MiR-106a facilitates the sensorineural hearing loss induced by oxidative stress by targeting connexin-43

ABSTRACT Sensorineural hearing loss (SNHL) is a common clinical side effect resulted from the overusing of aminoglycoside antibacterial drugs, such as gentamicin. Oxidative stress is recently evidenced to be an important inducer for SNHL, which is reported to be associated with the knockdown of connexin-43. MiR-106a is recently found as a regulator of connexin-43. The present study aims to investigate whether miR-106a is a vital mediator in the development of SNHL. Firstly, upregulated miR-106a was observed in the peripheral blood sample of SNHL patients. Glucose oxidase (GO) was utilized to induce oxidative injury in isolated rat cochlear marginal cells (MCs), followed by introducing the miR-106a inhibitor. We found that the declined proliferation ability, increased apoptosis, and activated oxidative stress in GO-stimulated MCs were dramatically abolished by the miR-106a inhibitor, accompanied by the upregulation of connexin-43. The targeting correlation between miR-106a and connexin-43 was predicted and confirmed by the dual luciferase gene reporter assay. Furthermore, the regulatory effect of miR-106a inhibitor against the proliferation, apoptosis, and oxidative stress in GO-treated MCs were dramatically abolished by the knockdown of connexin-43. Gentamicin was utilized to establish the SNHL model in rats, followed by the treatments of antagomir-106a and antagomir-106a combined with carbenoxolone, an inhibitor of connexin-43. The alleviated pathological state, reduced apoptosis, and ameliorated oxidative stress in cochlea tissues were observed in antagomir-106a treated SNHL rats, which were dramatically reversed by the co-administration of carbenoxolone. Collectively, miR-106a facilitated the SNHL induced by oxidative stress via targeting connexin-43.


Introduction
Sensorineural hearing loss (SNHL) is one of the most common clinical diseases of ophthalmology and otorhinolaryngology, which is the abnormality of the auditory system to feel and process the sound resulted from various factors [1]. With the acceleration of population aging, hearing loss is the third chronic disease affecting the healthy life of the elderly [2] and the incidence of SNHL is still on the rise [3]. The main effect of adult hearing loss is communication obstruction, even psychological disorders. Currently, the treatment methods are limited and the application of assistive listening devices is an available solution. However, the therapeutic effect of auxiliary devices is associated with the number and role of residual cells in the cochlea [4]. Maintaining sufficient number of normal auditory cells is the prerequisite for keep normal auditory function. It is particularly important to prevent and treat the damage of cochlear cells. The pathogenesis of SNHL is complex and oxidative stress injury has been a hot topic in recent years [5]. Studies have shown that cochlea cells can be protected by antioxidants by preventing reactive oxygen species (ROS) from triggering caspase-mediated apoptosis and protecting mitochondrial membranes [6]. In SNHL animal models established by gentamicin, sharply increased level of ROS is observed in a short period of time, which can not be removed by timely antioxidant enzymes. Subsequently excessive free radicals are accumulated to produce malondialdehyde (MDA), which induces metabolic disorders on glucoses, proteins, and nucleic acids. The damages on cochlear tissues and cellular structures are finally induced to disrupt the normal functions [7,8]. Therefore, improving the activity of antioxidant enzymes, scavenging ROS, alleviating oxidative stress damage, and maintaining redox balance will be effective methods to mitigate the damage of cochlear cells, which is of great significance for the protection of hearing function [9]. MicroRNAs(miRNAs) are a group of RNA molecules with gene regulatory functions that have been discovered in recent years, the length of which is approximately 22 nucleotides [10,11]. Recently, several researches claimed the important role miRNAs play in the pathogenesis of SNHL. It is predicted by Li [12]that by screening differentially expressed miRNAs and mRNAs in samples from SNHL patients and health subjects, genes regulated by miR-34a, miR-548, miR-15a, miR-210, and miR-23a are found to paly critical roles in developing SNHL. Ebeid further claimed the function of miR-183 in the regeneration of sensory hair cells (HCs), the loss of which is an important inducer of SNHL, by targeting and regulating the expression level of Atoh1 [13]. However, rare sufficient experimental evidences for the specific function and regulatory mechanism of miRNAs in SNHL have been provided in recent years. It is reported that the downregulation of protein connexin-43 is closely associated with the pathogenesis of SNHL [14] and connexin-43 is considered as an important inhibitor of oxidative stress [15]. Recently, miR-106a is reported to be a targeting miRNA and regulator of connexin-43 [16].
Therefore, we suspected that miR-106a might be a critical mediator in the development of SNHL by impacting connexin-43. The present study will explore a potential pathological mechanism of SNHL and offer a biomarker for the diagnosis of SNHL by investigating the specific function of miR-106a in mediating glucose oxidase (GO) induced oxidative stress injury in rat cochlear marginal cells (MCs) and pathological symptoms in gentamicininduced SNHL rat model.

