Exploratory seminal plasma metabolomics and independent HPLC-UV evaluation in men with impaired semen quality

Impaired semen quality may involve concurrent abnormalities in sperm concentration, total sperm number, motility, morphology, and DNA integrity. The corresponding alterations in the seminal plasma metabolome remain incompletely characterized.

To characterize seminal plasma metabolic alterations in men with multiple concurrent semen abnormalities and identify candidate signals associated with semen quality parameters.

Untargeted LC-MS/MS metabolomics was performed in a discovery cohort comprising 10 men with impaired semen quality and 9 normozoospermic controls. Candidate metabolites were screened using nominal p value and fold-change thresholds, followed by pathway and correlation analyses, post hoc ROC analysis, and qualitative comparisons with public datasets. External-calibration-derived estimates assigned to cytidine and 2-hydroxy-3-methylbutyric acid (2-HMBA) were subsequently evaluated by HPLC-UV in an independent cohort of 18 men with impaired semen quality and 18 controls.

Eighty-nine putatively annotated metabolites met the nominal screening criteria, but no individual metabolite remained significant after Benjamini-Hochberg false discovery rate correction. Discovery-level pathway analyses highlighted nucleotide- and nucleoside-related processes as hypothesis-generating signals. Qualitative comparison with heterogeneous publicly available reproductive metabolomics datasets provided limited directional context but was not considered formal external validation. In the independent HPLC-UV cohort, external-calibration-derived estimates assigned to 2-HMBA were nominally lower in the group with impaired semen quality (p = 0.030), but the difference did not remain significant after correction across the two candidate comparisons (q = 0.060). Cytidine showed no between-group difference (p = 0.812; q = 0.812).

The discovery-level nucleotide-related pathway signal and the independent observation concerning 2-HMBA represent distinct lines of evidence; the latter does not validate the former. Given the small sample sizes in both the discovery and independent HPLC-UV evaluation cohorts, the absence of FDR-significant metabolites, heterogeneity across the public datasets, and limitations of the HPLC-UV assay, these findings remain exploratory and hypothesis-generating. They do not constitute quantitative validation, establish a specific mechanism, or support clinical biomarker use. Larger multicenter cohorts and targeted quantitative metabolomic analyses using validated methods are needed.