Protective Effect of Ginsenoside Rd on Military Aviation Noise-Induced Cochlear Hair Cell Damage


 Background Soldiers are often exposed to high-intensity noise produced by military weapons and equipment during activities, and the incidence of noise-induced hearing loss (NIHL) in many arms is high. Oxidative stress has a significant role in the pathogenesis of NIHL, and research has confirmed that ginsenoside Rd (GSRd) suppresses oxidative stress. Therefore, we hypothesized that GSRd may attenuate NIHL and cochlear hair cell loss, induced by military aviation noise stimulation, through the Sirtuin1/proliferator-activated receptor-gamma coactivator 1α (SIRT1/PGC-1α) signaling pathway.Methods Forty-eight male guinea pigs were randomly divided into four groups: control, noise stimulation, GSRd, and glycerol. The experimental groups received military helicopter noise stimulation at 115 dB (A) for 4 h daily for five consecutive days. Hair cell damage was evaluated by using inner ear basilar membrane preparation and scanning electron microscopy. Terminal dUTP nick end labeling and immunofluorescence staining were conducted. Changes in the SIRT1/PGC-1α signaling pathway and other apoptosis-related markers in the cochleae, as well as oxidative stress parameters were used as readouts.Results Loss of outer hair cells, more disordered cilia, prominent apoptosis, and elevated free radical levels were observed in the experimental groups. GSRd treatment markedly improved morphological changes and apoptosis through decreasing Bcl-2 associated X protein (Bax) expression and increasing Bcl-2 expression. In addition, GSRd upregulated superoxide dismutase (SOD) and glutathione peroxidase (GSH-Px) levels, decreased malondialdehyde (MDA) levels, and enhanced the activity of SIRT1 and PGC-1α messenger ribonucleic acid and protein expression.Conclusion GSRd can improve structural and functional damage to the cochleae caused by noise. The underlying mechanisms may be associated with the SIRT1/PGC-1α signaling pathway.


Introduction
Noise-induced hearing loss (NIHL) is a type of sensorineural hearing loss caused by overexposure to Considerable efforts exist to elucidate the pathogenesis of NIHL. However, the mechanisms underlying NIHL are not completely understood. The current belief is that the causes of NIHL include environmental and genetic factors. Environmental factors include noise intensity, spectrum characteristics, and the length of time of exposure to noise. Genetic factors primarily refer to NIHL-related susceptibility genes such as CDH23 [2], CASP3 [3], GRM7 [4], and GJB2 [5]. Three main theories related to mechanical damage, metabolic damage, and vascular changes explain the pathogenesis of NIHL [6-8]. Increasing evidence indicates the possible role of dysregulation of free radicals and oxidative stress [9][10][11].
Oxidative stress is an imbalance between cell damage and cell adaptive changes that result from reactive oxygen species (ROS). The excessive production of ROS may be a major cause of NIHL, and antioxidants have a protective role against NIHL [12].
Mitochondrial dysfunction may result in oxidative stress, which contributes to the pathogenesis of NIHL and possibly in uences SIRT1 and numerous cellular antioxidant defense mechanisms [12,19]. In response to oxidative stress, SIRT1 is redistributed at the chromatin level, and causes deregulation at the transcription level [20]. SIRT1 has a protective role against cochlear injury in age-related hearing loss (AHL) [21,22], which is another type of sensorineural hearing loss that has a similar pathogenesis as NIHL. In vitro, inhibition of SIRT1 expression also increases the apoptosis rate of House Ear Institute-Organ of Corti 1 ( HEI-OC1) inner ear cells, which suggests that SIRT1 has a role in the pathogenesis of AHL [23], as well as in protecting against NIHL [24].

Functional enrichment analyses of GSRd and SIRT1 via bioinformatics analyses
We retrieved essential information of GSRd from the Encyclopedia of Traditional Chinese Medicine (ETCM) [32]. The gene ontology (GO) enrichment analysis of GSRd pharmacological effects and the single gene enrichment analysis of SIRT1 were conducted by using the STITCH database [33].