Collection of clinical SNHL samples
Peripheral blood samples were obtained from 10 pairs of SNHL patients resulted from overcommit of gentamicin and health subjects in Beijing University of Chinese Traditional Medicine Dongfang College, respectively. All SNHL patients and health subjects were aware of the purpose of sample collection and consented the project orally.

The isolation of rat cochlear MCs [17]
Following performed with inhaled anesthesia, 6 SD rats were administered with 75% ethanol, followed by removing the bilateral temporal bones and dissecting the stria vascularis from the apical to basal. Subsequently, the stria vascularis was cut into 0.5 mm slides, followed by being digested using the 0.1% type II collagenase. After 30 min incubation, cell suspension was centrifugated at 300 g for 5 min and cells were resuspended using the Epithelial Cell Medium-animal (EpiCM-animal, ScienCell, California, USA) under the condition of 37°C and 5% CO 2 . The isolated MCs were identified by determining the expression level of cytokeratin-18 (CK18) using the method of immunofluorescence [18].

Immunofluorescence assay [17]
MCs were collected and fixed utilizing the 4% paraformaldehyde on microslide for 10 min, which were further treated with triton-100 for 15 min. Following incubating with 5% BSA, cells were added with the mouse anti rat CK18 antibody (1:1000, biorbyt, Wuhan, China) at 37°C for 1.5 h, which were further added with the solution diluted with FITC Conjugated goat anti mouse antibody (1:500, Abcam, Cambridge, USA) at 37°C for 45 min. Lastly, the slides were sealed using nail polish after adding the DAPI solution. The fluorescence was observed under the fluorescence microscope (KEYENCE, Tokyo, Japan).

CCK-8 assay [18]
MCs were implanted in the 24-well plate, followed by treated with different strategies for 24 hours. Subsequently, each well was added with the CCK-8 solution to be incubated for 4 hours and the optical density (OD) at 450 nm was determined using the microplate reader (Bio-Chain, Shanghai, China).

Real time PCR [18]
TRIzol (Invitrogen, California, USA) was utilized to extract total RNAs from cells, followed by centrifugating at 16,000 g for 10 min. After being dissolved in ddH 2 O, the concentration of RNAs was quantified by determining the absorption at 260 nm, followed by transcribing RNAs to cDNAs with the reverse transcription kit (Applied Biosystems, Massachusetts, USA). Then, the reaction of RT-PCR was performed with the ABI 7500 Real-time PCR system (Applied Biosystems, Massachusetts, USA) utilizing the SYBR green (Invitrogen, California, USA). The procedure for the reaction was shown as below: 95°C for 1 min, 40 cycles of 95°C for 30 sec, 62°C for 60 sec, 72°C for 30 sec, and 72°C for 90 sec. The 2 −ΔΔCtn method was utilized to measure the expression level of mRNAs, with GAPDH for the normalization.The primers of human miR106a:XXXXXXXXXXXX; rat miR106a: XXXXXXX Analysis on apoptosis by using the flow cytometry [19] 3 × 10 6 cells were collected and the 5× Binding Buffer was diluted to 1× Binding Buffer with double steam water, followed by being used to resuspend cells. After adding 3 μL Annexin V-APC and 5 μL 7-AAD into each tube, cells were slightly mixed and incubated for 10 min. Then, 200 μL pre-cooled 1× Binding Buffer was added to each tube, followed by being loaded onto the ACEA NovoCyteTM (ACEA Biosciences, Hangzhou, China) for analysis.

