Apoptosis-associated genetic mechanisms in the transition from rheumatoid arthritis to osteoporosis: A bioinformatics and functional analysis approach

Abstract

This study explores the mechanisms of glucocorticoid-induced osteoporosis (OP) and Rheumatoid arthritis (RA), focusing on apoptosis and its role in the progression from RA to OP. Using microarray data from the GEO database, differential gene expression analysis was conducted with the limma package, identifying significant genes in RA and OP. Weighted Gene Co-expression Network Analysis (WGCNA) further examined gene relationships with the disease status, identifying co-expression patterns. Key genes were pinpointed by intersecting differentially expressed genes from RA and OP datasets with WGCNA module genes. Functional enrichment analysis using the "clusterProfiler" package focused on Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathways. Machine learning methods, including Lasso and Random Forest, refined the selection of key genes related to apoptosis. Immune infiltration analysis using CIBERSORT assessed immune cell differences between disease and normal samples. The study highlighted two crucial genes: ATXN2L and MMP14. These genes were identified through various analyses and found to be significantly associated with the progression of RA and OP. Gene Set Enrichment Analysis of ATXN2L and MMP14 revealed their involvement in specific biological processes and pathways. Correlation analysis between these key genes and immune cell infiltration showed significant associations. The ROC analysis evaluated the diagnostic performance of ATXN2L and MMP14, with miRNA regulatory networks related to these genes also predicted. In summary, this research provides valuable insights into the molecular mechanisms of RA and OP, emphasizing the importance of apoptosis and immune processes.

FIG. 1.

Differential analysis and WGCNA Results for RA; (a) volcano plot of differentially expressed genes; (b) heatmap showing differentially expressed genes; (c) display of threshold for WGCNA analysis; and (d) display of various modules in WGCNA.

FIG. 2.

Identification of key genes: (a) heatmap displaying differentially expressed genes in OP patients; (b) volcano plot showing differentially expressed genes in OP patients; and (c) determination of key genes in RA and OP.

FIG. 3.

Enrichment analysis results of key genes; (a) and(b) results of the GO enrichment analysis and (c) and (d) results of the KEGG enrichment analysis.

FIG. 4.

Protein–protein interaction analysis. (a) Intersection of key genes and (b) display of the protein–protein interaction network.

FIG. 5.

Machine learning screening. (a) and (b) Key genes selected using the LASSO method based on the OP dataset and (c) and (d) key genes selected using the Random Forest method based on the OP dataset.

FIG. 6.

Determination of key genes; (a) and (b) key gene selection based on the RA dataset using LASSO method; (c) intersection of key genes; (d) and (e) key gene selection based on the RA dataset using Random Forest method.

FIG. 7.

GSEA results; (a) GO analysis results for ATXN2L; (b) KEGG analysis results for ATXN2L; (c): GO analysis results for MMP14; and (d) KEGG analysis results for MMP14.

FIG. 8.

Immune infiltration analysis, (a) correlation between ATXN2L and immune infiltration results; (b) correlation between MMP14 and immune infiltration results; and (c) box plot of immune infiltration results across different groups.

FIG. 9.

Key protein ROC and interaction networks. (a) ROC graphs of key proteins. (b) Protein–protein interaction (PPI) network of key proteins. (c) miRNA interaction network of MMP14. (d) miRNA interaction network of ATXN2L