In the following interview, Professor Dr. Dirk Haller explains his research on specific gut bacteria of the microbiome and a possible connection with metabolic diseases such as type 2 diabetes mellitus. The interview was published in the September issue of the scientific magazine of the German Nutrition Society (DGE) DGEwissen.
In the 2020 study “Arrhythmic Gut Microbiome Signatures Predict Risk of Type 2 Diabetes,” Haller and colleagues demonstrated that specific gut bacteria are subject to rhythmic fluctuations and that this rhythm is interrupted in type 2 diabetes mellitus.
This arrhythmic signature of certain bacterial species contributes to the classification and prediction of type 2 diabetes mellitus on an epidemiological level and highlights that there are links between circadian rhythms and the microbiome in metabolic diseases.
DGEwissen:
Professor Haller, you have demonstrated that in people with type 2 diabetes mellitus, the circadian rhythm in the gut microbiome is disrupted. Can you explain what a circadian rhythm is?
Dirk Haller:
The circadian system is a biological rhythm adapted to the day and night rhythm that regulates many body functions. However, there is extremely little data on this for microbial ecosystems so far. We have identified signatures of bacteria in cohort studies to infer whether or not there are associations with specific diseases. The goal here is to find specific patterns. We were able to demonstrate and also validate for the first time in controlled human studies that we find diurnal variations in bacterial abundance [ed. note: number of individuals of a species] in the gut.
This also makes sense if we assume that certain functionalities are mediated at certain times of the day.
DGEwissen:
How did you find out that this rhythm is interrupted in diabetes mellitus type 2 and also in obesity?
Dirk Haller:
First, we looked in large human cohorts to see if there were diurnal variations in the microbiome. We then validated these in three different cohorts. We are quite sure that we have diurnal variations at the richness level – that is, how many bacterial species you can detect at a given time. This is associated with fluctuations of up to 10% in the dominant phyla [ed. note: strains], which is a lot.
As the studies progressed, we then looked at the incidence of disease. To demonstrate statistically viable findings, you also have to find enough cases of a particular disease in a cohort. That’s where diabetes mellitus type 2 and obesity come in, because they are relatively common.
DGEwissen:
How do you explain that this rhythm disruption occurs in these two metabolic diseases of all diseases? Does this possibly also apply to other metabolic diseases such as lipometabolic disorders?
Dirk Haller:
It is a problem that in many human studies we can only obtain one sample, and accordingly not all time of day can be covered in the form of stool samples. Therefore, large numbers of subjects are needed to be able to demonstrate diurnal fluctuations or the loss of these rhythms with statistical certainty.
Surprisingly, in obesity and type 2 diabetes mellitus, the bacteria that were arrhythmic in obesity were not necessarily those that were also arrhythmic in type 2 diabetes mellitus, although there were small overlaps. This means that the bacteria are definitely disease-specific.
In diabetes mellitus type 2 and obesity, we see that different profiles of bacteria become arrhythmic. So I think for a whole range of diseases there could be such correlations. But for rarer diseases, such as Crohn’s disease, we would need enormously large human cohorts.
DGEwissen:
Are there specific bacterial species that are responsible for these rhythmic fluctuations?
Dirk Haller:
We created a classification and prediction for type 2 diabetes mellitus using mathematical models as part of the KORA study. There, we had collected samples in 2013 and 2018/19. Using the 2013 profiles, we were able to predict to some extent who would develop type 2 diabetes mellitus from the 2018/19 cohort. Our marker species (signature) consisted mainly of Faecalibacteria, specifically F. prausnitzii. This bacterial species is always associated with good properties. Faecalibacteria are so-called butyrate producers, i.e. they ferment dietary fiber in the intestine. And we were able to show that these were the ones that lost the rhythm. To F. prausnitzii belong 10 of 13 organisms that we identified as signature.
DGEwissen:
What do the results mean for the future? Can we predict who will develop type 2 diabetes mellitus based on variations in the gut microbiome? Would it even be possible to use this as a medical marker?
Dirk Haller:
We were able to use mathematical models to distinguish relatively well between people with and without type 2 diabetes mellitus using this bacterial risk signature, but that only applies to statements made in the context of large cohorts. At the individual level, we cannot distinguish patients from healthy individuals with a discriminatory power of 70-80%, so this does not allow us to predict a personal risk.
