IDDF2023-ABS-0091 Integrated multi-cohort analysis of gut metagenome in type 2 diabetes mellitus

Background

Type 2 diabetes mellitus (T2DM) characterized by glucose intolerance accounts for approximately 90% of all diabetic patients worldwide. Although studies have monitored the gut microbiota and investigated their relationships with T2DM, inconsistent results describing microbial differences have been reported between T2DM and healthy individuals in different populations.

Methods

In this study, we systematically collected five independent shotgun metagenomic sequencing datasets from Pubmed as well as the literature. Raw metagenomic data were analyzed using Kraken2 and HUMANn2 to generate tables with relative abundances of microbial species and functions, respectively.

Results

We performed an integrated meta-analysis of stool microbial profiles to discern and describe microbial dysbiosis in T2DM and to explore heterogeneity among 5 studies, including 581 normal glucose tolerance (NGT) and 319 T2DM samples. Through the random effects model, we observed significant differences in beta diversity, but not alpha diversity, between individuals with T2DM and NGT. We identified various species and pathways with significant odds ratios for T2DM. At the species level, a total of 270 species were identified as significantly associated with T2DM. Four species had significant ORs lower than 1.0 representing enrichment in diabetes by both the RE and FE models, including [Clostridium] bolteae, Epibacterium mobile, Corynebacterium diphtheriae, and Ochrobactrum sp. A44. Other species for the absence of diabetes in the RE and FE models include Clostridium butyricum, Lactococcus lactis, Veillonella parvula and so on. In terms of microbial functions, two of twenty-nine pathways were depleted in T2DM patients, including L-isoleucine biosynthesis IV (PWY-5104) and Bifidobacterium shunt (P124-PWY), while other pathways such as lactose and galactose degradation I (LACTOSECAT-PWY) and gluconeogenesis III (PWY66-399) are enriched in T2DM. By stepwise feature selection, the AUROC for T2DM ranged from 0.75 to 0.96 at pathway levels and 0.72 to 0.95 at species levels, respectively (IDDF2023-ABS-0091 Figure 1 (A)The principal coordinates analysis (PCoA) of metagenomic species of all samples from the 5 studies (B) Forest plot of the pathway metrics across the 5 studies (C) ROC curves of individual studies and all samples together based on the selected species).

Conclusions

We integrated 900 metagenomic samples from 5 cohorts worldwide and identified gut microbiome and functional differences between T2DM patients and NGT across countries. These results help better understand dysbiosis characteristics of the gut microbiome in T2DM patients, develop a complementary approach for the risk assessment of T2DM, and provide new insights into diabetes prevention and improvement.