RGS17 mice with LoxP site gene modifications were crossed with tamoxifen-induced Atoh1-CreER mice from Brandon Cox, Ph.D. (SIU School of Medicine), to generate inducible hair cell-specific RGS17 knockout. We have generated C57BL/6J-RGS17^loxP/loxP mouse by flanking RGS17 exon 2 in the hair cells with LoxP sites, Jackson Laboratory (Bar Harbor, ME). No such line is available commercially or from private sources. Heterozygous C57BL/6J-RGS17^loxP/+ mice were mated to generate homozygous C57BL/6J-RGS17^loxP/loxP mice which were subsequently interbred with Atoh1-CreER to generate C57BL/6J:Atoh1-CreER: RGS17^loxP/+ . These C57BL/6J:Atoh1-CreER: RGS17^loxP/+ mice were then mated with RGS17^loxP/loxP or RGS17^loxP/+ to generate C57BL/6J:Atoh1-CreER: RGS17^loxP/+ or C57BL/6J:Atoh1-CreER: RGS17^loxP/loxP ( Figure 1A ). Both male and female mice were used in all studies. Animal protocols were approved by the Southern Illinois University School of Medicine, Institutional Animal Care and Use Committee (IACUC).

To knockout RGS17 in the hair cells, C57BL/6J:Atoh1-CreER: RGS17^loxP/loxP and C57BL/6J:Atoh1-CreER: RGS17^loxP/+ mice were injected with tamoxifen [3 mg/40 g intraperitoneally (IP)] at P0 and P1 (20–24 h apart). Control mice did not receive tamoxifen injection and were kept separately from the tamoxifen-treated mice.

Tail snip was collected from mice at P10 and used to genotype mice by TransnetYX company (Cordova, TN). In brief, DNA was extracted from the tail snip, and RGS17 null was determined using the following primers: sense 5′-CGTATAGCATACATTATACGAAGTTA TGTTGCGTT-3′, antisense 5′-CAGAAAATATTAAGTCATGAAGAGCCTGG-3′, and probe 5′-AAGAGAAACACGGTTAGACA-3′. Atoh1-CreER was identified using the following primers: sense 5′-TTAATCCATATTGGCAGAACGAAAACG-3′, antisense 5′-CAGGCTAAGTGCCTTCTCTACA-3, and probe 5′-CCTGCGGTGCTAACC-3.

Mice were housed in accordance with the standards established by the Division of Laboratory Animal Medicine (DLAM) facility of SIU School of Medicine. They were kept in a controlled environment with a 12:12-h light:dark cycle, with ad libitum access to food and water. Each group consisted of eight animals. Mice were anesthetized using a mixture of ketamine (90 g/kg) and xylazine (17mg/kg), administered intraperitoneally, and the depth of anesthesia was confirmed by the absence of reflex to the toe pinch. Auditory brainstem responses (ABRs) were examined in a soundproof chamber. Cisplatin (3.5 mg/kg) or equivalent volumes of vehicle were injected for two cycles, each cycle consisted of a 4-day injection followed by a 10-day recovery period (the cumulative dose is 28 mg/kg). Pre-ABRs were recorded on day 0, while post-ABRs were recorded on days 14 and 28. The mice used in our experiments were 6 weeks old at the time of study started.

High-Frequency Intelligent Hearing Systems (HIS) was used to conduct ABRs in mice as described previously (34). Mice were anesthetized using a ketamine/xylazine mixture and placed in a soundproof chamber, with electrodes placed as follows: negative electrodes were inserted under the pinna of each ear, a ground electrode was inserted in the hind flank muscle, and the positive electrode was inserted between the two ears at the vertex in the skull. Earphones positioned in each ear provided a source for sound impulses. The acoustic stimuli were applied as tone bursts at 8, 16, and 32 kHz with a 5-ms plateau and 1-ms rise/fall time at a rate of 5/s. The impulse intensity ranged from 10 dB sound pressure level (SPL) to 90 dB SPL, with 10 dB increments. ABR threshold was defined as the lowest intensity capable of evoking a reproducible and visually detectable response of wave II/III complex. ABRs were recorded from wild-type (RGS17 ^+/+), inducible hair cell-specific RGS17 knockdown (RGS17 ^+/−), and inducible hair cell-specific RGS17 knockout (RGS17^−/−) mice after one and two cycles of vehicle or cisplatin treatment.

