Biomarkers – why medicine is about to undergo the biggest single change ever
Professor Tony Freemont
Proctor Professor of Pathology, University of Manchester and Director of MMPathIC (Manchester MRC/EPSRC Molecular Pathology Innovation Centre).
A brief history of medical advancement
If you look at medical care over the past few centuries through the eyes of learned people living at the time, you gain a better appreciation of how the things we now take for granted impacted on their understanding of disease, illness and therapeutics.
The concept of the “seats of disease” (the relationship between abnormalities of organs and patient’s symptoms) had only just been proposed in the middle of the 18thcentury, when the physician and campaigner for public health, Thomas Percival (1740-1804), was embarking on his medical training. Even the classification of diseases as we might understand it today (infections, cancer, trauma, endocrine etc.) was only coming into acceptance when John Fielden (1784-1849), the politician and social reformer from Todmorden, was in his 20s.
Imagine how Joseph Whitworth (1803-1887), the engineer and entrepreneur who bequeathed much of his fortune to the people of Manchester, would have followed the development of anaesthetics in the 1840s, 50s and 60s, lest he might one day require surgery. In fact, the mid 19th century saw a mini boom in medical care. This included, in addition to anaesthesia, identifying the role of bacteria in causing infections, the discovery of the “cell” as the fundamental component of the body (displacing in this context the organ), and the whole idea of the aetiology of disease (diseases having understandable causes), which went way beyond bacterial infections.
Over the next century there was a steady acquisition of medical advances, such as the development of x-rays for diagnosis and treatment, until the next boom in medical knowledge around the middle of the 20thcentury, when poet and author Ted Hughes (1930-1998) began publishing his work. Developments during this period still influence our own views of disease and medical care. They include the discovery of DNA, the development of antibiotics, the invention of joint replacements, and the use of CT and MRI scanners.
The next medical boom
If history is to repeat itself as we move towards the middle of the 21stcentury, we might expect to see the earliest stirrings of the next “medical boom”. I would like to suggest that we are about to see just such a technologically driven change in medical practice that is so fundamental, exciting and disruptive, it will cause a complete revision of our views of disease and the ways in which medicine thinks of patients and prevents and manages disease.
Like the baby steps at the start of all scientific advance, the “new medicine” has evolved slowly and steadily. It is based on an increasing realisation over the last 50 years that disease can be understood at the molecular level (rather than that of the organ or cell). However, it is only through recent advances in engineering, such as rapid DNA sequencing and mass spectrometry, that it has become possible to identify and quantify large numbers of complex molecule. It is this that has opened the door to the utilisation of molecular information for informing diagnosis and patient management.
The “OMICS” revolution
It is now possible to have a snapshot of the molecular make up of blood, other body fluids and tissues in a way that defines individuals by their genome(DNA), transcriptome (expressed genes), proteome(protein), lipidome(fats) and metabolome(small molecules). The same applies to defining diseases, bacteria, and cancers. This capacity is truly game-changing, hence the term “omics revolution”.
The problem we now face is that there are many orders of magnitude more molecules than cells, and so to understand the make-up of individual patients and how that interacts with the molecular biology of disease, requires that we handle a huge volume of data. We are each composed of thousands of fats, tens of thousands of sugars, hundreds of thousands of proteins, and these can occur in a myriad of combinations.
The need to handle such vast amounts of “omics” data has led to the development of electromechanical computers and a new science, “informatics”, designed to handle “big data”. These technological advances have allowed us to think about experimental bioscience in a new way. The fact that we no longer need to do hypothesis-driven science is a case in point. Not long ago this would have been considered treasonable!
Traditional science relies on hypotheses. A hypothesis is a statement, based on available evidence, that predicts something that is not yet known but can be supported or refuted through carefully crafted experimentation or observation. With “big science” we now have the alternative option of revealing knowledge from the analysis of vast amounts of data.
Take the ‘old-fashioned’ hypothesis:Some cancers kill, I think it might be because they spread, I have discovered cancers that spread express molecule ‘x’, therefore my hypothesis is that ‘expression of molecule ‘x’ makes the outlook worse and I will look for molecule ‘x’ in these cancers’.
And the new:How do I predict who will die from their sarcoma? Some sarcomas kill because they have traits defined by genetic mutations in the sarcoma cells. Therefore if I start by knowing all the genes abnormally expressed in sarcoma cells I should be able to identify one or more that affect survival.
Biologics
Understanding disease at the molecular level has also spawned a new type of therapeutics based on the concept that most of us spontaneously launch molecular defences against diseases, which are more or less successful. If we could harness these naturally occurring molecules, or produce them outside the body, we would have treatments composed of naturally occurring molecules that have been tried and tested by nature for tens of thousands of years. This is now possible and has initiated a new area of therapeutic science based on the use of “biologics”.
Precision medicine
So the world of medicine has suddenly changed. By understanding individuals and diseases at the molecular level there is now the possibility to tailor treatments to the needs of specific patients. Presently a group of patients with the same disease will often receive the same initial treatment, yet not all will benefit from it. Of every 10 patients, say, 7 may respond positively to different degrees, 2 may not benefit and 1 may be made worse by the treatment. Now there lies the prospect, albeit currently distant, of all 10 receiving a slightly different treatment so that all benefit optimally. This is known as “personalised” or “precision medicine”. It would be wrong to raise hopes prematurely, but the concepts and technologies are around now that will one day bring this goal to fruition.
But the technology extends further. There has been a drive for acquiring molecular data on patients (and healthy people) when they are going about their daily activities, or recovering from surgery, illness etc., and certainly not restricting it to when they are in hospital or the doctor’s surgery. This is no futuristic dream, but simply an extension of the FitBit, the finger prick test used by diabetics for monitoring their blood sugar or 24 hour heart rate monitoring.
