Joint Program of Nanchang University and Queen Mary University of London, Medical College of Nangchang University, Nanchang, 330006, China.
Department of Hematology, The First Affiliated Hospital of Nanchang University, Nanchang, 330006, China.
Clin Exp Med. 2024 Oct 29;24(1):249. doi: 10.1007/s10238-024-01499-6.
Multiple myeloma (MM) is a highly heterogeneous hematological malignancy that is currently incurable. Individualized therapeutic approaches based on accurate risk assessment are essential for improving the prognosis of MM patients. Nevertheless, current prognostic models for MM exhibit certain limitations and prognosis heterogeneity still an unresolved issue. Recent studies have highlighted the pivotal involvement of mitochondrial autophagy in the development and drug sensitivity of MM. This study seeks to conduct an integrative analysis of the prognostic significance and immune microenvironment of mitophagy-related signature in MM, with the aim of constructing a novel predictive risk model. GSE4581 and GSE47552 datasets were acquired from the Gene Expression Omnibus database. MM-differentially expressed genes (DEGs) were identified by limma between MM samples and normal samples in GSE47552. Mitophagy key module genes were obtained by weighted gene co-expression network analysis in the Cancer Genome Atlas (TCGA)-MM dataset. Mitophagy DEGs were identified by the overlap genes between MM-DEGs and mitophagy key module genes. Prognostic genes were selected through univariate Cox regression and least absolute shrinkage and selection operator (LASSO) analysis, and a risk model was subsequently constructed based on these prognostic genes. Subsequently, the MM samples were stratified into high- and low-risk groups based on their median risk scores. The validity of the risk model was further evaluated using the GSE4581 dataset. Moreover, a nomogram was developed using the independent prognostic factors identified from the risk score and various clinical indicators. Additionally, analyses were conducted on immune infiltration, immune scores, immune checkpoint, and chemotherapy drug sensitivity. The 17 mitophagy DEGs were obtained by intersection of 803 MM-DEGs and 1084 mitophagy key module genes. Five prognostic genes (CDC6, PRIM1, SNRPB, TOP2A, and ZNF486) were selected via LASSO and univariate cox regression analyses. The predictive performance of the risk model, which was constructed based on the five prognostic genes, demonstrated favorable results in both TCGA-MM and GSE4581 datasets as indicated by the receiver operating characteristic (ROC) curves. In addition, calibration curve, ROC curve, and decision curve analysis curve corroborated that the nomogram exhibited superior predictive accuracy for MM. Furthermore, immune analysis results indicated a significant difference in stromal scores of two risk groups categorized on median risk scores. And four immune checkpoints (CD274, CTLA4, LAG3, and PDCD1LG2) showed significant differences in different risk groups. The analysis of chemotherapy drug sensitivity revealed that etoposide and doxorubicin, which target TOP2A, exhibited superior treatment outcomes in the high-risk group. A novel prognostic model for MM was developed and validated, demonstrating significant potential in predicting patient outcomes and providing valuable guidance for personalized immunotherapy counseling.
多发性骨髓瘤（MM）是一种高度异质性的血液恶性肿瘤，目前无法治愈。基于准确风险评估的个体化治疗方法对于改善 MM 患者的预后至关重要。然而，目前用于 MM 的预后模型存在一定的局限性，预后异质性仍然是一个未解决的问题。最近的研究强调了线粒体自噬在 MM 的发展和药物敏感性中的关键作用。本研究旨在对 MM 中与噬线粒体相关的特征的预后意义和免疫微环境进行综合分析，构建一种新的预测风险模型。从基因表达综合数据库（GEO）中获取 GSE4581 和 GSE47552 数据集。在 GSE47552 中，通过 limma 在 MM 样本和正常样本之间鉴定 MM 差异表达基因（DEG）。通过癌症基因组图谱（TCGA）-MM 数据集的加权基因共表达网络分析获得噬线粒体关键模块基因。通过 MM-DEG 和噬线粒体关键模块基因之间的重叠基因鉴定噬线粒体 DEG。通过单因素 Cox 回归和最小绝对值收缩和选择算子（LASSO）分析选择预后基因，并基于这些预后基因构建风险模型。然后，根据中位数风险评分将 MM 样本分为高风险组和低风险组。使用 GSE4581 数据集进一步评估风险模型的有效性。此外，还使用从风险评分和各种临床指标中确定的独立预后因素开发了列线图。还进行了免疫浸润、免疫评分、免疫检查点和化疗药物敏感性分析。通过 803 个 MM-DEG 和 1084 个噬线粒体关键模块基因的交集获得了 17 个噬线粒体 DEG。通过 LASSO 和单因素 cox 回归分析选择了 5 个预后基因（CDC6、PRIM1、SNRPB、TOP2A 和 ZNF486）。基于这 5 个预后基因构建的风险模型在 TCGA-MM 和 GSE4581 数据集的受试者工作特征（ROC）曲线中表现出良好的预测性能。此外，校准曲线、ROC 曲线和决策曲线分析曲线表明，基于中位风险评分的风险模型在预测 MM 方面具有更高的预测准确性。此外，免疫分析结果表明，根据中位数风险评分将两个风险组分类时，基质评分存在显著差异。并且在不同的风险组中，有四个免疫检查点（CD274、CTLA4、LAG3 和 PDCD1LG2）显示出显著差异。化疗药物敏感性分析表明，针对 TOP2A 的依托泊苷和阿霉素在高危组中具有更好的治疗效果。本研究构建并验证了一种新的 MM 预后模型，该模型在预测患者预后方面具有显著潜力，并为个性化免疫治疗咨询提供了有价值的指导。
