Neonatal (P2) Sprague-Dawley (SD) rats were provided by the Laboratory Animal Center of the Air Force Medical University. All the animals were housed in an environment with a temperature of 22 ± 1°C, relative humidity of 50 ± 1%, and a light/dark cycle of 12/12 h (lights on at 8 a.m. and off at 5 p.m.). All animal studies, including the mice euthanasia procedure, were done in compliance with Air Force Medical University institutional animal care regulations and guidelines and conducted according to the AAALAC and the IACUC guidelines.

At P2, the organs of Corti dissection from the rat’s cochlear tissue were performed. The specimens were placed in Hank’s balanced salt solution (H1045, Solabio, Beijing, China). The entire organs of Corti were cultured in Hanging Cell Culture Inserts (MCSP24H48, EMD Millipore Corp, Billerica, MA, United States). The culture medium consisted of Minimum Essential Medium supplemented with Earle’s salt and L-glutamine (11095080, Gibco, Grand Island, NY, United States), 3 mg/mL glucose, and 0.3 mg/mL penicillin (P3032, Sigma-Aldrich, St. Louis, MO, United States). The organs of Corti were cultured at 37°C in a humid environment with 5% CO₂. The final concentrations of gentamicin (E003632-1G, Sigma-Aldrich, St. Louis, MO, United States) were 0.3 mM, 0.6 mM, and 1 mM. The HO-1 activator Co (III) protoporphyrin IX chloride (CoPPIX) (Co654, Frontier Scientific, Logan, UT, United States) was used at 5 μM in the CoPPIX group with a 12 h incubation. Tissues undergoing both CoPPIX and gentamicin treatments were first cultured with CoPPIX for 12 h, after which the medium was replaced by gentamicin-loaded medium for 24 h incubation. Meanwhile, the HO-1 suppressor Zn (II) protoporphyrin IX (ZnPPIX) (Zn625, Frontier Scientific, Logan, UT, United States) was prepared based on a previous report (Francis et al., 2011). ZnPPIX was then diluted to 20μM in culture medium for 12 h, as described in a previous study (Kim et al., 2006). Nrf2 inhibitor ML385 (846557-71-9, Selleck Chemicals, Houston, TX, United States) was used at 15 μM in the medium for 24 h. The gentamicin + ML385 group received both gentamicin (0.6 mM) and ML385 (15 μM) for 24 h.

Approximately 20 organs of Corti were dissected from postnatal (P2) SD rats and cultured in medium with or without 0.6 mM gentamicin for 24 h. Total RNA extraction from the collected tissues was performed with the RNeasy Micro Kit (74004, QIAGEN, Hilden, Germany). The RT² First Strand Kit (330401, QIAGEN, Hilden, Germany) was utilized for cDNA preparation from 1 μg of total RNA. The expression of 84 oxidative damage-related genes was assessed with the Oxidative Stress RT² Profiler™ PCR Array kit (PARN-065ZC, QIAGEN, Hilden, Germany). The RT² SYBR Green/ROX qPCR Mastermix (330522, QIAGEN, Hilden, Germany) was employed for quantitative real-time PCR (qPCR), as directed by the manufacturer, on an ABI ViiTM7 Real-Time PCR System in 25 μL reactions. The following protocol was used for amplification: 95°C, 10 min; 40 cycles of 95°C (15 s) and 60°C (60 s). The experiments were carried out in triplicate. The results of qPCR were uploaded on the RT² Profiler™ PCR Array Data Analysis website.² According to the instructions of the software, Ct cut-off value was set at 35. 2^–ΔCt was used to calculate the fold-change of mRNA expression. Student t-tests were used for the assessment of statistical significance. A P-value < 0.05 was considered statistically significant. The genes with P < 0.05 and more than 2-fold difference in expression were considered to be differentially expressed genes with biological significance.

