Results

Gadd45a knockout inhibits cisplatin-induced cell death

Based on DNA topoisomerase II alpha (TOP2A), baculoviral IAP repeat-containing 5 (BIRC5), and GADD45A in platinum resistance and cell survival pathways [45], [46], [47], [48], we measured TOP2Α, BIRC5, and GADD45A protein expression levels after treatment with 25 μmol/L cisplatin at various time points in OC1 cells. All this showed a similar pattern, TOP2Α and GADD45A by an initial increase followed by a decrease, and an increasing trend in BIRC5 (Fig. 1a, ¹: ¹). To determine whether Top2a, Birc5, and Gadd45a genes contribute to cisplatin-induced cell death, we knocked out the expression of these genes using CRISPR-Cas9 and verified the knockouts by Western blotting, confirming the knockout of expression of each gene (¹: ¹). We then assessed cell viability using a CCK-8 assay in Top2a_-/-, Birc5-/-, and Gadd45a-/-_ cells following cisplatin treatment. Top2a or Birc5 knockout did not significantly affect cell viability after cisplatin treatment, whereas Gadd45a knockout significantly increased cell viability (Fig. 1b).

Fig. 1: Gadd45a knockout inhibits cisplatin-induced apoptosis. a OC1 cells were treated with 25 μmol/L cisplatin for different durations, and TOP2A, BIRC5, and GADD45A protein expression levels were visualized by Western blotting analysis. b Cisplatin-induced cell death rates of the Top2a, Birc5, and Gadd45a knockout cell lines. c Schematic illustration of the SD (P3) neonatal rat cochlear explant culture model and representative Western blotting analysis of GADD45A expression following treatment with 25 μmol/L cisplatin for 24 and 48 h. d Immunofluorescence staining of GADD45A protein expression in hair cells after intraperitoneal injection of cisplatin in adult C57BL/6 J mice. Asterisks indicate missing hair cells. Scale bar=20 μm. e Representative flow cytometry plots and statistical analysis of early apoptotic cells in control and Gadd45a knockout cells with or without 25 μmol/L cisplatin treatment for 24 h. f Representative images of TUNEL staining and statistical analysis of TUNEL+ cells in control and Gadd45a knockout cells following treatment with 25 μmol/L cisplatin for 24 h. Scale bar=100 μm. g Representative images of TOM20 staining and quantification of cells with abnormal mitochondrial morphology in control and Gadd45a knockout cells following treatment with 25 μmol/L cisplatin for 24 h. Scale bar=20 μm. h Representative images of Live/Dead staining and quantification of PI+ cells in control and Gadd45a knockout cells following treatment with 25 μmol/L cisplatin for 24 h. Scale bar=100 μm. ∗∗P<0.01, ∗∗∗P<0.001; horizontal bars indicate the groups compared. TOP2A. DNA topoisomerase II alpha; BIRC5. Baculoviral inhibitor of apoptosis repeat-containing 5; GADD45A. Growth arrest and DNA damage-inducible alpha; SD. Sprague–Dawley; P3. Postnatal day 3; PI. Propidium iodide; TUNEL. Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling; TOM20. Translocase of outer mitochondrial membrane 20; DAPI. 4’,6-diamidino-2-phenylindole.

Based on this protective phenotype, we selected GADD45A as a molecular target for further investigation. Additionally, cochlear organotypic cultures derived from SD rat pups were treated with cisplatin for 24 and 48 h. Cisplatin treatment increased GADD45A protein expression (Fig. 1c). To further explore the role of GADD45A expression in cisplatin-induced cochlear hair cell death in vivo, adult C57BL/6 J mice were intraperitoneally injected with cisplatin at a dose of 4 mg/kg, as described in previous studies [5]. After cisplatin treatment, GADD45A protein expression was found to be elevated by immunofluorescence analysis (Fig. 1d). Consistent with the initial screening results, Gadd45a−/− cells exhibited greater resistance to cisplatin across a range of concentrations, maintaining significantly higher viability than control cells (¹: ¹).

