New RNA diagnostics could become the crystal ball of diabetes
Physical activity may be one of the golden rules of diabetes prevention, but its effectiveness can vary from one person to another. In addition to clarifying the genetic mechanisms behind these differences, results from METAPREDICT provide the basis for what could be the cornerstone of future diabetes prevention — a programme best suited to our very own molecular make-up. Diabetes associations across the world have been stressing the importance of physical activity as a preventive measure against diabetes for a long time, and with good reason. Benefits include better physical fitness, improved lipid levels, improved glycaemic control, reduced risk of cardiovascular disease, decreased adiposity and enhanced physiological well-being. But what these associations also insist on is the fact that exercise is not without risk, and that physical activity should only be considered when the benefits outweigh these risks. The latter observation begs a fundamental question: how do we know if, and what sort of physical activity is good for us? Surely answering that question before a patient starts exercising would prove useful, but we currently know too little about the causes of adverse or absent reactions to exercise. Scientists from the EU-funded METAPREDICT (Developing predictors of the health benefits of exercise for individuals) project, who have already demonstrated how 30 % of patients show no improvement in insulin sensitivity when performing endurance training, are making waves with the development of new RNA diagnostics. The latter are capable of anticipating differences in responses to exercise and could potentially estimate the risk of a patient developing diabetes by designing individual exercise plans around the molecular make-up of the individual. Prof. Jamie Timmons, Chief Scientific Officer at project partner XRGenomics, highlights the value of the project results and their potential for creating the prevention programme of the 21st century — one that will be tailored to each patient’s genes. Your project was inspired by the observation that aerobic fitness is ineffective for some people. How is this possible? This is a remarkable observation I agree. We see no change in aerobic capacity in about 20 % of all people with fully supervised exercise training (training three to five days per week). In fact we were able to replicate this observation in the new METAPREDICT HIT (High Intensity Training) study (one of three modalities of exercise in the study). Those who failed to change their aerobic capacity were ‘trying’ harder on the bike in the end, and ventilating more, but this was not resulting in more oxygen being consumed by the mitochondria in their muscles. When studying the mechanisms behind this failure in gaining aerobic fitness, we originally found that people who do not gain in aerobic fitness fail to activate gene-expression programmes relating to blood vessel growth. Then we demonstrated, in a new animal model developed by Lauren Koch, that if you selectively breed for good or poor aerobic fitness responses to exercise it is possible to reproduce this same pattern of failed matrix remodelling, that you see in humans that are non-responders for gains in aerobic capacity. The latter relates to alterations in SMAD signalling and matrix remodelling, just as we noted in human subjects. The question remains, why? Why no improvement in aerobic fitness, and what triggered the selection of this common ‘genetic’ phenotype in humans? We could speculate, for example, that they will also be poor at growing blood vessels in tumours and that may be advantageous. But that is currently speculation, which we can address once we produce good diagnostics for this phenotype. What else did you learn from your studies on patient response to exercise? The METAPREDICT project aimed to go far beyond the idea of ‘non-responders’ to aerobic capacity. In total we have studied three types of supervised exercise training (HIT, classic public health high-volume aerobic training and resistance training) and this has involved over 1 000 people. We have also measured over 40 000 insulin samples in Prof. Rooyackers’ laboratory and we are building RNA diagnostics for muscle and liver insulin resistance as well as how each of the three modes of exercise training improve or fail to improve insulin sensitivity (as well as aerobic fitness and blood pressure). So far we have shown that, on average, HIT is highly effective at improving cardiovascular health biomarkers with a less than 15-minute of actual exercise per week. It was actually more effective than current public health guidelines, which suggest that people do more than 150 minutes a week of moderate intensity training. In a scenario where fitness is not a solution, how can diabetes be prevented? A key concept in life-style risk factors is that each biomarker for health plays its own role in determining your long-term health. So, for example, you may not get any improvement in aerobic fitness but you may reduce your blood pressure and reduce your insulin resistance. However, which of the three types of exercise will work for you best? And how often should you do exercise to optimise improvements? Answers to these questions are the long-term aims of the METAPREDICT project. Once we complete the development of our RNA diagnostics, we will be able to begin studies allocating people to the programme best suited to their ‘genes’. However, a small percentage (although that’s still a rather high number of people worldwide!) will not improve their aerobic capacity, show no change in insulin sensitivity and may even get an increase in blood pressure from performing exercise training three days a week. Should they then get prioritised for pharmacology treatment? This is quite possible. Should they try training less often? Also quite possibly. What we are talking about here is what the future of preventative health care will look like. Were you successful in your efforts to find biomarkers that would help find the most suitable lifestyle for each patient? We have been able to make prototype biomarkers from RNA using machine learning approaches for a number of clinical phenotypes and we will progress this work into 2016. Was it a difficult process? The project was very demanding, with a large number of logistical challenges. It was very difficult to ensure that all clinical centres remained motivated and engaged in the project, and in the end a small number of people within the consortium have continued to dedicate a great deal of time to the analysis process. I would definitely recommend that similar projects, in the future, focus as much resources on data management and computing resources, since the data-generation process can be overwhelming. Luckily we have made good progress now and will submit several patents and make several publications in 2016. What were the other main achievements of your project? I think with such a project, it is a major achievement just to gather all the clinical data and have the laboratory analysis carried out properly. In this sense the major achievement was getting to the point of a fully QC-checked novel ‘exercise’ database. In recent months we have developed the first predictors of metabolic status, and of course, when completed, that will be the project’s main legacy — a 21st century approach to metabolic physiology, if you like. What are your plans now that the project has ended? We are focused on patent applications and producing the publications. In that sense the project will live on for a few more years and of course it needs to find new funding to keep it going. In the United States, the NIH has just announced a new USD 100M funding to initiate a six-year study that replicates the aims of METAPREDICT. It is clearly seen as a topic of major importance. How do you see your research eventually benefiting patients? The METAPREDICT project was both basic science and translational research and so we always focused on producing something that was useful for patients or preventative medicine. Our approach has been radically different, we stopped looking for ‘master regulators’ of complex physiological processes from trying to isolate mechanisms in cells or mice, and accepted that biology in out-bred humans is stochastic; that molecular markers have to come together and make a ‘majority vote’ on any sort of physiological response. This means you are looking for variable molecules, molecules that help explain differences between people, not molecules that they all regulate in common. At the moment we have a huge problem with Type 2 diabetes, partly driven by obesity, partly by inactivity but also by factors that I think remain unclear or which are non-medical. A major component of diabetes is failure of the muscle and liver to respond to insulin, yet this parameter is rarely measured in real life because it’s too cumbersome and costly. We would hope to have a simple blood test that could provide an ‘early warning’ system for insulin resistance as well as guidelines on how to best counter your diabetes risk factors within the next couple of years. For further information please visit: METAPREDICT project website
Countries
Sweden