Unveiling Hypoxic Regulatory Networks by Bioinformatics: Mechanisms of Hypoxia-Related Hub Genes Driving Rituximab Resistance and Poor Prognosis in DLBCL

Jiayi Yao^1,2Zizhen Xu^1,2,3*Yuanfei Mao^1,3*

• ¹School of Medicine, Shanghai Jiao Tong University, Shanghai, China
• ²Department of Laboratory Medicine, Ruijin Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, Shanghai Municipality, China
• ³Ruijin Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, Beijing, China

Background Diffuse large B-cell lymphoma (DLBCL), an aggressive subtype of non-Hodgkin lymphoma, exhibits heterogeneous clinical outcomes. While rituximab, a CD20 inhibitor, combined with chemotherapy has improved survival in some patients, resistance remains prevalent, particularly in hypoxic tumor microenvironments. Understanding hypoxia-related genes (HRGs) and their role in rituximab resistance is critical to addressing therapeutic challenges in high-risk DLBCL. Methods Gene expression profiles from GEO datasets (GSE56315: DLBCL tumor vs. normal; GSE104212: hypoxia-treated DLBCL cell lines) were analyzed to identify overlapping genes between DLBCL-signature genes (DSGs) and HRGs. protein interaction network topology analysis and Lasso regression modeling of TCGA-DLBC dataset were employed to screen regulator and hub genes. Hub genes linked to rituximab response and survival were validated in DLBCL patients receiving rituximab therapy. Functional enrichment analysis was used to explore associated pathways. The expression of the identified regulator and hub genes was validated using reverse transcription quantitative polymerase chain reaction (RT-qPCR). Results 58 overlapping genes were identified between DSGs and HRGs. PPI network and Lasso regression revealed 5 MS4A1 regulator genes and 10 hub genes. Among these, LGALS1 (HR = 0.588, p = 0.00085), TIMP1 (HR = 0.591, p = 0.00098), ANXA1 (HR=0.614, p=0.0024) and STAP1 (HR=0.633, p=0.0035) were significantly associated with overall survival and GPNMB (AUC=0.869), CDCA7 (AUC=0.686), and STAP1 (AUC=0.663) associated with treatment response in rituximab-treated patients. Functional analysis implicated these genes in B-cell receptor (BCR) and PI3K-AKT signaling pathways, suggesting their mechanistic roles in therapeutic resistance. Conclusions This study identifies hypoxia-associated genes critical to rituximab resistance in DLBCL, highlighting potential therapeutic targets. Their involvement in BCR and PI3K-AKT pathways underscores novel vulnerabilities for overcoming refractory disease. Our findings provide a foundation for developing strategies to improve outcomes in high-risk DLBCL patients with hypoxic microenvironments.