Flow cytometry revealed that the proportion of early apoptotic cells decreased from 34.13% in the cisplatin group to 9.87% in the Gadd45a−/− group following cisplatin treatment (Fig. 1e). TUNEL staining further confirmed these findings, with significantly fewer TUNEL+ cells observed in the Gadd45a−/− group following cisplatin treatment than in the cisplatin group (Fig. 1f). Furthermore, translocase of outer mitochondrial membrane 20 (TOM20), a receptor subunit of the TOM complex involved in mitochondrial protein transport, exhibited marked aggregation and abnormal distribution following cisplatin treatment, whereas these alterations were substantially attenuated in Gadd45a−/− cells exposed to cisplatin (Fig. 1g). Live/dead staining further confirmed these findings, with numerous propidium iodide-positive (PI+) dead cells observed in the cisplatin-treated control group, whereas markedly fewer dead cells were detected in the Gadd45a−/− group following cisplatin treatment (Fig. 1h). Statistical analysis revealed that the percentage of PI+ cells decreased from approximately 27.33% in the cisplatin group to about 5.66% in the Gadd45a−/− group following cisplatin treatment (Fig. 1h). Finally, mitochondrial reactive oxygen species (ROS) levels were assessed using MitoSOX red and compared with the cisplatin group, the Gadd45a−/− group following cisplatin treatment exhibited markedly reduced ROS accumulation (¹: ¹).

Gadd45a knockout activates autophagy signaling

To elucidate the mechanism by which Gadd45a knockout protects cells from cisplatin-induced toxicity, we collected protein samples from both the control and Gadd45a_-/-_ groups for Western blotting analysis at various time points after cisplatin treatment (Fig. 2a). The results demonstrated that the expression of phosphorylated H2A histone family member X (γ-H2A.X), a sensitive marker of DNA damage, progressively increased in the control group as the duration of cisplatin exposure increased, indicating increased DNA damage. In contrast, γ-H2A.X induction was markedly attenuated in the Gadd45a_-/-_ group. Additionally, the apoptosis-related protein, cleaved caspase-3, was significantly activated in the control group after 24 h of cisplatin treatment but remained inactive in the Gadd45a_-/-_ group. Consistent with reduced apoptosis, cleaved caspase-3 and BCL2-associated X protein (BAX) expression increased following cisplatin treatment in control cells but showed minimal changes in Gadd45a−/− cells. In contrast, B-cell lymphoma 2 (BCL-2) expression was significantly higher in Gadd45a−/− cells than in control cells (¹: ¹). Notably, the expression of AMP-activated protein kinase (AMPK) and phosphorylated AMPK (p-AMPK) was significantly elevated in the Gadd45a_-/-_ group. To further assess the autophagic activity in Gadd45a−/− cells, we examined the expression of LC3B-II and SQSTM1 (p62). Compared with control cells, Gadd45a−/− cells exhibited increased LC3B-II expression and reduced SQSTM1 (p62) levels, supporting enhanced autophagic activity (Fig. 2b). Moreover, immunofluorescence staining revealed a marked increase in the number of autophagosomes in Gadd45a−/− cells, particularly following cisplatin treatment, supporting enhanced autophagic activity (Fig. 2c).

Fig. 2: Gadd45a knockout activates autophagy signaling. Representative Western blotting images and quantification of GADD45A, γ-H2A.X, cleaved caspase-3, and LC3B-II expression in control and Gadd45a knockout cells following treatment with 25 μmol/L cisplatin for 0, 6, 12, and 24 h. b Representative Western blotting images and quantification of γ-H2A.X, p62, and LC3B-II expression in control and Gadd45a knockout cells following treatment with 25 μmol/L cisplatin for 24 h. c Representative images of tandem fluorescent mRFP-GFP-LC3 staining and quantification of autophagosomes (yellow puncta) in control and Gadd45a knockout cells following treatment with or without 25 μmol/L cisplatin for 24 h. Scale bar=20 μm.∗P<0.05, ∗∗P<0.01, ∗∗∗P<0.001; horizontal bars indicate the groups compared. GADD45A. Growth arrest and DNA damage-inducible alpha; γ-H2A.X. Phosphorylated H2A histone family member X; BAX. BCL-2-associated X protein; BCL-2. B-cell lymphoma-2; AMPK. AMP-activated protein kinase; p-AMPK. Phosphorylated AMP-activated protein kinase; LC3B-II. Microtubule-associated protein 1 light chain 3 beta-II; p62. Sequestosome 1; mRFP-GFP. Monomeric red fluorescent protein-green fluorescent protein; DAPI. 4’,6-diamidino-2-phenylindole.

