 through ERs, and the activity of AQP4 can be regulated through the activation of protein kinase A, calcium/calmodulin dependent protein kinase and protein kinase C, which are also affected by estrogen (Menardo J et al., 2012).
In the cell experiment, we found that the apoptosis rate of the spiral ganglion cells increased after D-gal treatment, while the apoptosis rate decreased after the AQP4 inhibitor was added, and the qRT-PCR and ow cytometry results suggested reduced apoptosis. We speculated that the increased expression of AQP4 might lead to apoptosis of spiral ganglion cells. In ammatory factors in the cochlea increase during aging (Zhao Cq et al., 2007). In ammatory factors, such as IL-1β and TNF-α, can induce apoptotic cascade reactions, further activating Caspase-3 and promoting cell apoptosis (Cheng W et al., 2013).
Other studies have shown that the downregulation of AQP4 can inhibit in ammatory factors and thus reduce nerve injury (Liu S et al., 2017). This mechanism may be a way for APQ4 to participate in the apoptosis of spiral ganglion cells. Glutamic acid accumulation in the cochlea can cause loss of organelle function and reduced expression of apoptosis factors, thus causing hearing loss (Wang Y et al., 2010).
However, inhibiting the upregulation of AQP4 can reduce the excessive glutamate release caused by ischemia. We assume that changes in AQP4 may cause the accumulation of glutamate, thus affecting the apoptosis of spiral ganglion cells. In this experiment, estrogen downregulated the expression of AQP4 in the spiral ganglion. In a rat model of traumatic brain injury, the upregulation of AQP4 can aggravate neuronal apoptosis in the brain, while estrogen can reduce apoptosis by the downregulation of AQP4 . Therefore, we suggest that estrogen may inhibit the apoptosis of spiral ganglion cells by downregulating AQP4 in the spiral ganglion cells. The speci c interaction between AQP4 and apoptosis and the mechanism by which estrogen regulates AQP4 need further study.

Conclusion
In summary, this experiment found that estrogen de ciency can lead to increased apoptosis of spiral ganglion cells and decreased hearing in C57BL/6J mice during aging and that estrogen treatment can reverse these changes. In the D-gal-induced cell model of aging, estrogen downregulates the expression of AQP4 in the spiral ganglion of the cochlea, thus inhibiting cell apoptosis. The loss and apoptosis of spiral ganglion cells is an irreversible phenomenon. Our research proves that estrogen can inhibit the apoptosis of spiral ganglion cells in elderly mice by downregulating AQP4, thus exerting its protective effect on hearing. This study will provide a new experimental basis for estrogen-mediated hearing protection. However, the mechanism by which estrogen regulates AQP4 still needs further research.