Exploratory analysis of a novel obesity-related gene-based prognostic model as a potential prognostic biomarker in multiple myeloma

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Abstract Background This study aimed to elucidate the genetic interplay between obesity and multiple myeloma (MM) and to develop a novel obesity-related gene-based prognostic model using an integrated multi-omics framework combining single-cell RNA sequencing (scRNA-seq) and bulk RNA sequencing (RNA-seq). Methods A differential expression analysis of the GSE132604 exploratory discovery cohort was performed using the limma package to identify obesity-related differentially expressed genes (DEGs). Weighted gene co-expression network analysis was conducted on the MMRF-CoMMpass training cohort from GDC Data Portal to identify prognosis-related modules. Obesity-related DEGs were intersected with prognosis-related genes to identify obesity–MM prognosis-related genes (OMMPRGs). A univariate Cox regression analysis was used to link the OMMPRGs to overall survival (OS). Four machine-learning algorithms—StepCox, CoxBoost, Lasso regression, and random survival forest—were employed to refine these genes and construct a prognostic model. The model’s performance was evaluated by receiver operating characteristic (ROC), and the GSE57317 external validation cohort was used for independent validation. Further analysis of the GSE199359 scRNA-seq cohort included copy number variation assessment, CytoTRACE for cell stemness, Monocle2 for pseudotime trajectory analysis, malignant plasma cell markers identification, and immune-related gene set enrichment analysis to accurately identify malignant plasma cell subtypes. Virtual knockout of the four model genes was performed using scTenifoldKnk in malignant plasma cells, and the resulting perturbed transcriptional profiles were subjected to gene set enrichment analysis to evaluate their potential effects on malignant plasma-cell biology. Results A total of 39 OMMPRGs were identified by integrating differential expression results with prognosis-related modules. A four-gene prognostic signature (MCM4, PHF19, CCT2, and MAGEA1) was established and effectively stratified MM patients into high- and low-risk groups, with the high-risk group exhibiting significantly poorer OS in the MMRF-CoMMpass training cohort. ROC analyses and validation supported the potential robustness of the model, although the obesity-stratified discovery cohort was limited in size. Single-cell RNA-seq analysis identified malignant plasma-cell states and showed that the four core prognostic genes were preferentially expressed in these cells. Virtual knockout and GSEA further suggested that these genes may contribute to the maintenance of malignant plasma-cell programs. Conclusions This study developed a novel obesity-related gene-based prognostic model that reliably predicts survival outcomes in MM and provides a multi-omics foundation for risk stratification and future precision-medicine studies in MM. Future studies in larger obesity-stratified MM cohorts with comprehensive clinical, cytogenetic, and genomic annotations are warranted to validate the clinical relevance and biological specificity of this obesity-related prognostic signature.