DCFH-DA assay [20]
In brief, MCs were added with DCFH-DA reagent (ADANTI, Wuhan, China) to be incubated for 10 min and DMEM medium was used to wash the remaining reagent. The release of ROS was quantified with the ImageJ software after visualization utilizing the inverted fluorescence microscope (KEYENCE, Shanghai, China).

MDA, SOD and connexin43 measurement
The production of MDA and SOD in cells and tissues was determined using the commercial MDA and SOD assay kit (Spbio, Wuhan, China) referring to the manufacturer's instruction. The connexin43 in serum of human was detected by ELISA kit (Xinyu, Shanghai, China) referring to the manufacturer's instruction. [21,22] Proteins were obtained from cells and were then quantified with a BCA kit (Absin, Shanghai, China), followed by loaded onto a 12% SDS PAGE. After separating for 60 min, proteins were transferred onto the PVDF membrane, which was further incubated with 5% BSA. Then, the primary antibody against connexin-43 (1:800

Auditory brainstem-evoked response testing (ABR) [25]
The recording electrode was placed in the middle of the parietal skull, while the reference electrode and the ground electrode were placed under the skin of the donor ear and the contralateral ear, respectively. Stimulus parameters and test parameters were set as following: the stimulus sound was set as filtered click; the frequency of band-pass filter was 100-3000 Hz; the times of overlay was 512; the scanning duration was 10 ms; headset output; repetition rate was 10 times/second; the distance to the rat ear canal mouth was 0.5 cm. ABR was determined based on wave III.

H&E staining on cochlear tissues
Cochlear tissues were fixed with 4% paraformaldehyde and dehydrated using different concentrations of ethyl alcohol solution, followed by transparentized with xylene for 15 min. The paraffin was used to embed tissues for 60 min, followed by cutting into slides. After baking, dewaxing, and hydration, sections were dyed with hematoxylin solution for 3 min, which were further incubated with the hydrochloric acid for 15s and 3 washes. Samples were dyed with eosin solution for 3 min, which were washes, dehydration, transparentizing, and sealing. Lastly, images were taken under the inverted microscope (SUNNY OPTICAL, Zhejiang, China).

TUNEL assay
Slides were incubated with the working solution for 15 min at 37°C, followed by adding the 50 µL TUNEL reaction solutions (Solarbio, Beijing, China) to be incubated for 60 min. Then, samples were added with diaminobenzidine (DAB) for 30 minutes. Lastly, slides were visualized using the inverted microscope (SUNNY OPTICAL, Zhejiang, China).

Statistical analysis
Data achieved in the present study were presented as mean ± SD. The Student's t-test was utilized to analyze the difference between 2 groups and the one-way ANOVA method was utilized to analyze the differences among groups. P < 0.05 was taken as a significant difference.

Results
In this study, we explore the potential pathological mechanism of SNHL and offer a biomarker for the diagnosis of SNHL by investigating the specific function of miR-106a in mediating glucose oxidase (GO) induced oxidative stress injury in rat cochlear marginal cells (MCs) and pathological symptoms in gentamicin-induced SNHL rat model. Results showed that miR-106a might be a critical mediator in the development of SNHL by impacting connexin-43.

MiR-106a was significantly upregulated, but connexin43 was decreased in SNHL patients
We firstly measured levels of miR-106a and con-nexin43 in the blood sample collected from health subjects and SNHL patients. We found that a significantly higher levels of miR-106a (Figure 1(a)) were observed in SNHL patients (*p < 0.05), however, the protein of connexin43 (Figure 1(b)) was decreased compared with health subjects(p < 0.01).

The cytotoxicity of GO on isolated MCs
Firstly, the isolated MCs were identified by the positive expression of CK18 using the immunofluorescence assay (Figure 2(a)). The cell viability was further detected after MCs were added with GO (0, 1, 2.5, 5, 10, 20, 40, 80, and 160 mU/mL) to determine the optimized incubation concentration. When the concentration of GO was larger than 20 mU/mL, significantly declined cell viability of MCs was observed (Figure 2(b), **p < 0.01, ***p < 0.001). In the subsequent experiments, 20 mU/mL GO were introduced for the investigations as 20 mU/mL GO was the lowest concentration that induced the declined proliferation in MCs.

