The HCM dataset GSE180313(17) was obtained from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). This dataset contains heart tissues from 7 healthy individuals and 13 patients with HCM. Preprocessed and merged data were integrated into a unified dataset. The ‘limma’ package (18) in R software (v.3.6.3) (19) was used to identify differentially expressed genes (DEGs) between the HCM and control groups, with the screening criteria being set as a |log₂FC| value of <0.5 and a P-value of <0.05. The ‘ggplot2’ package (20) in R was used to generate heat maps and volcano plots. Information regarding ARGs was obtained from the Human Autophagy Database (http://www.autophagy.lu/index.html) and the Gene Set Enrichment Analysis (GSEA) website (http://software.broadinstitute.org/gsea/index.jsp). The extracted ARGs were processed and integrated into a gene set known as ARGs. GSEA was used to investigate the overall association between AT and HCM and identify potential biological processes involving ARGs that were associated with the pathogenesis of HCM. Genes in the GSE180313 dataset were scored using the AT dataset from GSEA, resulting in the calculation of normalized enrichment scores (NESs). P<0.05 was considered to indicate significant enrichment. The workflow of the present study is shown in Fig. S1.

DEARGs were obtained by intersecting the DEGs identified in the GSE180313 dataset with the integrated ARG set. The overlapping genes (DEARGs) were visualized on a Venn diagram. Subsequently, the cluster profiler package (v4.8.3) (21) in R and DAVID (david.ncifcrf.gov/) were used to implement Gene Ontology (GO; https://geneontology.org/) and Kyoto Encyclopedia of Genes and Genomes (KEGG; https://www.genome.jp/kegg/) enrichment analyses of DEARGs. P<0.05 was considered to indicate significant enrichment.

DEARGs were imported into the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database (v11.09) (https://string-db.org/) for PPI analysis with default settings. The resulting PPI network was visualized using Cytoscape software (v.3.7.2) (22). Subsequently, the MCODE plug-in (V 3.7.1) (20) was used to filter and visualize key PPI networks, resulting in the identification of key modules containing hub genes.

The ROC curves of hub genes were constructed using data from the GSE180313 and GSE36961(23) datasets. The area under the curve (AUC) was quantified for comparison, and only genes with AUC values of >0.6 were considered to have statistically significant diagnostic potential.

To investigate the immune microenvironment of HCM and key DEARGs, CIBERSORTx (https://cibersortx.stanford.edu/) was employed to analyze the differences in immune infiltration between patients with HCM and healthy controls. Additionally, this algorithm was utilized to determine the proportions of various immune cell types.

Key differentially expressed AT-associated genes and compound interaction data from the Drug Signatures Database (DSigDB; http://dsigdb.tanlab.org/) were extracted using the Enrichr online tool (http://amp.pharm.mssm.edu/Enrichr). Drugs for the treatment of HCM were predicted based on hub genes.

The H9c2 immortalized rat cardiomyocyte-like cell line (The Cell Bank of Type Culture Collection of The Chinese Academy of Sciences) is commonly used in cardiac research in vitro (24). H9c2 cells were cultured in high-glucose DMEM (Hyclone; Cytiva) supplemented with 10% fresh fetal bovine serum (HyClone; Cytiva) and 1% penicillin-streptomycin (Gibco; Thermo Fisher Scientific, Inc.) under standard conditions (37˚C, 5% CO₂, 95% humidity and 21% oxygen). Cells from passages 3-10 were used for subsequent experiments. Angiotensin II (AngII; GlpBio Technology, Inc.) was used to induce hypertrophy in H9c2 cells. The cells were incubated with AngII at different concentrations (50, 100, 200 and 400 nM) at 37˚C for 24 h, and the optimal concentration was determined based on the mRNA and protein expression of atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP). For further experimentation, the cells were incubated with the determined optimal concentration of AngII at 37˚C for 12, 24, 36 and 48 h.

