Identification of potential biomarkers for the study of the mechanisms of immune thrombocytopenia using bioinformatics and metabolomics.

This study integrates bioinformatics and metabolomics methodologies to identify potential biomarkers and elucidate the pathogenesis of immune thrombocytopenia (ITP). Analysis of the ITP patient gene expression dataset (GSE112278) obtained from the GEO database identified 472 differentially expressed genes (DEGs), comprising 116 upregulated and 358 downregulated genes. Functional enrichment analysis revealed that these DEGs are predominantly associated with nucleotide metabolism and bile secretion pathways. Complementary metabolomic profiling of mouse serum samples detected 133 differentially expressed metabolites, including 53 upregulated and 80 downregulated compounds. Notably, metabolites such as xanthosine, 2'-deoxyadenosine, lithocholic acid 3-O-sulfate, and 20-HETE exhibited strong correlations with key genes-PNP, XDH, CA2, and SLC2A1. Receiver operating characteristic (ROC) curve analysis demonstrated that these four metabolites possess high diagnostic specificity for ITP, with area under the curve (AUC) values exceeding 0.85. Validation using an animal model, wherein mice were administered anti-CD41 antibodies, revealed significant thrombocytopenia and splenomegaly (P < 0.05), accompanied by aberrant expression of hub genes in spleen and liver tissues: PNP, CA2, and SLC2A1 were upregulated, whereas XDH was downregulated. Collectively, these findings suggest that ITP pathogenesis may be driven by perturbations in purine metabolism and bile acid-mediated immune regulation. The identified metabolites and gene networks present promising targets for non-invasive diagnostic strategies, metabolic subtype classification, and therapeutic interventions aimed at modulating purine metabolism or bile acid signaling pathways in ITP.