Society: AGA

Background The gut microbiome is central to Crohn’s disease (CD) pathogenesis but the relative contribution of different environmental factors to disease compositional variation is unknown. We determined the impact of environmental factors and diet on microbiome variation in this international study.
Methods 480 stool samples from 240 CD patients and 240 healthy controls from regions of different epidemiologic stages: Australia (Compounding prevalence; AUS: 97 CD; 77 controls), Hong Kong (Acceleration in incidence; HK: 90 CD; 63 controls) and Mainland China (Emergence in incidence; MC: 53 CD;100 controls) were studied. Demographic, clinical and 3 food intake questionnaires (processed food intake in early life, recent (last 12 months) and current (last 3-days)) were obtained from all subjects. Fecal microbiota was assessed by shotgun metagenomic sequencing. Dysbiosis score was defined as the median Bray-Curtis dissimilarity to the reference control group. Adonis was performed to rank the effect of disease-associated parameters, clinical metadata, dietary intake and geography on overall microbial composition.
Results Microbiota diversity was significantly reduced in CD compared with controls in all three geographical cohorts (all p<0.05). Beta diversity analysis showed a significant disease-associated shift along principal components for active CD followed by inactive CD away from controls (p<0.001). Gut microbiome structure of HK subjects were intermediate between that of AUS and MC. CD patients in MC had the highest dysbiosis score compared with AUS and HK, but dissimilarity scores were no different between controls in AUS, HK and MC (Fig 1A). Geography had the greatest impact on microbiota variance followed by disease diagnosis and bowel resection history. Additionally, disease complications, feeding pattern and processed food exposure at early age were significantly correlated with microbiota variance. CD patients who consumed more processed food at early age had a higher dysbiosis score than those who did not (p<0.05). Amongst all diet types consumed, current fiber intake was the most dominant in contributing to microbiome variance. Other major dietary factors that correlated with microbiome variance belonged to fat (polyunsatuated fat, saturated fat, monounsaturated fat and cholesterol) and vitamins (Fig 1B). The proportion of dietary factor-associated species in CD was lower than that of controls, suggesting that gut microbiota in controls were more susceptible to the effects of diet.
Conclusion Geography especially regions transiting through different epidemiologic stages of disease is a dominant driver of gut microbiome variation, followed by Crohn’s disease diagnosis. The Crohn’s disease microbiota is also shaped and influenced by current dietary pattern and early life food additive intake.

This work is supported by Helmsley Charitable Trust

<b>Fig1. (A)</b> Distribution of stool dysbiosis scores in relation to presence of Crohn’s disease (median Bray-Curtis dissimilarity between a sample and healthy control samples). <b>(B) </b>Current dietary factors impact on microbiota variance. Adjusted permutational multivariate analysis was performed to rank the variances. *<i>p</i><0.05; **<i>p</i><0.01; ***<i>p</i><0.001.

Fig1. (A) Distribution of stool dysbiosis scores in relation to presence of Crohn’s disease (median Bray-Curtis dissimilarity between a sample and healthy control samples). (B) Current dietary factors impact on microbiota variance. Adjusted permutational multivariate analysis was performed to rank the variances. *p<0.05; **p<0.01; ***p<0.001.

Background
It is widely accepted that high-fat diet (HFD) has been considered as one of the risk factors of inflammatory bowel disease(IBD), while the mechanism is little known. We aimed to examined whether dietary high fat promotes colonic inflammation through modulating gut microbial tryptophan metabolites.
Methods
The C57BL/6 female mice were fed with chow diet (12kcal% fat) or HFD (60kcal% fat) for 2 weeks. Fresh fecal samples were collected before administration of 2% DSS. LC-MS/MS analysis was performed to detect fecal microbial tryptophan metabolites. Next, 8-week-old mice were administrated 3-indoleacetic acid (IAA) for 7 days prior to 2.5% DSS treatment. The severity of colitis was assessed and the colonic sections were collected. RNA sequencing was performed to analyze the differential genes in intestinal tissue of mice fed with IAA or not. Real-time PCR and Western blotting were performed to detect the transcription and protein expression of mucin sulfation-related genes. High iron diamine-alcian blue (HID-AB) staining of colon tissues and LS174T cells was used to detect the abundance of sulfated mucin. Additionally, the Chromatin immunoprecipitation (ChIP) sequencing is performed to confirm the binding of aryl hydrocarbon receptor(AHR) to 3’-phosphoadenosine 5’-phosphosulfate (PAPS) synthase 2 (PAPSS2).
Results
Short-term HFD increased the susceptibility to colitis of mice. Mice fed with HFD have impaired tryptophan metabolism with remarkable lower level of IAA. IAA supplementation resulted in ameliorative colitis symptoms. Mechanistically, the transcriptome sequencing detected that IAA supplementation enriched the the cytosolic sulfonation pathway and significantly increased the expression of PAPSS2 of colon tissue, a rate-limiting enzyme on the mucin sulfation pathway. Next, in vivo and in vitro experiments verified IAA increased the expression of PAPSS2 and its downstream genes and promoted the mucin sulfation in golgi apparatus of colon through AHR. ChIP sequencing and ChIP-PCR verified IAA can enhance the binding of AHR to the promoter region of PAPSS2 , which further promoted the expression of PAPS transporter 2 (PAPST2 or Slc35b3) and ultimately promotes the mucin sulfation.
Conclusion
HFD altered the gut microbial tryptophan metabolites and disrupt the protective effect of IAA on the intestinal barrier. IAA could relief colitis through AHR-PAPSS2-PAPST2-mucin sulfation axis, which may present a potential approach for precise prevention and treatment of IBD.