Animal groups
Forty-eight male guinea pigs, weighing approximately 250-300 g, were purchased from the Experimental Animal Center of Air Force Medical University (Xi'an, China). All animals had a sensitive auricle re ex, normal tympanic membrane, and no history of noise contact. They were housed in an environment with natural light, room temperature of 20-25ºC, air relative humidity of 60-65%, and free access to food and water. Adaptive feeding was administered for 5 d before the start of the experiment. All experimental procedures were approved by the Animal Ethics Committee of our university. Ginsenoside Rd (Tai-He Biopharmaceutical, Guangzhou, China) was kindly provided by the Department of Neurology in Xijing Hospital (Xi'an, China).
Guinea pigs were randomly assigned to one of four groups, each containing 12 animals: the control group (Con), which received no noise stimulation nor treatment; the noise group (NE), which was exposed to military helicopter noise but did not receive drug treatment; the GSRd treatment group (Rd), which was exposed to military helicopter noise and injected intraperitoneally with 30 mg/kg GSRd dissolved in glycerol; and the experimental control group (Vehl), which was exposed to military helicopter noise and injected intraperitoneally with just glycerol at 30 mg/kg. GSRd/glycerol was injected from 5 days before noise stimulation to the end of noise stimulation for a total of 10 days.

Noise stimulation and procedures
The ambient noise of a military helicopter was collected and input to a speaker (Soundtop SF-12; Jiasheng Audio Equipment, Co., Ltd., Guangzhou, China) through a power ampli er (Soundtop QA-700; Jiasheng Audio Equipment) for cyclic playback. Noise stimulation was conducted in a soundproof room with an air-conditioned fan to moisten the air and strengthen local ventilation. The guinea pigs from the experimental groups were placed in a rat cage on which a speaker was placed. The noise intensity was measured by an A-weighted sound level (Heng-sheng Electronics, Jiaxing, China) to ensure that the difference in the sound pressure level in the activity range of guinea pigs was less than 3 dB. The animals were exposed to 115 dB (A) noise stimulation for 4 hours daily for 5 consecutive days [34]. The Con group was not exposed to noise stimulation. The background noise in the cage was less than 20 dB, and the other conditions were the same as those in the experimental groups.

Tissue preparation
After detecting the hearing threshold levels [35], the animals were euthanized. The bilateral temporal bones of guinea pigs were removed, and the bilateral cochleae were separated immediately. There were 24 cochlear specimens in each group of guinea pigs. The specimens were distributed as shown in Table 2.
The cochlear tissues used in the surface preparation of the basilar membrane and section staining were removed and then soaked in 4% paraformaldehyde, and those used for scanning electron microscopy (SEM) were soaked in 2.5% glutaraldehyde. The stapes were removed with microscopic forceps under an anatomical microscope. A small hole was created at the tip of the cochlea by a syringe needle. The corresponding xed uid was slowly injected into the cochlea through the perforated cochlear tip and the oval window more than three times. The specimens were then transferred into an Eppendorf tube containing the xed solution overnight at 4ºC. Beginning on the second day, the 10% EDTA solution was replaced for decalci cation for 14 days. The freshly decalci ed solution was replaced every day until the cochlear bone softened. The decalci ed cochlear tissues were used for surface preparation of the basilar membrane or were embedded in para n to prepare 5 µm sections parallel to the direction of the modiolus for terminal dUTP nick end labeling (TUNEL) and immuno uorescence staining. The rest of the specimens were stored at -80ºC.

Phalloidin staining
Phalloidin staining was conducted, as described by Qi and colleagues [36]. The spiral case and spiral ligament were removed with microscopic forceps under an anatomical microscope. The basilar membrane was separated step-by-step from the tip to the bottom of the cochlea and was moved into a 96-well plate. The removed basilar membranes were rinsed three times with 0.1M phosphate-buffered saline (PBS) for 1 min each time. A 100 µL of 1% Triton X-100 (MP Biomedicals, Solon, OH, USA) was subsequently added to each hole for 10 min. Specimens were rinsed three times with 0.1M PBS for 5 min each time. A phalloidin dilution (1:1000; ab176753; Abcam, Cambridge, MA, USA) was added to each well for 30 min without light. The specimens were rinsed three times with 0.1M PBS and 4,6-diamino-2phenylindole (DAPI) (1:1000; ab104139; Abcam) and staining was conducted for 8 min without light to display the nucleus. The specimens were then rinsed with 0.1M PBS three times for 5 min each time. The basilar membrane was drawn onto the glass slide by using a pipette. The positive and negative sides of the basilar membrane were observed under a microscope (SMZ745; Nikon Corporation, Tokyo, Japan). The slides were carefully covered with an 80% glycerin seal to prevent uorescence quenching. A confocal microscope (LSM 800; Zeiss, Oberkochen, Germany) was used to observe the experimental results.

