Metabolomics for monitoring dietary exposure

There is strong evidence that dietary choices modulate risk of developing a range of chronic diseases and strategies for reducing disease burden emphasize the importance of changing dietary patterns. Current public health guidelines have common themes that encourage consumption of specific food groups, for example, oily fish, wholegrain foods, fruits and vegetables whilst reducing intakes of fatty and sugary foods. To maximise the effectiveness healthy eating interventions in the UK (and elsewhere), we need validated population screening methods with which to measure food consumption before, during and after the intervention. Conventional dietary survey methods are a considerable burden for study participants and are subject to bias. In addition, the processing and analysis of dietary data is extremely labour intensive with the net result that such dietary survey information may be difficult to validate. Such difficulties limit the ability to interpret the outcomes of healthy eating interventions.

Significant advances have been made recently in developing 'biomarkers' of dietary intake based on metabolites found in urine or blood but, currently, chemical biomarkers are available for only a relatively small number of specific foods and food components and most are of uncertain validity.

Funded largely by the UK Food Standards Agency the research teams in Aberystwyth, Newcastle and Imperial College have used metabolomics approaches to develop a novel strategy for discovery of new biomarkers of food intake. "Metabolomics" describes the measure of all (or many) of the metabolites (small molecules) in biological fluids such as blood or urine. Following digestion, absorption and metabolism, foods give rise to thousands of different metabolites in the human body and the appearance of certain metabolites is characteristic of particular foods. In our earlier work, we have developed urine sampling methods potentially suitable for use in population studies and have explored the use of comprehensive, non-targeted chemical analysis of urine (metabolomics technology) to identify dietary biomarker leads. Using this approach, we have identified potential biomarkers for several foods of high public health significance including oily fish such as salmon, citrus fruit, other fruit such as raspberries and the green vegetable broccoli. We thus now intend to test the hypothesis that the chemical composition of urine samples from free-living individuals will accurately reflect dietary intake.

The research programme will have the following major aims:

1. To identify chemicals in urine associated with recent consumption of specific foods deemed to have high public health importance.

2. To demonstrate a quantitative relationship between the amount of each specific food eaten and each potential chemical 'biomarker' in urine samples.

3. To validate a simple urine sampling procedure that is minimally intrusive to the every day activities of individuals participating in dietary surveys.

4. To assess the performance of dietary exposure biomarkers in the prediction of habitual diet in the UK.

The research strategy will be designed to integrate where possible with major surveys of UK dietary habits which will offer future opportunity to determine whether the adoption of biomarker technology improves the cost effectiveness and accuracy of population-level dietary surveys. In addition there will be distinctive advantages in future longitudinal studies aiming to generate definite links between disease risk and individual habitual exposure to specific foods. The results of this research will provide data to allow statuary bodies to refine dietary advice provided to consumers as well as offering further opportunity to interact with food producers, manufacturers and retailers to encourage provision of foods targeted to improve the overall health of the UK population.

In collaboration with colleagues at the MRC Human Nutrition Research, we have identified a panel of foods of high public health importance suitable for exposure monitoring in urine using biomarker technology. Free-living volunteers will consume these specific foods as part of everyday eating patterns over 3 days before collection of urine and blood for identification of potential biomarkers.

The timings of specific meals and of biofluid collection will be pre-determined to facilitate investigation of the duration of food-derived metabolite signals and the most informative urine collection strategies. In addition, 400 study participants consuming their usual diets will be recruited. We will estimate habitual dietary intake using the EPIC FFQ and collect "spot" urine samples. This population sample will be used to test the utility of putative metabolite biomarkers of dietary intake.

NMR and LC-MS technology will be used to generate comprehensive metabolite profiles of urine samples. Supervised data mining methods will identify signals explanatory of exposure to specific foods. Targeted NMR and accurate mass FT-ICR-MS will be used to structurally annotate metabolites with biomarker potential. Targeted analytical methods will be established to simultaneously measure several biomarker signals (e.g. using MRM approaches) in urine.

In controlled residential food intervention studies, volunteers will be asked to consume known amounts of test foods for which we have identified potential biomarkers and urine and blood will be collected at defined intervals for up to 3 days after consumption. We aim to describe the kinetics of signal decay and to determine the quantitative relationships between food intake and urinary metabolite concentrations.

Biomarkers proving informative will be tested for utility using dietary data and urine samples from both recent National Diet and Nutrition Survey studies (24hr urines) and the 400 "spot" urines from the present study.

Who will benefit from this research?

This proposal aims to help future studies to investigate more efficiently the challenging question of the effect of dietary exposure on disease prevention - health - or causation. We therefore anticipate that this research will have a wide range of beneficiaries (in addition to the academic community) including:

- Regulators with responsibilities in the area of food and health e.g. the European Food Safety Authority and the UK's Advisory Committee on Novel Foods and Processes,
- The food industry including those developing novel foods or food ingredients aimed at personalised nutrition and those advising consumers,
- Manufacturers of analytical platforms and software tools for metabolomics analyses,
- The general public.


How will they benefit from this research?

Regulators with responsibilities in the area of food and health are currently exerting considerable effort to get across the message to consumers about the benefit to health and well being of including, for example, sufficient amounts of fruit and vegetables (F & V) in their diet. However, their effort is pre-empted by the conflicting conclusions resulting from current epidemiological and dietary intervention studies in humans. Within the next 4-5 years this project will provide food biomarker technology that will allow future researchers to generate objective quantitative measures of habitual dietary exposure at the population level in the UK. Quicker and cheaper high throughput methods to assess dietary exposure will support the scale up of longitudinal population surveys to allow links to be made between exposure to specific food and future health risks.

Thus any health outcomes associated with dietary trends can allow regulators to develop even stronger messages to persuade the general public to change eating habits.

It is anticipated that there would be incremental integration of biomarker technology into population studies, starting with validation of the present dietary recording tools, particularly for monitoring accurately exposure to the individual components of composite meals. As the 'coding' of meals from diet diaries is extremely time consuming (as well as a source of much experimental inaccuracy) it is expected that the availability of validated food biomarker technology would allow organisations responsible for diet habit monitoring to simplify the data collection protocols, making the surveys more cost effective.

Outcomes from this research will provide the food industry with more robust measures and tools for R&D in the area of personalised nutrition and it is anticipated that impact will be realised within the next 5 years.

A deep understanding of metabolites in urine will help manufacturers of mass spectrometers and NMR instruments to develop new data processing and data annotation software tools more adapted to human nutritional metabolomics.

The general public are expected to benefit from this project in at least 3 ways i) through provision of objective information about the complex interactions of humans with their diet; ii) by enhancing the quality and robustness of evidence which will be required to elaborate and support dietary recommendation on consumption of healthy foods, and iii) by encouraging the food industry to focus on the marketing of food products with proven health benefits.

What will be done to ensure that they have the opportunity to benefit from this research?

We will disseminate the outcomes from the research project via the conventional route of academic publications using open access journals where possible. This will ensure widespread access by all stakeholders to the project outcomes following the important quality assurance step of peer review.
In addition, we will present findings on an on-going basis via conference presentations targeting conferences and workshops attended by diverse user groups including industry and regulators.