Dietary fat exacerbated DSS–induced colitis and reduced gut microbial tryptophan metabolites especially IAA

Dietary fat exacerbated DSS–induced colitis and reduced gut microbial tryptophan metabolites especially IAA

IAA ameliorated the severity of DSS-induced colitis and strengthened the mucosal barrier function through promoting mucin sulfation

IAA ameliorated the severity of DSS-induced colitis and strengthened the mucosal barrier function through promoting mucin sulfation

Background: The etiology of Crohn’s disease (CD) is unknown but may arise from complex interactions between genetic susceptibility and abnormal host immune responses related to environmental exposures, including the diet. Dietary intake is known to influence the gut microbiome along with multiple host cellular pathways, together resulting in perturbation of the host metabolic profile. We recently identified serum-derived metabolites associated with CD onset. We hypothesized that nutrients could impact these specific serum-derived metabolites.

Methods: Healthy first-degree relatives of CD patients were recruited as part of the Crohn’s Colitis Canada- Genetic, Environmental, Microbial (GEM) Project. We focused on a subset of 389 subjects (n=78 who developed CD and n=311 healthy controls), matched for age, sex, postal code, and time in the study with available dietary and serum metabolomics data. A total of 150 dietary constituents were derived from baseline food frequency questionnaires using the 2015 release of the StatsCan Nutrient database. Serum metabolites were measured using the combination of four ultrahigh Performance Liquid Chromatography-Tandem Mass Spectroscopy. In total, we identified 63 metabolites associated with CD onset, these metabolites were utilized for this exploratory analysis.
We used general estimating equation models to identify associations between dietary constituents and these CD-associated metabolites. Benjamini-Hochberg corrected p values, q<0.05, defined significance.

Results:
In total, 52 nutrients were associated with 53 of the 63 CD-associated metabolites (2.04^-11<q<0.045). Of all significant associations, three dietary constituents, docosenoic acid (a monounsaturated fatty acid found in rapeseed oil), alcohol, and aspartame had the highest number of associations with CD-associated metabolites (17, 13, and 12, respectively). For example, increased consumption of docosenoic acid, alcohol, and aspartame were associated with increases in sphingomyelin-related metabolites (2.33^-10<q<0.046), (2, 2, and 4 respectively). Furthermore, increased docosenoic acid and alcohol consumption was also associated with increases in a lactosylceramide-related metabolite, q = 0.004, and q = 0.038, respectively. Finally, vitamin D related nutrients were also associated with bile acids related metabolites (2.10^-11<q<0.044).

Conclusions: The association of dietary constituents with specific CD-associated metabolites may suggest a link between dietary constituents and CD pathogenesis. Identifying these associations may help in defining strategic dietary interventions in treating or preventing CD.

Funding
The Leona M. and Harry B. Helmsley Charitable Trust
Kenneth Croitoru is the recipient of the Canada Research Chair in Inflammatory Bowel Diseases

Background: Inflammatory bowel diseases (IBD) are the culmination of a complex interplay of environmental factors that interact with gut microbial communities in genetically susceptible individuals. Previously, we demonstrated that a de novo dietary pattern (the sulfur microbial diet) was associated with the relative enrichment and increased transcriptional activity of pro-inflammatory sulfur metabolizing bacteria (Nguyen et al, Gastroenterology 2020). Whether this pattern of intake is associated with risk of IBD is unknown.