SEM observation
The experimental steps of SEM corresponded to those proposed in articles by Sung et al. [37] and Santi et al. [38]. The basilar membrane was soaked overnight in 2.5% glutaraldehyde at 4ºC. The specimens were rinsed three times with 0.1M PBS for 15 min each, and then post xed in 1% osmium tetroxide for 2 h at a room temperature of 20°C-25°C. The specimens were then dehydrated with 30%, 50%, 70%, 80%, and 90% gradient ethanol and anhydrous ethanol; critical-point-dried by using liquid carbon dioxide; and then sputter-coated with gold-palladium for 30 s. The results were observed under a scanning electron microscope (S-3400; Hitachi, Tokyo, Japan).
TUNEL staining TUNEL staining, as described by Liu and colleagues [39], was carried out on sections of the cochlea by using the In Situ Cell Death Detection Kit (11684795910; Roche, Basel, Switzerland), based on the manufacturer's instructions. Sections of the cochlea were depara nized using xylene, hydrated with anhydrous, 90%, 80%, and 70% gradient ethanol. After rinsing with distilled water, the sections were covered with proteinase K solution for 15 min. They were subsequently rinsed with 0.1M PBS and distilled water for 6 min each and xed with 3% methanol-hydrogen peroxide solution for 20 min. Following three rinses with 0.1M PBS for 6 min each, the sections were covered with 3% bovine serum albumin-PBS solution for blocking, followed by incubation with the TUNEL reaction mixture in the dark for 1 h at a room temperature of 20°C-25°C. Finally, the sections were rinsed with PBS-Tween (PBST) buffer solution and 0.1M PBS for 6 min each. The results were examined and photographed under uorescence microscopy (DP71; Olympus, Tokyo, Japan).
Immuno uorescence staining for 4-hydroxy-4-hydroxynonenal and 3-nitrotyrosine Immuno uorescence staining was carried out, as described by Tian and colleagues [40]. Sections of cochlea were depara nized using xylene; hydrated with anhydrous, 90%, 80%, and 70% gradient ethanol; and then rinsed three times with 0.1M PBS for 10 min each. A 1% Triton X-100 (MP Biomedicals) was added for 10 min. Three rinses were carried out using 0.1M PBS. The sections were incubated with 5% BSA (MP Biomedicals) for 1 h. The antibodies used for immuno uorescence staining included rabbit polyclonal anti-4-hydroxy-4-hydroxynonenal (anti-4-HNE) (1:1000; ab46545; Abcam) and mouse monoclonal anti-3-nitrotyrosine (anti-3-NT) (1:1000; ab61392; Abcam). Antibodies were diluted and were added to the sections. The sections were placed in a black wet box and stored overnight in a refrigerator at 4ºC. They were removed on the second day and reheated at a room temperature of 20°C-25ºC for 1 h before the three washes with 0.1M PBS. The secondary antibodies of cy3-labeled goat antirabbit immunoglobulin G (IgG) (red, 1:1000; Zhuangzhi Biotechnology, Xi'an, China) and 488-labeled goat antimouse IgG (green, 1:1000; Zhuangzhi Biotechnology) were added and the sections were incubated at room temperature at 20°C-25°C without light for 2 h. Another three rinses with 0.1 M PBS were applied and 8 min of DAPI (1:1000, ab104139, Abcam, USA) staining was applied without light to display the nucleus. Glycerin (80%) was used to prevent uorescence quenching. A confocal microscope (LSM 800; Zeiss) was used for observation.