Cochleae were extracted from the temporal bone and perfused with 4% paraformaldehyde through the oval and round windows. They were then stored overnight at 4°C in the same solution. The cochleae were decalcified in 120 mM EDTA for 72 h at room temperature with constant stirring, and the EDTA solution was changed every 24 h. After decalcification, they were washed three times in PBS for 5 min each. Microdissection was then performed to isolate the cochlear turns (base, middle, and apex) for immunohistochemistry.

To prepare mid-modiolar sections, cochleae were removed from the EDTA solution and washed thrice in PBS for 5 min each. The cochleae were then submerged in 10% of sucrose with gentle rotation at room temperature for 30 min. Next, they were transferred to 15% sucrose and incubated at room temperature for 30 min before being moved to 4°C, where they were kept on rotation overnight. The cochleae were subsequently transferred to a 1:1 solution of 15% sucrose and OCT, incubated at 4°C overnight with rotation, and frozen at −80°C after being positioned in a cryomold.

Tile scans of cochlear whole mounts, apical, middle, and basal turns were imaged using a Zeiss confocal microscope. The same setting parameters were used for all images for counting OHCs and ribbon synapses. Myosin VIIa antibody was used to stain hair cells, whereas CtBP2 and GluR2 antibodies were used to stain presynaptic ribbons and postsynaptic glutamate receptors, respectively. The nucleus was stained with DAPI. Missing OHCs were manually counted in each cochlear turn using ×20 magnification, with the data presented as percentage loss relative to the total number of OHCs. Ribbon synapses from inner hair cells (IHCs) were imaged using ×60 magnification to include at least 12 IHCs per image. Functional (paired) ribbon synapses were defined as those possessing both CtBP2 + GluR2 immunolabeling apposing each other, whereas orphan (unpaired) synapses were defined as either CtBP2 or GluR2 puncta alone. Ribbon synapses were counted manually from three random microscopic fields for each turn and presented as the number of synapses per IHC.

Immunolabeling for whole-mount and mid-modiolar sections was initiated by incubating the samples in ice-cold 100% methanol for 10 min at −20°C. Following this, the sections were blocked with blocking buffer (10% normal horse or goat serum, 1% BSA, and 1% Triton X-100) for 2 h at room temperature. The sections were then incubated with primary antibodies diluted with antibody dilution buffer as indicated in Table 1 and incubated overnight at 4°C. On the next day, sections were washed thrice with 1× PBS and incubated with secondary antibodies as indicated in Table 1 . Afterward, the sections were then counterstained with Hoechst (DAPI) (1:2,000) at room temperature for 20 min and mounted with Prolong^® Diamond Antifade Mountant (Invitrogen, Saint Louis, Missouri). All the sections were imaged using the same laser illumination settings for all the groups (n ≥ 6 cochleae/group). The tissue sections were imaged using a Zeiss LSM 800 confocal microscope (Zeiss Inc., USA). All the images were processed using Zen 2.5 (blue edition). The immunofluorescent intensity was measured using ImageJ software (version/Fiji). Briefly, intensity values were obtained by selecting the region of interest (ROI) for three samples per group. The nearest region to the ROI that exhibited no fluorescence was selected to determine the background intensity. This background value was subtracted from the ROI value to calculate the RGS17 intensity. The results were then presented as a percentage of the control group, with the control set at 100%.

Data are presented as mean ± standard error of the mean (SEM). Statistical significance of differences among groups used one-way analysis of variance (ANOVA) depending on the experiments, followed by Bonferroni’s multiple comparison test using GraphPad Prism version 6.07 for Windows. P-value <0.05 was considered significant.

