**Background:** Extracellular vesicles (EVs) are released in response to physical exercise and carry bioactive molecules (miRNAs, purines) that mediate intercellular communication and systemic adaptations. Urine is a noninvasive source of EVs that may reflect both local urogenital tract and systemic (e.g., skeletal muscle) status. This study aimed to characterize the shape, miRNA content, and purine profile of urinary EVs (uEVs) from long-term endurance-trained triathletes versus inactive individuals, integrating nutritional and network analyses.
**Methods:** 13 inactive males (IN) and 13 male triathletes (TR) were enrolled. IN group: mean age 42±3.4 years, BMI 27.3±2.6 kg/m², WtHR 0.56±0.05, weekly training 1.4±1.4 h. TR group: mean age 47±8.8 years, BMI 24.1±1.8 kg/m², WtHR 0.48±0.05, weekly training 10±5.8 h (moderate-to-vigorous, ~6 MET). Food consumption was assessed via 3-day diaries analyzed with FOODCONS software. Small EVs were isolated from first-morning spot urine by differential ultracentrifugation (100,000×g). EV characterization included: atomic force microscopy (AFM) for shape/roughness, transmission electron microscopy (TEM), dynamic light scattering (DLS) for size distribution, and zeta potential. Twelve miRNAs (miR-16-5p, miR-486-5p, miR-378a-5p, miR-126-3p, miR-27a-3p, let-7b-5p, miR-23a-3p, miR-133a-3p, miR-133b, miR-206, miR-34a-5p, miR-92a-3p) were quantified by RT-qPCR (normalized to miR-16-5p). Purine content (adenosine and guanosine series) was measured by HPLC. Statistical analyses included t-tests/Welch, Mann-Whitney, ANCOVA (corrected for BMI, WtHR, age), and Spearman correlation networks.
**Key Results:** Nutritional analysis showed TR had higher total energy intake (3160±889 vs. 2505±688 kcal), higher water intake (p=0.017, g=1.315), higher protein per kg (1.8±0.6 vs. 1.2±0.2 g/kg/day, p=0.026, g=1.481), and near-significantly higher carbohydrate intake (p=0.051, g=1.273). Mediterranean Adequacy Index was low in both groups (IN 2.7±1.5, TR 3.6±2.2, p=0.326). Urinalysis showed no differences in creatinine or albumin; protein was lower and osmolality slightly higher in TR. DLS revealed broad size distributions: IN had two peaks at ~100 nm and ~200 nm; TR had a dominant peak <40 nm with a tail >100 nm. TEM showed TR EVs as sphere-shaped (50–200 nm), IN EVs as spheroid-like with aspect ratio ~2. AFM demonstrated significantly greater roughness in IN vs. TR: mean roughness (p<0.001, g=2.897), grain-wise RMS (p<0.001, g=3.008), RMS roughness (p<0.001, g=2.727). EV count per AFM area was 13±9 (IN) vs. 155±53 (TR). Zeta potential was similar (~−25 mV). miRNA analysis: miR-27a-3p (p<0.001, η²p=0.580), miR-92a-3p (p<0.001, η²p=0.459), and miR-378a-5p (p=0.014, η²p=0.225) were significantly lower (lower ΔCt) in IN vs. TR. miR-23a-3p (p<0.001, r=0.858), miR-133a (p<0.001, η²p=0.653), miR-206 (p<0.001, η²p=0.548), and miR-34a-5p (p<0.001, η²p=0.679) were significantly lower (lower ΔCt) in TR vs. IN. Differences remained largely after correcting for BMI, WtHR, and age. Pathway analysis (DIANA-miRPath) of differentially expressed miRNAs (miR-92a-3p, miR-27a-3p, miR-23a-3p, miR-133a, miR-206, miR-34a-5p) identified top KEGG pathways: cell cycle (p=5.01e-10), p53 signaling (p=5.77e-08), Hippo signaling (p=9.75e-08), adherens junction (p=1.62e-07), and AMPK signaling (p=1.10e-02). Purine analysis: guanosine (GUA) was the only purine detectable in ~50% of uEV samples (concentration in hundreds nM), with no significant difference between groups. Adenosine and ATP were below limit of quantification in all samples. Network analysis showed denser connections in IN vs. TR; miR-27a-3p was linked to GUA and beer in IN but disconnected in TR; GUA was linked to meat in TR.
**Clinical Implications:** This study provides evidence that long-term endurance training produces a distinct uEV signature—smaller, smoother, spheroid-shaped vesicles with altered metabolic miRNA expression and detectable guanosine—that is largely independent of BMI, WtHR, and age. uEV profiling offers a noninvasive, accessible approach to monitor exercise-induced metabolic adaptations and distinguish physically active from sedentary individuals. The differential miRNA expression (e.g., miR-27a-3p, miR-133a, miR-206, miR-34a-5p) and their predicted targeting of pathways like p53, Hippo, and AMPK signaling suggest molecular mechanisms linking exercise to metabolic health. The presence of guanosine in uEVs opens new perspectives on muscle-brain crosstalk and exercise-dependent neurotrophic effects. Limitations include small sample size, lack of EV subtype sorting, and absence of EV core protein confirmation. Larger studies are needed to validate these findings and develop uEV-based biomarkers for monitoring physical activity and preventing dysmetabolic diseases.