DAVID 6.8³ online analysis platform was used to annotate the screened differential genes in the GO and to classify processes or functions the differential genes mainly affect. R language was used to convert the data into a visual bubble chart. The protein interaction relationship of differentially expressed genes was analyzed through the STRING protein interaction database V11.0.⁴

PCR was performed on the selected genes in order to verify the results of RT² Profiler PCR Array. The entire organs of Corti were treated with 0 mM, 0.3 mM, 0.6 mM, and 1 mM gentamicin, after which they were collected. The total RNA was extracted with the RNeasy Micro Kit (74004, QIAGEN, Hilden, Germany). The RT² First Strand Kit (330401, QIAGEN, Hilden, Germany) was utilized for cDNA preparation from 1 μg of total RNA. The following primers were employed: HO-1, sense 5′-TTTCACCTTCCCGAGCATC-3′ and antisense 5′-TCTTAGC CTCTTCTGTCACCCT-3′; β-actin, sense 5′-GAAGAGCTATG AGCTGCCTGA-3′ and antisense 5′-TGATCCACATCTGCTGG AAGG-3′; Nrf2, sense 5′-TTCCTCTGCTGCCATTAGTCAGTC-3′ and antisense 5′-GCTCTTCCATTTCCGAGTCACTG-3′; NQO1, sense 5′- GCGAGAAGAGCCCTGATTGTACTG -3′ and antisense 5′- TCTCAAACCAGCCTTTCAGAATGG -3′; GCLC, sense 5′- ACATCTACCACGCAGTCAAGGACC -3′ and antisense 5′- CTCAAGAACATCGCCTCCATTCAG -3′. β-actin was utilized for normalization. Data were obtained in triplicate and presented as mean.

Supporting cells were separated from the organs of Corti and cultured in vitro. The organs of Corti of three newborn rats (P2) were minced into small pieces and digested with PBS containing 0.5 mg/mL collagenase IV (17104-019, Gibco, Grand Island, NY, United States) for 20 min. DMEM/F-12 medium (11330032, Gibco, Grand Island, NY, United States) with 10% fetal bovine serum (FBS; 10091148, Gibco, Grand Island, NY, United States) and 1% antibiotic cocktail (penicillin and streptomycin; 15140-12, Gibco, Grand Island, NY, United States) were used to culture the cells in suspension for 24 h. Adherent-sphere cells were cultured in culture medium including factor DMEM/F-12, B-27™ Supplement (1:50, 17504044, Gibco, Grand Island, NY, United States), human EGF (5 ng/mL, AF100-15, PeproTech, Rocky Hill, NJ, United States), human FGF-basic (2.5 ng/mL, 100-18B, PeproTech, Rocky Hill, NJ, United States) and 1% antibiotic cocktail, in a humid environment containing 5% CO₂ at 37°C for 24 h. Adherent cells were then transferred to DMEM/F-12 medium with 10% FBS and 1% antibiotic cocktail and cultured for additional 2 days.

HEI-OC-1 cells were cultured at 37°C with 5% CO₂ in DMEM (C11995500BT, Gibco, Grand Island, NY, United States) containing 10% FBS (10099141, Gibco, Grand Island, NY, United States) and 1% penicillin (SV30010, HyClone, South Logan, UT, United States). The cells were subcultured at 80% confluence using 0.25% trypsin/EDTA (25200056, Gibco, Grand Island, NY, United States). When cells were cultured to a suitable density, the serum was removed, and cells were washed with PBS three times. CoPPIX (5 μM) and ZnPPIX (20 μM) were then added in the medium for 12 h.

