A Proof of Concept for a Systems Approach to Biologically Characterize the Maternal Gut Microbiome, Immune Patterns and Mental Distress During Pregnancy

ABSTRACT Importance Screening for anxiety and depression is important but especially difficult during pregnancy. Identification and quantification of the variation of maternal biologic dimensions during the dynamic pregnancy period, e.g. immune and microbiome profiles, have the potential to greatly improve mental heath screening and patient stratification. Objective To determine if specific immune and microbiome profiles align with features of mental distress during pregnancy. Design Prospective case-control study from two cohorts in US followed from 2017 to 2019 in the second and third trimesters. Setting For the urban cohort, recruitment from an obstetric clinic; for the suburbian group, recruitment from advertisements in the hospital and surrounding communities. Participants A total of 90 pregnant women from two distinct geographic regions, with contrasting race, age, marital status, education and urban vs suburban settings. Exposures Self-reported mental health symptoms (anxiety, preceptions of stress and depression) and sociodemographic characteristics; blood was obtained for cytokine profiles and stool for microbiota analysis. Main Outcomes and Measures Dimensionality reduction clustered the symptom scores from the three self-report tools. Phenotypes based on these features were then defined. Results Factor analysis revealed four features: Burn-out, Low Self-Esteem, Mixed State, and Suicidal/Self-Harm; six phenotypes: severe, with and without suicidal thoughts, depressed with low self-steem, hyperactive, and healthy. Factors and phenotype groups were statistically associated with immune and microbiota profiles and sociodemography. The hyperactive phenotype, despite having low self-reported symptoms, had elevated inflammatory cytokines and a microbiota profile similar to the severe phenotype. Elevated T-helper cell 17 (Th17), inducer of inflammation, is related to anxiety in these cohorts. Variance in socioeconomic factors and and the gut microbiota profile can account for 50% of the variance in phenotype. Conclusion Better stratification of biological charateristics is essential for understanding complex perinatal mental disorders and can be achieved by including microbiota and immune data. Our preliminary results suggest that variance in microbial and immune factors can describe differences in mental health phenotype, and could be predictive of the women’s current or future risk for future diagnosis. These findings form testable hypotheses for larger longitudinal studies. KEY POINTS Question Does self-reported mental distress in late pregnancy align with specific immune and microbiome patterns? Findings We show potential limitations in sole reliance on self-reported symptoms of mental distress in 90 subjects from two prospective cohorts. The most different groups based on symptom-based factors presented statistically similar inflammation and microbiome profiles that may reflect underreporting of mental distress symptoms. Meaning A systems biology approach identified risk immune and microbiome profiles with potential advantages to augment self-report. These biological charateristics high risk for mental distress in pregnancy should be confirmed in larger cohorts and may improve understanding of perinatal mood disorders.