The deletion of RGS17 in the hair cells was evaluated by immunolabeling of hair cells in the organ of Corti with the anti-RGS17 antibody. The deletion of RGS17 in the inducible hair cell-specific RGS17 knockout was verified by the absence of RGS17 immunolabeling in the cochlear hair cells, while the wild type showed immunolabeling of hair cells with anti-RGS17 ( Figure 1B ). We have designated the wild-type mice as (RGS17^+/+), the inducible hair cell-specific RGS17 knockdown mice as heterozygous (RGS17^+/−), and the inducible hair cell-specific RGS17 knockout mice as homozygous (RGS17^−/−). We did not use globally manipulated RGS17 genetic mouse strains. All the mice described in the manuscript were either inducible hair cell-specific RGS17 knockdown (RGS17^+/−), knockout (RGS17^−/−), or wild type (RGS17^+/+).

Previously, our laboratory has shown that RGS17 knockdown in the inner ear using siRNA attenuates cisplatin ototoxicity (18). To validate the integral role of RGS17 in cisplatin-induced hearing loss, wild-type mice, as well as inducible hair cell-specific RGS17 knockdown (RGS17^+/−) and inducible hair cell-specific RGS17 knockout mice (RGS17^−/−), were injected with cisplatin (3.5 mg/kg IP) or vehicle over two cycles ( Figure 1C ). RGS17 antibody was used to examine the expression of RGS17 in the hair cells. We observed increased RGS17 immunoreactivity in OHCs and IHCs in RGS17^+/+ (wild-type) mice treated with cisplatin compared with RGS17^+/+ treated with vehicle. Inducible hair cell-specific RGS17 knockdown (RGS17^+/−) mice showed a slight increase of RGS17 immunolabeling in the hair cells after cisplatin injection, while no RGS17 immunolabeling was observed in the inducible hair cell-specific RGS17 knockout mice (RGS17^−/− ) following cisplatin administration ( Figure 1D ). In mid-modiolar sections, RGS17^+/+ mice treated with cisplatin showed strong immunolabeling 2of RGS17 in the OC, SV, SGN, and type II spiral ligament compared to RGS17^+/+ treated only with vehicle. Hair cells heterozygous RGS17^+/− mice showed a weak increase of RGS17 immunolabeling after cisplatin injection, while complete knockout of RGS17^−/− gene in the hair cells showed no increase of RGS17 immunolabeling after cisplatin administration, the immunolabeling of RGS17 in RGS17^−/− mice was similar to that observed in wild type treated with vehicle ( Figure 2A ). RGS17 immunofluorescence intensity for the OC, SV, and SGN in mid-modiolar sections was quantified in RGS17^+/+ mice treated with vehicle and normalized to 100% ( Figure 2B ), while cisplatin administration significantly increased RGS17 fluorescence intensities in RGS17^+/+ mice to 138.2 ± 2.5, 140.5 ± 3.6, and 132.4 ± 6.0 in the OC, SV, and SGN, respectively. In inducible hair cell-specific RGS17 knockdown RGS17^+/− mice treated with vehicle, RGS17 fluorescence intensities were 48.8 ± 1.8, 54.5 ± 1.9, and 49.9 ± 4.4 in the OC, SV, and SGN, respectively, which increased in RGS17^+/− mice following cisplatin administration to 66.3 ± 2.7, 64.5 ± 1.2, and 59.1 ± 5.7 in the OC, SV, and SGN, respectively. Meanwhile, inducible hair cell-specific RGS17 knockout (RGS17^−/− ) mice were significantly protected against cisplatin-induced RGS17 expression, as detected by immunolabeling. The fluorescence intensities of RGS17 in homozygous RGS17^−/− mice treated with cisplatin were 18.3 ± 2.0, 58.3 ± 4.1, and 17.5 ± 3.2 in the OC, SV, and SGN, respectively, compared to RGS17 fluorescence intensities in RGS17^−/− mice treated with vehicle which were much lower at 17.9 ± 0.8, 58.5 ± 6.3, and 11.4 ± 2.2 in the OC, SV, and SGN, respectively ( Figures 2A, B ). The fluorescence intensity in the organ of Corti in homozygous RGS17^−/− likely reflects RGS17 immunolabeling of supporting Deiters cells (DCs) ( Figure 2B ). Higher magnification images of RGS17 in the OC, SV, and SGN are in the Supplementary Figures S1A–C . In the organ of Corti, wild-type RGS17^+/+ mice treated with vehicle displayed immunolabeling in OHCs, IHCs, and DCs, which demonstrated the highest degree of RGS17 immunolabeling in the OHCs, IHCs, and DCs following two cycles of cisplatin injection; RGS17^+/− mice showed less expression; and RGS17^−/− mice showed the least or non-immunolabeling and therefore the greatest protection against cisplatin-induced RGS17 expression in OHCs, IHCs, and DCs ( Supplementary Figure S1A ). Similar trends were observed in the SV and SGN for RGS17^+/+, RGS17^+/−, and RGS17^−/− mice ( Supplementary Figures S1B, C ).

