Smelling the disease - Nosing our way towards earlier diagnosis and new biomarkers

Could the sense of smell drive future biological detection methods? In university and drug company labs which develop assays for diseases and for measuring drug effects, you will find a range of readers for measuring changes in biological systems.

Typically, they rely upon detection techniques which are about ‘seeing’ biological changes, analogous to our sense of sight. These techniques include colorimetric changes, fluorescence labelling of proteins in cells to see where they are, luminescence assays, and harder to describe techniques such as fluorescence resonance energy transfer, or fluorescence lifetime. All of these detect light changes, by photomultiplier tubes or cameras as the equivalent of human eyes.

But of course, humans don’t just rely on the sense of sight. Similarly, there are many biological changes in systems which can be detected using methods which mimic other senses. We all know for example that the standard approach for detecting if a pint of milk has ‘gone off’, is to smell it (Or if you are like me, to ask your partner to smell it). Smell is a highly sensitive method for detecting complex biological changes, usually detecting volatile organic compounds (or VOCs) that are being emitted from a biological system. Emerging scientific literature is demonstrating that measuring these VOCs can provide sensitive and accurate measures of the biological system. A benefit of this is allowing researchers to avoid needing liquid or tissue biosamples for testing – that is, bioassays that are non-invasive in nature.

Animals have long been used to detect biological signatures. Giant rats have been trained in Tanzania to smell landmines and more recently to smell Tuberculosis (TB) within sputum of patients. This is now used as an extra check for TB in remote settings in which smear tests alone may miss individuals unable to produce much sputum or having low bacterial levels in their sputum samples. As more of us encounter dogs than giant rats, there are also numerous reports of dogs smelling biological changes associated with human disease. It has been reported for example that trained dogs are able to detect colorectal cancer and lung cancer from samples of patient breath.

A high-profile example is the work of Joy Milne and Perdita Barran in demonstrating that Joy (a so-called ‘super smeller’ individual) can detect individuals with Parkinson’s Disease by smelling VOCs in their sebum (a natural oily secretion of the skin’s sebaceous glands). Joy discovered this ability following her experience with her husband, on whom she detected a specific smell twelve years before he was diagnosed with the disease, and subsequent exposure to a wider Parkinson’s patient group. The research with Barran’s group in Manchester­ - an expert group in gas chromatography mass spectrometry - showed that when Joy smelled the same samples being heated (to release volatile compounds) and sent to the mass spectrometer, molecular VOC signals could be identified both by Joy and by the multiple sclerosis.

Can these early findings, from Joy Milne to the giant rats, stimulate the development of new technologies to measure such changes? As laboratory assay technologies catch up with nasal evolution in sensitive measurement of VOCs, changes in VOCs could form a growing opportunity to determine disease and biomarkers of disease (the ‘volatilome’). It is exciting that such assays can be performed in highly accessible samples (sebum, breath), which could reduce cost and hospital time for patients.

As well as the mass spectrometry technology referred to above, numerous companies are now developing ‘e-nose’ platforms. These typically use pattern recognition and artificial intelligence to detect changes in disease samples without needing to identify the specific biological molecules that underpin it. This can make them cheaper and easy to use and so appropriate for point of care settings. In the UK, Owlstone Medical have been developing a breath-based e-nose platform for VOC biomarker discovery for early disease detection, therapeutic development, and disease monitoring. Other e-nose approaches are under development in UK universities.

We are, then, seeing a growing understanding that smell-based biological detection can be affordable, reliable, and enabling the detection of disease at an earlier time point. The impact of this for biobanks (what biosamples we should store?), clinical trial design (presymptomatic patients being determined that enable early-intervention clinical trials), and technology companies seeing new markets, could in time be huge.

By Dr Peter Simpson, Medicines Discovery Catapult