Abstract
BACKGROUND: Adolescence is the peak period of newly-onset bipolar disorder (BD). Accumulating studies have revealed disturbed gut microbiota can interfere with neurodevelopment in adolescents. In this study, we aimed to characterize the gut microbiota in adolescents with BD and its correlation with brain dysfunction. METHODS: Thirty unmedicated BD adolescents within depressive episode were recruited and underwent four-week quetiapine treatment. Twenty-five age-, gender-, and BMI-matched healthy controls (HCs) were recruited. Fecal samples were collected from HCs and all BD adolescents before and after treatment and analyzed by metagenomic sequencing. Resting-state cranial functional magnetic images were collected from 21 BD adolescents before treatment. Random forest models were used to evaluate the discriminative power of gut microbiota and neuroimaging data for BD and the predictive power of treatment effect. RESULTS: Although no significant difference was found in alpha-diversity, intra- and inter-group differences in beta-diversity were observed among HCs, pre- and post-treatment patients. Compared to HCs, unmedicated BD adolescents presented a differentiated gut microbial communities, which correlated to the short-chain fatty acids, choline, lipids, vitamins, polyamines, aromatic amino acids metabolic pathways. Four-week quetiapine treatment improved the abundance of specific genus, such as Odoribacter splanchnicus, Oribacterium sinus, Hafnia alvei, Fusobacterium periodonticum, Acidaminococcus interstini and Veillonella rogosae. Neuroimaging analysis revealed sensor-emotional brain regions were associated with BD severity. Finally, random forest models based on gut microbial biomarkers can well distinguish unmedicated BD from HCs (AUC = 91.12%) and predict the treatment effect (AUC = 91.84%). The random forest model integrating gut microbiota and neuroimaging data exhibited a better predictive efficacy than using microbiota data alone. CONCLUSION: This study first characterized the gut microbiota architecture in adolescent BD. Combining gut microbiota and brain function biomarkers may benefit disease diagnosis and predict treatment outcome. Nonetheless, these findings should be carefully interpreted considering the limitations of a modest sample size and the absence of detailed mechanistic explorations. Trial registration NCT05480150. Registered 29 July 2022-Retrospectively registered, https://clinicaltrials.gov/study/NCT05480150 .