How the H3K9me2-FOXG1-microRNA Pathway May Protect Ear Hair Cells from Age-Related Hearing Loss by Controlling Cell Cleanup

Male C57BL/6J mice at 5 weeks of age were sourced from SPF Biotechnology Co., Ltd. (Beijing, China). Baseline hearing was confirmed via auditory brainstem response (ABR) measurements. After a 1-week habituation phase, the mice were randomly allocated into two groups: a control group and a D-gal-induced aging group. Mice in the aging group received daily subcutaneous administrations of D-gal (200 mg/kg; #G0625; Sigma-Aldrich) over an 8-week period, while control mice received an equal volume of saline (200 mg/kg/day) under identical conditions. After completion of the injection protocol, all mice were maintained until 27 weeks of age. ABR testing was subsequently performed to evaluate hearing thresholds. Euthanasia was performed via intraperitoneal administration of pentobarbital sodium at a dose of 150 mg/kg. All animal experiments in this study were designed and reported in accordance with the ARRIVE guidelines. All procedures complied with institutional ethical standards and were approved by the Institutional Animal Care and Use Committee of Zhongnan Hospital, Wuhan University (ZN2023226).

The mouse auditory hair-cell line HEI-OC1 was propagated in Dulbecco’s modified Eagle’s medium (DMEM; Gibco) enriched with 10% fetal bovine serum (FBS; Gibco) and maintained at 33°C within a humidified incubator with 5% CO2. Upon attaining 80–90% confluence, cells were detached using 0.25% trypsin-EDTA (Gibco) for subculturing. For injury induction, HEI-OC1 cells were exposed to D-gal (30 mg/mL) for 72 h to mimic aging-associated stress.

After receiving 50 mg/kg pentobarbital sodium administered intraperitoneally for anesthesia, mice were placed on a thermoregulated heating pad within a soundproof enclosure. Subdermal electrodes were positioned at the vertex (recording electrode), the test ear (reference), and the contralateral ear (ground). ABRs were acquired using a Tucker Davies Technologies (TDT, Gainesville, FL, United States) setup. Click and tone-burst stimuli were delivered at frequencies ranging from 4 to 32 kHz (4, 8, 16, 24, and 32 kHz). Stimulus intensity began at 90 dB sound pressure level (SPL) and were reduced in 10-dB increments until Wave II was no longer detectable. The hearing threshold for each frequency was determined as the minimal sound level at which a reproducible ABR waveform could be observed.

Cochlear specimens or HEI-OC1 cells were fixed with 4% paraformaldehyde for 20 min, followed by three 5-min washes with phosphate-buffered saline (PBS). Samples were permeabilized using 2% Triton X-100 for 15 min and nonspecific binding was blocked with 10% donkey serum for 1 h at room temperature. Primary antibodies included those targeting FOXG1 (ab18259; Abcam), LC3B (L7543; Sigma-Aldrich), H3K9me2 (ab176882; Abcam), and Myosin7a (25-6790; Proteintech). Microfilaments were visualized with rhodamine-phalloidin (Yeasen), and nuclei were labeled with DAPI (Solarbio). After overnight incubation with primary antibodies at 4°C, samples were treated with appropriate fluorescent secondary antibodies and DAPI under light-protected conditions. Excess dye was removed by PBS washing, and samples were mounted using an anti-fluorescence quencher. Imaging was conducted using a confocal microscope.

HEI-OC1 cells were seeded into 96-well plates at the appropriate density and exposed to graded concentrations of the indicated agents. After treatment, culture medium was aspirated, and 100 μL of 10% CCK-8 reagent solution was introduced into each well. Following incubation at 37 °C for the specified duration, absorbance at 450 nm was recorded with a microplate reader to assess cellular viability.

Apoptosis was assessed via an Annexin V-FITC/propidium iodide (PI) double-staining assay (BD Biosciences). Following collection, cells were washed twice with PBS, and reconstituted in binding buffer. The cell suspension was stained with Annexin V-FITC and PI, gently mixed, and incubated for 15 min at room temperature under dark conditions before immediate flow cytometric analysis. Fluorescence was detected on a CytoFLEX flow cytometer (Beckman Coulter, Brea, CA, United States) equipped with a 488 nm argon laser. Annexin V-FITC was detected using a 525/40 nm bandpass filter (FL1 channel), and PI was detected using a 585/42 nm bandpass filter (FL2 channel).