The entire organs of Corti underwent fixation with 4% paraformaldehyde for 8 h at 4°C. Supporting cells and HEI-OC1 cells underwent fixation with 4% paraformaldehyde for 20 min at room temperature. Blocking was carried out in phosphate-buffered saline (PBS) with 1% Triton X-100 (V900502, Sigma-Aldrich, St. Louis, MO, United States) and 5% bovine serum solution for 1 h. The tissue specimens underwent overnight incubation with rabbit anti-HO-1 (1:100, GTX101147, Gene Tex, Alton Pkwy Irvine, CA, United States), mouse anti-Nrf2 (1:200, ab89443, Abcam, Cambridge, MA, United States), goat anti-Sox2 (1:100, AF2018-SP, R&D Systems, Minneapolis, MN, United States), mouse anti-p27^Kip1 (1:100, sc-1641, Santa Cruz, Santa Cruz, CA, United States), and rabbit anti-Myosin VII-a (1:800, 25-6790, Proteus Bioscience, Ramona, CA, United States) primary antibodies, respectively, at 4°C. Following four PBS washes, Alexa Fluor 488-linked donkey anti-mouse (AP192F, Millipore, Billerica, MA, United States), Alexa Fluor 594-linked donkey anti-rabbit (AP182C, Millipore, Billerica, MA, United States), and Alexa Fluor Plus 647 Donkey anti-Goat (H + L; A32849, Invitrogen, Carlsbad, CA, United States) secondary antibodies (1:200) were added to the specimens, respectively, for 12 h at 25°C. Rhodamine Phalloidin (1:200, PHDR1, Cytoskeleton, Denver, CO, United States) was used to mark HCs for 20min. Counterstaining was performed using 4′, 6′-diamidino-2-phenylindole (DAPI; 1:1000, D9542, Sigma-Aldrich, St. Louis, MO, United States). The entire specimen was evaluated in different turns under a confocal microscope (Nikon, Tokyo, Japan). 3D reconstructions were made by using Imaris (×64) software (Version: 9.0, Bitplane).

The cultured organs of Corti were placed under a confocal microscope. HCs of each turn were separately enumerated from micrographs acquired.

Mitochondrial ROS amounts were assessed by MitoSOX Red staining (M36008, Invitrogen, Carlsbad, CA, United States). MitoSOX Red was used for the detection of mitochondrial reactive oxygen species by live-cell imaging. After organs of Corti were cultured with gentamicin (0.6 mM), CoPPIX (10 μM) or ZnPPIX (100 μM) for 48 h, the specimens underwent PBS washing and incubation with MitoSOX Red (5 μM) at 37°C in the presence of 5% CO₂ for 10 min. Next, 4% paraformaldehyde was used for cell fixation at room temperature for 30 min before immunofluorescent staining.

The entire organs of Corti were treated in different groups. Total protein from cochlear specimens was obtained with the RIPA buffer that contained 1% PMSF. Nuclear–cytoplasmic fractionation was conducted using the NE-PER Nuclear and Cytoplasmic Extraction Reagents kit (Pierce, Thermo Fisher Scientific, Waltham, MA, United States) according to the manufacturer’s protocol. After clearing the lysate by centrifugation, the proteins were resolved by 8% SDS-PAGE and electro-transferred onto PVDF membranes. Upon blocking with 5% non-fat milk (1h at ambient), the membranes underwent overnight incubation with anti-HO-1 (1:100, GTX101147, Gene Tex, Alton Pkwy Irvine, CA, United States), anti-Nrf2 (1:100, ab89443, Abcam, Cambridge, MA, United States), and anti-GAPDH (1:1000, 10494-1-AP, Proteintech) and β-actin (1:1000, sc-47778, Santa Cruz, Santa Cruz, CA, United States) primary antibodies, respectively, at 4°C. This was followed by incubation with HRP-linked secondary antibodies (1:2000, SA00001-2, Proteintech) for 1h at ambient. The ECL kit (Pierce, Thermo Fisher Scientific, Waltham, MA, United States) was used for visualization.

SPSS 26 (SPSS Software, Chicago, IL, United States) and GraphPad 5.01 (GraphPad Software, San Diego, CA, United States) were used for data analysis. Data were compared by one-way analysis of variance (ANOVA). A P-value < 0.05 indicated statistically significant differences.

The organs of Corti of the inner ear from 24 P2 rats were cultured in a medium with or without 0.6 mM gentamicin. After 24 h of culture, the RT²Profiler™ PCR Array Rat Oxidative Stress kit was used to investigate the expression of 84 known oxidative-related genes. Re-collect the organs of Corti and conduct three individual experiments. Gene expression assessment of triplicate assays was carried out as described on the RT² Profiler™ PCR Array Data Analysis website (see text footnote 2), based on cycle threshold (Ct). Hierarchical clustering was used to evaluate the transcriptional levels of 84 oxidative stress-related genes in the cochlea (Figure 1A); volcano plots were used to compare gene expression between the control and gentamicin groups (Figure 1B). Figure 1C shows a scatter plot highlighting several notable genes based on large differences in expression fold between the control and gentamicin groups (Figure 1C). Of the 84 genes, 18 were up-regulated by more than 2-fold (Supplementary Table 1), and three genes were down-regulated by ≥2-fold (Supplementary Table 2). HO-1 was the gene with a relatively high up-regulation multiple which ranked the 3rd among these 21 genes, while with the smallest p-value in the whole group.

