Microbiome directly affects glucose metabolism, type 2 diabetes risk

The micro-organisms in our gut, the microbiome, have a profound effect on our health. However, even though dozens of new associations between gut bacteria and health are published each week, it is difficult to establish their causal relationship. Scientists from the University Groningen Medical Centre, the University of Oxford and colleagues have now used a Mendelian Randomization approach to causally link gut microbiota with glucose metabolism. The results were published in Nature Genetics on 18 February 2019.

Micro-organisms in our gut play an important role in the digestion of food and have been associated with a number of traits such as obesity. ‘But all studies on the microbiome only correlate certain bacterial species with a phenotype, like obesity. The causal link is not proven’, says Serena Sanna, assistant professor at the University of Groningen Medical Centre (UMCG). Sanna graduated in mathematics and is now working in the UMCG Department of Genetics in the group led by Prof Cisca Wijmenga.

The problem is that the microbiome can affect health, but health can also affect the microbiome. ‘And correlations can even turn out to be just statistical correlations’, says Sanna. That is why she is interested in finding a way to prove causal links between our gut bacteria and certain phenotypes. In the new study, she and her colleagues, including leaders in diabetes research Mark I. McCarthy, Natalie R. van Zuydam and Anubha Mahajan from the University of Oxford, show how changes in the microbiome can lead to increased insulin secretion after glucose challenge and also to an increased risk for Type 2 diabetes.

Sanna and her colleagues first analysed data on genetics and the microbiome in almost one thousand participants of the Dutch LifeLines DEEP cohort study. ‘We looked for microbiome features – bacterial species, or the function bacteria play in the gut – that associated with changes in glucose metabolism and obesity, risk factors for type 2 diabetes’, explains Sanna.

The next step was to determine cause and effect, which was done using a technique called Mendelian Randomization. The microbiome features could be directly responsible for the effects on glucose metabolism, or alternatively, impaired glucose metabolism could affect microbiome composition. To differentiate between these two hypotheses, the scientists first checked to see which genetic variants in the host would predict higher abundance of these microbiome features.

They then looked at participants in different European cohort studies (including the large UK Biobank study that has enrolled 500,000 participants) that have collected data on host genetics. Participants were grouped based on the genetic variants that predict each of these microbiome features, and Sanna and colleagues looked to see if there was any difference in the efficiency of glucose metabolism between groups.

Some of the microbiome features led to groups with different glucose metabolism. In contrast, when grouping Lifelines DEEP participants based on genetic variants associated to glucose metabolism parameters, no difference in microbiome features was seen between groups. So the microbiome features determined the risk, and not the other way around.