Half of the cells in the human body belong to its microbiome, including the gut microbiome, and there are huge differences between different parts of the microbiome in terms of composition, pH and metabolic activities. The microbiome contains 30 million genes, which is 1000 times more than in the human genome. It is a metabolically active organ in its own right, which is in contact with large pools of immune and neural cells and may be associated with a wide range of different diseases, including fatty liver, stroke, cardiovascular disease, type 2 diabetes and obesity. “In healthy individuals, the gut microbiome is affected by transit time, age, BMI, medication and diet,” said Professor Marju Orho-Melander of Lund University, Sweden. “It is dynamic and amenable to modifications and novel therapeutic applications.”
Challenges in microbiome research
Professor Orho-Melander went on to review the knowledge gaps that must be addressed if microbiome research is to deliver for diabetes. First, study size is always important for statistical power and reproducibility. Given that there is huge variability between individuals, large cohort studies are needed, while the dynamic nature of the microbiome means that a cross-sectional study only offers a snapshot and follow-up studies will be necessary. Meanwhile, as the gut microbiome is affected by so many different factors, confounding is a major issue in working out if an association is causal or not.
Nevertheless, there are many new methods in microbiome epidemiology, including developments in how to profile the gut microbiome that have revealed hundreds of previously unknown microbial species, and research into the impact of the gut microbiome in type 2 diabetes is beginning to emerge. For instance, Professor Orho-Melander’s team has been exploring the impact of metformin on the gut microbiome. They have found an association between metformin and 113 different gut species, some of which were newly isolated from the human gut.
Large longitudinal studies are now needed with detailed cardiometabolic phenotyping. The ultimate aim is to understand the mechanism of the changes in the microbiome and how they might mediate the anti-diabetic effects of metformin. This could form the basis for a deeper understanding of the effect of other anti-diabetic medications on the gut microbiome.
The microbiome in metabolic disease
Dr Suzanne Devkota, Director of the Human Microbiome Research Institute at Cedars-Sinai Medical Center, Los Angeles, took to the stage to describe further research into the role of the microbiome in type 2 diabetes. “We live in a microbial world,” she said. “A vast quantity of life is microbial and we wouldn’t exist without it. We have found a way to have a symbiotic relationship with these microbes inside us and it’s really quite beautiful.”
For instance, gut microbes produce short chain fatty acids (SCFAs) like butyrate, which are key drivers for many metabolic effects, such as the stimulation of the production of GLP-1 from the enteroendocrine cells in the intestine. Butyrate can also stimulate the gut-brain-liver axis via GLP-1 to suppress hepatic glucose production. “The gut is one of the most highly innervated organs in the body and is connected to brain via the vagus nerve,” said Dr Devkota.
GLP-1 production needs butyrate, so it is a matter of persuading the gut to make it and other SCFAs. And the substrate they need to do this is fibre. Gut bacteria selectively promoted by the presence of dietary fibre help with the management of type 2 diabetes. In a randomised trial, 27 participants consumed a high-fibre supplement three times a day, while 16 followed their usual diet. All were on acarbose, which gives the gut more fermentable fibre in the form of starch. After 84 days, there was reduced HbA1c and glucose, and an increase in butyric acid, GLP-1 and PYY in faeces of those in the high-fibre supplement group. “This was a beautiful study in humans showing that just by adding fibre to the diet, you can support butyrate-producing bacteria,” she said.
Meanwhile, the SCFA propionate can also stimulate the production of PYY and GLP-1, but it needs to be bound to fibre in the form of inulin. So, in another trial, 60 overweight adults consumed 10 grams of inulin or placebo for 24 weeks. Propionate levels in plasma coincided with a spike in GLP-1 and PYY in the inulin group, who also had less weight gain and a decrease in hepatic lipids.
“Going forward, we need to look at high and low responders to medication, where the microbiome may play a role and may be targetable to make more drugs available to more people,” said Dr Devkota. “In the most stunning study I’ve seen in many years, it’s been shown that DPP-4 that is derived from gut bacteria escapes inhibition by sitagliptin, revealing a potential new target.”
