Unveiling the Interconnected Dynamics of Mitochondrial Dysfunction Associated With Age-Related Cardiovascular Risk: A Cross-Sectional Pilot Study

Aging, influenced by complex epigenetic mechanisms, significantly contributes to the progression of cardiovascular diseases (CVDs). This cross-sectional pilot study investigated mitochondrial-associated epigenetic stress responses in two age groups: Group I (18-38, n = 154), representing younger adults generally at lower risk for CVD, and Group II (39-65, n = 105), representing middle-aged and older adults with increased biological susceptibility. The age grouping was based on established physiological and cardiovascular risk transitions typically observed around age 40. To assess age-related molecular variations, we examined key mitochondrial and metabolic parameters, including mitochondrial DNA (mtDNA) damage repair capacity, mtDNA copy number (mtDNAcn), methylation status, mitochondrial dynamics (fusion/fission), telomere length, expression of respiratory complex genes, levels of pro-inflammatory cytokines, and N-terminal pro-B-type natriuretic peptide (NT-proBNP) concentrations. Our results indicated that the older group exhibited higher mtDNA methylation (r² = 0.5205, p < 0.0001), increased mtDNAcn, and elevated NT-proBNP levels, which also showed a weak positive correlation with mtDNA methylation (r² = 0.3218, p < 0.0001). Additionally, a strong negative correlation was observed between telomerase reverse transcriptase (TERT) expression and age (r² = 0.6070, p < 0.0001), suggesting a decline in telomeric maintenance with advancing age. Group II also showed altered inflammatory and telomeric profiles and a notable reduction in the expression of mitochondrial respiratory complex genes (ND6, COXI, ATPase 6 and 8), alongside increased expression of genes involved in mitochondrial stress response pathways. We employed four machine learning models - Logistic Regression, Decision Tree, Random Forest, and Support Vector Machine (SVM) - for CVD risk prediction, using selected mitochondrial and metabolic features. All models demonstrated high classification accuracy, ranging from 0.920 to 1.0, with the Random Forest model achieving the highest accuracy of 0.984. These preliminary findings highlight distinct age-related molecular signatures and illustrate the potential of combining biomarkers with machine-learning approaches to improve cardiovascular risk prediction and therapeutic targeting in aging populations.