microbiome in acne lesional versus healthy skin

The biological state of clinically non-lesional skin in acne remains incompletely characterized, despite its potential relevance to disease progression and early intervention. Using longitudinal multi-omics profiling, including 16S rRNA gene sequencing (V1–V3 and V4), shotgun metagenomics, and UPLC-MS/MS metabolomics, of facial skin from 10 individuals with acne (lesional and non-lesional sites) and 10 healthy controls across multiple time points, we identified microbial and metabolic differences across healthy, non-lesional, and lesional skin. Compositional tensor factorization showed that non-lesional acne skin occupied an intermediate position between healthy and lesional skin in both microbiome and metabolome profiles. Acne-associated skin displayed reduced microbial diversity with strain-level and temporal changes correlating with lesion severity. Machine learning classification distinguished healthy from acne lesional samples (AUC=0.67) and healthy from acne non-lesional samples (AUC=0.70). Metabolomics revealed widespread lipid alterations, including elevated N-oleoylethanolamine (a PPAR-a agonist related to sebum production) consistently present in both non-lesional and lesional acne skin. Amino acid perturbations, particularly of (iso)leucine, tryptophan, and phenylalanine, were pronounced even in non-lesional sites. Integrative microbe–metabolite association analyses revealed distinct co-occurrence patterns in acne, while healthy skin showed enrichment of protective metabolites including urocanic acid. These findings indicate that clinically normal-appearing skin in individuals with acne exhibits measurable microbial and metabolic differences compared to healthy skin, providing a foundation for future studies evaluating candidate biomarkers and microbiome-informed intervention strategies.