Gadd45a knockout downregulates the expression of CXCL family genes via NF-κB1 suppression

To investigate transcriptomic alterations associated with Gadd45a depletion, control and Gadd45a_-/-_ OC1 cells were subjected to transcriptome sequencing. Hierarchical clustering analysis demonstrated a clear separation between the two groups, indicating distinct transcriptional profiles (Fig. 3a). Correlation analysis showed higher intra-group correlation within each group and reduced similarity between Gadd45a_-/-and control (Fig. 3b). Differential expression analysis was visualized using a volcano plot, which revealed a set of upregulated and downregulated genes in Gadd45a-/-_ cells compared with control cells (Fig. 3c). GO enrichment analysis showed that differentially expressed genes were predominantly associated with immune-related biological processes (Fig. 3d). KEGG pathway analysis revealed that the downregulated genes were enriched in chemokine-related pathways, including tumor necrosis factor (TNF) and interleukin-17 (IL-17) signaling pathways (Fig. 3e). To further explore the specific genes contributing to this enrichment, we examined the differentially expressed gene set and identified multiple chemokine-related genes with marked changes, including members of the CXCL family (Cxcl1, Cxcl3, and Cxcl10) (Fig. 3f). Based on these gene-level alterations, we focused on the CXCL family as potential downstream targets of Gadd45a depletion.

Fig. 3: Gadd45a knockout downregulates CXCL family gene expression by decreasing expression of transcription factor NF-κB1. a Hierarchical clustering analysis of gene expression data. b Pearson correlation of transcriptome sequencing in control and Gadd45a knockout groups. c Volcano map of transcriptome sequencing in control and Gadd45a knockout groups. d Gene function analysis based on GO revealed the most altered pathways based on transcriptome sequencing in control and Gadd45a knockout groups. e Upregulation and downregulation of pathway activation are exhibited based on KEGG pathway analysis. f Ranking of genes related to the TNF signaling pathway. g Venn diagram of CXCL1/3/10 transcription factors. h Relative mRNA expression of Nf-κb1, Rlea, and Cxcl1/3/10/12 to that of β-actin in control cells compared with Gadd45a knockout cells as determined by qRT-PCR. i Representative Western blotting images and quantification of NF-κB1, RELA, and p-RELA expression in OC1 cells following treatment with 25 μmol/L cisplatin for the indicated times. j Representative Western blotting images and quantification of NF-κB1 expression in control and Gadd45a knockout cells following treatment with 25 μmol/L cisplatin for 24 h. k The expression of CXCL1/3/10 proteins in the control group and Gadd45a knockout group after BFA or cisplatin alone or cotreatment was determined by Western blotting. l Representative Western blotting images and quantification of CXCL1, CXCL3, and CXCL10 expression in concentrated cell culture supernatants. m The expression levels of GADD45A, NF-κB1, RELA, CXCL1, CXCL3, and CXCL10 in control and Gadd45a knockout groups were detected by Western blotting. n Volcano map of protein sequencing in the control and Gadd45a knockout groups. o Heatmap of protein sequencing in the control and Gadd45a knockout groups. p Up and downregulation of pathway activation were exhibited between the control and Gadd45a knockout group. ∗P<0.05, ∗∗P<0.01, ∗∗∗P<0.001; horizontal bars indicate the groups compared. GADD45A. Growth arrest and DNA damage-inducible alpha; Gadd45a−/−. Gadd45a knockout; NF-κB1. Nuclear factor kappa B subunit 1; RELA. RELA proto-oncogene, NF-κB subunit; p-RELA. Phosphorylated RELA; CXCL. C-X-C motif chemokine ligand; KEGG. Kyoto Encyclopedia of Genes and Genomes; GO. Gene Ontology; BFA. Brefeldin A.

