In this study we hypothesized that intestinal microbiota in humans with type 2 diabetes is different from non-diabetic persons. The hypothesis was tested on adults with a broad range of ages and BMIs, using pyrosequencing of the V4 region of the 16S rRNA gene and qPCR. To our knowledge no related studies on humans with diabetes type 2 have been published so far.
We demonstrated in this research that type 2 diabetes is associated with compositional changes in the intestinal microbiota mostly apparent at phylum and class levels. The relative abundance of Firmicutes
was significantly lower, while the proportion of Bacteroidetes
was somewhat higher in diabetic persons compared to their non-diabetic counterparts. Accordingly, the ratios of Bacteroidetes
significantly and positively correlated with reduced glucose tolerance. Assuming that diabetes and impaired glucose tolerance are linked to obesity, our results are in agreement with the recent evidence obtained for overweight persons by Schwiertz and colleagues 
, though contradict with other studies 
. Furthermore, based on the assumption above, a positive correlation between ratios of Bacteroidetes
and BMI could be expected. However, the reverse tendency was observed (), indicating that overweight and diabetes are associated with different groups of the intestinal microbiota.
Bacterial groups that distinguished the diabetic from the non-diabetic microbiome included Bacteroides
group versus class Clostridia
and C. coccoides
group, which ratios were significantly higher in diabetic persons. These results are supported by previous studies showing reduction in Bacteroides
spp. related to a strong decrease of metabolic endotoxemia and inflammation in type 2 diabetes mice 
. Accordingly, a significant reduction in Clostridium
ssp, C. coccoides
and an increase in the Bacteroides
group along with body weight loss have been observed in human studies 
. The significantly higher levels of Bacilli
and the Lactobacillus
group in diabetic subjects compared to controls in the present study, have recently been reported in relation to type 2 diabetes in mice models 
and to obesity in human adults 
. Genus Lactobacillus
represents a heterogeneous group with well documented immunomodulating properties 
and might potentially contribute to chronic inflammation in diabetic subjects.
The tendency of increased Chao1 diversity concurrently with BMI observed in this study, might be related to the negative correlation between BMI and Bacteroidetes/Firmicutes
ratios () as Firmicutes
is a highly diverse division 
. This observation is, however, in disagreement with recently published data on the obese twin pairs showing reduced bacterial diversity in obese individuals 
. The reduced individual variation in diversity of the fecal microbiota observed in the diabetic group compared to the controls probably reflected the differences in diet, lifestyle or other factors 
which are not possible to specify in the present study.
In an obesity study, using mice models, Cani and coworkers 
proposed a hypothesis connecting metabolic diseases with the presence of Gram-negative bacteria in the gut, also offering a likely explanation of the differences between the diabetic and non-diabetic microbiomes in this study. The intestinal microbiota across the subjects with type 2 diabetes was relatively enriched with Gram-negative bacteria, belonging to the phyla Bacteroidetes
. The main compounds of outer membranes in gram-negative bacteria are lipopolysaccharides (LPS), known as potent stimulators of inflammation, which can exhibit endotoxaemia 
. Consequently, LPS will continue to be produced within the gut, which might trigger an inflammatory response and play a role in the development of diabetes.
In conclusion, our data suggest that the levels of glucose tolerance or severity of diabetes should be considered while linking microbiota with obesity and other metabolic diseases in humans. It is especially important for developing the strategies to modify the gut microbiota in order to control metabolic diseases, since obesity and diabetes might be associated with different bacterial populations.