Could individually tailored diets help improve health outcomes for people with diabetes, hypertension, and other chronic diseases? Prof. Elaine Holmes believes they can. With the help of nuclear magnetic resonance (NMR), she can obtain a molecule-level understanding of the differences in the way people metabolize food. We talk to her about her hopes for making ‘personalized nutrition’ a reality.
Guidance on healthy eating has for many years been the same: “less salt, less sugar, less fat, more fiber.” Although it remains good advice, says Prof. Elaine Holmes at Murdoch University’s Health Futures Institute, there’s a growing understanding that a ‘one-size-fits-all’ approach doesn’t work for everyone, and that in certain situations health outcomes can be improved by taking a more nuanced approach.
Getting the big picture of diet
This more detailed approach to understanding diet and health is one that Prof. Holmes and her team are strong advocates for. “Unlike many researchers, we don’t do studies on individual foods – people’s diets and metabolism are more complex than that,” she says. As an example of this complexity, she cites a study carried out at Imperial College London, in which a group of subjects received exactly the same diet under controlled conditions, but some were found to excrete far fewer calories in their urine than others, and different metabolic pathways were activated in different people.
So instead of looking at individual foods or nutrient classes, her team studies the effect of the diet as a whole, by identifying the full range of compounds that emerge at the end of the digestive process – the ‘metabolic phenotype’. These compounds, she says, fall into three main categories depending on where they originate: from the foods themselves, from human cellular processes, or from gut microbial transformation of chemicals in the diet or from other sources. “By understanding the processes that give rise to compounds in all three classes, we can obtain a detailed picture of what’s going on inside the body – with the impact of the gut microflora now being recognized as particularly important,” she says.
Taking account of sampling subtleties
The most useful body fluid for studying these compounds, according to Prof. Holmes, is urine: “No single body fluid tells you everything, but urine is the most useful because it’s not held under such tight homeostatic control as blood plasma. Also, waste products are present at concentrations that are an order-of-magnitude higher than in plasma, so they’re easier to detect. Urine is also preferred over stool samples because many small chemicals produced by the gut bacteria get re-absorbed, and subsequently eliminated from the body in the urine.”
Her team is careful to ensure the quality of the samples they’re using – “In general, we prefer to collect samples ourselves rather than use biobanks, as these were mostly built for the purposes of genetic profiling rather than metabolic profiling. Also, at least historically, biobank samples have either not been collected under the optimal conditions for preserving biochemical stability, or the associated dietary reporting has been questionable. But having said that, the picture is improving, and we’re currently collaborating on several large-scale population studies, as they’re a very good way of getting a better understanding of how diets and metabolism vary across populations globally.”
The role of NMR in understanding metabolism
For pinpointing and identifying compounds in biofluids, Prof. Holmes instinctively turns to NMR. “It has been central to my research since 1987, when I first encountered the technology as part of my PhD on renal toxins in rodent models. At the time, I quite literally knew nothing about it, so I was fascinated to see how much information you could extract from a single analysis of a biofluid.”
Since then, Prof. Holmes has come to value numerous characteristics of NMR: “On a day-to-day basis, it’s simply the efficiency of our Bruker NMR instruments that makes them such good tools, in terms of their reproducibility, ease of use, and broad applicability.” These features lend themselves very well to her work, she says: “With NMR, we get a quick snapshot of a sample, and we can repeat this at intervals of weeks, months or even years, to follow a person’s metabolism and see their response to a particular dietary intervention.”
She says that the atomic-level data provided by NMR gives it a particular advantage over mass spectrometry (MS) – a popular tool for analyzing the metabolome: “To give just one example, we can apply diffusion-weighted pulse sequences that target certain glycoproteins, allowing us to obtain biomarker profiles that detail individual compounds, rather than broader compound classes.” (Figure 1).
Improving health by metabolic phenotyping
This ability to make links between the detailed molecular profile of a biofluid, metabolic pathways, gut microflora and dietary choices has been backed-up by a recent peer-reviewed paper in Nature Food, co-authored by Prof. Holmes. In this study, they showed that NMR-acquired metabolite profiles in urine can be correlated to the levels of dietary nutrients, and so accurately predict healthy and unhealthy diets.
One of the most exciting applications of such ‘deep’ metabolic phenotyping, says Prof. Holmes, is personalized nutrition: “In the same way that a clinician currently asks for a blood test to test for glucose, we can envisage them requesting a urine test for a suite of dietary biomarkers. This would provide them with information on the appropriateness of that diet for the patient’s metabolism, and hence how their food intake might be adjusted to reduce the risk of a poor health outcome.”
There are many ways in which this concept could be applied, she explains: “It could be really useful for people needing to lose weight – for example, those with cardiovascular disease, hypertension, or pre-diabetes. In such cases, the biomarker profile could potentially be used to recommend a probiotic or prebiotic intervention, with the aim of promoting a gut flora that is less efficient at harvesting calories.” Of course, this doesn’t solve the disease issue, but it can contribute to a tailored lifestyle management plan. The simplicity and speed of such an NMR-based test, she says, would mean that patients could be tracked over time, and the recommendations modified accordingly. To help make this a reality, Prof. Holmes and colleagues, Isabel Garcia Perez and Gary Frost, have recently formed a spin-off company, Melico, to help implement NMR profiling of biofluids as part of routine clinical dietary advice.
