Abstract
Asthenozoospermia and teratozoospermia are common causes of male infertility. Despite their prevalence, the underlying metabolic mechanisms remain poorly understood. In this study, we conducted targeted metabolomic profiling of sperm samples from 131 Chinese reproductive-age men (48 normozoospermic controls, 40 asthenozoospermic patients, and 43 teratozoospermic patients) to identify distinct metabolic signatures associated with these conditions. We identified 47 significantly altered metabolites in asthenozoospermia compared to normozoospermia (18 downregulated and 29 upregulated). In teratozoospermia, 25 metabolites showed significant changes compared to normozoospermia (10 downregulated, 15 upregulated). These differential metabolites, along with others identified through machine learning as relevant, are implicated in key pathways including energy metabolism, lipid metabolism, and amino acid metabolism. Alterations, such as those in corticosterone and hypoxanthine, suggest potential impacts on redox balance and stress responses. Using machine learning approaches, we developed diagnostic models for asthenozoospermia and teratozoospermia diagnosis, with the Glmnet model showing strong performance (AUC = 0.99 for asthenozoospermia, AUC = 0.9997 for teratozoospermia), although these results require external validation. Two metabolites emerged as candidate diagnostic biomarkers: corticosterone reveals common metabolic disturbances in both pathological conditions, and arachidate specifically indicates impaired lipid metabolism in asthenozoospermia. These findings improve our understanding of the metabolic basis of sperm dysfunction, and highlight the potential of sperm metabolomics as a powerful tool for diagnosing and managing male infertility.