To determine whether Gadd45a functions as a transcription factor, we searched for its binding motifs and structural domains in the JASPAR database. No DNA-binding motif or transcription factor-binding profile for GADD45A was identified in the JASPAR database. However, the coordinated downregulation of Cxcl1/3/10 following Gadd45a knockout suggests the existence of a shared upstream transcriptional regulator. Analysis of predicted transcriptional regulators for CXCL1, CXCL3, and CXCL10 identified 50 shared transcription factors (Fig. 3g). Among the top 3 transcription factors ranked by their scores, all were related to the NF-κB signaling pathway, prompting us to focus on whether NF-κB mediates the Gadd45a knockout-induced reduction in CXCL family gene expression (¹: ¹). Correspondingly, we validated decreased mRNA expression of Nf-κb1 and CXCL family genes after Gadd45a knockout, whereas Rela expression was increased (Fig. 3h). Similar trends were observed in cells treated with cisplatin for 24 h (¹: ¹). The protein expression of NF-κB1, RELA, and p-RELA also gradually increased after cisplatin treatment (Fig. 3i; ¹: ¹). Notably, after 24 h of cisplatin treatment, Gadd45a knockout reduced NF-κB1 protein expression, whereas RELA expression remained unchanged (Fig. 3j; ¹: ¹).

Considering that CXCL1/3/10 are secreted proteins and cannot be directly measured through cell lysis, cells were treated with brefeldin A (BFA) (¹: ¹) to block protein secretion and facilitate intracellular detection of CXCL proteins. Following BFA treatment, intracellular CXCL1, CXCL3, and CXCL10 proteins accumulated in control cells but remained barely detectable (Fig. 3k; ¹: ¹), indicating that Gadd45a knockout inhibits CXCL protein expression. ELISA analysis revealed that CXCL1 secretion was markedly reduced in Gadd45a−/− cells under both basal and cisplatin-treated conditions (¹: ¹).

To further evaluate secreted CXCL family proteins, concentrated culture supernatants and cell lysates were analyzed separately. CXCL1, CXCL3, and CXCL10 protein levels were increased in the supernatant of cisplatin-treated control cells. Still, they remained barely detectable in Gadd45a−/− cells regardless of cisplatin treatment (Fig. 3l). Consistent with the preceding findings, Gadd45a knockout was associated with reduced NF-κB1 expression and concomitant decreases in CXCL1, CXCL3, and CXCL10 protein levels, whereas RELA expression remained largely unchanged (Fig. 3m; ¹: ¹). To further characterize changes in the secreted proteome, culture supernatants were subjected to proteomic analysis. Volcano plot analysis identified CXCL1 as one of the most prominently altered proteins in the secretome (Fig. 3n). Heatmap analysis revealed distinct protein expression patterns between control and Gadd45a−/− samples (Fig. 3o). A Venn diagram showed the distribution of functional annotation results across different databases. Functional annotation analysis revealed overlapping and database-specific annotations among GO, KEGG, eukaryotic orthologous groups (KOGs), and interPro (IPR) datasets (¹: ¹). KEGG pathway analysis revealed that the downregulated pathways were mainly related to viral protein interactions with cytokines and cytokine receptors, as well as the TNF signaling pathway (Fig. 3p). Gene set enrichment analysis (GSEA) revealed that extracellular protein enrichment decreased when Gadd45a was knocked out (¹: ¹).