Total RNA was extracted from H9c2 cells using TRIzol reagent (Beijing Solarbio Science & Technology Co., Ltd.) and reverse transcribed using the PrimeScript RT reagent kit (Monad Biotech Co., Ltd.) according to the manufacturer's instructions. qPCR was conducted using the SYBR Green qPCR Master Mix (cat. no. B21203; Bimake.com) on a BIO-RAD CFX Connect Real-Time PCR Detection System (Bio-Rad Laboratories, Inc.). The qPCR protocol included an initial denaturation step at 95˚C for 10 min, followed by 40 cycles of thermal cycling with denaturation at 95˚C for 15 sec and annealing at 60˚C for 1 min. The primer sequences used for qPCR are shown in Table SI. ACTB served as the internal reference, and the relative mRNA expression of target genes was calculated using the Cq (2^-ΔΔCq) method (25).

To extract total proteins, H9c2 cells were lysed in RIPA buffer supplemented with protease inhibitors (Beijing Solarbio Science & Technology Co., Ltd.) on ice. The extracted proteins were quantified using a BCA assay kit (GlpBio Technology, Inc.) and denatured by boiling for 5 min, and subsequently separated by on 12% gels by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, with 40 µg of protein loaded per lane. The separated proteins were transferred to a PVDF membrane (MilliporeSigma), which was incubated with 5% skimmed milk powder at room temperature for 2 h on a shaker. Subsequently, the membrane was incubated with primary antibodies against ANP (1:500; cat. no. 27426-1-AP; Proteintech Group, Inc.), BNP (1:500; cat. no. A2179; ABclonal, Inc.), LC3 (1:1,000; cat.no. 381544; Chengdu Zen-Bioscience Co., Ltd.), P62 (1:1,000; cat. no. 380612; Chengdu Zen-Bioscience Co., Ltd.) and β-actin (1:1,000; cat. no. 66009-1-Ig; Proteintech Group, Inc.) at 4˚C overnight. The following day, the membrane was incubated with horseradish peroxidase-conjugated mouse and rabbit secondary antibodies (1:5,000; cat. nos. 511103 and 511203, respectively; Chengdu Zen-Bioscience, Co., Ltd.) at room temperature for 2 h. Protein bands were visualized using the Ultra High Sensitivity ECL kit (catalog no. GK10008; GlpBio Technology, Inc.) and captured using the FluorChem FC3 System (ProteinSimple). ImageJ software (v1.8.0.345; National Institutes of Health) was used to semi-quantify the optical density of protein bands.

To assess the level of AT in cells, LysoTracker Red was used to label intracellular lysosomes, as autophagosomes can bind to lysosomes. Briefly, H9c2 cells were incubated with 50 nM LysoTracker Red working solution (Beyotime Institute of Biotechnology) at 37˚C for 30 min in the dark and subsequently examined using a fluorescence microscope.

LC3 expression in H9c2 cells was detected through immunofluorescence staining. Briefly, the cells were washed twice with PBS, fixed with 4% paraformaldehyde (biosharp life sciences) at room temperature for 10 min, blocked with 2% BSA (cat. no. CAS#9048-46-8; Shanghai Yuanye Bio-Technology Co., Ltd.) at room temperature for 30 min and incubated with anti-LC3 antibody (1:200; cat. no. 381544; Chengdu Zen-Bioscience Co., Ltd.) at 4˚C overnight. The following day, the cells were incubated with a fluorescently labeled secondary antibody (1:200; cat. no. A32732; Thermo Fisher Scientific, Inc.) at room temperature for 1 h. Thereafter, nuclei were stained with DAPI at room temperature for 5 min and the cells were examined using a fluorescence microscope.

Statistical analysis was performed using the GraphPad Prism 8 software (Dotmatics). The unpaired two-tailed Student's t-test was used to compare the differences between two groups. One-way analysis of variance and Dunnett's multiple comparison test were employed to assess the differences among multiple groups. Data are expressed as the mean ± standard deviation of 3 repeats. P<0.05 was considered to indicate a statistically significant difference.

