Bioinformatics analysis and experimental verification of the multiple myeloma microenvironment and treatment response prediction on the basis of the ecDNA-associated gene signature

Abstract Objective This study was conducted to investigate the microenvironment of multiple myeloma (MM) and predict treatment response based on the extrachromosomal DNA (ecDNA) signature. Methods Bioinformatics was implemented to determine the expression characteristics of ecDNA-related genes to construct a prognostic signature for MM, analyze its association with the tumor immune microenvironment (TME), predict treatment regimens, and assess the core gene expression patterns at the single-cell level. The background expression of core genes in the MM cell lines RPMI and GM12878 was evaluated. Bone marrow samples were collected from six patients with MM, after which plasma cells were isolated using magnetic bead technology, and eight healthy participants were placed in the control group. QPCR analysis was conducted to detect differences in the expression of core genes between the two groups. Results Three key ecDNA-related genes (AKAP9, TYMS, and YEATS4) were identified via differential analysis, univariate Cox regression analysis, Kaplan–Meier (KM) survival analysis, and LASSO-Cox regression based on the bulk transcriptome to construct an ecDNA-related prognostic signature. Patients were divided into high-risk and low-risk groups based on their prognosis score, and patients with high scores had a poorer prognosis. The results of univariate and multivariate Cox regression analyses revealed that the constructed prognostic signature was an independent prognostic factor; these findings were confirmed in the independent validation set. By analyzing a single-cell subset of C5 cells that had a high expression of the AKAP9 gene and was present in low proportions in MM patients, it was found that this subset of cells may be related to the pathogenesis of MM. Cell experiments confirmed that RPMI cells expressed significantly higher levels of the AKAP9, TYMS, and YEATS4 genes than GM12878 cells. No significant difference in the expression of AKAP9 was found between the plasma cells of MM patients and those of healthy controls. The expression of TYMS and YEATS4 genes was significantly higher in the plasma cells of MM patients than in healthy controls. Conclusions The genes AKAP9, TYMS, and YEATS4 are highly expressed in MM cells. The level of expression of the AKAP9 gene in plasma cells of MM patients was not significantly different from that in the bone marrow of healthy controls. The level of expression of the TYMS and YEATS4 genes was significantly higher in MM patients than in the healthy control group. Therefore, the prediction model based on ecDNA-related features can predict the treatment response and prognosis of MM. The feasibility of this method was preliminarily determined by experiments.