Quantitative real-time polymerase chain reaction analysis
The protocols used in this study were conducted, as described by Chen et al [35]. RNAiso (TaKaRa, Kyoto, Japan) was added to the homogenized cochlear specimens to extract total ribonucleic acid (RNA), based on the manufacturer's instructions. The extracted RNA was diluted to 300-500 ng/µL, and then combined with 2 µL 5 × Primer Script RT Master Mix (TaKaRa). The 10 µL reverse transcription reaction system was lled with diethyl pyrocarbonate water. The mixture was placed in a real-time system (Applied Biosystems, Waltham, MA, USA) at 37ºC for 15 min and then at 85ºC for 5 s to synthesize the complementary DNA (cDNA) template. Two microliters of the cDNA template were mixed with 20 µL of SYBR Premix Ex Taq™ II (2×) (TaKaRa), 14 µL of diethyl pyrocarbonate water, and 2 µL of the forward and reverse primers (Table 3; Sangon, China). The con gured reaction system was fully blown and mixed and then was added to an octagonal tube with 9 µL per hole. The parameters were set as follows: 95ºC for 30 s, 40 cycles of 95ºC for 5 s and 60ºC for 34 s, and 65ºC for 15 s. Glyceraldehyde-3-phosphate dehydrogenase was used as an internal control and the 2 −ΔΔCt method was used to calculate the relative expression of the target gene.

Western blot analysis
The protocols used in this study were conducted as described by Su et al [41]. The left and right cochlea of the same guinea pig were placed together in a tissue homogenizer. A 200 µL protein extraction reagent (78505; Thermo Scienti c, Waltham, MA, USA) containing 2mM phenylmethylsulfonyl uoride was used as the protein lysate to extract the total cochlear proteins. The total protein concentration was calculated using the BCA Protein Assay Kit (23250; Thermo Scienti c). A total of 30 µg of each protein sample was denatured, separated on 4-12% Bis-Tris PAGE gels, and then transferred to polyvinylidene uoride membranes (0.45 µm; Millipore, Darmstadt, Germany). The membranes were blocked in 5% fat-free milk powder for 2 h at a room temperature of 20°C-25°C and were then incubated with rabbit polyclonal antibody against Bax (1:500; WL01637; Wanleibio, Shenyang, China), Bcl-2 (1:500; WL01556; Wanleibio, China), SIRT1 (1:500; WL02995; Wanleibio), PGC-1α (1:500; WL02123; Wanleibio), or β-actin (1:1000; WL01845; Wanleibio) overnight at 4ºC. After six rinses with PBST for 5 min each, the membranes were incubated in PBST with horseradish peroxidase-conjugated goat antirabbit IgG (1:5000; WLA023; Wanleibio) for 1 h at a room temperature of 20°C-25°C and were detected using enhanced chemiluminescence detection reagents (Millipore) in a Gel Image Analyzing System (Tanon Science & Technology, Shanghai, China). The band intensity was measured using Image J v1.51 (National Institutes of Health, Bethesda, MD, USA), and the values were normalized to β-actin.
Detection of superoxide dismutase activity, malondialdehyde level, and glutathione peroxidase level The left and right cochlea of the same guinea pig were placed in the same Eppendorf tube, and the samples were prepared with 0.9% normal saline as a 10% homogenate. Superoxide dismutase (SOD) assay kit (WST-1 method; A001-3; Jiancheng Biotechnology, Nanjing, China), malondialdehyde (MDA) assay kit (TBA method; A003-1; Jiancheng Biotechnology), and glutathione peroxidase (GSH-Px) assay kit (colorimetric method; A005-1; Jiancheng Biotechnology) were used to detect SOD activity, MDA levels, and GSH-Px levels, based on the manufacturer's instructions. Each group's chromaticity was assessed using a microplate reader (Thermo Fisher) at 450 nm for SOD activity, 532 nm for the MDA level, and 412 nm for the GSH-Px level.