To integrate the transcriptomic and proteomic datasets, we performed RNA-protein intersection analysis. The analysis revealed intersecting elements, 32 of which were upregulated and 31 of which were downregulated (Fig. 4a). KEGG pathway analysis revealed that the intersecting downregulated pathways were predominantly associated with inflammatory and immune responses, including chemokine signaling, cytokine-cytokine receptor interaction, TNF signaling, IL-17 signaling, and NF-κB signaling (Fig. 4b). In contrast, the intersecting upregulated pathways were mainly involved in RNA metabolism and repair-related processes, such as RNA degradation and mismatch repair. Among the downregulated overlapping molecules, CXCL1, CXCL12, colony stimulating factor 1 (CSF1), C-C motif chemokine ligand 2 (CCL2), CXCL10, and lipopolysaccharide-binding protein (LBP) exhibited consistent decreases at both the transcriptomic and proteomic levels (Fig. 4c). Immunofluorescence staining further demonstrated that Gadd45a knockout attenuated cisplatin-induced NF-κB1 nuclear translocation, whereas RELA localization remained largely unchanged (Fig. 4d). Together, these findings suggest that GADD45A regulates inflammatory signaling through NF-κB1-dependent mechanisms.

Fig. 4: Gadd45a knockout affected NF-κB1 localization and its degradation by the lysosomal pathway. a Co-upregulated and -downregulated genes identified by RNA-protein intersection analysis. b KEGG pathway enrichment analysis of co- upregulated and downregulated genes identified by transcriptome-proteome intersection analysis. c Heatmaps showing the expression profiles of downregulated inflammatory cytokines at the transcriptomic (upper panel) and proteomic (lower panel) levels. d Representative image of NF-κB1 and RELA staining of control and Gadd45a knockout cell lines after treatment with 25 μmol/L cisplatin for 24 h. Scale bar=20 μm. e NF-κB1 protein expression after 10 μmol/L MG132 treatment for 1, 2, 4, 8, and 12 h. f NF-κB1 protein expression after 10 μmol/L MG132 treatment and 25 μmol/L cisplatin treatment. g NF-κB1 protein expression after 10 μmol/L MG132 treatment for 4 h and 8 h. h NF-κB1 protein expression in control and Gadd45a knockout groups treated with CQ at different concentrations and different time points. i NF-κB1 protein expression after 10 μmol/L MG132 treatment, 20 μmol/L CQ, and 25 μmol/L cisplatin treatment. j Protein expression of ubiquitination levels of NF-κB1 after treatment with MG132, CQ, and cisplatin. NF-κB1. Nuclear factor kappa B subunit 1; RELA. RELA proto-oncogene, NF-κB subunit; CXCL. C-X-C motif chemokine ligand; RNA. Ribonucleic acid; KEGG. Kyoto Encyclopedia of Genes and Genomes; GO. Gene Ontology; CQ. Chloroquine; MG132. Carbobenzoxy-Leu-Leu-leucinal; IP. Immunoprecipitation.

GADD45A maintains NF-κB1 stability by preventing its lysosomal degradation

The process of ubiquitination, a type of post-translational modification involving the covalent binding of ubiquitin molecules to specific target proteins, is essential for maintaining cellular homeostasis. MG132, a well-characterized proteasome inhibitor, directly suppresses proteasomal activity, thereby preventing the degradation of ubiquitinated proteins. NF-κB1 protein levels did not accumulate following treatment with 10 µmol/L MG132 for the indicated durations (Fig. 4e; ¹: ¹), suggesting that NF-κB1 may not be a substrate for proteasomal degradation. To further determine whether the reduction in NF-κB1 induced by Gadd45a deficiency could be rescued by proteasome inhibition, MG132 was applied for 12 h to control and Gadd45a−/− cells (Fig. 4f; ¹: ¹). MG132 treatment failed to increase NF-κB1 expression in either control or Gadd45a−/− cells. To assess whether NF-κB1 protein levels were altered at additional MG132 treatment time points, NF-κB1 expression was examined after MG132 exposure for the indicated durations (Fig. 4g; ¹: ¹). NF-κB1 expression remained largely unchanged, further supporting that proteasome inhibition had limited effects on NF-κB1 protein stability under these experimental conditions.