A search for microbial isozymes found a DPP-4 produced by some Bacterioides, which is slightly different from human DPP-4. The study then looked at high and low responders to sitagliptin for the presence of bacterial DPP-4. In an elegant series of experiments, antibiotics were used to knock out microbial DPP-4. The high and low responders were stratified on the basis of their HbA1c post-treatment, and the faecal DPP-4 signature and genes for bacterial DPP-4 were increased in low responders. “So there is a real opportunity to mine the microbiome for metabolites that affect drug response,” she said. “Molecules that we target in humans, like DPP-4, we might need to target in microbes as well.”
Meanwhile, the type 1 diabetes microbiome is understudied, as is the small intestinal microbiome – both potentially fruitful avenues for research. Finally, microbiomes are communities and one microbiome equals one disease does not apply, except in infectious disease. “We need to combine deep metagenomic sequencing, microbial genetics and functional assays and readouts to get us where we need to go,” concluded Dr Devkota. “There is also an important opportunity to combine people’s genetic risk scores with their microbiome to explain heterogeneity in disease, for bacteria and humans both have single-nucleotide polymorphisms (SNPs). This is already being done in the inflammatory bowel disease field, which is similar to diabetes in terms of heterogeneity in response to medication.”
From faecal transplants to bacterial metabolites
Max Nieuwdorp, Chair of (Experimental) Vascular Medicine at Amsterdam UMC, noted that potential microbiome-based treatments for diabetes range from less precise interventions, such as faecal transplants to those that are more precise, such as bacterial metabolites, and include prebiotics and diet, bacterial consortia and probiotics. “Bacteria communicate with our bodies via their metabolites,” he said. “They send signals to the body via the liver.”
Professor Nieuwdorp has been working on faecal microbiota transplants. “This is a very crude way of changing our microbiome for a short amount of time,” he said. “The Chinese used to do faecal transplants for diarrhoea in the fourth century and research in this area has been underway for the last 50 years or so. But there was heightened interest in the 2000s when it became possible to sequence what was being transplanted and then study the clinical phenotype of the recipient.”
There are two ways of doing a faecal transplant – either infusing fresh faeces from a donor or make capsules, which can even be given to children and administered over a period of time. Currently, there are around 450 trials on faecal transplants, including five in type 1 diabetes and five to 10 in type 2 diabetes and obesity. Professor Nieuwdorp’s team is involved in a trial with 20 people with new onset type 1 diabetes, of whom 10 received a transplant of their own faeces and 10 received donor faeces. In the latter, there was stabilisation of residual beta cell function, along with changes in the small intestinal microbiome and plasma metabolites. “We already know from the TEDDY study that the microbiome is involved in type 1 diabetes, so if you alter it at the outset, maybe you can change the course of the disease,” he said.
In type 2 diabetes, there was no effect from a donor faecal transplant on BMI, but there was an improvement in insulin resistance and LDL-cholesterol. Another study used the faeces of a patient on the Mediterranean diet in capsule form and found that it could prolong the beneficial effects of the diet. There are also pilot studies – with new studies on the way – on the effect of donor faeces in NAFLD/NASH, which have shown a lowering of inflammation and fat in the liver.
However, Professor Nieuwdorp believes it is probably going to be impractical to use this approach on a large scale and findings so far have been limited. “The challenges are huge, especially with the lack of effect so far in many trials,” he said. “Different combinations of strains are needed and large trials. Microbiome therapies are personalised rather than blockbuster and they are hard to keep fresh and hard to deliver.”
Turning to the other end of the microbiome therapeutic spectrum, Professor Nieuwdorp has carried out research with a mysterious metabolite known as 6-bromotryptophan (6-BT). This was found to be associated with less autoimmunity when looking at responders to faecal transplant. Its role in human pathophysiology was previously unknown, but levels of 6-BT were lower in the plasma of people with type 1 diabetes. It turned out that 6-BT halves the inflammatory response in vitro and, in an animal model, it improves insulin production and produces fewer cytokines. “We are now going into human studies with this natural metabolite, where we will treat 36 patients with metabolic syndrome and we hope to move quickly into type 1 diabetes,” Professor Nieuwdorp concluded.
To learn more about arresting type 1 diabetes, enrol on the EASD e-Learning course ‘The pathogenesis of type1 diabetes’: https://easd-elearning.net/all-courses/the-pathogenesis-of-type-1-diabetes/
Any opinions expressed in this article are the responsibility of the EASD e-Learning Programme Director, Dr Eleanor D Kennedy.