Professor Judith Klein-Seetharaman, WMS

Published January 2014

“When people design and manufacture machines, the aim is to optimise everything about it to ensure it runs effectively at all times. For example a car has various in-built sensors that that provide continuous information to ensure it performs effectively. Similarly, the human body has sensors, signalling when we need to do something. Like the fuel gauge in a car, our appetite is an in-built sensor, designed to tell us when we need to ‘top up the tank’ and eat."

"However unlike the fuel gauge, our appetite no longer works effectively, like it did for our ancestors. We’re supported by technology and an abundance of resources, so we don’t need to spend time or energy foraging for food. Living in this partly artificial environment is preventing our appetite – one of our most important sensors – from functioning optimally. In other words, we cannot rely on natural instinct any longer when it comes to food."

Tackling obesity

"So how do we tackle this? In the UK steps have been taken to encourage people to eat a balanced diet. Most people will be familiar with the ‘eat five a day’ mantra, a Department of Health campaign which launched over ten years ago. The aim was to increase the number of people eating a healthy and balanced diet. However, only one in five people actually follow through with this recommendation."

"Even the people who are getting their ‘five a day’ may not necessarily be eating healthily. A recent BBC reality TV show shadowing a family illustrated this. The family ate fruit – but often it was covered in ice-cream, caramel sauce and topped with heavy whipping cream. They also ate some vegetables – but only accompanying a large fatty meal. After meal times the family frequently sat down to watch TV and have a few snacks. Within minutes of eating a meal they exceeded the recommended daily calorie intake, just by eating snacks. More worrying is the fact that the family couldn’t remember all the food they ate, only the camera revealed the true excess of their diet."

"Unfortunately, this case study is just one example of a phenomenon well known to clinicians dealing with obesity. People just don’t want to admit they overeat, or they simply forget, or underestimate the portion size."

Food sense

"Unreliable self-reporting [1] has been demonstrated in a number of clinical and epidemiological studies. Regardless of body weight, people systematically underestimate meal calories and food intake. The reasons for this are complex and many.

Just like we wouldn’t drive to the petrol station if the tank is full, we shouldn’t eat a Mars bar if we’ve recently had a meal. The discrepancy between what we need and the amount of food we actually eat is increasing to the extent that obesity levels are of global concern.

Many believe that obesity problems are related to sensing: since we cannot reliably figure out what and when we ate, we need feedback from an artificial sensor, so that we don’t have to rely on the clearly unreliable self-reporting. We need artificial sensors to help make the right decisions."

Will artificial sensors replace our appetite?

"What’s needed is an independent food intake sensor. Some technological solutions have been proposed such as cameras, even mounted onto people, and picture taking apps to recognize food. However, pictures alone will never be sufficient since cooking (some styles more than others) can render the ingredients unrecognizable. Clearly, a molecular sensor that identifies nutrients in body fluids such as blood or urine is needed.

An entire field of research, nutritional epidemiology, has been invented to address this task. Reliable biomarkers for salt (sodium) and protein (nitrogen) as well as energy expenditure (doubly labelled water) exist, as well as identifiers of food groups such as fruit (vitamin C) and specific foods such as salmon, broccoli, wholegrain wheat cereal, raspberry, coffee, tea and red meat. The relatively new area of metabolomics, in which all small molecules in a sample are simultaneously identified using either mass spectrometry or nuclear magnetic resonance techniques, promises to help in identifying new biomarkers of food intake. However, we’re still a way off before this type of research can be used to create artificial sensors that can be used by patients.

While a full automatic evaluation of food intake through independent, artificial sensors is still out of reach, theoretically we know it has the potential to work because specific molecules are already being sensed routinely by patients. The best known example is glucose, measured by people with diabetes to understand fluctuating sugar levels in their blood. Decades of research have gone into the optimization of this measurement to make it easy to use and sufficiently reliable for people to use at home. However, because blood sugar levels are normally constant, measuring glucose is only useful for patients who already have diabetes.

