Trans-ethnic gut microbiota signatures of type 2 diabetes in Denmark and India

Abstract

Background: Type 2 diabetes (T2D), a multifactorial disease influenced by host genetics and environmental factors, is the most common endocrine disease. Several studies have shown that the gut microbiota as a close-up environmental mediator influences host physiology including metabolism. The aim of the present study is to examine the compositional and functional potential of the gut microbiota across individuals from Denmark and South India with a focus on T2D. Many earlier studies have investigated the microbiome aspects of T2D, and it has also been anticipated that such microbial associations would be dependent on diet and ethnic origin. However, there has been no large scale trans-ethnic microbiome study earlier in this direction aimed at evaluating any "universal" microbiome signature of T2D.

Methods: 16S ribosomal RNA gene amplicon sequencing was performed on stool samples from 279 Danish and 294 Indian study participants. Any differences between the gut microbiota of both populations were explored using diversity measures and negative binomial Wald tests. Study samples were stratified to discover global and country-specific microbial signatures for T2D and treatment with the anti-hyperglycemic drug, metformin. To identify taxonomical and functional signatures of the gut microbiota for T2D and metformin treatment, we used alpha and beta diversity measures and differential abundances analysis, comparing metformin-naive T2D patients, metformin-treated T2D patients, and normoglycemic individuals.

Results: Overall, the gut microbial communities of Danes and Indians are compositionally very different. By analyzing the combined study materials, we identify microbial taxonomic and functional signatures for T2D and metformin treatment. T2D patients have an increased relative abundance of two operational taxonomic units (OTUs) from the Lachnospiraceae family, and a decreased abundance of Subdoligranulum and Butyricicoccus. Studying each population per se, we identified T2D-related microbial changes at the taxonomic level within the Danish population only. Alpha diversity indices show that there is no significant difference between normoglycemic individuals and metformin-naive T2D patients, whereas microbial richness is significantly decreased in metformin-treated T2D patients compared to metformin-naive T2D patients and normoglycemic individuals. Enrichment of two OTUs from Bacteroides and depletion of Faecalibacterium constitute a trans-ethnic signature of metformin treatment.

Conclusions: We demonstrate major compositional differences of the gut microbiota between Danish and South Indian individuals, some of which may relate to differences in ethnicity, lifestyle, and demography. By comparing metformin-naive T2D patients and normoglycemic individuals, we identify T2D-related microbiota changes in the Danish and Indian study samples. In the present trans-ethnic study, we confirm that metformin changes the taxonomic profile and functional potential of the gut microbiota.

Keywords: Danes; Gut microbiota; Indians; Metformin; Populations; Trans-ethnic; Type 2 diabetes.

Fig. 1

Country of origin is the main driver of gut microbial community structure. Danish and Indian microbiota profiles are clearly separated at OTU level in the principal coordinate analysis using Bray–Curtis dissimilarity measure. Country of origin also explains 12% of the beta-diversity variation (p < 0.001, PERMANOVA test). Ellipses represent 95% confidence intervals for the Danish and Indian gut microbiota profiles. Gut microbiota compositions do not separate based on T2D status. DK, Denmark; IN, India; NG, normoglycemic controls; T2D, type 2 diabetes

Fig. 2

Metformin alters gut microbial richness in combined groups of Danes and Indians. Microbial richness was significantly reduced in metformin-treated T2D patients (n = 166), when compared to NG controls (n = 275) or metformin-naive T2D patients (n = 132). Microbial richness did not significantly differ between NG controls and metformin-naive T2D patients. These differences were not reproduced when looking at Shannon index. Danish and Indians samples were included in the analysis. NG, normoglycemic controls; T2D, type 2 diabetes

Fig. 3

OTUs are differentially abundant in the microbiota from the combined Danish-Indian group of T2D patients after controlling for confounding factors. Relative abundances of differentially abundant OTUs are shown in log scale. Among T2D patients, metformin-treated individuals are represented by filled circles and metformin-naive individuals are represented by hollow circles. Danish and Indians’ samples are included in the analysis (also see Additional file 2: Figure S11 for separate visualizations for the countries). These four OTUs are consistently identified from six different analyses where country and metformin were always adjusted for, while each additional factor (BMI, age, gender, usage of sulfonyl urea, statins, and proton pump inhibitors) was adjusted for one by one. NG, normoglycemic controls (n = 275); T2D, type 2 diabetes (n = 298)

Fig. 4

Microbial functional modules are differentially abundant in the microbiota of the combined Danish-Indian group of T2D patients after controlling for confounding factors. Fold change of differentially abundant functional modules are shown in log scale. Both Danish and Indians’ samples were included in the analysis. These modules are consistently identified from six different analyses where country and metformin were always adjusted for, while each additional factor (BMI, age, gender, usage of sulfonyl urea, statins, and proton pump inhibitors) was adjusted for one by one. NG, normoglycemic controls (n = 275); T2D, type 2 diabetes (n = 298)

Fig. 5

OTUs in the gut microbiota of the combined Danish-Indian group of T2D patients are affected by metformin treatment. Relative abundances of differentially abundant OTUs are shown in log scale. Both Danish and Indians’ samples were included in the analysis. Comparisons were made between T2D metformin-treated (n = 166) and T2D metformin-naive (n = 132). These OTUs are consistently identified from six different analyses where country was always adjusted for, while each additional factor (BMI, age, gender, usage of sulfonyl urea, statins, and proton pump inhibitors) was adjusted for one by one. T2D, type 2 diabetes