With the inclusion of additional risk factors, such as BMI and other markers of metabolic disease, the sensitivity and specificity of these statements can be further improved. In addition, it can be assumed that we all have a quite different spectrum of bacteria, so that very individualized diagnostics would be needed here. This is not possible at the moment. After all, each person has an individual composition of the microbiome.
DGEwissen:
What role does this play in your study results?
Dirk Haller:
To say that with this risk signature of arrhythmic bacteria we are able to diagnose type 2 diabetes mellitus would be misleading. However, we could find correlations between microbiome and type 2 diabetes.
It could be that rhythmic bacteria play a role specifically in metabolic disease because they perform an additional function in metabolism at times when we are also eating main meals. That would make perfect sense. What relevance this will ultimately have remains to be seen. Correlations of this kind must be validated and understood in animal models. Then we will get closer to the truth.
DGEwissen:
Is there a composition of the microbiome that is particularly beneficial for health?
Dirk Haller:
Comparable to the state of human genome research 20 years ago, microbiome research today is still in its infancy. People are just now starting to slowly do larger studies in controlled human cohorts. It is a question of what is good and what is bad. I get requests again and again to interpret individual stool analyses. In most cases, however, it is impossible to establish a disease reference from stool analyses. Nevertheless, there are of course markers for a microbiome in a healthy person, e.g. the abundance of bacteria. If a healthy person has an average of 300-500 bacterial species and a Crohn’s disease patient has only 30, that is dramatically less.
This means that, first, richness is a measure of a stable microbiome. And secondly, metabolic processes such as the fermentation of dietary fiber are also a feature of a healthy microbial ecosystem in the gut. Here we come back to the diabetic signature, F. prausnitzii are part of these fiber fermenters, they produce the short chain fatty acid butyrate.
So far, it is not possible to say one person’s microbial ecosystem is better than another. When we look at the 4,500 people from our cohorts, no obvious patterns emerge.
If you just look at the distribution, you also can’t just say this one is healthy, that one is sick. You need more information to stratify meaningfully. We have just started to identify bacterial strains by sequencing. However, this requires a lot more effort and is also more expensive.
I am convinced that individual bacterial strains make the difference. However, we currently lack the data to establish a link to diseases. That will be the task of the next few years.
DGEknowledge:
In a plenary lecture at the 58th DGE Congress, it was said that diet has only a fairly small effect on the diversity of the microbiome composition. How do you see that? Can diet contribute to a more favorable microbiome? And if so, what does an optimal diet look like in your opinion?
Dirk Haller:
At first glance, you might say that’s true. We can distinguish extreme dietary differences like veganism compared to a mixed diet. Dramatic conversions also leave a visible fingerprint. But it is the case that with the methods we use, we stop at the resolution of 108. For anything less than 108, we don’t actually see anything. We call these the low-abundance bacteria.
For better resolution, as I just mentioned, we need much deeper sequencing. Bacteria adapt to the food that is available. This is true not only for fiber availability, but also for micronutrients. The spectrum of available macro- and micronutrients in different gut segments is certainly capable of influencing the number and spectrum of bacteria. But in the end, there is only room for a certain amount of bacteria in an ecosystem, i.e. ecosystems adapt to what they are offered to eat.
For example, if an omnivorous person suddenly goes vegan, then the bacterial composition will also adapt to this, because the nutrient supply is suddenly completely different. So we have a long-term adaptation to our dietary patterns. But to be more specific, we can assume that the type and amount of dietary fiber have a significant impact on the composition and function of the gut microbiome.
In cancer research, it is assumed that fiber amounts of 60 g or more daily have a protective effect on the colon. This is then usually accompanied by a reduced fat intake, which also affects the amount of bile acid. The microbiome is then more attuned to fermentation and forms fewer secondary bile acids.
In summary, diet affects bacterial composition, with long-term dietary patterns leaving a fingerprint. And that makes sense because bacteria can adapt. The second aspect is that each person has an individual microbiota.
So far, I don’t see how a clearly defined diet can influence the microbiome in a clearly defined direction. I always warn against such statements, because that does not do justice to the complexity.
DGEwissen:
Thank you very much for the interview Professor Haller.