Supposing we could harness molecular data from wearable or indwelling devices? How easy would it be to monitor a patient’s response to treatment, or a patient’s changing molecular profile to detect when a disease starts to recur. The challenge is designing and powering sensors to do just this and world-class engineers in Manchester are amongst those leading the field. Remote monitoring of patients has fostered another area of data science called “mobile” or “e-medicine”, and this is another area where Manchester is leading internationally.
Infinite combinations?
However, there is a problem. If all individuals are different (with unique molecular make-up) and if all diseases are different (not only in their complex molecular profiles, but also in the way their molecules interact with the molecules of the individual patient) how on earth do we harness all that data and turn it to diagnostic or therapeutic advantage in real time?
Whilst one day we might have the technology to allow this level of complexity to be assimilated and used, today we don’t. But nature comes to our aid. Molecules don’t work independently, but rather as a hierarchy.
Rheumatoid arthritis illustrates this point well. At the molecular level, rheumatoid arthritis is one of the most complicated diseases it is possible to imagine. In the 1980s a video was made of a woman with rheumatoid disease being treated with an antibody to a naturally occurring molecule called TNF-a. The effect was astonishing. Following treatment she was able to walk, when before she could only hobble. The experts behind the discovery had recognised that although the molecular profile of the disease was complex, and differed from patient to patient, TNF-awas a master regulator of the molecular events occurring in the joints of their patients. Turn it off and all the disease processes would be damped down. TNF-atherefore became a therapeutic target. All that was needed was the bullet. This came in the form of a “biologic”, a molecule, an antibody, against TNF-a.
Molecular biomarkers
Recognising there is a hierarchy of molecules involved in disease means that the expression of certain specific molecules (or small group of molecules) can be used diagnostically. Furthermore, changes in their expression can enable the success, or otherwise, of treatment to be monitored. These molecules, or groups of molecules, are called “molecular biomarkers”.
Identifying molecular biomarkers that assist diagnosis (diagnostic biomarkers), help monitor effects of treatment (theranostic biomarkers), predict outcomes (prognostic biomarkers) or could become, like TNF-a, a target for treatment (therapeutic biomarkers) is taking off as a biomedical discipline. As new biomarkers are discovered (for example in cancers, inflammatory diseases, and in healing or non-healing wounds), there is a drive for them to enter clinical use.
Manchester is at the forefront of the endeavour to bring novel molecular biomarkers to the clinic. Manchester’s lead has been made possible, not only by the superb biomedical, engineering and clinical science across its Universities and Healthcare Trusts, but also by local politicians who have embraced the devolution of the health and social care budget from central government. Their vision for a healthcare system driven by technological advance and tailored for the local population is unique and inspirational. The ‘COPD Salford Lung Study’, a collaboration between public and private sector organisations, which used near real-time data, is just one example of the pioneering health analytics work taking place in Manchester.
What, then, will this “molecular revolution” in medicine mean to us? Medicine will change of that there is no doubt.
Monitoring patients in the community will mean earlier detection of disease, earlier initiation of treatment, and a move towards real time management of disease in, or near, the patient’s home. This will reduce the likelihood of disease going out of control requiring admission to hospital to change treatment. Fewer hospitals, therefore, will be needed, making health services more cost efficient.
Reclassification of disease
There will be a reclassification of disease. Already large companies, working in conjunction with the NHS and Universities, are moving from a disease taxonomy based on changes at the organ and cellular level, to one based on the expression of molecules. Diseases which have until recently been known as being quite disparate (e.g. cancers and joint diseases) are suddenly seen as being close relatives in terms of deranged molecular function with similar therapeutic biomarker profiles. Put another way, there are, for instance, inflammatory and malignant diseases that will respond to the same biologic. This shouldn’t be a surprise, after all the drug methotrexate is used to treat cancer and inflammatory arthritis. So, all healthcare professionals will need to change the way they think about disease and how they investigate and manage patients.
Our increasing understanding of molecular biomarkers will enable us to define diseases in different ways and completely new diseases will be recognised.
For instance, single disorders will be found to have several subtypes that have different outcomes (prognosis) and will require different treatments, as illustrated by the recent reclassification of Type 2 Diabetes.
The future of health-related professions
Healthcare professionals of the future will be guided as much by the results of molecular biomarker tests as they will by the patient’s history and physical findings. New specialisms and roles will emerge, not only in healthcare delivery itself, but also in the array of scientific, computing and engineering disciplines which are already supporting developments in biomarker discovery, identification, measurement, monitoring and therapeutic targeting.
New branches of medicine will evolve and there will be increasing numbers of doctors working behind the scenes at the molecular level, guiding diagnosis and even directing treatment. Not only that, but machine learning and artificial intelligence will help to hone complex tests to gain the most from the smallest impact, whether that is measured in terms of cost, pain or inconvenience to the patient.
For the doctor sat with the patient in front of them, a key component of the next medical boom will be the creation of a platform to combine patient data (the “omics”, the biologics and monitoring) with the patient record. This will allow the doctor to combine the best “scientific” precision medicine with the “art” of medicine using their skills of empathy and communication.
This report is based on a paper written by Tony Freemont, which formed the basis for his talk to the Hebden Bridge Literary & Scientific Society.
With thanks to Tony Freemont for providing this paper and to Ingrid Marshall for adapting it to reflect what was said on the night whilst retaining bonus explanatory material from the paper itself.
This is a copy of the Prof_Tony_Freemont presentation