To classify the processes and functions among genes with differential expression between the control and gentamicin groups, gene ontology (GO) analysis was carried out to categorize the 21 differentially expressed genes for their molecular functions (Figure 2A). Based on biological processes, these genes were involved in response to oxygen-containing compounds, cellular response to chemical stimulus, response to oxidative stress, response to the drug, and response to chemicals. Genes with altered expression in the gentamicin group were highly enriched in functional categories such as oxidoreductase activity, organic cyclic compound binding, ion binding, heme binding, and cofactor binding. The molecular components of the genes were highly enriched in the plasma membrane, cell part, NADPH oxidase complex, cytoplasm, and cytosol. In addition, a STRING protein-protein interaction network was constructed to visualize the possible connections between differentially expressed genes (Figure 2B). Among them, HO-1 showed an interaction with eight genes; PTGS2 interacted with six genes, and APOE and NOS2 interacted with five genes.

The organs of Corti were cultured in a medium with or without 0.6 mM gentamicin. After 24 h of culture, immunofluorescence staining showed that HO-1 highly expressed in SCs, and only slightly expressed in HCs in the control group. Treatment with 0.6 mM gentamicin increased the expression of HO-1 in HCs and decreased its expression in SCs (Figure 3A). SCs were separated from control group rats for in vitro culture. SOX2 and p27^Kip1 are known markers of SCs. In vitro assays also confirmed that HO-1 was expressed in SCs in the control group (Figure 3B). In order to verify the expression of HO-1 up-regulated by gentamicin, the organs of Corti were treated with varying concentrations (0, 0.3, 0.6, and 1 mM) of gentamicin. The quantitative RT-PCR (qPCR) (Figure 3C) and immunoblot (Figures 3D,E) data indicated that gentamicin up-regulated HO-1 in a concentration-dependent manner (***P < 0.001, *P < 0.05, n = 3), which suggested that HO-1 participates in the process of gentamicin injury.

To explore the role of HO-1 in the process of gentamicin injury, the HO-1 inducer CoPPIX (Ferrandiz and Devesa, 2008) was utilized to determine whether activating HO-1 protects HCs from gentamicin damage. The immunostaining showed that HO-1 induced in both HCs and SCs by CoPPIX (Figure 4A). The organs of Corti were pretreated with CoPPIX (5 μM) in the presence or absence of the HO-1 inhibitor ZnPPIX (20 μM) for 12 h and then further incubated with 0.6 mM gentamicin for 24 h. Myosin VII-a was used as a marker of inner and outer HCs. In control cells, the four HC rows were neatly arranged. Gentamicin administration for 24 h resulted in elevated HC loss in basal turns versus apical turns, forming a gradient from apex to base (Figure 4B); compared to the control group, the number of HCs in the apical turns was reduced by 68.20%, 83.01% in middle turns and approximately 87% in basal turn in the gentamicin group (***P < 0.001, n = 5) (Figures 4C–E, Formula for calculating ratio was (X-Y)/X. X means the number of HCs in each turn of control group, 56.6, 62.4, and 60, respectively. Y means the number of HCs in each turn of gentamicin group, 18, 10.6 and 7.8, respectively). Meanwhile, compared to the gentamicin group, the number of HCs in the apical, middle, and basal turns respectively increased by 185.56, 447.17, and 369.239% in the gentamicin + CoPPIX group (Figures 4C–E, Formula for calculating ratio was (Z-Y)/Z. Z means the average number of HCs in each turn of gentamicin + CoPPIX group, 51.4, 58, and 36.6, respectively). To further confirm that the protective effects involved HO-1 activation, the HO-1 inhibitor ZnPPIX (Wong et al., 2011) was applied. The gentamicin + CoPPIX + ZnPPIX group showed no protection by CoPPIX in the apical, middle, and basal turns (Figures 4B–E). We confirmed that HO-1 was required for the protection against gentamicin-induced HC death. No loss of HCs was observed in the CoPPIX and ZnPPIX groups. Furthermore, HC counts showed the same trend in cell density changes from apical to basal turns in various groups (Figures 4C–E).

