Identification of immunological characteristics and cuproptosis-related molecular clusters in primary Sjögren's syndrome

Abstract

Background: Primary Sjögren's syndrome (pSS) is a chronic systemic autoimmune disease characterized by lymphocyte infiltration of the exocrine glands. The typical clinical symptoms of pSS include dryness of the mouth (xerostomia) and eyes (xerophthalmia), fatigue, and joint pain. Cuproptosis is a recently identified mode of programmed cell death that leads to the progression of multiple diseases, and the precise etiology and pathophysiology of pSS remain unknown. Consequently, the aim of our study was to explore cuproptosis-related molecular clusters and identify key genes in pSS.

Method: Gene expression profiles of the peripheral blood in the GSE84844 dataset were downloaded to identify the expression characteristics of cuproptosis regulators and immune cell infiltration. Subsequently, further exploration was conducted on the clusters involving cuproptosis-related genes (CRGs) and the corresponding immune cell infiltration, and the WGCNA algorithm was applied to explore the cluster-specific differentially expressed genes. Finally, the best machine prediction model was selected for candidate hub cuproptosis-associated genes and the accuracy of predictive efficiency was verified by the salivary gland in an external dataset (GSE143153) and enzyme-linked immunosorbent assay.

Result: Through a comparison of patients with pSS and controls, 7 CRGs and 4 types of immune cells were identified. Immune cell infiltration revealed significant immune heterogeneity in three cuproptosis-related molecular clusters in pSS. The random forest machine model showed the best discriminatory performance (area under the receiver operating characteristic curve (AUC) = 1.000) and built a predictive model based on 5 genes, which demonstrated satisfactory performance (AUC = 0.70) in the GSE143153 dataset. Based on serum samples, EED (AUC = 0.557), CBL (AUC = 0.635), and NFU1 (AUC = 0.655) showed lower expression levels in patients with pSS (p = 0.037, p = 0.000, p = 0.000, respectively).

Conclusion: In this study, we systematically analyzed the association between pSS and cuproptosis, established a predictive model that screened for high-risk genes linked to the advancement of pSS, and explored the pathogenic mechanisms and novel therapeutic strategies for pSS, targeting EED, CBL and NFU1.

Keywords: Cuproptosis; Immune cell infiltration; Machine learning model; Primary Sjögren's syndrome.