DOP008 Dietary manipulation of the healthy human and colitic murine gut microbiome by CD-TREAT diet and exclusive enteral nutrition; a proof of concept study

Abstract

Background: The extensive modulation of gut microbiome in children with Crohn's disease (CD) treated with exclusive enteral nutrition (EEN) offers clues about EEN's potential mode of action; but also on the development of novel therapies through dietary manipulation of the gut microbiota. This proof of concept study compared the effect of a novel “ordinary” food-based diet (CD-TREAT diet) and EEN on healthy human and colitic murine gut microbiota. Methods: A) Healthy adults followed two experimental diets for 7 days with a 15 day wash out period in between; EEN and CD-TREAT, an “ordinary” food diet which has similar nutrient and food ingredient composition to EEN (e.g., fibre, gluten, lactose content and fatty acid composition). Participants were randomly allocated to start with EEN or CD-TREAT first. Fresh faecal and urine samples were collected before and after each dietary intervention and 16s rRNA sequencing, untargeted faecal and urine metabolomics (using LC-MS) were performed; B) 10-month-old HLA-B27 and HLA-B7 trangenic rats received EEN, CD-TREAT diet or regular rat chow for 4 weeks. Faeces were collected at baseline, 1, 2, 3 and 4 weeks post treatment initiation. Gut contents, ileal and colonic tissue were harvested at sacrifice. Disease activity was quantified by blinded histological scores and gut microbiota metabolic activity was measured by faecal short chain fatty acids (SCFA) quantification. Results: A) 100 samples were collected from 25 healthy subjects. During EEN and CD-TREAT gut bacterial community structure (using OTUs) significantly changed after both EEN and CD-TREAT (R2=0.15, R2=0.05, p<0.004) and shifted towards the same direction. EEN's and CD-TREAT's impact on 3% OTU community structures was strongly correlated (Adj R2=0.38, p<2.2e-16). Similarly, untargeted faecal metabolomics revealed a strong correlation between the changes during EEN and CD-TREAT (Pearson's R=0.31, p<10–14)$; B) 100 faecal samples were collected from 12 HLA-B27 and 8 HLA-B7 adult transgenic rats. Both dietary interventions increased the body weight of the HLA-B27 rats (Median %weight change, EEN:+9.2 vs CD-TREAT:+15.7 vs Control:-2.1) and decreased the weight of caecum and colon contents. Faecal concentration of total SCFA, acetic, propionate decreased while iso-butyric and isocaproic increased during both dietary interventions [ΔMedian μmol/g, EEN: −324, −271, −44.8, +4.7, +2.2 vs CD-TREAT: −354, −292, −56.2, +3, +1.5]. Histopathology scores revealed that both dietary interventions benefited moderately ileal but not colonic inflammation. Conclusions: We have developed an “EN composition alike” food based diet which induces similar effects on gut microbiome with EEN. This proof of concept study supports a subsequent pilot trial in people with active CD.