MitoSOX Red, a redox fluorophore specifically measuring superoxide amounts in the mitochondria, was utilized to investigate the effects of HO-1 on ROS production after gentamicin treatment. Immunofluorescence staining revealed that the amount of ROS in cochlear HCs treated with gentamicin increased to 34.73-times of that in the control group. In the gentamicin +CoPPIX group, the amounts of ROS-positive HCs decreased by 73.55% compared to gentamicin group. The gentamicin + CoPPIX + ZnPPIX group exhibited more ROS-positive HCs than the gentamicin + CoPPIX group. ROS were seldom found in the control, CoPPIX and ZnPPIX groups (**P < 0.01 and *P < 0.05, n = 3) (Figures 5A,B).

To further investigate the pathway, through which gentamicin up-regulated HO-1, Nrf2 inhibitor ML385 was used. In the experiment, the organs of Corti were randomly divided into four groups: the control group, the gentamicin group, the ML385 group only received ML385 (15 μM), and the gentamicin + ML385 group. Immunoblot showed a significant increase in Nrf2 and HO-1 levels after 0.6 mM gentamicin. Nrf2 and HO-1 levels were not suppressed significantly in the organs of Corti following treatment with Nrf2 inhibitors ML385. However, compared with the gentamicin group, ML385 inhibited both Nrf2 and HO-1 induced by gentamicin (***P < 0.001 and *P < 0.05, n = 4) (Figures 6A,B). Immunofluorescence staining revealed that the ML385 + gentamicin group exhibited more ROS-positive HCs than the gentamicin group. ROS were seldom found in the control and ML385 groups (***P < 0.001, n = 3) (Figures 6C,D). These suggested that disrupting Nrf2 activity prevents the upregulation of HO-1 and the protection after gentamicin treatment.

Nrf2 transcriptionally activates several antioxidant genes, including NAD(P)H: quinone oxidoreductase 1 (NQO1) and gamma- glutamate cysteine ligase catalytic subunit (GCLC). mRNA levels of HO-1, Nrf2, and Nrf2 target genes were analyzed by qPCR. CoPPIX treatment of the organs of Corti remarkably up-regulated HO-1 mRNA expression. Further, mRNA levels of Nrf2, NQO1 and GCLC increased in response to CoPPIX treatment (***P < 0.001, *P < 0.05, n = 3) (Figure 7A). CoPPIX and ZnPPIX treatment of the organs of Corti for 12 h were analyzed for HO-1 and Nrf2 proteins by Western blot. In the nucleus, two HO-1 immunoreactive bands were observed upon Western analysis, with one band migrating at 28 kDa and the other migrating at 32 kDa. In the nucleus CoPPIX treatment of the organs of Corti, 32 kDa HO-1 and 28 kDa HO-1 were up-regulated at the same time. ZnPPIX treatment inhibited 28 kDa HO-1 but not 32 kDa HO-1 compared with the control group. There were no significant changes in nuclear Nrf2 whether CoPPIX or ZnPPIX treatment. After CoPPIX treatment, HO-1 in the cytoplasm appeared a band at 32 kDa and a faint band at 28 kDa, which was accompanied with a simultaneous increase of Nrf2. ZnPPIX treatment inhibited 32 kDa HO-1 in the cytoplasm significantly. However, neither 28 kDa HO-1 nor Nrf2 has changed significantly after ZnPPIX treatment (*P < 0.05, **P < 0.01, n = 3) (Figures 7B,C). Immunofluorescence staining of HEI-OC1 cells verified that the expression of Nrf2 was induced by the HO-1 induction. In addition, Nrf2 was mainly accumulated in the cytoplasm (Figure 7D). These data suggested that HO-1 may modulate the accumulation of Nrf2.