Statistical analysis
The results are presented as mean ± the standard error. The hearing level; immuno uorescence staining results; mRNA and protein expression; and SOD, MDA, and GSH-Px activities were statistically analyzed using one-way analysis of variance. Dunnett's t-test was used to compare the experimental and control groups. Differences were signi cant when P < 0.05. Statistical analyses were performed using GraphPad Prism 7.0 (GraphPad Software, San Diego, CA, USA) and SPSS 23.0 (IBM Corporation, Armonk, NY, USA).

Results
GSRd improved noise-induced hearing loss through the SIRT1/PGC-1α signaling pathway The retrieval results in the ETCM database suggest that the diseases treated with GSRd include sensorineural hearing impairment, progressive sensorineural hearing impairment, and dysregulation of apoptosis (http://www.tcmip.cn/ETCM/index.php/Home/Index/cf_details. html?id = 4640). Therefore, we initially hypothesized that GSRd may be a candidate drug for hearing impairment. In addition, GO enrichment analysis showed that the pharmacological effects of GSRd are primarily the regulation of oxidative stress and apoptosis and involvement in mitochondrial activity (Fig. 1). NIHL is a type of illness characterized by oxidative stress damage and apoptosis progression [42]. Previous results have demonstrated that GSRd can signi cantly reverse the increase in the auditory brainstem response threshold caused by noise stimulation (Supplementary Fig. 1) [35], which supports our initial hypothesis. The results of the GSRd GO enrichment analysis focused our attention on a molecule associated with the regulation of mitochondrial activity and oxidative stress, which is SIRT1 (Table 1). Furthermore, SIRT1 single gene enrichment analysis demonstrated that the SIRT1/PGC-1α signaling pathway is more prominently involved in gene expression, oxidative stress, and apoptosis regulation.

GSRd ameliorated damage to cochlear hair cells of guinea pigs
The cochlear basilar membrane preparation and SEM were used to observe morphological changes of the hair cells in the inner ear of guinea pigs after noise stimulation. Basilar membrane phalloidin staining ( Fig. 2A) revealed the micro lament cytoskeleton and manifested the distribution of actin in the cochleae. In the control group (Con), the outer hair cells of the guinea pigs were intact and neatly arranged, and the cilia were clearly visible and arranged in a "V" shape. The outer hair cells of the noise group (NE) and the experimental control group (Vehl) showed damage with different degrees of dot deletion. The degree of outer hair cell loss in the Rd treatment (Rd) group was less than that of the NE and Vehl groups.
SEM images (Fig. 2B) showed the structural damage to the cilia and supporting cells of inner ear hair cells. The SEM images of the Con group veri ed the results of phalloidin staining. In the NE and Vehl groups, the polarity of the outer hair cells disappeared, and the arrangement was disordered with lodging, fusion, and even falling off. Among them, the third layer of outer hair cells was most seriously damaged. Following Rd treatment, the outer hair cells in the Rd group showed slight lodging and loss of cilia.

Effect of GSRd on the expression of apoptotic factors in the cochleae of guinea pigs
The results of the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling assay (TUNEL) staining showed no TUNEL-positive cells in the cochleae of guinea pigs in the Con group. The NE and Vehl groups had TUNEL-positive cells, primarily distributed in the organ of Corti, vascular stria, and spiral ligament. The number of TUNEL-positive cells in the Rd group was lower than that in the NE and Vehl groups. The results of TUNEL staining are shown in Fig. 2C.
Effect of GSRd on the gene and protein expression of SIRT1 and PGC-1α in the cochleae of guinea pigs The mRNA content (Fig. 3B) and protein content (Figs. 3F and 3G) of SIRT1 in the NE group (P = 0.0003 and P = 0.0001, respectively), Rd group (P = 0.0192 and P = 0.0015, respectively) and Vehl group (P = 0.0003 and P = 0.0001, respectively) were signi cantly lower than those of the Con group, whereas they were signi cantly higher in the Rd group than in the NE group (P = 0.0255 and P = 0.0111, respectively) and Vehl group (P = 0.0221 and P = 0.0173, respectively). The mRNA content (Fig. 3B) and protein content (Figs. 3F and 3H) of PGC-1α likewise showed a similar relationship between groups: NE group (P = 0.0004 and P = 0.0001, respectively), Rd group (P = 0.0272 and P = 0.0003, respectively), and Vehl group (P = 0.0006 and P = 0.0001, respectively), compared to the Con group; NE group (P = 0.0246 and P = 0.0101, respectively) and Vehl group (P = 0.0454 and 0.0137, respectively), compared to the Rd group.
GSRd reduced the expression of free radicals in the cochleae of guinea pigs The common indices of the free radical level test are 4-hydroxy-4-hydroxynonenal (4-HNE) (i.e., an ROS marker) and 3-nitrotyrosine (3-NT) (i.e., a reactive nitrogen species [RNS] marker). In this experiment, 4-HNE was displayed with red uorescence, 3-NT was displayed with green uorescence, 4,6-diamino-2phenylindole (DAPI) staining nucleus was displayed with blue uorescence, and merge was synthesized in the three colors (Fig. 4A). Only a small amount of 4-HNE and 3-NT was expressed in the Con group.
Following the noise stimulation, the expression of 4-HNE (Fig. 4B) and 3-NT (Fig. 4C) was signi cantly increased in the NE group (P = 0.0001) and Vehl group (P = 0.0001) and was primarily located in the organ of Corti. The immuno uorescence expression of 4-HNE and 3-NT was decreased signi cantly in the Rd group (P = 0.0002).

