Dietary manipulation of the gut microbiome in inflammatory bowel disease patients: Pilot study

Abstract

Diet is a modifiable, noninvasive, inexpensive behavior that is crucial in shaping the intestinal microbiome. A microbiome "imbalance" or dysbiosis in inflammatory bowel disease (IBD) is linked to inflammation. Here, we aim to define the impact of specific foods on bacterial species commonly depleted in patients with IBD to better inform dietary treatment. We performed a single-arm, pre-post intervention trial. After a baseline period, a dietary intervention with the IBD-Anti-Inflammatory Diet (IBD-AID) was initiated. We collected stool and blood samples and assessed dietary intake throughout the study. We applied advanced computational approaches to define and model complex interactions between the foods reported and the microbiome. A dense dataset comprising 553 dietary records and 340 stool samples was obtained from 22 participants. Consumption of prebiotics, probiotics, and beneficial foods correlated with increased abundance of Clostridia and Bacteroides, commonly depleted in IBD cohorts. We further show that specific foods categorized as prebiotics or adverse foods are correlated to levels of cytokines in serum (i.e., GM-CSF, IL-6, IL-8, TNF-alpha) that play a central role in IBD pathogenesis. By using robust predictive analytics, this study represents the first steps to detangle diet-microbiome and diet-immune interactions to inform personalized nutrition for patients suffering from dysbiosis-related IBD.

Keywords: IBD; Microbiome; diet; dysbiosis.

Figure 1.

A) Participant inclusion and exclusion during the study duration. B) A schematic representation of the study design which involved bi-weekly stool samples collection and completion of 24-hour IBD-AID Food Querys up to three times a week throughout the study. At the beginning of the baseline and the end of the intervention, blood samples were collected.

Figure 2.

Participants adhere to the IBD-AID. A) Boxplot of the serving sizes reported for each food category consumed by CD and UC participants at baseline (BSL) and intervention (INT). B) Reported servings per week of foods with increased consumption during the intervention (Multiple T-test, p-value < 0.05). The mean servings per study period: BSL (in red) and INT (in blue), was calculated on the average intake per food category per week. Each circle represents the mean intake per food category grouped in 2 weeks intervals.

Figure 3.

Mixed effect random forest classification analysis identified microbes affected by the intervention. Bar plots show the variance of the importance of bacterial species found to be enriched (in green) or depleted (in gray) during the intervention in all (A and B), CD (C and D), and UC (E and F) subjects completing the intervention (BH p-value >0.05).

Figure 4.

The IBD-AID increases the microbiome capacity for SCFA production. Bar plots represent the variance importance of the: A) gene pathways, B) genes involved in butyrate production, or C) genes involved in acetate production; that were enriched during the intervention (BH p-value >0.05).

Figure 5.

Significant correlations of foods with bacterial species enriched at baseline (red) or intervention (blue) in A) CD participants and B) UC participants.

Figure 6.

Levels of serum cytokines correlate with specific food categories. We obtained serum samples from 9 participants at baseline and at the end of the intervention. Consumption of prebiotics correlated with high levels of GM-CSF (A) and negatively correlated with IL-6 (B and C) and IL-8 (D). Consumption of fatty acids (MUFAs and omega-3) correlated with lower levels IL-8 (E). Consumption of adverse foods positively correlated with IL8 (F) and TNF-alpha (J-G). Simple linear regression, p-values <0.05. Dotted lines represent 95% confidence intervals.