Bowen Wei,1,2,* Qingyun Yue,1,2,* Yuxiu Ka,1,2,* Chenyang Sun,1,2 Yuxing Zhao,1,2 Xiaomei Ning,1,2 Yue Jin,1,2 Jingyue Gao,1,2 Yuanhao Wu,1,2 Wei Liu1,2 

1Department of Rheumatism and Immunity, First Teaching Hospital of Tianjin University of Traditional Chinese Medicine, Tianjin, People’s Republic of China; 2National Clinical Research Center for Chinese Medicine Acupuncture and Moxibustion, Tianjin, People’s Republic of China

*These authors contributed equally to this work

Correspondence: Wei Liu, Email fengshiliuwei@163.com

Background: Primary Sjögren’s syndrome (pSS) is an autoimmune condition marked by lymphocyte infiltration in the exocrine glands. Our study aimed to identify a novel biomarker for pSS to improve its diagnosis and treatment.
Methods: The gene expression profiles of pSS were obtained from the Gene Expression Omnibus (GEO) database. The specific differentially expressed genes (DEGs) were screened by the Least Absolute Shrinkage and Selection Operator (LASSO), Random Forest (RF), and Recursive Feature Elimination with Support Vector Machines (SVM-RFE). A biomarker was picked out based on correlation and diagnostic performance, the connection between the biomarker and clinical traits and immune infiltrating cells was explored, and the biomarker’s protein expression level in the serum of pSS patients was detected by enzyme-linked immunosorbent assay (ELISA). The competitive endogenous RNA (ceRNA) network regulated by the biomarker was predicted to verify the reliability of the biomarker in diagnosing pSS.
Results: IFI44, XAF1, GBP1, EIF2AK2, IFI27, and IFI6 showed prominent diagnostic ability, with the high accuracy (AUC = 0.859) and significance (R ≥ 0.8) of IFI44 within the training dataset. IFI44 strongly exhibited a negative correlation with resting NK cells, macrophages M0, and eosinophils, and a positive correlation with activated dendritic cells, naive B cells, and activated CD4 memory T cells. Furthermore, IFI44 was significantly positively correlated with clinical traits such as IgG, SSA, SSB, ANA, and ESSDAI, with its protein expression level in the serum of pSS patients being notably elevated compared to controls (p < 0.001). Finally, the ceRNA regulatory network showed that hsa-miR-944, hsa-miR-9-5p, hsa-miR-126-5p, and hsa-miR-335-3p were significantly targeted IFI44, suggesting that IFI44 may serve as a dependable biomarker for pSS.
Conclusion: In this study, we dug out IFI44 as a biomarker for pSS, systematically studied the potential regulatory mechanism of IFI44, and verified its reliability as a biomarker for pSS.

Keywords: primary Sjögren’s syndrome, machine learning, immune cell infiltration, biomarker, IFI44