Discussion
In this study, we tested our hypothesis that GSRd supplementation would attenuate NIHL via regulating oxidative stress and apoptosis, mediated by the SIRT1/PGC-1α signaling pathway. Our results suggested that GSRd can alleviate hair cell loss and free radical production in the cochleae of guinea pigs. The possible mechanism may involve the activation of the SIRT1/PGC-1α signaling pathway, the upregulation of SIRT1/PGC-1α expression, increased SOD and GSH-Px activity, and decreased MDA level, which enhances the antioxidant capacity of the auditory system and hence reduces peripheral auditory system damage. The protective effect of GSRd on cochlear hair cell damage via the SIRT1/PGC-1α signaling pathway in guinea pigs are shown in Fig. 6.
Our results are consistent with previous ndings indicating that GSRd can scavenge free radicals, exert the effects of anti-oxidation, antiaging, and analgesia [43], and that it has a certain protective effect to some extent on the cardiovascular system [44,45], kidneys [46], immune system [47], and nervous system [48][49][50][51]. Ginsenosides (GSs) are the main active components of Panax ginseng, Panax quinquefolium, and Panax notoginseng. More than 50 types of GS monomers have been isolated and extracted from ginseng plants [52]. The separation of effective monomers from GS is helpful in clarifying its pharmacological effects. Among all GS monomers, Rb1, Rb2, Rc, Re, and Rg1 account for more than 80% of the total saponins, and are thus termed the main saponins. Rd, Rg3, Rh2, and other GS monomers are rare, although their pharmacological e cacy are better than those of the main saponins.
GSRd is a monomer extract of total saponins, which belongs to the protopanaxadiol type structure. Its molecular formula is C 48 H 82 O 18 ·3H 2 O and its molecular weight is 1001 Da. GSRd has high fat solubility, easy diffusion, and permeation through bio lm, but has poor water solubility [53]. GSRd has two glycosyl groups on C-3, which is the pharmacological basis of its antioxidant activity [54]. GSRd is a rare saponin monomer, which is low in ginseng and high in Panax notoginseng [55]. Intestinal enzymes can metabolize the main saponins Rb1 and Rc into Rd; therefore, Rd is one of the main forms of absorption and utilization of main saponins in the intestine after metabolism [56,57]. Among the more than 50 types of GS, GSRd is the main active ingredient in many pharmacological actions of Panax ginseng.
Our experimental results showed that GSRd treatment may have otoprotective effects by regulating the apoptosis pathway and oxidative stress. Following noise stimulation, the peripheral auditory system is rst damaged, including mechanical damage to the inner ear and a series of metabolic damage. The main manifestation of the mechanical and metabolic damage is high-frequency hearing loss. Results from previous research show that, after noise stimulation, the expression of activated Caspase-3 and Bax in the cochlea increases signi cantly, whereas the expression level of Bcl-2 decreases signi cantly, thereby resulting in a signi cant decrease in the Bcl-2/Bax ratio [58]. This nding is consistent with our results. In addition, excessive ROS and lipid peroxidation were stimulated in the inner ears. The aldehyde products are thus produced and may continue to contribute to the destruction of the normal physiological activities of hair cells, including 4-HNE, which form cross-links with proteins and nucleic acids.
