231. Multi-omic longitudinal profiling reveals functional and metabolic gut microbiota signatures predictive of antidepressant response in first episode, drug-naïve depression

Abstract

Abstract Background Antidepressant treatment response is highly heterogeneous, and the biological mechanisms underlying this variability remain insufficiently understood. Emerging evidence suggests that gut microbiota and microbially derived metabolites influence neurobiological pathways relevant to antidepressant efficacy; however, most studies have relied on cross-sectional sampling or single-omic approaches, limiting insight into temporal microbial remodeling, functional adaptation, or metabolic signaling during treatment. Critically, how dynamic changes in microbial composition, functional pathways, and fecal metabolites interact with the evolution of depressive symptoms remains unclear. A longitudinal, multi-omic framework is therefore essential to clarify the gut–brain–metabolite processes that shape therapeutic outcomes. Aims & Objectives This study aimed to systematically characterize temporal interactions among the microbiota, metabolites, and symptoms across the course of antidepressant therapy. Specifically, we sought to: (1) map longitudinal trajectories of gut microbial composition, functional pathways, and metabolite profiles in responders versus non-responders; (2) evaluate directional relationships between microbiota shifts and symptom change using cross-lagged panel modeling; (3) identify microbial taxa and pathways associated with treatment response; and (4) determine whether specific bile acids or short-chain fatty acids (SCFAs) mediate associations between microbial alterations and clinical improvement. These objectives address the multi-layered processes underlying heterogeneity in antidepressant effectiveness. Method First-episode, antidepressant-naïve individuals with major depressive disorder or bipolar depression were enrolled in an 8–10 week longitudinal study. Of 32 participants, 26 completed all visits, contributing 75 stool samples across baseline (BL), Time Point 1 (TP1), and Time Point 2 (TP2). Responders (n = 15, 57.7%) and non-responders (n = 11, 42.3%) were classified at the final visit. Shotgun metagenomics (MetaPhlAn4, HUMAnN3) provided species-level taxonomic and pathway-level functional profiles. Targeted metabolomics quantified 42 bile acids and 11 SCFAs. Mixed-effects models assessed temporal patterns and group differences; cross-lagged panel models evaluated bidirectional relationships; and mediation analyses tested whether metabolite changes mediated microbe–response associations. Results Several microbial species showed longitudinal abundance shifts that preceded—and statistically predicted—subsequent improvements in depressive symptoms, supporting a directional effect from microbiota to symptoms. Responders displayed increasing metabolic redundancy, more balanced species contributions, and resilient pathway organization. In contrast, non-responders developed progressively centralized and less redundant functional structures, indicating fragility. Metabolomic profiling revealed distinct trajectories: responders preserved or increased multiple secondary bile acids and recovered butyric acid, whereas non-responders showed coordinated declines across these metabolites. Mediation analyses further identified glycolithocholic acid, deoxycholic acid, lithocholic acid, and 3-ketocholanic acid as significant intermediaries linking targeted species (e.g., Clostridium saudiense, Intestinibacter bartlettii, Megasphaera sp. NM10, Blautia hansenii) to treatment response. Discussion & Conclusions Our multi-omic longitudinal findings show that antidepressant response reflects not broad diversity shifts but coordinated, fine-scale taxonomic remodeling and metabolite recovery. Responders maintained higher functional redundancy and balanced pathway contributions, indicating a resilient microbial ecosystem, whereas non-responders developed centralized and fragile functional structures. Divergent bile acid and SCFA trajectories also differentiated the groups, and mediation analyses identified key bile acids as intermediaries linking microbial taxa to symptom improvement. Overall, these results highlight a dynamic gut–brain–metabolite axis that shapes treatment outcomes and support microbiota-informed, metabolite-targeted approaches to improve antidepressant efficacy.