MiR-106a inhibitor repressed the apoptosis in GO-treated MCs
To investigate the impact of miR-106a on GOinduced apoptosis in MCs, MCs were treated with 20 mU/mL GO with or without inhibitor NC and miR-106a inhibitor, respectively. The OD values (Figure 3(a)) was declined in MCs by the stimulation of GO, which was greatly elevated by miR-106a inhibitor. The apoptotic rate of MCs (Figure 3(b)) was promoted from 7.1% to 39.3% by GO, which was slight changed to 40.2% by inhibitor NC and declined to 17.9% by miR-106a inhibitor. Additionally, Bax and caspase-3 (Figure 3(c)) were dramatically upregulated and Bcl-2 was greatly downregulated in the GO group and GO+ inhibitor NC group, which were significantly reversed by miR-106a inhibitor (**p < 0.01 vs. control, ##p < 0.01 vs. GO+ inhibitor NC). These data collectively implied that apoptosis in GO-treated MCs was alleviated by the miR-106a inhibitor.

MiR-106a inhibitor mitigated oxidative stress in GO-treated MCs
We further explored the influence of miR-106a on oxidative stress in MCs induced by GO. Significantly higher ROS level (Figure 3(d)) was observed in the GO group and GO+ inhibitor NC group, which was greatly repressed by the introduction of miR-106a inhibitor. Additionally, the concentration of MDA was dramatically elevated and that of SOD was greatly declined by the stimulation of GO with or without inhibitor NC, which were greatly reversed by miR-106a inhibitor (Figure 3(e-f)). Subsequently, the activity of the key pathway regulating oxidative stress was evaluated. Nrf2 and HO-1 (Figure 3(g)) were dramatically downregulated and Keap1 was greatly upregulated in the GO group and GO+ inhibitor NC group, which were abolished by the miR-106a inhibitor (**p < 0.01 vs. control, ##p < 0.01 vs. GO+ inhibitor NC). These results implied that GO-stimulated oxidative stress in MCs was mitigated by the miR-106a inhibitor.

MiR-106a regulated the expression of connexin-43 by targeting the 3ʹUTR region
Firstly, we checked the state of miR-106a/connexin-43 in above grouping. MiR-106a (Figure 4 (a)) was found significantly upregulated and connexin-43 (Figure 4(b)) was dramatically downregulated by the treatment of GO with or without inhibitor NC, which were greatly reversed by miR-106a inhibitor (**p < 0.01 vs. control, ##p < 0.01 vs. GO+ inhibitor NC). We suspected that a correlation between miR-106a and connexin-43 might be available, which was further identified using the dual-luciferase gene reporter assay. The predicted binding sequence between miR-106a and connexin-43 were shown in Figure 4(c). We found that in cells transfected with connexin-43-WT, no obvious difference was seen between the miR-NC and miR-106a mimic group. However, dramatically declined luciferase activity was observed in the miR-106a mimic group (**p < 0.01 vs. miR- Figure 2. The isolated rat cochlear MCs were identified and the incubation concentration of GO was determined. A. The isolated rat cochlear MCs were identified by checking the expression of CK18 using the immunofluorescence assay. B. The OD value was measured using the CCK-8 assay (**p < 0.01, ***p < 0.001). Scale bar: 50 μm. Figure 3. The apoptosis and oxidative stress in GO-treated MCs were ameliorated by the knockdown of miR-106a. MCs were treated with 20 mU/mL GO with or without inhibitor NC and miR-106a inhibitor, respectively. A. The OD value was checked utilizing the CCK-8 assay. B. The apoptotic rate was investigated by the flow cytometry assay. C. The level of Bcl-2, Bax, and caspase-3 was evaluated by the Western blotting assay. D. The DCFH-DA assay was utilized to determine the ROS level in MCs. E. The production of MDA was investigated utilizing the commercial kit. F. The production of SOD was investigated utilizing the commercial kit. G. The level of Nrf2, Keap1, and HO-1 was confirmed by the Western blotting assay (**p < 0.01 vs. control, ##p < 0.01 vs. GO+ inhibitor NC). NC). Considering above observation, we concluded that miR-106a targeted connexin-43 mRNA to regulate the expression of connexin-43.

