Dynamic human gut microbiome and immune shifts during an immersive psychosocial intervention program

Abstract

Background: Although depression is a leading cause of disability worldwide, the pathophysiological mechanisms underlying this disorder-particularly those involving the gut microbiome-are poorly understood.

Method: To investigate, we conducted a community-based observational study to explore complex associations between changes in the gut microbiome, cytokine levels, and depression symptoms in 51 participants (M_age = 49.56, SD = 13.31) receiving an immersive psychosocial intervention. A total of 142 multi-omics samples were collected from participants before, during, and three months after the nine-day inquiry-based stress reduction program.

Results: Results revealed that depression was associated with both an increased presence of putatively pathogenic bacteria and reduced microbial beta-diversity. Following the intervention, we observed reductions in neuroinflammatory cytokines and improvements in several mental health indicators. Interestingly, participants with a Prevotella-dominant microbiome showed milder symptoms when depressed, along with a more resilient microbiome and more favorable inflammatory cytokine profile, including reduced levels of CXCL-1.

Conclusions: These findings reveal a potentially protective link between the Prevotella-dominant microbiome and depression, as evidenced by a reduced pro-inflammatory environment and fewer depressive symptoms. These insights, coupled with observed improvements in neuroinflammatory markers and mental health from the intervention, may highlight potential avenues for microbiome-targeted therapies for managing depression.

Keywords: CXCL-1; Gut microbiome; Neuroinflammation; Psychosocial intervention.

Fig. 1.. Study Design, Depression Trajectories, and Microbiome Composition Analysis.

(A) Cartoon representation of the study’s methodology (Image adapted from Pixabay). (B) Beck Depression Inventory-II (BDI-II) scores of participants across four time points, with each dashed line corresponding to one individual. The color coding indicates whether the individual was depressed (red) or non-depressed (blue) at baseline. (C) Principal Coordinates Analysis (PCoA) of microbiome composition at the four time points. The first two axes are plotted, with the variance captured annotated along each axis. Red and blue colors denote the initial depression status of the participants, with red indicating depressed and blue indicating non-depressed individuals. (D) Intra-individual dissimilarity between T1 and the rest of the program.

Fig. 2.. Enterotype Analysis Reveals Potential Beneficial Role of Prevotella in Depression.

(A) Enterotype Classification of Microbiome Samples. Samples were clustered into three enterotypes. Each cluster is represented by a predominant genus, detailed in Supplementary Fig. S2A. Colors represent different enterotypes: Ruminococcus (R, blue), Bacteroidetes (B, green), and Prevotella (P, purple). (B) PCA of Psychometric Parameters Based on Depression Status. Samples are colored based on each individual’s initial depression status (red for depressed, blue for non-depressed). The first two principal components (PCs) are plotted, with the variance explained by each PC annotated. (C) Pairwise Comparison of Psychometric Data by Enterotype and Depression Status. Comparison across Ruminococcus (Ru, blue), Bacteroidetes (Ba, green), and Prevotella (Pe, purple) enterotypes. A two-sided Student’s T-test was used for each comparison. Significance levels are indicated as follows: p < 0.1 (.), p < 0.05 (*), p < 0.01 (**), p < 0.005 (***).

Fig. 3.. Association of Prevotella Enterotype with Reduced Inflammation and Increased Gut Microbiome Stability.

(A) Hierarchical Clustering of Cytokine Values. Clustering based on enterotype and depression status, illustrating the cytokine profiles across different groups. (B) Bayesian mixed-effects model identifying significant correlations between microbial genera and cytokine levels across three enterotypes (Bacteroides, Ruminococcus, and Prevotella). The left column represents significant cytokine-microbe associations identified in the Bacteroides enterotype. The middle and right columns show how these associations are either maintained or altered in the Ruminococcus and Prevotella enterotypes, respectively. A positive or negative sign indicates the direction of the correlation (positive or negative) in the Bacteroides enterotype and how this relationship changes in the Ruminococcus and Prevotella enterotypes. “Ns” denotes no significant change in the given enterotype. The color-coded flows (green, blue, red, and yellow) represent the consistency or divergence of these associations across the enterotypes. (C) Co-occurrence Network of Microbial Families. Network representations comparing depressed and non-depressed individuals. The left network represents the non-depressed group, and the right network represents the depressed group. Unique microbial families to each network are color-coded (blue for non-depressed, red for depressed). (D) Immediate Changes in Cytokines Post-Intervention. Significant Cytokine Changes Between T1 and T2. Identification of cytokines that showed significant differences between T1 and T2 in at least one group (depressed or non-depressed). (E) Delayed Changes in Cytokines Post-Intervention Significant Cytokine Changes Between T2 and T3: Identification of cytokines that showed significant differences between T2 and T3 in at least one group (depressed or non-depressed). A pairwise two-sided Student’s T-test was used for comparing cytokine levels between different time points. Significance levels are indicated as follows: BH-adjusted p < 0.1 (.), BH-adjusted p < 0.05 (*), BH-adjusted p < 0.01 (**), BH-adjusted p < 0.005 (***).

Fig. 4.. Mediation Linkage between Microbiome, Cytokine and Mental Health.

(A) The top five bacterial genera identified through mediation analysis for their influence on mental health outcomes, mediated by cytokine levels. (B) The mediation association involving the psychometric parameter ‘Enticing,’ the genus Prevotella, and the cytokine CXCL1. (C) The mediation association involving the psychometric parameter ‘Safe,’ the genus Escherichia/Shigella, and the cytokine CXCL13.