The overall protocol of the present study is shown in Fig. S1. The patient data used in the present study were derived from the GSE180313 dataset in the GEO database.

After pre-processing and normalization of the GSE180313 dataset, a total of 966 DEGs were identified between the HCM and control groups. Of these 966 DEGs, 510 genes were upregulated and 456 genes were downregulated. A volcano plot and a heat map were generated to visualize the DEGs (Fig. 1A and B).

GSEA was used to compare ARGs between the HCM and control groups. Fig. 1C demonstrates the significant differences in ARG expression between the two groups (|NES|=1.521; P<0.05). This indicates that ARGs could be a crucial characteristic of HCM and that their dysregulation supports a link between HCM and autophagy. Furthermore, a total of 1,167 ARGs were identified after pre-processing of the integrated gene set. These ARGs were intersected with DEGs to obtain 58 DEARGs (Fig. 1D). A heat map was generated to visualize the expression patterns of these DEARGs in the HCM and control groups (Fig. 1E).

To investigate the functions and pathways of DEARGs, GO and KEGG enrichment analyses were performed using DAVID. The results demonstrated that DEARGs were significantly enriched in biological processes such as ‘peptidyl-serine phosphorylation’, ‘cellular response to oxidative stress’, ‘regulation of autophagy’ and ‘response to UV’; molecular functions such as ‘protein serine kinase activity’, ‘tau protein binding’ and ‘death domain binding’; and cellular components such as the ‘mitochondrial outer membrane’ and ‘organelle outer membrane’ (Fig. 2A-C). KEGG analysis demonstrated that the DEARGs were notably enriched in the ‘AGE-RAGE signaling pathway in diabetic complications’, ‘PI3K-Akt signaling pathway’, ‘measles’, ‘hepatitis C’, ‘AMPK signaling pathway’ and ‘EGFR tyrosine kinase inhibitor resistance’ (Fig. 2D). Interactions were identified between DEARGs and the aforementioned functions and pathways through gene and pathway cross-talk mapping, suggesting that multiple genes and pathways may be involved in the regulation of DEARGs in HCM (Fig. 2E-H). Overall, these findings suggested that ARGs regulate the progression of HCM through intricate interplay among multiple gene functions and pathways.

A PPI network of 58 DEARGs was constructed using the STRING database (Fig. 3A). After the network was imported into the Cytoscape software, the Maximal Clique Centrality algorithm was used to identify a sub-network comprising 32 hub genes. To filter these hub genes, the MCODE plug-in was used to identify important functional modules within the PPI network. Notably, a key cluster in the network consisted of a functional module with 10 nodes and 14 edges, including EIF4EBP1, MCL1, PIK3R1, CCND1, PPARG, SMPD1, RICTOR, NOS3, SNCA and UBB. Fig. 3B and C illustrate the interactions between DEARGs and hub genes.

ROC curves were generated to evaluate the diagnostic efficacy of the 10 hub genes (Fig. 4A). In the GSE180313 dataset, all hub genes exhibited AUC values of >0.8, indicating a significant association with HCM and promising diagnostic potential. An external dataset (GSE36961) was used to validate these findings (Fig. 4B). Notably, discrepancies were observed in the results of hub gene analysis between the two datasets. During the validation of external datasets, it was observed that the AUC values for EIF4EBP1, MCL1, and SMPD1 were <0.7, indicating that these three hub genes exhibit limited diagnostic performance for HCM. By contrast, most other hub genes demonstrated robust diagnostic capabilities (AUC >0.7), suggesting a potential association between autophagy and HCM. Nonetheless, the significance of these hub genes warrants further investigation in future studies