Knockdown of connexin-43 canceled the inhibitory effects of miR-106a inhibitor against GO-induced apoptosis in MCs
To verify that miR-106a exerted the facilitating function in GO-induced injury in MCs, MCs were treated with 20 mU/mL GO and miR-106a inhibitor with or without a siRNA targeting connexin-43 (siR-connexin-43). Firstly, the declined OD value ( Figure 5 (a)) in the miR-106a inhibitor group was dramatically promoted by the co-introduction of siR-connexin-43. The apoptotic rate was slightly changed from 40.0% to 40.1% in GO-stimulated MCs by the treatment of inhibitor NC combined with siR-NC, while the apoptotic rate was greatly reduced to 16.5% in the GO+ miR-106a inhibitor group, which was reversed to 28.2% in the GO+ miR-106a inhibitor+ siR-connexin-43 group (Figure 5(b)). Additionally, in the miR-106a inhibitor group, significantly declined expression level of Bax and caspase-3, as well as the increased level of Bcl-2 ( Figure 5(c)), were observed, which were greatly reversed by siR-connexin-43 (**p < 0.01 vs. GO+ inhibitor NC+ siR-NC, ##p < 0.01 vs. GO+ miR-106a inhibitor). These data implied that miR-106a regulated the GO-induced apoptosis in MCs by mediating the expression of connexin-43.

Knockdown of connexin-43 abolished the inhibitory effects of miR-106a inhibitor against GO-induced oxidative stress in MCs
In additional to apoptosis, we suspected that the regulatory effect of miR-106a on oxidative stress was also associated with its inhibitory effect on the expression of connexin-43. We found that the significantly reduced ROS level ( Figure 5(d)) observed in the GO+ miR-106a inhibitor group was dramatically promoted by the co-treatment of siR-connexin-43. Additionally, the reduced production of MDA ( Figure 5(e)) and elevated concentration of SOD ( Figure 5(f)) in the GO+ miR-106a inhibitor group were dramatically reversed by siR-connexin-43. Lastly, upregulated Nrf2, connexin-43, and HO-1 and downregulated Keap1 ( Figure 5(g)) in the GO+ miR-106a inhibitor group were greatly abolished by siR-connexin-43 (**p < 0.01 vs. GO+ inhibitor NC+ siR-NC,  A. The OD value was checked utilizing the CCK-8 assay. B. The apoptotic rate was investigated by the flow cytometry assay. C. The level of Bcl-2, Bax, and caspase-3 was evaluated by the Western blotting assay. D. The DCFH-DA assay was utilized to determine the ROS level in MCs. E. The production of MDA was investigated utilizing the commercial kit. F. The production of SOD was investigated utilizing the commercial kit. G. The level of Nrf2, Keap1, and HO-1 was confirmed by the Western blotting assay (**p < 0.01 vs. GO+ inhibitor NC+ siR-NC, #p < 0.05 vs. GO+ miR-106a inhibitor, ##p < 0.01 vs. GO+ miR-106a inhibitor). #p < 0.05 vs. GO+ miR-106a inhibitor, ##p < 0.01 vs. GO+ miR-106a inhibitor). These results implied that miR-106a regulated the GO-induced oxidative stress in MCs by mediating the expression of connexin-43.