The 4-HNE level is closely associated with oxidative stress, apoptosis, and in ammation, and is considered a marker of ROS. In addition, excessive RNS stimulated by noise nitrify the free tyrosine or tyrosine residues in the protein, thereby resulting in the formation of 3-NT. It attacks the normal structure of the protein, affects its normal physiological function, and causes cell damage, which is a marker of RNS. GSRd treatment downregulated the expression of Bax and free radical markers such as 4-HNE and 3-NT and upregulated the expression of Bcl-2, thereby reducing the level of apoptosis and oxidative stress in the inner ear. Considering the pathogenic mechanisms underlying NIHL, which is associated with ROS, the aforementioned ndings suggest that GSRd has protective therapeutic effects against NIHL by regulating the level of oxidative stress in the cochleae.
In this experiment, we observed that the damage to the third-row outer hair cells was the most severe, whereas the damage to inner hair cells was relatively mild. Ren et al. [59] believe that this result may be attributable to the location. Third-row outer hair cells are in the center of the basilar membrane where the largest vibration amplitude occurs. However, the inner hair cells are on the edge of the bony spiral lamina, which is affected by less vibration from the basilar membrane.
SEM revealed that GSRd may improve the lodging and loss of the cilia of hair cells, but the results of basilar membrane preparation were not signi cantly different between the NE and Vehl groups. We concluded that the abnormal morphology of the cilia and ultrastructural changes in the hair cells generally occurred after noise stimulation. Only when the damage factors reached a certain degree did death and loss of hair cells occur and manifest in the basilar membrane preparation [60].
SIRT1 can activate the antioxidant pathway and signi cantly reduce ROS levels [61]. Xiong et al. [22] explored SIRT1 expression in the cochlea and found that it was primarily expressed in the spiral ganglion neurons, outer hair cells, inner hair cells, and some supporting cells. A small amount of SIRT1 was expressed in the vascular stria of the lateral wall of the cochlea. The expression of SIRT1 was less pronounced in the outer hair cells than in the inner hair cells [62]. Previous studies have shown that SIRT1 has a protective effect on AHL-induced cochlear damage by activating downstream PGC-1α signaling [63]. Previous publications from our laboratory also indicated that GSRd can ameliorate auditory cortex damage of the central auditory system associated with military aviation noise-induced hearing loss by activating the SIRT1/PGC-1α signaling pathway [35]. Considering the similarities in the pathogenesis between NIHL and AHL, we hypothesized that the upregulation of the SIRT1/PGC-1α signaling pathway also has a role in the otoprotective effects of GSRd on the peripheral auditory system. Our results showed that the mRNA and protein levels of SIRT1/PGC-1α in the cochlea of guinea pigs were downregulated after noise stimulation, which indicated that they may be involved in the noise-induced damage in the inner ear. Moreover, GSRd can upregulate the SIRT1/PGC-1α signaling pathway, which helps to maintain the oxidative balance of the peripheral auditory system, thereby exerting its otoprotective effects.

Conclusions
Our results suggested that GSRd may improve damage to the peripheral auditory system of guinea pigs, caused by military helicopter noise, by activating the SIRT1/PGC-1α signaling pathway. Further studies on the dose-response curve of GSRd in the treatment of NIHL and the therapeutic effect of GSRd after prolonged noise stimulation are recommended. We also recommend further investigation to determine whether GSRd directly regulates oxidative stress levels by regulating SOD and MDA activity, or indirectly regulates these parameters through other signaling pathways.

Availability of data and materials
All data generated or analyzed during this study are available from the corresponding author on reasonable request. All materials are commercially available.

Competing interests
The authors declare no con icts of interest.   Figure 1 GO enrichment analysis about GSRd via STITCH database. CETA, cysteine-type endopeptidase activity