Methods: We leveraged data from three prospective cohorts (Health Professional Follow-Up Study, Nurses’ Health Study [NHS] I, and NHS II) biennially followed since 1984 with greater than 90% follow-up for lifestyle habits, medication use, and medical diagnoses. Diet was assessed every four years through validated food frequency questionnaire through 2015. Consistent with prior efforts, we calculated sulfur microbial diet scores using the weighted sum of select food groups most strongly linked to gut sulfur economy, such as low-calorie beverages, red/processed meats (each linked to increased abundance), and whole grains and leafy/cruciferous vegetables (decreased abundance). Incident IBD was confirmed and classified by subtype (ulcerative colitis [UC] and Crohn’s disease [CD]) and phenotype based on blinded medical record review. Time-varying multivariable Cox proportional hazards models were used to assess the association between the sulfur microbial diet and risk of IBD. Models were adjusted for age, sex, body mass index (BMI), and family history of IBD, among other risk factors.

Results: Among 218,509 participants followed for 6,271,468 person-years, 901 were diagnosed with IBD, 516 with UC, and 385 with CD. At baseline, participants with greater adherence to the sulfur microbial diet were generally younger, had higher BMIs, and engaged in less physical activity. We observed that compared to the lowest, the highest quartile of sulfur microbial diet scores (indicating greater adherence to this pattern of intake) was associated with increased risk for IBD (HR_Q4vs.Q1 1.27, 95% CI: 1.03-1.56; P_trend=0.02), largely driven by increases in incident CD (HR_Q4vs.Q1 1.40, 95% CI: 1.02-1.92; P_trend=0.04). No clear trend was observed with UC (HR_Q4vs.Q1 1.17, 95% CI: 0.89-1.55; P_trend=0.17). Among CD phenotypes, greater adherence to the sulfur microbial diet appeared to be associated with stricturing/penetrating CD behavior compared to non-stricturing/penetrating, but this finding did not reach statistical significance (P_{heterogeneity}>0.05).

Conclusion: We observed that adherence to a sulfur microbial diet was associated with increased risk of IBD, largely driven by increased risk of CD. Our findings highlight the possible interaction between diet and gut microbial communities with risk and severity of IBD.

<b>Figure 1. Multivariable Cox proportional hazards models of sulfur microbial diet score and (A) incident inflammatory bowel disease, Crohn’s disease, and ulcerative colitis and by (B) Crohn’s disease phenotypes.</b> Models were adjusted for age (years), sex (male/female), body mass index (calculated as weight in kilograms/height in meters squared, continuous), total calorie intake (kcal/day, continuous), physical activity (MET-hours/week, continuous), family history of IBD (yes/no), ibuprofen use (yes/no), oral contraceptive use (yes/no), and hormone replacement therapy among women (never/former/current). Linear trend was tested by modeling the median value of each diet score quartile as a continuous variable.

Figure 1. Multivariable Cox proportional hazards models of sulfur microbial diet score and (A) incident inflammatory bowel disease, Crohn’s disease, and ulcerative colitis and by (B) Crohn’s disease phenotypes. Models were adjusted for age (years), sex (male/female), body mass index (calculated as weight in kilograms/height in meters squared, continuous), total calorie intake (kcal/day, continuous), physical activity (MET-hours/week, continuous), family history of IBD (yes/no), ibuprofen use (yes/no), oral contraceptive use (yes/no), and hormone replacement therapy among women (never/former/current). Linear trend was tested by modeling the median value of each diet score quartile as a continuous variable.

Inflammatory bowel disease (IBD) is on the rise in Hispanics, but little is known about factors influencing disease. Diet may influence IBD inflammation including by modification of the intestinal microbiome. We examined diet patterns associating with disease activity among Hispanics and their relationship to inflammation and gut microbiome.

Methods
This was a 1-year longitudinal study of Hispanics with IBD. Clinical activity and diet were assessed at two time points using validated indices and food frequency questionnaires. Medications and surgeries were captured as covariates. Stool was collected at both visits for fecal calprotectin (FeCal) and stool microbiome. DNA was extracted from stool on a subset of patients and shallow shotgun microbiome sequencing to 20 million reads was performed. Unsupervised machine learning using sidClustering was performed to form diet clusters. Gap statistics was used to determine the appropriate number of clusters. Clusters were compared to clinical activity using Fisher’s exact test as clinical validation (p<0.05 for significance). Mixed effect models were used in examining the association between microbial changes over time and repeated measure outcomes (FeCal and clinical activity) using a random intercept, adjusting for covariates. Results were visualized in relative abundance. Analyses were performed using R (ver. 4.1.1) and R package “cluster” (ver. 2.1.2).

