J Cell Mol Med. 2026 Sep;30(18):e71379. doi: 10.1111/jcmm.71379.
ABSTRACT
Dilated cardiomyopathy (DCM) is a leading cause of heart failure. Due to its complex pathogenesis, effective treatment strategies remain limited. Therefore, it is particularly important to identify novel candidate biomarkers from the pathogenesis level. In recent years, lactylation modification plays an important role in various cardiovascular diseases; however, its specific involvement in DCM remains unclear. Therefore, it is of great significance to identify lactylation-related biomarkers associated with DCM to provide insights for future mechanistic investigations. The DCM gene expression datasets were obtained from the Gene Expression Omnibus (GEO) database. Differentially expressed genes and co-expression modules were identified using differential expression analysis and weighted gene co-expression network analysis (WGCNA). Lactylation-related gene sets (LRGs) were obtained from the MSigDB database. The intersection of these results was further analysed using machine learning algorithms: Least Absolute Shrinkage and Selection Operator (LASSO), Random Forest (RF), and Support Vector Machine (SVM). CIBERSORT algorithm was used to analyse the immune infiltration characteristics of DCM, and DCM was divided into two subtypes by unsupervised consensus clustering. Mendelian randomization (MR) analysis was employed to evaluate potential causal associations between the candidate genes and DCM or heart failure progression. Finally, the candidate genes were verified by the GSE57338 database and β-AA-induced heart failure model with DCM characteristics. Through comprehensive transcriptomic analysis integrating WGCNA and LRGs, 38 characteristic genes associated with DCM and lactylation were finally identified. Subsequently, 18 key genes were selected by machine learning. The risk prediction model based on these genes demonstrated good predictive performance for DCM. Immune infiltration analysis revealed dysregulated inflammatory responses in DCM patients, and 18 key genes were significantly related to CD8+ T cells, CD4+ T cells, neutrophils, and Natural killer T cell infiltration. Unsupervised consensus clustering divided DCM patients into two distinct subtypes (C1 and C2), with marked differences in gene expression. Subtype C1 was characterized by activation of CD8+ T cells and Natural Killer T cells, whereas C2 mainly exhibited CD4+ T cells and neutrophils activation. Furthermore, Mendelian randomization analysis suggested that EXT1 may be associated with heart failure, while IER3 was potentially associated with DCM. Finally, the external validation set GSE57338 and the heart failure model with DCM characteristics were used to verify the two genes identified. The results showed that EXT1 was up-regulated and IER3 was down-regulated. Based on the deep excavation of public database, combined with machine learning algorithm and Mendelian randomization, this study identified EXT1 and IER3 as potential candidate biomarkers for DCM, providing a new theoretical foundation and hypotheses for future mechanistic and translational research.
PMID:42791204 | DOI:10.1111/jcmm.71379