To quantify mitochondrial ROS, cells were incubated with the mitochondrial superoxide probe Mito-SOX Red at a final concentration of 500 nM in PBS for 15 min at 37°C in darkness. After washing twice with pre-warmed PBS, cells were resuspended in PBS and fluorescence was immediately detected on a CytoFLEX flow cytometer (Beckman Coulter, Brea, CA, United States) using a 488 nm excitation laser and a 585/42 nm bandpass filter (FL2 channel).

HEI-OC1 cells from the control and D-gal groups were collected separately. Total DNA was extracted using the TIANGEN Cell Genomic DNA Extraction Kit according to the manufacturer’s instructions. The mtDNA common deletion ratio was measured by Taqman PCR. The primers and probes used for TaqMan PCR detection of the D-loop region were as follows: Forward: 5’-GGT TCT TAC TTC AGG GCC ATC A; Reverse: 5’-GAT TAG ACC CGT TAC CAT CGA GAT; probe: 5’-FAM-TTG GTT CAT CGT TAC GTT CCC CTT A-TAMRA. The primers and probes used for TaqMan PCR detection of the common deletion were as follows: Forward: 5’-AAG GAC GAA CCT GAG CCC TAA TA; Reverse: 5’- CGA AGT AGA TGA TCC GTA TGC TGT A; probe: 5’-FAM- TCA CTT TAA TCG CCA CAT CCA TAA CTG CTG T-TAMRA.

HEI-OC1 cells were harvested and fixed in 2.5% glutaraldehyde for 24 h, followed by fixation with 1% osmium tetroxide for 1–2 h. Specimens underwent dehydration through a graded ethanol series before being embedded in Araldite CY 212 epoxy resin (TAAB). Ultrathin sections were cut, sequentially stained with uranyl acetate and alkaline lead citrate, rinsed with distilled water, and visualized using a transmission electron microscope (JEM-1230).

Small interfering RNA (siRNAs) targeting Foxg1 and Foxg1 overexpression plasmid were custom-synthesized by GenePharma Co. (Shanghai, China). MiRNA inhibitors and mimics were acquired from RiboBio Co., Ltd. (Guangzhou, China).

HEI-OC1 cells were plated in 6-well plates and grown to approximately 50–60% confluence before transfection. Transfections were performed using Lipofectamine 3000 (Thermo Fisher Scientific) following the supplier’s instructions. In brief, siRNA (or plasmid DNA), miRNA mimics, or inhibitors were diluted in Opti-MEM, combined with the transfection reagent, and allowed to incubate for 15 min at room temperature. The resulting complexes were gently added to the cells in a dropwise manner. After 6–8 h, the transfection mixture was replaced with fresh complete medium, and cells were cultured for an additional 24 h before subsequent treatments.

Proteins were extracted by lysing cells in pre-chilled radioimmunoprecipitation assay (RIPA) buffer containing phenylmethylsulfonyl fluoride (PMSF), protease inhibitor cocktails II and III, and phosphatase inhibitors (Roche, Basel, Switzerland). Tissue samples from mouse cochleae were homogenized in the same buffer using a mechanical grinder, followed by incubation on ice for 30 min. Lysates were centrifuged at 12,000 × g for 15 min at 4 °C to obtain the supernatants. Protein concentrations were quantified using a bicinchoninic acid (BCA) assay kit (Thermo Fisher Scientific).

Equivalent protein amounts were resolved by SDS-PAGE and electroblotted onto PVDF membranes (Millipore). Membranes were blocked in 5% nonfat milk in TBST for 1 h at room temperature and subsequently incubated overnight at 4°C with the following primary antibodies: anti-H3K9me2 (ab176882; Abcam), anti-FOXG1 (ab18259; Abcam), anti-LC3B (L7543; Sigma-Aldrich), anti-G9a (ab185050; Abcam), and anti-GAPDH antibody (10494-1-AP; Proteintech). Following three washes with TBST, membranes were probed with HRP-conjugated secondary antibodies (Antgene; ANT020) diluted 1:5,000 for 1 h at room temperature. Band intensity was detected with an enhanced chemiluminescence (ECL) kit (E423; Vazyme), and analyzed with ImageJ software.

