A multi-tissue integration of immunocytes and inflammaging biomarkers predicts biological age through LASSO-optimized modeling.

Abstract

Immunosenescence, a recognized hallmark of aging, is characterized by imbalances in immunocyte populations and a state of chronic inflammation. However, the tissue-specific dynamics of these changes and their potential as predictive biomarkers for aging remain poorly characterized. In this study, we established a multi-tissue immunological signature as a robust predictor of biological age by integrating immunocyte and cytokine profiling. Using Sprague-Dawley (SD) rats from five age groups (1-12 months), we systematically quantified 45 immunocyte subsets across peripheral blood, mesenteric lymph nodes, thymus, and spleen using flow cytometry, and profiled 22 serum cytokines/chemokines via Flexible Multi-Analyte Profiling (xMAP). Firstly, classic age-dependent shifts were observed across our rat samples, including progressive thymic involution and depletion of peripheral T-cells. Cytokine levels exhibited age-related chronic inflammation progression, marked by elevated IL-1α, granulocyte colony-stimulating factor (G-CSF), and TNF-α. To integrate these multidimensional datasets into a predictive aging metric, we employed Least Absolute Shrinkage and Selection Operator (LASSO) regression, selecting 22 biomarkers through regularization (λ = 0.111). The integrated model combining cellular and cytokine data demonstrated superior performance (training R