A45 FIBER-FREE DIET REDUCES BACTEROIDES ABUNDANCE AND PREVENTS MUCUS DEGRADATION IN MICE COLONIZED WITH MICROBIOTA FROM PATIENTS WITH GENERALIZED ANXIETY DISORDER

Background

Generalized anxiety disorder (GAD) is a debilitating condition with a lifetime prevalence of 4-7% worldwide. We have previously found that compared to healthy controls, GAD patients had lower reported fiber intake, increased gastrointestinal symptoms; and enrichment of Bacteroides genus as well as carbohydrate metabolism pathways (as determined by PICRUSt2, correlated to Bacteroides abundance). Bacteroides are known for its ability to degrade a wide variety of host polysaccharides, such as the intestinal mucus, which could lead to local and systemic inflammation. In this regard, GAD patients had higher C-reactive protein (CRP) compared to healthy controls (p=0.049).

Purpose

To investigate whether a fiber-free diet could decrease Bacteroides abundance and prevent damage of the mucus layer reducing anxiety-behavior in mice with GAD microbiota.

Method

Two germ-free NIH Swiss mouse breeding pairs were colonized with GAD microbiota using patients’ stool samples and kept on either fiber-free or 10 % inulin (fiber) diet. Offspring were weaned at week 3 and psychometric tests were performed at 10 weeks of age. After sacrifice, samples for histology (mucus layer thickness determination), blood (CRP ELISA determination) and stool (Illumina 16S rRNA gene sequencing) were collected. The microbiota data was analyzed following the pipelines of dada2 and by mean comparisons, correlation, AncomBC using R software (v.1.2.1335). Multiple comparison results were corrected allowing 5% of FDR.

Result(s)

Beta diversity analysis showed that parent and offspring’s (n=24 fiber-supplemented and n=14 fiber-free groups) microbiota was similar to the GAD donor. The most differentially abundant bacterial taxon was Bacteroides uniformis, which was decreased in the fiber-free group (p.adj= 0.003). Furthermore, fiber-free diet reduced the overall Bacteroides abundance by half compared to the donor and fiber-supplemented mice group. This led to a thickening (p.adj=0.027) of the mucus layer, increasing from 30 µm (fiber-supplemented diet) to 60 µm in the fiber-free diet group. B. uniformis was negatively correlated to the mucus layer thickness (R= -0.81; p.adj=0.26), although not statistically significant, likely due to a low n number (n=4). We only found a statistical trend for higher CRP levels and anxiety-like behavior in the fiber-supplemented group. This might be because fiber supplementation has several beneficial effects that can counteract the proposed increase in anxiety-like behavior fromBacteroides. Despite that, we found a significant correlation between B. uniformis and time mice spent in the dark (indicative of anxiety-like behavior) in the light preference test.

Conclusion(s)

Our data suggests that Bacteroides abundance, specifically Bacteroides uniformis, contributes to the degradation of the mucus layer and potentially triggers low grade gut inflammation and anxiety-like behavior.

Please acknowledge all funding agencies by checking the applicable boxes below

CIHR

Disclosure of Interest

None Declared

A229 GUT MICROBIOTA PROFILES, DIET AND SHORT-CHAIN FATTY ACIDS AS PREDICTORS OF GENERALIZED ANXIETY DISORDER

Background

Generalized anxiety disorder (GAD) is a debilitating chronic condition with a lifetime prevalence of 4–7% worldwide. Both diet and gut microbiota have been previously associated with anxiety.

Aims

To investigate whether bacterial taxa and/or nutrients associate with GAD, and whether they differ from those of healthy controls (HC).

Methods

Patients with GAD (n=82) and matched HC (n=97) were assessed by validated questionnaires for anxiety (DASS-21), gastrointestinal (GI) symptoms (Rome III, Short-Form Leeds Dyspepsia), and dietary profiles by the Dietary Questionnaire for Epidemiological Studies. We quantified several blood and stool biomarkers, including inflammatory and neuroactive metabolites, as well as short-chain fatty acids. Stool microbiota profiles were assessed by16S rRNA gene sequencing through Illumina. The data was then analyzed following the pipelines of dada2 and by multiple factor analysis (MFA), mean comparisons, correlation, LEfSe and XGBoost using R software (v.1.2.1335). Multiple comparison results were corrected allowing 5% of FDR.