Total RNA was isolated from HEI-OC1 cells and mouse cochlear tissues with the TRIzol reagent (R411-01; Vazyme) following the supplier’s instructions. Complementary DNA (cDNA) synthesis was carried out using a reverse transcription kit (Takara). qRT-PCR was conducted with the Takara qPCR Master Mix on a real-time PCR detection instrument. GAPDH served as the endogenous reference gene for normalization of target gene expression. All the primer sequences are provided in. Primers sets for miRNAs were purchased from RiboBio Co. (Guangzhou, China). Supplementary Table S1

HEI-OC1 cells (control, D-gal, and si-FOXG1 groups) and mouse cochlear (control and aging groups) were collected for proteomic profiling (Novogene, Beijing, China), with three independent biological replicates per condition. Differentially expressed proteins (DEPs) were defined as those with a p-value < 0.05 and | log2 fold change (FC)| > 1.2. Functional enrichment analyses, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway assessments, were executed using the ClusterProfiler package in R to evaluate the biological roles and signaling networks linked to the identified DEPs.

HEI-OC1 cells from control, D-gal, and si-FOXG1 groups were collected for small RNA sequencing (Novogene, Beijing, China), with triplicate biological samples per group. Total RNA was isolated using the TRIzol, and libraries were prepared and sequenced in accordance with Illumina standard workflows. Differentially expressed miRNAs (DEmiRNAs) were identified via the DESeq2 package in R, applying thresholds of p < 0.05 and | log2 FC| > 1.2. Potential miRNA targets were predicted using both miRanda and RNAhybrid algorithms, and intersecting targets were selected for downstream analysis. Functional enrichment of target genes was conducted through GO and KEGG pathway assessments with the clusterProfiler package in R, applying a significance cut-off of p < 0.05.

HEI-OC1 cells from the control, D-gal, and D-gal combined with si-FOXG1 groups were collected for untargeted metabolomic profiling (Novogene, Beijing, China), with three independent replicates per group. Metabolite annotation was performed using the KEGG, the Human Metabolome Database (HMDB), and the LIPID Metabolites and Pathways Strategy (LIPIDMaps) databases. Differential metabolites were identified based on the following criteria: variable importance in projection (VIP) > 1, p < 0.05, and FC > 1.2. K-means clustering (k = 4) was performed on the differential metabolites.

All data are expressed as mean ± standard deviation (SD). Statistical evaluations were conducted with Microsoft Excel and GraphPad Prism 8. Intergroup comparisons involving two conditions were analyzed using two-tailed unpaired Student’s t-tests, whereas comparisons across three or more groups were assessed by one-way ANOVA with Dunnett’s multiple comparison test. Statistical significance was defined as p < 0.05.

The D-gal-induced aging model is widely used to recapitulate phenotypes characteristic of natural aging. Following established protocols (He et al., 2021; Wang et al., 2023; Chen F. et al., 2024), C57BL/6J mice received daily subcutaneous D-gal injection (200 mg/kg) for 8 weeks; control mice received matched saline (Figure 1A). ABR showed a significant elevation of click thresholds in D-gal-treated mice (Figure 1B) and a significant threshold shift at 32 kHz on tone-burst ABR (Figure 1C). Low-frequency changes trended upward but were not significant (Figure 1C). These findings align with the reports of hearing loss in naturally aging mice. Consistent with functional deficits, immunofluorescence for Myosin7a/Phalloidin/DAPI revealed a significant loss of OHCs in D-gal-treated cochleae (Figures 1D,E).

HEI-OC1 cells, which express multiple inner-ear HC markers, were exposed to D-gal (0.2–50 mg/mL) for 72 h to establish a cell-based aging model (Figure 1F). The CCK-8 assays demonstrated a concentration-dependent loss of viability, with a significant reduction at 30 mg/mL (Figure 1G); this dose and duration were used for subsequent senescence phenotyping. D-gal increased the mtDNA common deletion load (Figure 1H), elevated apoptosis and cell death by Annexin V/PI flow cytometry (Figures 1I–K), and raised mitochondrial ROS by Mito-SOX flow cytometry (Figures 1L,M). TEM revealed canonical senescence-associated ultrastructural changes, including nuclear degeneration, chromatin condensation, cytoplasmic accumulation, and mitochondrial pathology (Figure 1N). Collectively, the D-gal models reproduced key functional and structural features of natural aging.

