Su Zhang,1,2,* Weitao Hu,3,* Chunyan Huang,4 Xinxin Lin,4 Xiaoqing Chen2 

1The Second Clinical College of Fujian Medical University, Quanzhou, People’s Republic of China; 2Department of Rheumatology, The Second Affiliated Hospital of Fujian Medical University, Quanzhou, People’s Republic of China; 3Department of Gastroenterology, The Second Affiliated Hospital of Fujian Medical University, Quanzhou, People’s Republic of China; 4Department of General Practice, The Second Affiliated Hospital of Fujian Medical University, Quanzhou, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Xiaoqing Chen, Email chenxiaoqing202203@163.com

Introduction: Lupus nephritis (LN) is a common and severe complication of systemic lupus erythematosus (SLE). Ferroptosis is a form of iron-dependent cell death induced by lipid peroxidation. However, its specific role in LN remains unclear.
Methods: We utilized the GSE32591 obtained from the GEO database to identify differentially expressed genes (DEGs) and conduct enrichment analysis. Differentially expressed ferroptosis-related genes of LN (LNDE-FRGs) were derived by taking the intersection of DEGs and ferroptosis-related genes (FRGs). Three machine learning algorithms were applied to screen candidate key LNDE-FRGs. The expression of the key LNDE-FRGs was validated using external validation cohorts and clinical samples. The diagnostic value of the key LNDE-FRGs was then assessed by receiver operating characteristic curve (ROC) analysis. In addition, we investigated the correlation between the key genes and glomerular filtration rate (GFR), urinary protein and serum creatinine (Scr) in LN patients via the Nephroseq V5 database. Subsequently, we performed immune infiltrating cell analysis of LN kidney tissue using Cibersort. Finally, we validated the expression of the key gene CYBB by clinical samples and in vivo experiments.
Results: A total of 377 DEGs and 20 LNDE-FRGs were identified. Machine learning algorithms selected four candidate key LNDE-FRGs (CD44, CYBB, TCF4, and PARP12). However, only CYBB exhibited consistent expression trends in both the training and validation cohorts (P < 0.05). Immune infiltration analysis revealed that the expression levels of monocytes and M0 macrophages were significantly higher in the LN group than in the normal control group. In addition, there was a correlation between key genes and GFR, urinary protein and Scr. Finally, the expression level of CYBB was verified in lupus mice.
Conclusion: CYBB may be a ferroptosis-related biomarker in LN. This may provide new ideas for the clinical treatment and pathogenesis of LN.

Keywords: lupus nephritis, ferroptosis, CYBB, immune infiltration, bioinformatics, machine learning