Pathological state in SNHL rats were alleviated by antagomir-106a
To confirm the function of miR-106a in SNHL, SNHL rats were administered with antagomir-NC, antagomir-106a, or antagomir-106a combined with carbenoxolone, respectively. Firstly, the increased ABR value (Figure 6(a)) in the Model+ antagomir-NC group was significantly reduced in the antagomir-106a group, which was greatly reversed by the co-treatment of carbenoxolone. Next step, the level of miRNA106a in SNHL rats was detected, the level of miR106a (Figure 6(b)) was increased in the Model+ antagomir-NC group, which was significantly reduced in the antagomir-106a group (p < 0.01). In the control group, the spiral ganglion cells in cochlea tissues were complete in shape and neatly arranged. The cell membrane was round or oval, the nucleoli was clear, and the chromatin was uniform. In the antagomir-NC group, the number of spiral ganglion cells in cochlea tissues was significantly reduced with different sizes, irregular and disordered arrangements, almost disappeared nucleus, and vacuolated degeneration. These pathological changes were significantly ameliorated in the antagomir-106a group. However, by the coadministration of carbenoxolone, the disordered arrangements, disappeared nucleus, and vacuolated degeneration were anew observed in spiral ganglion cells (Figure 6(c)). Additionally, increased TUNEL positive staining was observed in the antagomir-NC group, which was reduced by the treatment of antagomir-106a. However, compared to the antagomir-106a group, the TUNEL positive staining was anew enhanced in the antagomir-106a+ carbenoxolone group (Figure 6(d)). We further found that the upregulated Bax and caspase-3 ( Fig 8E) and the downregulated Bcl-2 in the antagomir-NC group were greatly reversed by antagomir-106a. However, compared to the antagomir-106a group, the expression level of Bax and caspase-3 was largely promoted and the expression level of Bcl-2 was greatly reduced in the antagomir-106a+ carbenoxolone group (**p < 0.01 vs. control, ##p < 0.01 vs. antagomir-NC, && p < 0.01 vs. antagomir-106a). These data suggested that the pathological state in SNHL rats were alleviated by antagomir-106a through regulating connexin-43.

Oxidative stress in SNHL rats were mitigated by antagomir-106a
We lastly evaluated the state of oxidative stress in cochlear tissues. We found that the elevated release of MDA (Figure 6(f)) and declined concentration of SOD (Figure 6(g)) in the antagomir-NC group were dramatically reversed by antagomir-106a. However, compared to the antagomir-106a group, significantly increased MDA level and reduced SOD level were observed in the antagomir-106a+ carbenoxolone group. Additionally, as the Figure 6(h) showed, the downregulated Nrf2 and HO-1, as well as the upregulated Keap1, were observed in the antagomir-NC group, which were greatly reversed by antagomir-106a. Compared to the antagomir-106a group, dramatically declined expression level of Nrf2 and HO-1 and promoted level of Keap1 were observed in the antagomir-106a+ carbenoxolone group (**p < 0.01 vs. control, ##p < 0.01 vs. antagomir-NC, & p < 0.05 vs. antagomir-106a, && p < 0.01 vs. antagomir-106a). These data suggested that the oxidative stress in SNHL rats were mitigated by antagomir-106a through regulating connexin-43.