To examine the molecular alterations induced by D-gal in the cochlea, we analyzed the expression of FOXG1, H3K9me2, and autophagy-related proteins. Western blotting revealed that FOXG1 and LC3 protein levels were significantly reduced in the cochlea of D-gal-treated mice compared with controls (Figures 2A–C). Immunofluorescence confirmed the downregulation of FOXG1 (Figures 2D,E) and LC3 (Figures 2G,I), consistent with western blot results. Because epigenetic dysregulation contributes to age-related oxidative and neurodegeneration processes, we further examined histone methylation changes in the cochlea. H3K9me2, a key epigenetic mark involved in chromatin silencing, transcriptional regulation, and redox-responsive cellular adaptation (Yu et al., 2013; Tang et al., 2016; Xiong et al., 2019), plays an essential role in maintaining euchromatin structure and regulating gene repression. Consistent with its reported role in neurodegenerative diseases, H3K9me2 expression was markedly elevated in the cochleae of D-gal-treated compared with controls (Figures 2F,H). Next, we treated HEI-OC1 cells with graded concentrations of D-gal and examined FOXG1, H3K9me2, and LC3 expression by western blotting. FOXG1 expression increased at 2–15 mg/mL but declined significantly when the D-gal concentration exceeded 15 mg/mL. In contrast, H3K9me2 levels exhibited the opposite trend (Figures 2J–L). LC3 expression paralleled the pattern of FOXG1 (Figures 2M,N), suggesting that autophagic activity followed FOXG1 regulation under redox imbalance. These data collectively indicate that FOXG1, H3K9me2, and autophagy respond dynamically to D-gal-induced aging.

G9a (euchromatic histone-lysine N-methyltransferase 2; EHMT2) and GLP (G9a-like protein; EHMT1) are the primary functional methyltransferases responsible for H3K9me2 methylation. BIX01294, a selective inhibitor of these enzymes, is commonly used to investigate the biological functions of H3K9me2. HEI-OC1 cells were treated with increasing concentrations of BIX01294 (0.5, 1, 2, 3, and 5 μM) for 24 h. CCK-8 assays revealed a dose-dependent decrease in cell viability following BIX01294 treatment (Supplementary Figure S1A). Consistently, flow cytometry demonstrated a concentration-dependent increase in apoptosis and cell death ratios (Supplementary Figures S1B–D). Western blotting showed that the expression of G9a and H3K9me2 decreased progressively with increasing BIX01294 concentrations (Figures 3A–C). To determine whether H3K9me2 affects FOXG1 and autophagy, we analyzed FOXG1 and LC3 expression in BIX01294-treated HEI-OC1 cells. The results revealed a dose-dependent upregulation of both FOXG1 and LC3, suggesting that suppression of H3K9me2 enhances FOXG1 expression and autophagic activity (Figures 3D–F). Collectively, these findings indicate that pharmacological inhibition of G9a by BIX01294 effectively reduces H3K9me2 levels and concomitantly increases FOXG1 and LC3 expression, implying that H3K9me2 exerts an inhibitory effect on FOXG1 and autophagy in HEI-OC1 cells. Subsequent experiments were performed using 2 μM BIX01294 for 24 h.

To elucidate whether H3K9me2 is associated with FOXG1 and autophagy, we pretreated HEI-OC1 cells with BIX01294 for 24 h prior to D-gal exposure and found that BIX01294 attenuated the D-gal-induced reduction in both FOXG1 and LC3 expression levels (Figures 3G–J). We then silenced FOXG1 expression in HEI-OC1 cells using siRNA and examined the levels of FOXG1 and LC3 by western blotting. The results showed that the levels of LC3-II were significantly decreased following FOXG1 knockdown (Supplementary Figures S2A–C). Next, we constructed a FOXG1-overexpression plasmid. Following transfection, western blotting confirmed markedly elevated FOXG1 expression, accompanied by a concomitant increase in LC3-II levels (Supplementary Figures S2D–F). We also found that FOXG1 overexpression mitigated the D-gal-induced suppression of autophagy in HEI-OC1 cells (Figures 3K–M). These findings suggest that FOXG1 positively regulates autophagy during D-gal-induced oxidative injury. To further explore whether H3K9me2 regulates autophagy through FOXG1, we knocked down FOXG1 and treated the cells with BIX01294. Western blotting showed that autophagy was enhanced by BIX01294 treatment alone but markedly reduced when silenced prior to BIX01294 exposure (Figures 3N–P). These results indicate that H3K9me2 regulates autophagy via a FOXG1-dependent mechanism, and FOXG1 knockdown abolishes the BIX01294-induced activation of autophagy.

