# Identification of anoikis-related genes and immune infiltration characteristics in Sjögren’s syndrome based on machine learning

## TL;DR

This study identifies genes linked to anoikis in Sjögren’s syndrome and explores their role in immune responses and potential drug targets.

## Contribution

The study introduces a machine learning-based approach to identify novel anoikis-related biomarkers and immune infiltration patterns in Sjögren’s syndrome.

## Key findings

- 35 differentially expressed anoikis-related genes were identified, linked to inflammation and immune pathways.
- Five key genes (MAPK3, IL15, S100A9, IFI27, CXCL10) were validated as potential biomarkers for SS progression.
- Eight candidate drugs and five regulatory miRNAs were identified as potential therapeutic targets in SS.

## Abstract

Anoikis, a recently identified type of programmed cell death analogous to apoptosis, has been implicated in the pathogenesis of Sjögren’s syndrome (SS). Although accumulating evidence indicates its involvement in modulating immune responses and contributing to SS progression, the precise role of anoikis in SS remains inadequately understood. This study aimed to explore anoikis-related genes (ARGs) and their molecular mechanisms in SS using public databases.

SS datasets (GSE23117, GSE84844 and GSE12795) were retrieved from the GEO database. In total, 924 ARGs were extracted from the GeneCards and Harmonizome databases, followed by differential expression gene (DEGs) analysis and weighted gene co-expression network analysis (WGCNA). Machine learning algorithms were utilized to screen candidate biomarkers, and their diagnostic effectiveness was assessed using receiver operating characteristic (ROC) curve analysis. Concurrently, a mouse model of SS was established and validated through in vivo experiments. Immune cell infiltration in SS tissues was evaluated using CIBERSORT, and correlations between characteristic genes and immune cell profiles were analyzed. Potential drug candidates targeting these genes were identified using the DGIdb database. Subsequently, an lncRNA-miRNA-mRNA network associated with these genes was constructed, and preliminary experimental validation was conducted.

A total of 35 differentially expressed anoikis-related genes (DEARGs) were identified. GO and KEGG enrichment analyses demonstrated that DEARGs were primarily associated with inflammation, viral infections, and the necroptosis signaling pathway. Machine learning analysis pinpointed 14 feature genes, among seven were associated with cancer (NAT1, BIRC3, EZH2, MAD2L1, ATP2A3, HMGA1, and BST2). Given the unclear roles of SKI and PRDX4 in SS, the study focused specifically on five relevant genes, MAPK3, IL15, S100A9, IFI27, and CXCL10, which were validated by in vivo experiments. Immune cell analysis revealed increased proportions of B cells, T cells, macrophages, and other immune cells in SS tissues. Furthermore, ceRNA and drug-gene interaction networks were established, underscoring the regulatory significance of five key miRNAs (miR-30b-5p, miR-148a-3p, miR-130a, miR-483-5p, and miR-486-3p) in SS. In addition, eight candidate drugs were identified with potential for modulating SS pathogenesis.

This study substantiates the significant involvement of anoikis in SS and suggests that MAPK3, IL15, S100A9, IFI27, and CXCL10 may serve as critical biomarkers in the inflammatory progression of SS. These genes likely mediate their effects by influencing immune cell infiltration, participating in immune regulation, and modulating inflammatory responses. Our findings offer new insights into drug selection and immunotherapeutic strategies for SS.