In addition to proteasomal degradation, lysosomal degradation represents another major pathway. Chloroquine (CQ), a classical lysosomal inhibitor, exerts its effects by increasing intra-lysosomal pH, thereby inhibiting the activity of lysosomal hydrolases. CQ treatment resulted in the accumulation of NF-κB1 protein in control cells and was not observed in the Gadd45a knockout group (Fig. 4h; ¹: ¹). These findings suggest that NF-κB1 is more likely to be degraded through the lysosomal pathway than through the proteasome pathway. We next compared the effects of MG132 and CQ in control and Gadd45a−/− cells. CQ treatment, but not MG132 treatment, resulted in marked accumulation of NF-κB1 protein in control cells (Fig. 4i; ¹: ¹). Co-immunoprecipitation (co-IP) analysis showed that CQ treatment increased the ubiquitin signal associated with NF-κB1, whereas MG132 treatment had little effect, and revealed ubiquitinated NF-κB1, with increased ubiquitin signals following CQ treatment and cisplatin treatment (Fig. 4j; ¹: ¹). Combined with the results of MG132 and CQ treatment, these findings indicate that ubiquitinated NF-κB1 is primarily degraded through the lysosomal pathway rather than the proteasomal pathway.

Although Flag_-Gadd45a_ expression was detectable after transfection, it progressively decreased during three consecutive passages and was nearly undetectable by passage 3, preventing the establishment of a stable Gadd45a-overexpressing cell line (¹: ¹-h). Transient overexpression followed by co-IP assays did not detect a stable association of GADD45A with NF-κB1, RELA, or phosphorylated RELA (¹: ¹), suggesting the regulation of NF-κB1 by GADD45A may not involve stable direct binding.

Injection of siGadd45a through the posterior semicircular canal reduces cisplatin-induced ototoxicity

Given the protective effects of Gadd45a deficiency observed in vitro, we next evaluated whether siRNA-mediated Gadd45a knockdown could alleviate cisplatin-induced ototoxicity in vivo. Accordingly, Gadd45a-targeting siRNA was delivered into the mouse inner ear through the PSC. The procedure for PSC injection is shown in ¹: ¹. Following injection of 5’-FAM-labeled siRNA, strong fluorescence signals were detected in the dissected inner ear specimen, confirming successful delivery of siRNA into the inner ear. In vivo fluorescence imaging demonstrated persistent retention of 5’-FAM-siRNA in the injected ear region for up to 7 d after administration (¹: ¹).

To further improve the efficacy of cochlear Gadd45a expression suppression, we used Invivofectamine 3.0 Reagent, a non-animal-derived lipid nanoparticle. This reagent encapsulates siRNA for in vivo delivery, facilitating efficient uptake by living mouse cells without eliciting toxicity or stress responses. Western blotting (¹: ¹) and immunofluorescence staining (¹: ¹) confirmed that siRNA treatment significantly reduced Gadd45a expression in the cochlea.

To evaluate the efficacy of si_Gadd45a_ knockdown in mitigating cisplatin ototoxicity, we first performed audiometry on mice to establish a baseline for subsequent experiments. Mice with qualified baseline audiometric profiles were selected for si_Gadd45a_ injection via the PSC. On the second day, the mice were intraperitoneally injected with cisplatin. Subsequent audiometry was performed (Fig. 5a), followed by euthanasia and cochlear basement membrane immunofluorescence staining. si_Gadd45a_ treatment significantly reduced cisplatin-induced ABR threshold shifts and preserved OHCs, particularly in the middle and basal turns of the cochlea (Fig. 5b, c). To investigate whether si_Gadd45a_ injected into the cochlea could alleviate cisplatin-induced ototoxicity by modulating the inflammatory signaling pathway, we examined the expression of proteins associated with the inflammatory signaling pathway. Our results demonstrated that si_Gadd45a_ effectively reduced the expression of the transcription factors NF-κB1(Fig. 5d), which subsequently led to the downregulation of the expression of the inflammatory cytokines CXCL1, CXCL3, and CXCL10 (Fig. 5e). Furthermore, immunofluorescence staining showed increased NF-κB1 and CXCL3 expression in the cochleae of cisplatin-treated mice, whereas this increase was reduced following si_Gadd45a_ administration (Fig. 5f, g). GADD45A immunostaining showed markedly reduced fluorescence intensity in the cochlea following si_Gadd45a_ treatment. The knockdown efficiency of si_Gadd45a_ in the cochlea was confirmed by GADD45A immunostaining (¹: ¹).