This would be different for insulin. Insulin is the hormone responsible for the regulation of glucose levels, and its concentration varies with meal intake, time and metabolic state. While insulin concentration cannot be used to quantify a given meal’s composition, its concentration does provide valuable information on the consequence of a meal intake for a body. There are many advances in biotechnology that can be exploited to build artificial sensors of biological molecules such as insulin and others. One of these advances is the new field of nanobiotechnology, which provides novel platforms to develop sensitive and inexpensive biosensors. Researchers at Warwick are currently optimizing assays for insulin using the nanobiotechnology approach to make them accessible to people to use at home routinely, rather than requiring expensive and time-consuming laboratory tests. By coating carbon nanotubes with receptors – biological molecules whose structures are sensitive to the binding of a ligand, here insulin, a change in electrical signal is read to indicate the quantity of ligand present in a solution such as blood or urine. We hope to optimize this signal to develop a sensitive and specific biosensor for insulin.

Finally, the area of digital health is increasingly becoming important. Digital health aims to improve people's health and wellbeing through the use of digital technologies. For example, the glucose test can be combined with a digital device, like an iphone or a computer, to give a visual of the results. Other examples of digital health include activity trackers and heart rate monitors. In all cases, the data is stored in digital format, allowing it to be manipulated, shared and mined for information. In principle, any sensor – including the insulin sensor - can be connected to a digital device, and the data can be used to provide the user with feedback. Digital health has the potential is to connect artificial sensors with our own natural sensors helping us to make sure we eat the right type of food at the right time.

Artificial sensors could be of benefit not only to morbidly obese people, they could help everyone who struggles to control their weight. We could envision a world where everyone will use the sensors to help them decide what to eat. You are more likely to switch off alight if you see a counter tracking electricity consumption, and it’s the same with food. You are more likely to wait to eat a meal until your phone gives you the green light to indicate that insulin levels have dropped below a threshold. This could provide people with the extra-motivation needed to eat only when your body actually needs it. Research currently taking place at Warwick and other institutions could see this idea become a reality within the next few years.”

Find out more about Warwick’s Medical Research and how you can support this work.

References on unreliable self-reporting include:

• Chandon and Wansink (2007) J. of Marketing Res. 44, 84-99; Hudson et al. (2006) J. of the American College of Nutrition 25, 370-381
• Schatzkin et al. (2003) International J. of Epidemiology 32, 1054-1062; Tooze et al. (2004) The American Journal of Clinical Nutrition 79,795-80
• Trabulsi and Schoeller (2001) American Journal of Physiology – Endocrinology and Metabolism 281, E891-E899. Kaaks and Riboli (2005) Int J Cancer 116, 662-664
• Michels (2005) Int J Cancer 116, 665-666; Kipnis et al. (2002) Public Health Nutr 5, 915-923; Bingham (1991) Ann Nutr Metab 35, 117-127.

We strive to break new ground every day.

Keeping up this momentum is a challenge but, by maintaining our research excellence, we continue to go beyond boundaries. That’s why we’re proud of our strong performance in the Government’s Research Excellence Framework (REF) 2014, for both overall grade point average (GPA) and intensity.

We’ve strengthened our position amongst the UK’s ten best research universities. Warwick’s intensity also achieved a top ten ranking, signifying the strength-in-depth of our exceptional body of research staff.

Judith Klein-Seetharaman is Professor of Biomedicine and Systems Biology jointly appointed by the Division of Metabolic and Vascular Health, WMS and the Warwick Systems Biology Centre. Prior, she held professorial posts in the Department of Structural Biology at the University of Pittsburgh, USA and the School of Biological Sciences at Royal Holloway University of London.

She received her PhD from the Massachusetts Institute of Technology, USA. She has published 112 scientific articles, and her work has received numerous prizes, including the Sofya Kovalevskaya Award from the Humboldt Foundation, the Margaret Oakley Dayhoff Award from the Biophysical Society, a Bill and Melinda Gates Foundation Grand Challenge Award and a Marie Curie Incoming Fellowship.