Integration of Autophagy-Related Genes and Immune Dysregulation Reveals a Prognostic Landscape in Multiple Myeloma

Yibo Xia¹Dong Zheng²Xinyi Zhang²Shuxia Zhu²Enqing Lan³Hansen Ying³Zixing Chen³Bingxin Zhang²Shujuan Zhou²Yu Zhang²Xuanru Lin²Qiang Zhuang²Honglan Qian²Xudong Hu²Yan Zhuang²Qianying Zhang²Xiangjing Zhou⁴Zuoting Xie⁴Songfu Jiang²Ma Yongyong^2,4*Zhouxiang Jin^3*Sisi Zheng^1*

• ¹Department of Medical Laboratory Technology, School of Laboratory Medicine and Life Science, Wenzhou Medical University, Wenzhou, China
• ²Department of Hematology, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou, China
• ³Department of Hepatobiliary Surgery, The Second Affiliated Hospital and Yuying Children’s Hospital of Wenzhou Medical University, Wenzhou, China
• ⁴Department of Blood Transfusion, The First Hospital, Wenzhou Medical College, Wenzhou, China

Background: Autophagy is a self-renewal mechanism in which cells degrade damaged organelles or abnormal proteins through lysosomes. This process eliminates harmful components within the cell and maintains energy homeostasis. Multiple myeloma (MM) is a hematological malignancy characterized by uncontrolled plasma cell proliferation. Autophagy plays a dual role in tumorigenesis, yet its prognostic implications in MM remain underexplored. Methods: Transcriptomic and clinical data from 1,386 MM patients (training cohort: GSE136337, n = 415; validation cohorts: GSE24080, n = 558; GSE4581, n = 413) were analyzed. A seven-gene signature (ATIC, CDKN1A, DNAJB9, EDEM1, GABARAPL1, RAB1A, VAMP7) was identified using LASSO-Cox regression. Predictive performance of the autophagy-related model was assessed via Kaplan-Meier analysis, ROC curves, and nomograms. Immune infiltration, drug sensitivity, and functional pathways of the autophagy-related model were evaluated using CIBERSORT, ESTIMATE, and GSEA. The gene expression in the autophagy prognostic model was verified by qRT-PCR in the U266 and RPMI8226 cell lines and blood samples of multiple myeloma patients from the First Affiliated Hospital of Wenzhou Medical University. Results: The autophagy-related risk score stratified patients into high-risk and low-risk groups with distinct survival outcomes (high-risk HR=0.391, 95%CI:0.284-0.540, p < 0.001). The model demonstrated robust predictive accuracy (5-year AUC = 0.729) and was independently validated. High-risk patients exhibited elevated immune checkpoint expression (CD48, CD70, BTLA), stromal infiltration, and drug resistance. Functional enrichment linked high-risk profiles to MYC activation and oxidative phosphorylation. Through qRT-PCR, the accuracy of the autophagy-related model has been verified in the U266 and RPMI8226 cell lines, as well as in the blood samples of multiple myeloma patients from the First Affiliated Hospital of Wenzhou Medical University. Conclusion: This autophagy-related gene signature provides a reliable prognostic tool for MM, highlighting immune dysregulation and therapeutic resistance mechanisms. Its integration with clinical parameters enhances risk stratification and treatment planning.