Identification Of An Oxidative Phosphorylation Metabolism Related Risk Signature For Multiple Myeloma

Objective: We aimed to identify the feature of oxidative phosphorylation(OXPHOS)phenotype and construct an OXPHOS multigene model by the prognostic related OXPHOS genes in multiple myeloma(MM)through a comprehensive bioinformatic analysis of transcriptome data,the risk signature generated by this model was used to analyze their clinical significance.Methods: Firstly,we conducted differentially expressed gene(DEG)and weighted gene co-expression network analysis(WGCNA)to obtain the MM-related genes and modules and used the enrichment analyses to explore their biological function.Then,univariate Cox,least absolute shrinkage and selection operator(LASSO),and multivariate Cox regression analysis were performed to screen OXPHOS genes associated with the overall survival(OS)of MM patients and build an optimal prognostic prediction model.After calculating the risk score of each patient by the expression value and regression coefficient of the model genes,we divided the patients into low-and high-risk group by the median risk score.The prognostic performance of this OXPHOS-related signature was validated by Kaplan-Meier survival analysis and time-dependent receiver operating characteristic(ROC)curves.We also assessed the independence of this risk score by univariate and multivariate Cox regression analysis.Furthermore,a nomogram contains the risk signature and classic adverse prognostic factors of MM was built and decision curve analysis(DCA)and ROC curve analysis was performed to explore the efficacy of this integrated predictive model and their clinical significance in MM patients.Finally,we analyzed the expression and interaction of model genes,and verified the relative expression of model genes in human MM cell line(HMCL)by real-time quantitative polymerase chain reaction(RT-q PCR).Results: The enrichment analysis of DEG and 3 MM-related modules showed that OXPHOS-related pathways or function were significantly enriched in MM patients.After multi-steps variable screening and Cox regression analysis,an OXPHOS metabolic risk model consisting of 9 genes was obtained,and the corresponding risk score could effectively divide MM patients into two groups with significantly different outcomes,which means the patients of high-risk group have significantly shorter OS and event-free survival time than those of low-risk group(both P<0.0001).The AUC for 3-year and 5-year OS were 0.728 and 0.739 in the training set,respectively,suggesting that the risk score has medium predictive efficiency.The results of survival analysis in the two external test sets were similar with the training set.The Cox analysis suggested that the risk score is an independent risk factor.Comparative analysis showed that patients with poor prognostic characteristics typically have higher level of risk score.The nomogram model could effectively predict the OS of MM patients,and DCA and ROC analysis showed that the clinical benefits and predictive accuracy of the nomogram model that incorporates risk score were significantly higher than the model including purely classical poor prognostic parameters.RT-q PCR showed that the expression of six dangerous genes were significantly upregulated in the HMCL,which was basically consistent with the results from the dataset.Conclusion: In conclusion,our study found that OXPHOS metabolism was significantly enriched in MM.The OXPHOS risk signature predicted by our study could effectively distinguish patients with different survival outcomes,as well as improve the predictive power and clinical benefit of classical prognostic parameters,which might contribute to the future risk stratification and rational therapy selection in clinical practice of MM patients.Moreover,those genes associated with poor prognosis may be candidate therapeutic targets for MM,which worth further analysis.