OP20 The gut microbiota during biological therapy for inflammatory bowel disease

Abstract

Background The expansion of therapeutic options in IBD brought forward a need to personalise treatment. Gut inflammation in inflammatory bowel disease (IBD) patients has been associated with reduced microbial richness and abundance of SCFA producers. We aimed to explore the longitudinal impact of treatment on the inflammatory burden and faecal microbiota in patients with CD and UC, treated with anti-tumour necrosis factor (anti-TNF) therapy, vedolizumab (VDZ) or ustekinumab (UST). Methods We analysed faecal samples from 349 IBD patients (112 UC, 237 CD) initiating biological therapy (anti-TNF; VDZ; UST), regardless baseline disease activity, between 2010 and 2019 at a tertiary referral centre. Samples were collected at week 0, 14 and 24. Microbiota phylogenetic profiling was conducted by 16S rRNA gene sequencing and faecal cell counts were determined using flow cytometry. Moisture levels were measured using lyophilisation. Disease activity and response were assessed by faecal calprotectin (FCal). Statistics were performed in R, v3.5.1. Enterotyping was based on the Dirichlet multinomial mixtures approach. Results Faecal microbiota profiles showed high diversity, with samples classified into all four enterotypes (Fig1), although Bact2 was 6- to 10-fold more prevalent in patients compared with controls (Fig2). The variation in faecal microbiota composition was explained (multivariate dbRDA) by diagnosis (R2 = 1.2%, p = 1.00E−04), timepoint (R2 = 0.52, p = 0.006), significantly by age, gender and faecal moisture. The full model only explained 2.85% of the microbiota variation. A slight non-significant decrease in the Bact2 enterotype prevalence was observed in the CD, but not in the UC cohort, during treatment. A significant decrease in FCal concentrations (w0 vs. 14, UC p = 5.12E-08, CD p = 2.56E−08; week 0 vs. 24: CD p = 4.86E-08) was observed with treatment, accompanied by a significant increase in cellcounts (w0 vs. 14: UC p = 0.024, CD p = 0.0022; week 0 vs. 24: CD p = 0.0201) (Figure 3). No significant change in Bact2 prevalence was found during treatment. Only treatment-associated variables—week of treatment (p = 2.4E−18), diagnosis (0.0005) and timepoint (p = 0.0073)—were significant predictors for response, while microbiota-associated variables (enterotype, cell counts and faecal moisture) not. Baseline samples were associated with higher FCal levels. This suggests that the response time of the microbiota to treatment may be higher than the host inflammatory response. Conclusion The prevalence of the inflammatory Bact2 enterotype was 5- to 10-fold higher in CD and UC patients as compared with controls. Although initiation of biological therapies leads to a decrease in inflammation levels as witnessed by faecal calprotectin and increase in microbial richness, a shift in enterotypes did not occur.

Dysbiosis in inflammatory bowel diseases: egg, not chicken

“…Similarly, in ulcerative colitis, the principal components of the microbiome normalize after achieving long-term, but not short-term remission (36). In a very recent study, it appeared that microbial normalization upon achieving remission was much more pronounced following anti-TNF treatment compared to vedolizumab and ustekinumab treatment (16). Some studies, including the latter, found an association between the pre-therapy microbiome and therapy response but a solid prediction using the bacterial composition data alone is still not available (16).…”

“…In a very recent study, it appeared that microbial normalization upon achieving remission was much more pronounced following anti-TNF treatment compared to vedolizumab and ustekinumab treatment (16). Some studies, including the latter, found an association between the pre-therapy microbiome and therapy response but a solid prediction using the bacterial composition data alone is still not available (16). When taken together, a large part of the alterations observed in the IBD microbiome appear to be reversible and secondary to inflammation.…”

“…The standard suspicion of an "overshooting immune response" may simply reflect a perfectly adequate defense against this intrusion. In IBD, this mutually aggressive host/microbe interaction dysbiosis has been suggested to play a primary role in leading to pathology (15,16), answering the chicken and egg question in favor of dysbiosis acting as the chicken. The present narrative minireview, after critically screening the data, might convince the reader of the opposite, i.e., the altered microbiome may indeed be secondary (arguments listed in Table 1).…”