The DSigDB in the Enrichr platform was used to identify small-molecule drugs targeting hub genes for the treatment of HCM. A total of 1,332 drugs with potential therapeutic value were identified based on the degree of gene-compound match and median number, with the screening criteria being set as a false discovery rate of <0.05 and composite scores of >5,000. The top 10 small-molecule drugs with the most significant impact on the expression of hub genes are shown in Fig. 5A. Among these, rapamycin and melatonin are the top two candidates. Fig. 5B and C show the molecular structure of rapamycin and melatonin (Mel; N-acetyl-5-methoxytryptamine). Research has demonstrated that rapamycin exerts a notable effect on the treatment and prevention of HCM (26,27). Mel possesses antioxidant properties and exerts protective effects against various cardiovascular diseases, including diabetic cardiomyopathy and myocardial hypertrophy (28,29). To gain insights into the binding between hub genes and predicted drugs, molecular docking simulations were performed using rapamycin and Mel as examples (Fig. 5D and E).

Cellular and humoral immune functions serve a crucial role in the development of HCM (30). The relative proportions of infiltrating immune cells in heart samples from the GSE180313 dataset were assessed and quantified using the CIBERSORT algorithm (Fig. 6A). The resulting heatmap illustrates the relationships between various infiltrating immune cells (Fig. 6B). In addition, violin plots were generated to visualize the expression profiles of the 20 immune cell subtypes in the control and HCM groups (Fig. 6C). Only the infiltration levels of T follicular helper (Tfh) cells were significantly different between the HCM and control groups, with lower levels being observed in the HCM group. Therefore, Tfh cells were identified as differentially infiltrating immune cells. Furthermore, the correlation between immune cells and the 10 hub DEARGs was examined (Fig. 7). Nine hub genes, except for CCND1, exhibited varying correlations with six types of immune cells, namely Tfh cells, monocytes, neutrophils, regulatory T cells, resting natural killer cells and T cells CD4 memory resting. These findings suggested that the hub genes serve an important role in the immune response to HCM.

To validate the expression of the hub genes EIF4EBP1, MCL1, PIK3R1, CCND1, PPARG, SMPD1, RICTOR, NOS3, SNCA and UBB in vitro, H9c2 cells were stimulated with AngII to induce HCM. Western blotting and RT-qPCR were used to determine the optimal concentration and duration of AngII treatment. The results demonstrated that when H9C2 cells were pretreated with AngII for the same duration, the highest protein expression levels of ANP and BNP were observed in the 100 nM AngII group (Fig. 8A-C). Additionally, when H9C2 cells were pretreated with 100 nM AngII for different time periods, the highest ANP and BNP protein expression levels were found at 24 h of pretreatment (Fig. 8D-F). Consistent with these changes in protein expression, the mRNA levels also corroborated this finding (Fig. 8G-J). Therefore, a pretreatment of H9C2 cells with 100 nM AngII for 24 h was selected to induce hypertrophy. Western blotting was used to evaluate the expression levels of the AT-associated proteins LC3 and P62 in the AngII and control groups (Fig. 9A). As shown in Fig. 9B and C, the LC3/β-actin ratio was significantly lower and the protein expression of P62 was higher in the AngII group. Furthermore, autolysosome acidification was detected, and immunofluorescence analysis was performed to assess the level of AT. H9c2 cells stained with LysoTracker Red showed reduced fluorescence intensity in the AngII group compared with that in the control group (Fig. 9D). Immunofluorescence analysis revealed a decrease in LC3 fluorescence intensity in the AngII group compared with that in the control group (Fig. 9F). The aforementioned results indicate that autophagy is reduced in the AngII-induced H9C2 cell hypertrophy model. To ensure the reliability of the results, RT-qPCR was performed to evaluate the expression levels of hub genes in both groups. Based on PPI network analysis, EIF4EBP1, MCL1, PIK3R1, CCND1 and PPARG were identified as the most significant hub ARGs associated with HCM. RT-qPCR revealed that EIF4EBP1 and PPARG were downregulated, while MCLl, PIK3R1 and CCND1 were upregulated after treatment with AngII (Fig. 10). Therefore, we propose that EIF4EBP1, MCL1, PIK3R1, CCND1 and PPARG play roles in the regulation of autophagy in HCM.