Integrated bioinformatics and in vitro analyses identify BHLHE40 as a candidate regulator associated with senescence-like growth arrest in multiple myeloma

Background Multiple myeloma exhibits significant molecular heterogeneity and is continuously influenced by the bone marrow immune microenvironment. Cellular senescence can restrict tumor expansion through stable growth arrest, while also remodeling the local microenvironment via the senescence-associated secretory phenotype. Methods By integrating the GSE6477 and GSE72213 microarray cohorts, candidate genes were screened using differential expression analysis, weighted gene co-expression network analysis, and three machine learning algorithms; subsequently, BHLHE40 was evaluated by combining two-sample Mendelian randomization, immune infiltration analysis, gene set enrichment analysis, single-cell transcriptomics, and in vitro overexpression experiments. Results A total of 1,928 differentially expressed genes and 34 cellular senescence-related candidate genes were identified. LASSO, SVM-RFE, and random forest collectively identified CXCL12, ENO2, and BHLHE40. In the same combined discovery cohort used for feature selection, BHLHE40 showed an apparent AUC of 0.911. Mendelian randomization analysis suggested a nominal inverse association between genetically predicted circulating BHLHE40 levels and MM risk. Single-cell transcriptomic and ssGSEA analyses further showed that BHLHE40 expression was decreased in MM plasma cells and was associated with relative enrichment scores for selected immune signatures. In vitro experiments showed that BHLHE40 overexpression reduced MM cell viability and EdU incorporation, accompanied by upregulation of p16, p21, and p53 and changes consistent with canonical NF-κB pathway activation (increased p-p65 and decreased IκBα). Conclusion Multilevel evidence supports BHLHE40 as a candidate regulator associated with senescence-like growth arrest and immune microenvironment status in MM. Given the lack of external diagnostic cohorts, loss-of-function or rescue experiments, comprehensive senescence phenotyping, and in vivo validation, the current findings are insufficient to establish it as a clinical diagnostic biomarker or therapeutic target.