Previously, we have shown that CXCL1 increases in male Wister rat cochlea after cisplatin administration (11 mg/kg IP) (4). To test whether RGS17 regulates CXCL1 expression, wild-type RGS17^+/+, heterozygous (RGS17^+/−), and homozygous (RGS17^−/−) mice were treated with either vehicle or cisplatin (3.5 mg/kg) for two cycles ( Figure 1C ). The cochleae were then collected and processed for whole-mount dissection and immunolabeled with CXCL1 antibody. CXCL1 immunoreactivity was low in the OHCs of RGS17^+/+ mice. However, cisplatin administration increased CXCL1 immunolabeling in the organ of Corti cells (OHCs and IHCs). CXCL1 immunolabeling was ameliorated in RGS17^+/− and RGS17^−/− mice following cisplatin injection ( Figure 3A ). Thus, depletion of RGS17 prevented cisplatin-induced CXCL1 expression. The fluorescence intensity of CXCL1 immunolabeling was expressed as a percentage in the control, which was set at 100%. Cisplatin administration significantly increased CXCL1 fluorescence intensity in the hair cells of RGS17 ^+/+ mice (224.6 ± 6.1). However, in RGS17 ^+/− and RGS17 ^−/− mice, CXCL1 fluorescence intensity was not highly elevated following cisplatin treatment, remaining at 45.0 ± 0.4 and 47.4 ± 2.9, respectively, compared to untreated RGS17 ^+/− and RGS17 ^−/− mice where CXCL1 fluorescence intensity was 15.2 ± 0.4 and 8.1 ± 0.5, respectively ( Figure 3B ).

CD45 (leukocyte common antigen) is a receptor-linked tyrosine phosphatase present on white blood cells, playing an important role in the activation of T lymphocyte, neutrophil, and macrophage adhesion. We used a CD45 antibody to label the immune cells in the cochlea, as previously described (4). CD45 immunostaining was increased in the SGN and SV of RGS17^+/+ mice treated with cisplatin, while CD45 immunolabeling was suppressed in both inducible hair cell-specific RGS17 heterozygous RGS17^+/− and homozygous RSG17^−/− mice after cisplatin injection for two cycles ( Figure 4A ). The fluorescence intensities of CD45 immunolabeling in RGS17^+/+ mice significantly increased following cisplatin administration, reaching 293 ± 25.8, 357 ± 10.8, and 116.4 ± 4.1 in the SGN, SVA, and OC, respectively, compared to the untreated wild-type group, where CD45 fluorescence intensities were normalized to 100%. The fluorescence intensities of CD45 in RGS17^+/− mice were not induced following cisplatin treatment, remaining at 102.8 ± 1.7, 102.8 ± 1.8, and 99.8 ± 1.4 in the SGN, SVA, and OC, respectively. This was consistent with untreated RGS17 ^+/− where CD45 fluorescence intensities were 102.3 ± 1.5, 103.3 ± 1.25, and 101.9 ± 2.3 in the SGN, SVA, and OC, respectively. Similarly, CD45 fluorescence intensities in RGS17^−/− mice were not induced following cisplatin treatment, remaining at 103.4 ± 1.5, 103.5 ± 1.5, and 98.9 ± 0.8 in the SGN, SVA, and OC, respectively. This was compared to untreated RGS17 ^+/−, where CD45 fluorescence intensities were 102.3 ± 1.6, 103.0 ± 1.8, and 98.3 ± 1.2 in the SGN, SVA, and OC, respectively ( Figure 4B ).