While our previous studies suggested a connection between FOXG1 and autophagy, the precise molecular mechanism driving this relationship remains to be fully elucidated. MiRNAs are known to regulate different stages of autophagy, and their expression is regulated by transcription factors. We hypothesized that miRNAs act as potential mediators between FOXG1 expression and autophagy. To verify this hypothesis and based on previous publications (Mu et al., 2023), we examined the expression of miRNAs that not only correlate with autophagy but also function in the inner ear, and analyzed their expression in D-gal-induced aging models, including miR-34a, miR-96, miR-182, and miR-183. We extracted RNA from the cochlea, hippocampus, and cortex of D-gal-induced aging mice and detected the expression of these miRNAs via qRT-PCR. Compared with the control group, qRT-PCR analysis revealed a significant downregulation of miR-34a, miR-96, miR-182, and miR-183 in the cochlea of D-gal-induced aging mice (Figure 4A), whereas their expression remained largely unaltered in the hippocampus and cortex (Figures 4B,C). We then silenced FOXG1 expression in HEI-OC1 cells via siRNA transfection and observed a significant reduction in the expression of miR-34a, miR-96, miR-182, and miR-183 (Figure 4D). These results suggest that these autophagy-related miRNAs are positively regulated by FOXG1. We next treated HEI-OC1 cells, pretreated with either 2 μM BIX01294 or FOXG1-overexpression plasmids, with D-gal. The qRT-PCR results showed that D-gal treatment downregulated the expression of miR-34a, miR-96, miR-182, and miR-183; however, pretreatment with BIX01294 or FOXG1-overexpression reversed the D-gal-induced decrease in miRNA expression (Figure 4E). To further investigate the functional role of these miRNAs in regulating autophagy, we inhibited their expression in HEI-OC1 cells using miRNA inhibitors (Supplementary Figure S3A) and measured LC3 expression. Following miRNA inhibition, a significant decline in LC3-II expression was observed (Figures 4F,G), indicating a reduction in autophagic activity.

Next, after inhibiting the four miRNAs in HEI-OC1 cells, we treated the cells with 2 μM BIX01294 and examined FOXG1 and LC3 protein levels. Western blot analysis demonstrated that relative to BIX01294 treatment alone, co-treatment with BIX01294 and miRNA inhibitors resulted in a significant reduction in LC3-II, whereas FOXG1 levels remained unchanged (Figures 4H–J). This suggests that inhibition of these miRNAs suppresses autophagy under BIX01294 treatment, while FOXG1 expression is unaffected by miRNA inhibition. We next overexpressed miR-34a, miR-96, miR-182, and miR-183 in HEI-OC1 cells by transfection with miRNA mimics (Supplementary Figure S3B), followed by D-gal treatment, and examined FOXG1 and LC3 expression. Western blot analysis revealed that, relative to the D-gal group, LC3-II levels were increased in the D-gal + miRNA mimics group, whereas FOXG1 expression remained unaffected by miRNA mimics (Figures 4K–M). These results indicate that overexpression of these four miRNAs restores autophagy under D-gal-induced stress. Subsequently, we used flow cytometry to quantify apoptosis and cell death in HEI-OC1 cells following inhibition of these miRNAs, both with and without D-gal treatment. The proportion of apoptotic cells and cell death increased after inhibition alone and was further elevated under combined miRNA inhibition and D-gal exposure (Figures 5A–C). Consistently, Mito-SOX flow cytometry revealed a similar increase in mitochondrial ROS levels (Figures 5D,E). These observations underscore a critical protective function of these miRNAs in preserving HEI-OC1 cell viability and mitochondrial homeostasis.

In summary, these findings demonstrate an association between miR-34a, miR-96, miR-182, miR-183 and autophagic processes, and FOXG1-mediated autophagy activation may be mediated by these miRNAs.

