Identification and validation of differentially expressed disulfidptosis-related genes in hypertrophic cardiomyopathy

Abstract

Hypertrophic cardiomyopathy (HCM) is one of the most common cardiovascular diseases with no effective treatment due to its complex pathogenesis. A novel cell death, disulfidptosis, has been extensively studied in the cancer field but rarely in cardiovascular diseases. This study revealed the potential relationship between disulfidptosis and hypertrophic cardiomyopathy and put forward a predictive model containing disulfidptosis-associated genes (DRGs) of GYS1, MYH10, PDMIL1, SLC3A2, CAPZB, showing excellent performance by SVM machine learning model. The results were further validated by western blot, RNA sequencing and immunohistochemistry in a TAC mice model. In addition, resveratrol was selected as a therapeutic drug targeting core genes using the CTD database. In summary, this study provides new perspectives for exploring disulfidptosis-related biomarkers and potential therapeutic targets for hypertrophic cardiomyopathy.

Fig. 1

Schematic diagram of this study. The figure is created with BioRender.com

Fig. 2

Identification of DRG expressions and immune analysis in HCM patients (A) The expression pattern of DRGs in control people and HCM patients. (B) Heatmap of 16 differentially expressed DRGs between control and HCM patients. (C) The correlation pattern of 16 differentially expressed DRGs. (D) Heatmap of immune cell infiltration of control people and HCM patients. (E) The immune cell infiltration pattern of control people and HCM patients. (F) The correlation analysis between 16 DRGs and immune cells. * p< 0.05, ** p< 0.01

Fig. 3

Identification of disulfidptosis-related clusters in HCM. (A) Consensus clustering matrix when k = 2. (B, C) Representative cumulative distribution function (CDF) curves (B), CDF delta area curves (C). (D) The score of consensus clustering. (E) The distribution of two clusters. (F) The different expressions of 16 DRGs between cluster 1 and cluster 2. (G) The expression pattern of 16 DRGs is presented in the heatmap

Fig. 4

Construction and evaluation of four machine learning models. (A) Cumulative residual distribution of SVM, RF, XGB, GLM models. (B) Boxplots representing the residuals of each machine learning model. (C) Feature importance of top 10 DRGs in each machine learning model. (D) Receiver Operating Characteristic (ROC) analysis of four machine learning models

Fig. 5

Validation of the 5-gene-based SVM model. (A) Construction of a nomogram for predicting the risk of HCM based on the 5-gene-based SVM model. (B) Calibration curve for assessing the predictive efficiency of the nomogram model. (C, D) ROC analysis of the prediction nomogram based on 5-fold cross-validation in validation dataset GSE141910. (E, F) ROC analysis of the prediction nomogram based on 5-fold cross-validation in validation dataset GSE160997

Fig. 6

CeRNA network based on five-core DRGs. The yellow nodes represent mRNAs, purple nodes represent miRNAs and green nodes represent lncRNAs

Fig. 7

Prediction of chemicals targeting the five core markers. (A) Drug-gene network. Chemical ID in red color represents resveratrol. (B) Molecular docking schematic diagrams of resveratrol docked with five core markers respectively

Fig. 8

In vivo validation of the five disulfidptosis-related markers in HCM mice. (A) Masson and HE staining of control and TAC mice. (B) Heatmap of the relative mRNA expressions of Gys1, Myh10, Capzb, Slc3a2 and Pdlim1 in control and HCM mice heart tissues (n = 3). (C) Typical images of immunohistochemical staining of GYS1, MYH10, CAPZB, SLC3A2 and PDLIMI in heart tissues of control and HCM mice. (D) Protein expressions of GYS1, MYH10, CAPZB, SLC3A2 and PDLIMI in control and HCM heart tissues detected by Western Blot. (E) Quantitative analysis of the Western Blot results. Data are representative of three independent experiments and presented as mean ± SD. Statistical significance was analyzed by unpaired Student’s t-test. ns, no significant, * p< 0.05, ** p< 0.01