Fig. 5: Injection of siGadd45a through the PSC reduces cisplatin-induced ototoxicity. a Single cycle cisplatin injection pattern in mice. b ABR threshold shift results before and after cisplatin injection. **P<0.01, ***P<0.001 vs. cisplatin group; ###P<0.001 vs. sicontrol group. c Representative immunofluorescence images and quantification of outer hair cell (OHC) survival following cisplatin administration. Scale bar=20 μm. d Representative Western blotting images and quantification of GADD45A, NF-κB1, and RELA expression in cochlear tissues following cisplatin administration. e Representative Western blotting images and quantification of CXCL1, CXCL3, and CXCL10 expression in cochlear tissues from control and siGadd45a-treated mice following cisplatin administration. f Immunofluorescence staining of NF-κB1 in the cochlea. Scale bar=100 μm. g Immunofluorescence staining of CXCL3 in the cochlea. Scale bar=100 μm. ∗P<0.05, ∗∗P<0.01, ∗∗∗P<0.001; horizontal bars indicate the groups compared. ABR. Auditory brainstem response; SPL. Sound pressure level; i.p. Intraperitoneal injection; NF-κB1. Nuclear factor κB subunit 1; RELA. RELA proto-oncogene, NF-κB subunit; CXCL. C-X-C motif chemokine ligand; GAPDH. Glyceraldehyde-3-phosphate dehydrogenase; SOX2. SRY-box transcription factor 2; TUJ1. Class III beta-tubulin; DAPI. 4’,6-diamidino-2-phenylindole.

Gadd45a knockout alleviates cisplatin-induced ototoxicity in vivo

A Gadd45a knockout allele was generated by inserting LoxP sites flanking exons 1–3. Cre-mediated recombination excises this region and disrupts the Gadd45a coding sequence (¹: ¹). Gadd45a cKO mice underwent two cycles of cisplatin treatment, with each cycle consisting of 4 consecutive days of cisplatin administration followed by a 2-day recovery period (Fig. 6a). Notably, compared with saline-treated mice, cisplatin-treated mice presented a substantial degree of body weight loss, regardless of whether they were conditional gene knockout mice (¹: ¹). We utilized immunofluorescence staining of cochlear cryosections to assess the efficiency of Gadd45a knockout. Since the GADD45A protein is predominantly localized to the nucleus, the absence of GADD45A in the nuclei of hair cells following Gadd45a knockout was evident. In contrast, GADD45A expression remained intact in the nuclei of supporting cells expressing SRY-box transcription factor 2 (SOX2) (¹: ¹). To determine whether Gadd45a knockout affects cisplatin-induced hearing loss, ABRs were measured in both control and Gadd45a-/- mice. Baseline auditory function was determined by conducting auditory tests before cisplatin pretreatment. After the second cycle of cisplatin injection, ABR threshold shifts were significantly lower in cisplatin-treated Gadd45a knockout mice than in cisplatin-treated mice (Fig. 6b). Representative ABR waveform images at 32 kHz are shown in ¹: ¹.

Fig. 6: Gadd45a knockout can alleviate cisplatin ototoxicity. a Double cycle cisplatin intraperitoneal injection pattern in mice. b ABR test results of baseline levels of hearing and after cisplatin injection. c Representative immunofluorescence images of cochlear hair cells and quantification of outer hair cell (OHC) and inner hair cell (IHC) survival following cisplatin administration. Scale bar=20 μm. d Representative images of synaptic immunofluorescence staining and quantification of IHC survival and paired synapses per IHC. Scale bar=20 μm. *P<0.05, **P<0.01, ***P<0.001 vs. cisplatin group; #P<0.05, ###P<0.001 vs. control group (b, c); *P<0.05, **P<0.01, ***P<0.001 (d). ABR. Auditory brainstem response; SPL. Sound pressure level; i.p. Intraperitoneal injection; OHC. Outer hair cell; IHC. Inner hair cell; CtBP2. C-terminal binding protein 2; GluR2. Glutamate receptor 2; GADD45A. Growth arrest and DNA damage-inducible alpha; Gadd45a−/−. Gadd45a knockout.