Discussion
Cochlea is an organ that maintains high aerobic metabolism and produces a large amount of ROS while providing ATP for energy supply, which is thereby vulnerable to ROS oxidation [26]. Stria vascularis is an important structure located in the lateral wall of the cochlea, which a special stratified epithelium containing capillaries composed of three types of cells: MCs in the outer layer, intermediate cells in the middle layer, and basal cells in the inner layer. MCs are adjacent to the endolymph cavity, with Na + -K + -ATPase and Na + -K + channels located on the cell membranes, which are of great significance for maintaining the endolymph ionic environment and the cochlear micro-potential [27]. Once MCs are damaged by ROS, the energy supply of the inner ear will be disrupted [28]. The present study used MCs isolated from the rat cochlea tissue to establish the oxidative stress model under the stimulation of Figure 6. Pathological state and the level of oxidative stress in SNHL rats were alleviated by antagomir-106a. SNHL rats were administered with antagomir-NC, antagomir-106a, or antagomir-106a combined with carbenoxolone, respectively. A. The ABR values were recorded. B. The level of miRNA106a in SNHL rats was detected by RT-PCR. C. The pathological changes in cochlea tissues were determined by HE staining. D. The apoptosis in cochlea tissues was evaluated by the TUNEL assay. E. The level of Bcl-2, Bax, and caspase-3 was evaluated by the Western blotting assay. F. The concentration of MDA was measured using the commercial kit. G. The concentration of SOD was measured using the commercial kit.H. The expression level of Nrf2, Keap1, and HO-1 was evaluated by the Western blotting assay (**p < 0.01 vs. control, ##p < 0.01 vs. antagomir-NC, & p < 0.05 vs. antagomir-106a, && p < 0.01 vs. antagomir-106a). Scale bar: 100 μm.
GO, a reagent utilized to mimic the oxidative state [27].
Under a normal physiological state, the balance between the production and elimination of ROS was maintained in endothelial cells. However, when the ability of eliminating ROS declines, oxidative stress will be activated [29]. There were many research showed that Oxidative stress plays a role in the pathogenesis of noise-related hearing loss, as well as in drug-and aging-related hearing loss [9,30,31].MDA is the terminal output of lipid peroxidation induced by the action of oxygen free radicals on polyunsaturated fatty acids on biofilms. The level of MDA reflects the severity of free radical attack and can be used as an indicator to evaluate degree of oxidative stress [32][33][34]. Superoxide Dismutase (SOD) is an enzyme that catalyzes the superoxide anion radical disproportionation to generate oxygen and hydrogen peroxide. Declined activity of SOD is considered as a critical biomarker of oxidative stress [35]. We found that under the treatment of GO, classic symptoms of oxidative stress were observed in MCs, such as increased release of ROS, elevated MDA level, and declined SOD concentration, which was also reported by Kumar in H9c2 cells [36]. Additionally, in gentamicin-induced SNHL animal model, significant oxidative stress was observed in cochlea tissues. Similar results have been reported by Jiang [37] in 2016. These data further confirmed the important role of oxidative stress in SNHL.
MicroRNAs are small, non-coding, singlestranded RNAs, which have approximately 21-23 nucleotides [38]. There is increasing evidence that microRNA could regulated of gene expression in the post-developmental inner ear and contributed to the development of acquired hearing loss [38][39][40]. Resent study also showed that circulating microRNAs maybe new diagnostic biomarkers of hearing loss [41].MiR-106a is a miRNA that has been evidenced to be an important regulator in multiple diseases, such as diabetic nephropathy [42], gastric cancer [43], and gestational hypertension [44]. However, the function of miR-106a in SNHL is not clear. Recently, several reports claimed the regulatory effects of miR-106a against oxidative stress. Hu reported [45] that the ox-LDL-induced oxidative stress in endothelial cells could be reversed by the knockdown of miR-106a, indicating a facilitating effect of miR-106a against oxidative stress. However, an inhibitory effect of miR-106a against oxidative stress was reported by Wu [46]. Therefore, the role of miR-106a in regulating oxidative stress is controversial. The present study revealed that knockdown of miR-106a not only repressed oxidative stress in GOtreated MCs but also reversed oxidative stress in cochlear tissues in SNHL rats, accompanied by the alleviation of pathological state of SNHL rats, suggesting that in the development of SNHL, miR-106a functioned as an inhibitory mediator against oxidative stress.
Gap junction (GJ) is a channel between the membranes of adjacent cells, which allows the passage of substances with molecular weight less than 1000 Da, including some ions, second messengers, and nutrient metabolites [47]. GJ are mainly formed by the interlacing of two junctions, which are half-channels consisting of connexins, on adjacent cell membranes. Connexin 43, a member of the connexins family, belongs to type II connexin, which is the first connexin that is confirmed associated with deafness [14]. Zhang reported that the mild hearing loss could be induced by the downregulation of connexin 43 [48], which was supported by Wang [49] using a postnatal mice model. We found that the declined expression level of connexin 43 was both observed in GO-treated MCs and cochlear tissues of SNHL rats, suggesting connexin 43 might be protective to the development of SNHL. The correction between miR-106a and connexin 43 was further identified, which was consistent with Wang's report [16]. Verification experiments implied that the inhibitory effects of miR-106a inhibitor against oxidative stress and alleviating effects of antagomir-106a against both GO-induced injury in MCs and SNHL symptoms were dramatically abolished by the knockdown of connexin 43. Actually, the inhibitory effects of connexin 43 against oxidative stress have already been reported in several researches [15,50], which together with our finding confirmed the important role of miR-106a/connexin 43 in oxidative stress injury in MCs and SNHL. In future work, the upstream regulatory mechanism of miR-106a will be further investigated to explore more information on the pathogenesis of SNHL.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This work was supported by 2022 National Double first-class discipline (otolaryngology combined TCM with western medicine) promotion project of Beijing University of Chinese Medicine.