Aging is a complex process characterized by progressive molecular and cellular changes. Given that ARHL cannot be attributed to alterations in a few genes, we employed proteomics to comprehensively analyze protein expression profiles in the aging mouse cochlea and D-gal-induced aging cells, aiming to obtain a more holistic understanding of protein alterations and associated signaling pathways involved in aging. We conducted tandem mass tag (TMT) proteomic analysis of the cochleae from 12-month-old (aging) and 8-week-old (young control) mice. To evaluate inter-group separation and data quality before differential analysis, we performed principal component analysis (PCA), which clearly distinguished the aged and young control groups (Figure 6A). Comparative analysis revealed 192 DEPs between the two groups. The expression distribution of these DEPs was visualized in the heatmap and volcano plots in Figure 6B and Supplementary Figure S4A, respectively. Specifically, 119 proteins were upregulated and 73 were downregulated in the cochleae of aging mice. GO and KEGG enrichment analyses were performed to explore the biological implications of these DEPs. GO analysis, focusing on biological processes (BP), showed that DEPs were predominantly implicated in tissue remodeling, antigen processing and presentation, fibrinolysis, vascular processes in the circulatory system, bone remodeling, and regulation of coagulation (Supplementary Figure S4B). KEGG pathway analysis indicated significant enrichment in the complement and coagulation cascades, protein digestion and absorption, extracellular matrix (ECM)-receptor interaction, lysosomes, fructose and mannose metabolism, PI3K-Akt, Hippo, and Wnt signaling pathways, alongside pyrimidine metabolism, and folate biosynthesis (Figure 6C).

We next employed rapid data-independent acquisition (DIA) proteomics to identify DEPs in HEI-OC1 cells between the D-gal and control groups and between the si-FOXG1 and control groups. The PCA results showed distinct clustering between the control and D-gal groups (Supplementary Figure S4C). Differential analysis identified 1,215 DEPs between these two conditions, of which 746 proteins were upregulated and 469 were downregulated in the D-gal group (Figure 6D). GO enrichment analysis categorized these DEPs by BP, cellular components (CC), and molecular functions (MF) (Supplementary Figure S4D). KEGG pathway analysis revealed that upregulated DEPs were notably enriched in lysosomes, peroxisomes, sphingolipid metabolism, glycolysis/ gluconeogenesis, amino sugar and nucleotide sugar metabolism, branched-chain amino acid degradation (valine, leucine, isoleucine), fatty acid degradation, and lysine degradation (Figure 6E). Conversely, the downregulated DEPs were significantly enriched in ribosome biogenesis, mitophagy, DNA replication, mRNA surveillance pathway, purine metabolism, ferroptosis, and the Hippo signaling pathway (Figure 6F). We then compared DEPs between the si-FOXG1 and control groups. PCA and volcano plots demonstrated clear separation and distribution of DEPs between the two groups (Figure 6G and Supplementary Figure S4E). A total of 941 DEPs were identified, comprising 649 upregulated and 292 downregulated proteins in the si-FOXG1 group. GO analysis revealed that these DEPs were enriched in multiple categories, including BP, MF, and CC (Supplementary Figure S4F). KEGG analysis indicated that upregulated DEPs were primarily associated with peroxisomes, glycolysis/gluconeogenesis, amino sugar and nucleotide sugar metabolism, branched-chain amino acid degradation, fatty acid degradation, and lysine degradation (Figure 6H), whereas downregulated DEPs were predominantly enriched in ribosome biogenesis, oxidative phosphorylation, autophagy, insulin signaling, mRNA surveillance, and purine metabolism (Figure 6I).

Next, we performed an overlap analysis of upregulated and downregulated DEPs in the D-gal and si-FOXG1 groups, followed by KEGG enrichment of the intersecting proteins. The overlapping upregulated DEPs were primarily enriched in peroxisomes, fatty acid degradation, branched-chain amino acid degradation, glycolysis/gluconeogenesis, lysine degradation, pentose phosphate pathway, DNA replication, lysosome, glutathione metabolism, and galactose metabolism (Figure 6J and Supplementary Table S2). These results suggest that both D-gal treatment and FOXG1 knockdown induce oxidative stress, metabolic reprogramming, and inflammatory activation in HEI-OC1 cells. The overlapping downregulated DEPs were enriched in ribosome biogenesis in eukaryotes, one-carbon pool by folate, purine metabolism, nucleocytoplasmic transport, mTOR signaling pathway, mRNA surveillance pathway, and the longevity-regulating pathway (Figure 6K and Supplementary Table S2). This pattern implies that either D-gal exposure or FOXG1 loss suppresses global protein synthesis, reduces genomic stability, and impairs anabolic metabolism in HEI-OC1 cells. Finally, Figure 6L presented qRT-PCR validation of the top four overlapping upregulated and downregulated proteins identified in both the D-gal and si-FOXG1 groups, confirming the reliability of the proteomic data.