Results
214 Hispanics with IBD were included: 127 with UC and 87 with CD. Most were US born (36.6%) or Cuban-born (30.52%). A total of 4 diet clusters were identified at baseline and visit 2. Dietary cluster 4 associated with the lowest clinical activity (Fig 1). Diet cluster 4 differed from other clusters in having the highest consumption of traditional Hispanic starchy vegetables (e.g. cocoyam, yucca), consumed on avg 4.5 days compared to 2.0 days per week (p=.006). A greater intake of soups consisting of mixed starchy vegetables was consumed more frequently in cluster 4 than the rest (p<.001). On longitudinal analyses, patients whose FeCal and clinical activity decreased over time had a greater intake of fruits and vegetables and a much lower intake of grains, monounsaturated fatty acids, and refined carbohydrates. In a subset of 42 UC patients sequenced, we found associations between increasing clinical activity and FeCal with microbiome changes at two time points (Figure 2). Five genera associate with both clinical and FeCal changes and two top genera replicate in all our baseline mean and repeated measures regression models (g__Rubrobacter (p=.01), g__Sellimonas (p=.001)).

Conclusions
A diet rich in Hispanic starchy vegetables at baseline associates with lower clinical activity in Hispanics with IBD. Changes in disease inflammation in Hispanics parallel changes in microbiome composition, which have the potential to predict clinical outcomes.

Dietary cluster 4 is characterized by a higher consumption of Hispanic foods and is associated with lower clinical disease activity.

Dietary cluster 4 is characterized by a higher consumption of Hispanic foods and is associated with lower clinical disease activity.

Longitudinal changes in fecal calprotectin associate with changes in microbiota composition

Longitudinal changes in fecal calprotectin associate with changes in microbiota composition

Background: The pathogenesis of microscopic colitis (MC) is attributed to a dysregulated immune response to perturbances in the gut luminal microenvironment. Diet plays a fundamental role in shaping the gut microenvironment. A pro-inflammatory diet has been associated with development of other inflammatory bowel diseases, but the association with MC remains unknown. We assessed the association between dietary inflammatory potential and risk of MC in two prospective cohorts.

Methods: We performed a cohort study of 247,954 women enrolled in the Nurses’ Health Study (NHS) and NHSII without a diagnosis of MC at baseline. Dietary exposures were ascertained using a validated semiquantitative food frequency questionnaire in 1990 in NHS and 1991 in NHSII, and were updated every 4 years. We derived Empiric Dietary Inflammatory Pattern (EDIP) scores at each assessment using the weighted sum of 18 food groups that predict plasma inflammatory markers, with higher scores representing more “proinflammatory” diets. Reported diagnoses of MC were confirmed through review of medical records. We used Cox proportional hazards modeling to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) according to quartiles of EDIP while adjusting for other potential confounders.

Results: Through the end of follow-up in 2014 in NHS and 2015 in NHSII, we confirmed 256 incident cases of MC over 1,912,029 person-years. There was an inverse association between cumulative average EDIP and risk of MC (P_trend=0.001; Figure 1). Compared to participants in the lowest quartile of EDIP, the HR of MC was 0.56 (95% CI 0.38-0.81) for women in the highest quartile. The association did not vary according to disease subtypes of collagenous colitis (Quartile 4 aHR 0.45; 95% CI 0.45 – 0.79) and lymphocytic colitis (Quartile 4 aHR 0.61; 95% CI 0.36 – 1.04; P_{heterogeneity}=0.26). We also examined the association between individual components of EDIP and risk of MC and observed a positive association with higher consumption of coffee, wine, and low-energy beverages (i.e. diet cola). Specifically, for each serving per day increase the aHR of MC was 1.19 (95% CI 1.09 – 1.31; p<0.001) for coffee consumption, 1.55 (95% CI 1.25 – 1.91; p<0.001) for wine consumption, and 1.15 (95% CI 1.02 – 1.30; p=0.022) for low-energy beverage consumption (Figure 2).

Conclusions: In two large prospective cohort studies, we identified an inverse association between EDIP and risk of MC. This association was in contrast to previous studies that examined EDIP and risk of classical inflammatory bowel diseases, in large part due to the positive associations between coffee, wine, and low-energy beverages and risk of MC. Future studies should assess if metabolic or gut microbiota changes mediate the associations between these dietary factors and risk of MC to further define disease pathogenesis.

Kristin Burke

Massachusetts General Hospital

Appearances