We also observed an increase in CD68 immunolabeling in the SGN and SV of RGS17^+/+ mice treated with cisplatin. However, both inducible hair cell-specific RGS17 heterozygous RGS17^+/− and homozygous RGS17^−/− mice showed no immunolabeling of the CD68 immune cell marker in the SGN after cisplatin ( Figure 4C ). The fluorescence intensities of CD68 immunolabeling in RGS17^+/+ mice were significantly increased after cisplatin administration, measuring 155 ± 3.6, 149.1 ± 5.2, and 124.5 ± 6.2 in the SGN, SVA, and OC, respectively, compared to the untreated wild-type group, where CD68 fluorescence intensity was normalized to 100%. In contrast, the fluorescence intensities of CD68 in RGS17^+/− mice were not increased following cisplatin treatment, remaining at 97.1 ± 3.9, 105 ± 2.6, and 106.5 ± 7.5 in the SGN, SVA, and OC, respectively, compared with untreated RGS17 ^+/−, which exhibited CD68 fluorescence intensities of 104.4 ± 4.4, 103.5 ± 1.3, and 108.8 ± 8.1 in the SGN, SVA, and OC, respectively. Similarly, the fluorescence intensities of CD68 in RGS17^−/− mice were not induced after cisplatin treatment, remaining at 101.7 ± 5.5, 105.8 ± 1.6, and 103.9 ± 7.3 in the SGN, SVA, and OC, respectively. This was consistent with untreated RGS17 ^+/− where CD68 fluorescence intensities were 103.3 ± 3.3, 103.9 ± 1.1, and 104.6 ± 9.9 in the SGN, SVA, and OC, respectively ( Figure 4D ). These data suggest that specific deletion of the RGS17 gene in hair cells reduces cochlear proinflammatory CXCL1 levels and prevents the entry of CD45 and CD68-positive immune cells into the cochlea. Higher magnification images were shown in Supplementary Figures S2A–C , S3A–C .