As key post-transcriptional regulators of gene expression, miRNAs regulate the aging process, and numerous miRNAs are dysregulated in a tissue-specific manner during mammalian aging. Although several miRNAs associated with ARHL have been identified, our understanding of their regulatory mechanisms remains limited. To expand current knowledge beyond previously identified miRNAs, we employed small RNA-seq to explore more changes in miRNAs during the aging process in HEI-OC1 cells and to identify those associated with the transcription factor FOXG1. We first identified DEmiRNAs between the control and D-gal groups, as well as between the control and si-FOXG1 groups. We then extracted the target DEPs corresponding to these DEmiRNAs and performed KEGG pathway analysis of the target DEPs. Finally, we filtered the pathways associated with aging and identified the DEmiRNAs and their corresponding DEPs within these aging-related pathways (Figure 7A). A total of 83 DEmiRNAs were identified between the control and D-gal groups, with 42 upregulated and 41 downregulated DEmiRNAs in the D-gal group (Figure 7B). Figures 7C,D showed the KEGG enrichment analysis of DEPs targeted by upregulated and downregulated DEmiRNAs, respectively. Notably, the targets of upregulated DEmiRNAs were significantly enriched in multiple aging-related pathways, including lysosome; fatty acid degradation; valine, leucine, and isoleucine degradation; glycolysis; gluconeogenesis; lysine degradation; pentose phosphate pathway; and amino sugar and nucleotide sugar metabolism. Meanwhile, the targets of downregulated DEmiRNAs were enriched in pathways associated with aging, such as ribosome biogenesis in eukaryotes, nucleocytoplasmic transport, Hippo signaling, mRNA surveillance, purine metabolism, one-carbon pool by folate, and mitophagy. To identify aging-related molecular targets more precisely, we further extracted specific DEmiRNAs and their corresponding DEPs in these pathways (Figures 7E,F).

We next identified 53 DEmiRNAs between the control and si-FOXG1 groups, comprising 22 upregulated and 31 downregulated miRNAs in the si-FOXG1 group (Figure 7G). KEGG enrichment analysis was conducted on DEPs targeted by the upregulated and downregulated DEmiRNAs. Similarly, multiple age-related pathways were identified among the enriched pathways. The upregulated pathways included fatty acid degradation, valine, leucine, and isoleucine degradation, DNA replication, glycolysis, gluconeogenesis, histidine metabolism, and the p53 signaling pathway (Figure 7H). The downregulated pathways included thermogenesis, oxidative phosphorylation, one-carbon pool by folate, ribosome biogenesis in eukaryotes, glutathione metabolism, and the longevity-regulating pathway (Figure 7I). To further explore miRNAs and proteins potentially associated with FOXG1 regulation, we identified the target DEmiRNAs and their corresponding DEPs within these aging-related pathways, which were presented in Figures 7J,K. These results further indicate that FOXG1 knockdown induces miRNA expression patterns resembling D-gal-induced aging, suggesting that FOXG1 regulates cellular metabolism, mitochondrial function, and redox balance through miRNA-mediated mechanisms.

