Abstract
Background: Resistance to the proteasome inhibitor bortezomib is a major obstacle to the treatment of multiple myeloma (MM) and a major cause of relapse and death. Investigating the role of bortezomib resistance genes in MM is crucial. This study aimed to evaluate the potential of bortezomib resistance-related genes (BRGs) as prognostic biomarkers in MM. Methods: The transcriptome data of bortezomib resistant myeloma cell lines, as well as the gene expression and clinical data of patients were downloaded from the Gene Expression Omnibus (GEO) database. Univariate Cox and least absolute shrinkage and selection operator (LASSO) Cox regression models were employed to screen variables and construct a multigene prognostic signature based on BRGs. Single-sample gene set enrichment analysis (ssGSEA) was performed to quantify the relative infiltration levels of immune cells. The pRRophetic algorithm was utilized to assess and infer the sensitivity of anti-MM chemotherapeutics. Results: We identified 129 differentially expressed BRGs, with 25 associated with MM prognosis. Using the LASSO Cox regression model, we identified five key genes (IFI16, ARID5B, LTBP1, PNOC, CRIP1) and developed a bortezomib resistance model for risk stratification and prognosis prediction. Multivariate Cox regression analysis revealed that the risk score was an independent prognostic factor for overall survival (OS). Based on pRRophetic results, high-risk patients may be more sensitive to other chemotherapeutic agents, such as doxorubicin and etoposide. Additionally, we constructed a nomogram incorporating patient age, LDH, International Staging System (ISS), and BRGs, which demonstrated robust prognostic prediction capabilities. The receiver operating characteristic (ROC) values for 1-, 3-, and 5-years survival rates were 0.730, 0.734, and 0.775, respectively. We also validated the expression patterns of the five key genes in MM. IFI16 and CRIP1 expression levels were upregulated in relapsed patients, whereas ARID5B expression was decreased. PNOC and LTBP1 showed no significant differences. Notably, lower ARID5B expression was associated with poorer OS in patients. Conclusions: The BRGs signature is a reliable biomarker for predicting the prognosis of MM and helps optimize clinical decision-making for treatment, and identifies key gene ARID5B downregulation as an adverse prognostic factor in multiple myeloma.