To assess whether RGS17 gene deletion in hair cells protects against cisplatin-induced hearing loss, we performed a functional study comparing wild-type mice with inducible hair cell-specific RGS17 knockdown (RGS17^+/−) and knockout (RGS17^−/−) mice. All genotypes were treated with either vehicle or cisplatin. A 28-day protocol was used to establish two cycles of cisplatin administration and recovery periods ( Figure 5A ). Pretreatment ABRs were recorded for all mice followed by intraperitoneal injection of either vehicle or cisplatin (3.5 mg/kg) for two cycles. Each cycle comprised 4-day injections followed by a 10-day recovery period. Pre-ABRs were recorded on day 0, while post-ABRs were recorded on day 14 (after the first cycle) and day 28 (after the second cycle, Figure 5A ). Administration of the vehicle produced slight changes in ABRs, but there was no significant difference compared to pretreatment ABRs. RGS17^+/+ mice showed a significant increase in ABR threshold shifts after the first cycle of cisplatin (11.2 ± 1.9, 12.1 ± 1.1, and 15.0 ± 1.9 dB at 8, 16, and 32 kHz, respectively). Similarly, after two cycles of cisplatin injection, ABR threshold shifts in wild-type RGS17^+/+ mice were increased to 23 ± 1.6, 24 ± 1.3, and 26.9 ± 2.0 dB at 8, 16, and 32 kHz, respectively. Meanwhile, inducible hair cell-specific RGS17 knockdown (RGS17^+/−) mice showed smaller ABR threshold shifts at all frequencies tested after cisplatin first cycle, with these values being 3.3 ± 1.5, 2.8 ± 1.6, and 5.0 ± 1.8, respectively, compared to 1.4 ± 0.9, 2.3 ± 1.1, and 2.7 ± 1.3, for RGS17^+/− vehicle-treated mice. ABR thresholds were elevated in RGS17^+/− mice treated with vehicle for two cycles, but these were close to 5 dB at all frequencies tested 4.3 ± 1.9, 5.0 ± 2.1, and 5.3 ± 2.4 dB, respectively. No increase in ABR threshold was observed for RGS17^+/− mice treated with cisplatin for two cycles, with thresholds measured at 4.3 ± 1.9, 5.0 ± 2.1, and 5.3 ± 2.4dB, respectively. This suggests that partial deletion of RGS17 significantly decreases cisplatin-induced ABR threshold shifts ( Figure 5B ). ABRs were also analyzed in inducible hair cell-specific RGS17 knockout (RGS17^−/−) mice treated with either vehicle or cisplatin. ABR thresholds in RGS17^−/− mice treated with vehicle or cisplatin remained the same at 8 kHz for both cycles (1.4 ± 1.3 dB). For 16 kHz, the ABR thresholds for homozygous RGS17^−/− mice treated with two cisplatin cycles were 2.8 ± 1.3 and 2.8 ± 1.3 dB, respectively, compared to the RGS17^−/− mice treated with vehicle, which had thresholds of 1.4 ± 1.0 and 1.6 ± 1.0 dB, respectively. Also, cisplatin-induced alteration in hearing threshold at 32 kHz was abolished by complete knockout of RGS17 in the hair cells, with ABR threshold shifts remaining below 5 dB in both the vehicle- and cisplatin-treated groups ( Figure 5B ). These data suggest that inducible hair cell-specific RGS17 knockout (RGS17^−/−) mice exhibit protective effects against cisplatin-induced ABR threshold. No significant ABR threshold shifts were observed among untreated groups, which included wild-type RGS17^+/+, heterozygous RGS17^+/−, and homozygous RGS17^−/− mice. Heterozygous RGS17^+/− mice treated with one cycle (blue bars) or two cycles (purple bars) of vehicle showed no significant ABR threshold shifts compared to their respective untreated wild-type (RGS17^+/+) mice at 8, 16, and 32 kHz. A similar trend was observed in homozygous RGS17 ^−/− mutant mice, where those treated with one cycle (deep turquoise bars) or two cycles (deep green bars) of vehicle showed no significant ABR threshold shifts compared to their respective untreated wild-type (RGS17^+/+) mice at 8, 16, and 32 kHz. This suggested that RGS17 is a non-factor for the survival or function of hair cells, and the mutant mice were not deaf prior to cisplatin treatments ( Figure 5B ).

To establish a temporal relationship between RGS17 expression and auditory cell apoptosis, expression of RGS17 and cleaved caspase-3 was measured and compared between RGS17 wild-type and knockout animals. After one cycle of cisplatin treatment, RGS17 wild-type mice showed an increased expression of RGS17 and caspase-3, while inducible hair cell-specific RGS17 knockout mice showed no change in RGS17 or caspase-3 expression ( Figures 6A, B ). This suggests that the expression of RGS17 has a temporal relationship with cleaved caspase-3 and the apoptotic pathway. By knocking out RGS17, the caspase-3 apoptotic pathway may be circumvented, thereby protecting auditory cells from apoptosis and reducing cisplatin ototoxicity.

