Aging-Related Transcriptomic Signatures Associated with Immune Heterogeneity in Tuberculosis: An Exploratory Multi-Cohort Analysis
Year: 2026
Type: article
Abstract: Background: Aging-related genes (ARGs) may contribute to immune heterogeneity in tuberculosis (TB), but their diagnostic and biological relevance remain incompletely defined. This study aimed to identify TB-associated ARG transcripts, characterize their immune and molecular correlates, and evaluate their exploratory diagnostic performance across prespecified cohorts. Patients and Methods: Public TB transcriptomic cohorts were analyzed together with a local bulk RNA-seq cohort comprising 5 TB patients and 5 controls and a local single-cell RNA-seq cohort comprising 3 active TB patients and 3 controls. The local cohorts were selected using random sampling. Age-adjusted differential expression, weighted gene co-expression network analysis, machine-learning-based feature selection, immune-infiltration analysis, consensus clustering, single-cell annotation, and Comparative Toxicogenomics Database-based compound–gene screening were performed. Results: GSE83456 included 45 TB patients and 61 controls. Age-adjusted differential expression identified 189 differentially expressed genes, and the intersection of DEGs, ARGs, and blue-module genes yielded 11 discovery candidates. Nested model selection in GSE83456 identified a naive Bayes classifier based on STAT1, PLAUR, JAK2, and P2RY14. Across five repeated stratified nested cross-validation runs, the mean AUC was 0.962, with a repeat-level range of 0.954– 0.969. Balanced accuracy, sensitivity, specificity, and accuracy were 0.926, 0.911, 0.941, and 0.928, respectively. Exploratory evaluations included GSE19439, GSE34608, GSE62525, GSE19444, and the local RNA-seq cohort. Consensus clustering separated TB samples into two ARG-associated subgroups with distinct immune-infiltration and pathway-activity profiles. CIBERSORT analysis identified significant TB-control differences in 9 of 22 inferred immune-cell types. Single-cell analysis demonstrated heterogeneous expression of STAT1, PLAUR, JAK2, and P2RY14 across annotated immune-cell subsets. CTD screening generated exploratory compound–gene associations that provide testable hypotheses for future experimental studies. Conclusion: Age-adjusted multi-cohort analysis identified candidate ARG-associated TB markers and immune patterns. STAT1, PLAUR, JAK2, and P2RY14 are candidate markers rather than clinically validated biomarkers and require independent experimental and clinical validation. Keywords: immunosenescence, transcriptomics, machine learning, single-cell RNA sequencing, immune heterogeneity, inflammatory biomarkers
Source: Journal of Inflammation Research
DOI: 10.2147/jir.s633793
Language: English
Subfield: Infectious Diseases
Field: Medicine
Domain: Health Sciences
Keywords: tuberculosis, aging-related genes, immune heterogeneity, gene expression signatures, STAT1, PLAUR, JAK2, diagnostic biomarkers, immune cell infiltration, single-cell RNA sequencing, consensus clustering, Naive Bayes classifier, toxicogenomics, immune cell subsets, gene co-expression network, multi-cohort analysis
Open Access status: gold
Estimated APC paid: $2,836
OpenAlex ID: W7225986487
