Gut microbial determinants of clinically important improvement in patients with rheumatoid arthritis

• DOI: 10.1186/s13073-021-00957-0
• Publication Date: 2021-09-13T23:04:59Z
• Authors: Vinod K. Gupta; Kevin Y. Cunningham; Benjamin Hur; Utpal Bakshi; Harvey Huang; Kenneth J. Warrin...; Veena Taneja; Elena Myasoedova; John M. Davis; Jaeyun Sung
• Keywords: Dysbiosis

Abstract Background Rapid advances in the past decade have shown that dysbiosis of gut microbiome is a key hallmark rheumatoid arthritis (RA). Yet, relationship between and clinical improvement RA disease activity remains unclear. In this study, we explored patients with to identify features are associated with, as well predictive of, minimum clinically important (MCII) activity. Methods We conducted retrospective, observational cohort study on diagnosed 1988 2014. Whole metagenome shotgun sequencing was performed 64 stool samples, which were collected from 32 at two separate time-points approximately 6–12 months apart. The Clinical Disease Activity Index (CDAI) each patient measured both assess achievement MCII; depending status, distinguished into groups: MCII+ (who achieved n = 12) MCII− did not achieve 20). Multiple linear regression models used microbial taxa biochemical pathways MCII while controlling for potentially confounding factors. Lastly, deep-learning neural network trained upon microbiome, clinical, demographic data baseline classify according thereby enabling prediction whether will follow-up. Results found age be largest determinant overall compositional variance ( R 2 7.7%, P 0.001, PERMANOVA). Interestingly, next factor identified explain most status 3.8%, 0.005). Additionally, by looking patients’ profiles, observed significantly different traits who eventually showed those not. Taxonomic include alpha- beta-diversity measures, several taxa, such Coprococcus , Bilophila sp. 4_1_30, Eubacterium 3_1_31. Notably, had higher alpha-diversity their microbiomes follow-up visits. Functional profiling fifteen pathways, involved biosynthesis L-arginine, L-methionine, tetrahydrofolate, differentially abundant groups. Moreover, groups fold-changes (from follow-up) eight seven pathways. These results could suggest that, course, only start ecological states, but also trajectories. Finally, proved highly effective predicting (balanced accuracy 90.0%, leave-one-out cross-validation), demonstrating potential utility profiles. Conclusions Our findings confirm presence taxonomic functional signatures patients. Ultimately, modifying enhance outcome may hold promise future treatment RA.