Results

Using MFA to analyze all variables, we identified 3 clusters: one mainly composed of HC (n=99, 91% HC, GI symptoms in 25% of subjects), a second mixed cluster (n=30, 80% GAD, GI symptoms in 80%) and a third cluster mainly composed of GAD patients (n=50, 98% GAD, GI symptoms in 86%). When focusing only on the HCs of cluster 1 (n=90) and GADs of cluster 3 (n=49), we found higher GI symptoms, body mass index, serum C-reactive protein and stool calprotectin levels (adj. p=1.3x10-9, 0.001, 0.017 and 0.017, respectively) and lower concentrations of propionate, butyrate and acetate in GAD compared to HC. GADs also reported overall lower caloric intake (kJ/day; adj. p=1.7x10-4) in the food frequency questionnaire. Fibre (g/day) was the macronutrient most negatively associated with anxiety scores (R=-0.44; adj. p=4.2x10-5). Bacteroides was the only bacterial taxon significantly associated with GAD, as well as with anxiety scores (R=0.31, adj. p=0.003). Interestingly, Bacteroides/fiber ratio was strongly correlated to anxiety scores (R=0.58, adj. p=2.7x10-09). Furthermore, demographic, biomarkers and bacterial taxa data were predictive of the patients’ disease state with 92.8% accuracy. The features that aid the model to predict disease state were Bacteroides/fiber ratio, GI symptoms and stool acetate levels.

Conclusions

Our results suggest that most GAD patients differ in dietary and microbiota profiles from HCs, and that the Bacteroides/fiber ratio, stool acetate and GI symptoms might be good predictors of disease state. Furthermore, these data strongly support the role of microbiota-gut-brain axis in genesis of psychiatric diseases, and they will inform mechanistic studies in gnotobiotic mouse models.

Funding Agencies

NIH

A261 MICROBIOTA PROFILES OF PATIENTS WITH MENTAL DISODERS DIFFER FROM THOSE OF HEALTHY CONTROLS

Background

Mental disorders are the leading cause of disabilities worldwide, with depression and anxiety among the most common ones, affecting up to 1/3 of the worldwide population at least once in their lifetime. In both preclinical models and clinical studies, gut microbiota has been associated with altered behavior and anxiety or depression, respectively.

Aims

To investigate 1) whether the microbial profiles of patients with generalized anxiety disorder (GAD) and major depression disorder (MDD) differ from those of healthy controls (HC), and 2) whether specific bacterial taxa associate with GAD or MDD.

Methods

118 patients with primary GAD (n=82, 83.3 % female) or MDD (n=36, 62.9 % female) and 99 matched HC (66.6 % female) were recruited through the Anxiety Treatment and Research Centre. Anxiety, depression and stress levels were assessed by DASS-21 questionnaire. Stool samples were collected anaerobically and analysed for 16S rRNA gene sequencing through Illumina technique. The data was divided in 4 groups: 1) mental health disorder (MHD) combining GAD and MDD, 2) GAD, 3) MDD, and 4) HC. The data was analyzed following the pipelines of dada2 and QIIME2. RandomForest plugin for QIIME2 was used to investigate predictive characteristics of MHD, GAD or MDD microbiota. SPSS software v.23 was used to perform Spearman correlations and logistic regressions between microbial taxa and clinical scores.

Results

The mean anxiety score was 16.2 (severe anxiety) for GAD patients and 9.8 (moderate anxiety) for MDD patients; the mean depression score was 19.2 (moderate depression) for MDD patients and 16.0 (moderate depression) for GAD patients, while healthy controls averaged only 1.5 (normal anxiety) and 1.7 (normal depression) for anxiety and depression, respectively. The microbiota profile of the MHD group was predictive of the patients’ disease state with an 83.3% accuracy. In particular, increased relative abundance of Bacteroides ovatus and Bacteroides spp. and decreased relative abundance of Dialister spp. (Veilonellaceae), Haemophilus parainfluenzae and Bifidobacterium adolescentis, were predictive of MHD. Neither the GAD or MDD group microbiota profiles alone were accurate in the prediction of the patients’ disease state. There was a positive correlation between the relative abundance of Bacteroides spp. and a negative correlation between the relative abundance of Clostridium sensu stricto spp. and Sutterella, and the clinical scores of combined MDH and HC groups.

Conclusions

Our data suggest that patients with mental health disorders have different microbiota profiles compared to healthy controls. We have identified specific bacterial signatures that will inform mechanistic studies in gnotobiotic mouse models to investigate further the role of microbiome in mental disorders.

Funding Agencies

NIH