It has been well established that metabolic dysregulation is a hallmark of aging and contributes significantly to the pathogenesis of age-related diseases (Yin et al., 2016; Mattson and Arumugam, 2018). Indeed, both our proteomic and small RNA-seq analyses revealed the disruption of metabolic homeostasis in HEI-OC1 cells subjected to D-gal-induced senescence. To delineate these metabolic alterations in greater depth, we performed untargeted metabolomic analysis on HEI-OC1 cells from three experimental groups: control, D-gal, and D-gal combined with si-FOXG1. PCA results revealed a distinct separation between the D-gal and control groups (Supplementary Figure S5A). Differential analysis identified 509 differentially expressed metabolites (DEMs) between the D-gal and control groups, with 324 upregulated and 185 downregulated metabolites in the D-gal group. The volcano plot and heatmap visually demonstrated the statistical significance of metabolic dysregulation (Figure 8A and Supplementary Figure S5B). To further investigate the biological implications of these metabolites, KEGG pathway enrichment analysis was performed on the upregulated and downregulated DEMs. As shown in Figures 8B,C, ATP-binding cassette (ABC) transporters, purine metabolism, glycerophospholipid metabolism, glutamatergic synapse, and γ-aminobutyric acid (GABAergic) synapse pathways were significantly upregulated in the D-gal group, while amino acid biosynthesis, phenylalanine, tyrosine, and tryptophan biosynthesis, tryptophan metabolism, protein digestion and absorption, and aminoacyl-tRNA biosynthesis were significantly downregulated.

We then analyzed the DEMs between the D-gal and D-gal + si-FOXG1 groups. PCA results revealed a clear separation between these groups (Supplementary Figure S5C). In total, 159 DEMs were identified, with 44 upregulated and 115 downregulated in the D-gal + si-FOXG1 group. As shown in Figure 8D and Supplementary Figure S5D, the expression patterns of these metabolites were visualized using volcano plots and heatmaps. KEGG analysis revealed that the upregulated DEMs highlighted pathways involved in arginine biosynthesis; alanine, aspartate, and glutamate metabolism; ABC transporters; purine metabolism; glutamatergic synapse; and histidine metabolism (Figure 8E). Downregulated metabolites were primarily enriched in tyrosine metabolism, glucocorticoid and mineralocorticoid receptor signaling, ascorbate and aldarate metabolism, fatty acid biosynthesis, unsaturated fatty acid biosynthesis, and cortisol synthesis and secretion (Figure 8F). These results suggested that under D-gal-induced aging conditions, FOXG1 knockdown exacerbates oxidative stress and causes a functional decline in several key protective and regulatory metabolic pathways.

Next, we performed an integrated metabolomic analysis of HEI-OC1 cells from control, D-gal, and D-gal + si-FOXG1 groups. The PCA results showed significant separation among the three groups (Figure 8G). A total of 1,948 metabolites, categorized into 10 chemical classes, were detected (Figure 8H). To elucidate the metabolic mechanisms underlying the exacerbation of aging phenotypes following FOXG1 knockdown, we performed K-means clustering analysis. The 705 DEMs among the control, D-gal, and D-gal + si-FOXG1 groups were stratified into four distinct clusters (Figure 8I). Clusters 1 (n = 154 metabolites) and 4 (n = 230 metabolites) displayed the most significant trends. The metabolites in Cluster 1 progressively increased from the control group to the D-gal group, with the greatest significant rise in the D-gal + si-FOXG1 group. Conversely, the metabolites in Cluster 4 exhibited a marked decrease in D-gal + si-FOXG1. These two clusters best reflected the metabolic dysregulation exacerbated by FOXG1 knockdown under aging conditions and were thus selected for subsequent KEGG pathway enrichment to clarify their biological significance. KEGG analysis of Cluster 1 metabolites revealed significant enrichment in ABC transporters, nucleotide metabolism, arginine biosynthesis, GABAergic synapses, alanine/aspartate/glutamate metabolism, purine metabolism, glutamatergic synapses, and histidine metabolism (Figure 8J). These pathways are closely associated with neuronal aging and oxidative signaling, indicating a disruption of the excitatory/inhibitory neurotransmission balance in neural networks (Gleichmann et al., 2012; Chen W. et al., 2024). Furthermore, the enrichment of ABC transporters and purine metabolism highlights energy homeostasis impairment and redox stress in HEI-OC1 cells. The metabolites in Cluster 4 were significantly enriched in amino acid biosynthesis; phenylalanine, tyrosine, and tryptophan pathways; protein digestion and absorption; phenylalanine metabolism; tryptophan metabolism; nucleotide metabolism; and purine metabolism (Figure 8K). Significant downregulation of these pathways indicates a functional decline in protein synthesis, amino acid turnover, and mitochondrial energy metabolism in D-gal-induced aging cells, which was further exacerbated by FOXG1 knockdown.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: http://www.proteomexchange.org/↗, PXD072393↗; https://ngdc.cncb.ac.cn/gsa/s/6pKKG6Mr↗, CRA035839↗.