The mice were sacrificed at the end of the hearing test, and the inner ear tissues were collected for immunofluorescence staining (Fig. 6c). The whole cochlea was stained with Myosin VIIa (green) and phalloidin (red) to observe hair cell integrity. In control and Gadd45a knockout mice without cisplatin treatment, the cochlear morphology was normal, with 3 rows of OHCs and 1 row of IHCs aligned and no hair cell loss. Compared with cisplatin mice, Gadd45a knockout mice showed increased OHC survival after cisplatin treatment (Fig. 6c). Notably, Cochlear mapping showed that cisplatin induced substantial OHC loss in control mice, with the greatest loss occurring in the middle-to-basal region, whereas Gadd45a knockout significantly reduced OHC loss at these locations. In contrast, IHC loss remained minimal throughout the cochlea in both cisplatin-treated control and Gadd45a−/− groups (Fig. 6c). Previous literature has shown that although IHCs remain intact after cisplatin treatment, there is a significant loss of specialized ribbon synapses [49]. To evaluate the impact of cisplatin on IHC synapses, whole-mount cochlear sections were stained for CtBP2 to label presynaptic ribbons and GluR2, an AMPA receptor subunit, to label postsynaptic glutamate receptors. The number of colocalized CtBP2 and GluR2 synapses per IHC was quantified and statistically analyzed. Significant synaptic loss was observed in cisplatin mice across all cochlear regions, whereas this reduction was not observed in cisplatin-treated Gadd45a knockout mice (Fig. 6d).

Cisplatin causes damage to the density of SGNs at the apex, whereas no such damage has been observed in the middle or basal regions of the cochlea [50]. To explore the impact of cisplatin on SGNs, cochlear sections were subjected to hematoxylin-eosin (H&E) staining. In our study, quantitative analyses revealed that the SGN density in the apex, middle, or base of the cochlea was not affected by cisplatin. Consistent with the hair cell survival analysis, H&E staining revealed more pronounced morphological alterations in hair cells than in SGNs or the stria vascularis following cisplatin treatment (¹: ¹). Additionally, the thickness of the stria vascularis was measured after cisplatin treatment, revealing that the stria vascularis was damaged, likely due to its role as the primary route for cisplatin entry (¹: ¹).

Footnotes

1. Supplementary data associated with this article can be found in the online version at doi:10.1016/j.mmr.2026.100063. ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶ ↩⁷ ↩⁸ ↩⁹ ↩¹⁰ ↩¹¹ ↩¹² ↩¹³ ↩¹⁴ ↩¹⁵ ↩¹⁶ ↩¹⁷ ↩¹⁸ ↩¹⁹ ↩²⁰ ↩²¹ ↩²² ↩²³ ↩²⁴ ↩²⁵ ↩²⁶ ↩²⁷ ↩²⁸ ↩²⁹ ↩³⁰ ↩³¹ ↩³² ↩³³ ↩³⁴ ↩³⁵ ↩³⁶ ↩³⁷ ↩³⁸ ↩³⁹ ↩⁴⁰ ↩⁴¹ ↩⁴² ↩⁴³ ↩⁴⁴ ↩⁴⁵ ↩⁴⁶ ↩⁴⁷ ↩⁴⁸ ↩⁴⁹ ↩⁵⁰ ↩⁵¹ ↩⁵² ↩⁵³ ↩⁵⁴ ↩⁵⁵ ↩⁵⁶ ↩⁵⁷ ↩⁵⁸ ↩⁵⁹ ↩⁶⁰ ↩⁶¹ ↩⁶² ↩⁶³ ↩⁶⁴ ↩⁶⁵ ↩⁶⁶ ↩⁶⁷ ↩⁶⁸