We also investigated whether RGS17 is implicated in cisplatin-induced OHC loss. Using two cycles of cisplatin injection protocol ( Figure 5A ), cochleae were collected for whole-mount dissection. Manual counting of OHCs in wild-type RGS17^+/+ mice showed an increase in OHC loss after cisplatin treatment, with percentage losses at 0.8% ± 0.0, 4.1% ± 0.4 and 32.5% ± 1.5 in the apical, middle, and basal turns, respectively. However, RGS17^+/− mice showed significant protection against cisplatin-induced OHC loss, limiting OHC loss to 0.0% ± 0.0, 0.8% ± 0.0, and 1.1% ± 0.2 from these respective regions, without change in IHC number. Following a similar trend, homozygous RGS17^−/− mice reduced the effect of cisplatin-induced OHC loss, with OHC loss limited to 0.0% ± 0.0, 0.2% ± 0.2, and 0.8% ± 0.0 from these respective regions. These results suggest that inducible hair cell-specific RGS17 reduction (RGS17^+/−) or depletion (RGS17^−/−) can reduce OHC loss after two cycles of cisplatin treatment ( Figures 7A, B ).

We also examined the missing OHCs after one cycle of cisplatin administration. OHC counting from wild-type RGS17^+/+ mice treated with cisplatin showed 3.6% ± 0.2 and 26.1% ± 0.3 loss from the middle and basal turns, respectively, while heterozygous RGS17^+/− significantly reduced the effect of cisplatin to 0.8% ± 0.0 and 1.1% ± 0.2 of OHC loss, without altering the IHC number. Similarly, complete knockout of the RGS17^−/− gene decreased the effect of cisplatin, limiting OHC loss to 0.3% ± 0.2 and 0.3% ± 0.2 from these respective regions ( Supplementary Figures S4A, B ).

Ototoxic drugs, aging, and noise are implicated in the loss of ribbon synapses and SGN (35–39). Cochlear synaptopathy decreases supra-threshold ABR wave I amplitudes and increases latencies as a result of losing synapses between type I SG afferent neurons and IHCs (40). Our laboratory has previously shown that cochlear synaptopathy occurs subsequent to cisplatin treatment (4). To assess synapse integrity, we labeled the presynaptic ribbons with an antibody against C-terminal-binding protein 2 (CtBP2) and the postsynaptic terminals with an antibody against glutamate receptor (GluR2). IHCs were labeled with antibodies against myosin VIIa (blue). Synapses were confirmed by immunolabeling whole-mount sections with these antibodies. Paired synapses were detected as yellow fluorescence [a combination of red (CtBP2) plus green (GluR2)]. We evaluated IHC-synapse for two cycles of cisplatin injection protocol ( Figure 5A ). The average of paired synapses showing both CtBP2 and GluR2 staining per IHC in the basal turn of wild-type RGS17^+/+ mice treated with vehicle was 15.6 ± 0.2, and this number significantly decreased to 5.4 ± 0.2 following two cycles of cisplatin treatment in wild-type RGS17^+/+. Paired synapses were significantly rescued in inducible hair cell-specific RGS17 knockdown (heterozygous, RGS17^+/−) mice after two cycles of cisplatin treatment, with an average of 15.2 ± 0.2. Similarly, inducible hair cell-specific RGS17 depletion (homozygous, RGS17^−/−) significantly prevented the loss of paired synapses after two cycles of cisplatin treatment, reducing the loss to 15.4 ± 0.1 ( Figures 8A, B ). The number of orphan synapses after CtBP2 or GluR2 staining not paired with each other increased from 0.5 ± 0.1 per IHC in wild-type RGS17^+/+ mice treated with vehicle to 5.8 ± 0.5 in wild-type RGS17^+/+ treated with two cycles of cisplatin. Both inducible hair cell-specific RGS17 knockdown and knockout protected against cisplatin-increased orphan synapses, which averaged 0.7 ± 0.1 and 0.5 ± 0.1, respectively ( Figure 8C ).