Putting personalized nutrition into practice
With these numerous possibilities, Prof. Holmes has plenty to keep her busy. “We’ve got lots of ideas around this concept of personalized nutrition, and we’ve even teamed-up with someone from the agricultural industry… and a chef! The idea is that, using NMR profiling, we can identify the optimum diet for a given person, and then provide them with ready-to-cook meals that will not only taste good, but be a good match for their metabolic profile. Essentially, by considering the whole chain from the farm to the clinic, it should help people to ensure they’re on the best possible diet – especially in cases where they’re not able to prepare meals themselves.”
Another application centers on food supplements and so-called ‘superfoods’. Prof. Holmes says: “Specialist honey is very popular here in Western Australia. With the help of NMR databases like those developed by Bruker, we’re already answering questions relating to product authenticity. But from the dietary side, we can now go further, and begin to investigate the health impact of products from different sources, or products of different quality. For example, if a particular honey has been adulterated with sugar water, could that have health consequences for the person taking it?”
Fine-tuning the ‘healthy eating’ message
Through her research, Prof. Holmes is still amazed at the diversity of metabolic processes: “In a study of individuals on a diet designed to improve cardiovascular risk , most participants responded with reduction in cholesterol and/or blood pressure. For a subset of people, the diet did not relate to a significant reduction in either blood pressure or blood lipids and this group was associated with differences in chemicals in the urine profile that are made by the gut bacteria.”
However, she points out, this metabolic diversity doesn’t mean that conventional wisdom on healthy eating is wrong: “Whatever someone’s metabolism, if someone on a healthy diet suddenly switches to binge-eating junk food, then they’re going to fare worse overall. But once you get down to the fine detail like we’re doing, then you might see significant differences in their response to that new diet – for example, in terms of the levels of high- and low-density lipoproteins, or the amount of cholesterol.” 
Understanding such differences is what drives her research. “Investigating the detailed, molecule-by-molecule response to a given dietary intervention for a particular person has the potential to make a major difference to health outcomes. By making use of the molecular insights offered by NMR, I see a future where obtaining personalized recommendations about diet moves out of the realm of research and becomes part of regular clinical practice.”
Prof. Elaine Holmes is the Professor of Molecular Phenomics at Murdoch University, Perth, Western Australia. She is also Director of the Centre for Computational and Systems Medicine and a Professor of Chemical Biology in the Department of Surgery and Cancer at Imperial College, UK. she has over 20 years’ experience in metabonomic technology and its application in the discovery and development of metabolic biomarkers of disease in personalized healthcare and nutrition.
 I. Garcia Perez et al., Objective assessment of dietary patterns by use of metabolic phenotyping: a randomised, controlled, crossover trial. Lancet Diabetes Endocrinol, 2017; 5(3): 184-195.
 J.M. Posma et al., Nutriome–metabolome relationships provide insights into dietary intake and metabolism, Nature Food, 2020, 1: 426–436.
 E. Holmes, A. Wijeyesekera, S.D. Taylor-Robinson and J.K. Nicholson, The promise of metabolic phenotyping in gastroenterology and hepatology, Nature Reviews Gastroenterology & Hepatology, 2015, 12: 458–471.
 R. L. Loo, et al Characterization of metabolic responses to healthy diets and association with blood pressure: application to the Optimal Macronutrient Intake Trial for Heart Health (OmniHeart), a randomized controlled study. Am J Clin Nutr. 2018; 107(3): 323-334.
 M.V. Holmes et al., Lipids, lipoproteins, and metabolites and risk of myocardial infarction and stroke, Journal of the American College of Cardiology, 2018, 71: 620–632.
About the Health Futures Institute
The Health Futures Institute at Murdoch University, Perth, focuses on clearly defined health-related research areas and pursues those in partnership with local healthcare providers, community groups and international collaborators. At the heart of the Institute is the Australian National Phenome Centre, which uses state-of -the-art technologies including NMR and mass spectrometry to transform and optimize disease prevention, diagnosis, and personalized healthcare.
For more information, please visit www.murdoch.edu.au/research/hfi.
About Bruker Corporation
Bruker is enabling scientists to make breakthrough discoveries and develop new applications that improve the quality of human life. Bruker’s high-performance scientific instruments and high-value analytical and diagnostic solutions enable scientists to explore life and materials at molecular, cellular, and microscopic levels. In close cooperation with our customers, Bruker is enabling innovation, improved productivity, and customer success in life science molecular research, in applied and pharma applications, in microscopy and nanoanalysis, and in industrial applications, as well as in cell biology, preclinical imaging, clinical phenomics and proteomics research and clinical microbiology.
For more information, please visit: www.bruker.com.
Bruker NMR Instruments are for research use only and not intended for Use in Clinical Diagnostic Procedures.