07 December 2009
ByAppeared in BioNews 537
The availability of inexpensive, fast and accurate genome sequencing technology has lead to so-called genome-wide association (GWA) studies becoming one of the most popular approaches to the genetic study of complex disorders such as diabetes, cancer and heart disease. GWA studies involve examining the entire genome, without a specific hypothesis in mind, merely looking for replicable associations between genetic variation and a given disorder. While this has lead to many new genetic variants that raise the risk of inheriting a particular complex disorder being identified, variants identified in this way often only account for a small proportion of the heritable component of such disorders. But does 'missing heritabilitysuggest that the genetic component of common complex disorders has been overestimated?
A panel of speakers set about answering this question at the Progress Educational Trust (PET) conference - 'Does Genetics Matter? Help Hype and the New Horizon of Epigenetics' - held at Clifford Chance on Wednesday 18 November 2009.
Professor Steve Humphries, Director of the Centre for Cardiovascular Genetics at University College London and chair of the session, set the agenda against the backdrop of the recently insolvent Icelandic firm, deCODE, which raised privacy fears among customers for whom it held genetic data. Meanwhile, competing genomics company 23andMe reportedly put up its prices, which Professor Humphries took as a potential sign of the 'struggling consumer genetics industry', casting further doubt over whether this emergent field is ready to enter the clinic.
But determining clinical value, at this stage, is difficult. Professor Steve Jones, Head of Genetics, Evolution and Environment at University College London, said that, on the whole, we know little about genes - and the public knows even less. He explained that among the surprises stemming from the sequencing of the human genome was the discovery that a mere 25,000 genes account for all the variation seen among humanity. This notion, he said, tells us we don't know much about how the human genome works. Given our current level of understanding, it would be foolish to expect to explain complex common diseases in terms of their genetics. Professor Jones reiterated the concerns he expressed in an earlier Daily Telegraph article about spending large sums of money on identifying genetic variants with little or no predictive value.
Professor Jones referred briefly to the eugenics movement of the early 1900's, whose members sought to banish intellectual 'simplicity' and other socially undesirable traits by restricting breeding. Our improved understanding today of the rudiments of genetics refutes the assumption of the eugenics movement that complex human traits are governed by simple genetics. And yet, he said, a narrative of genetic determinism still rages in today's media reporting of genetics and elsewhere. He recalled that a recent Google search for the phrase 'Scientists find the gene for...' returned some 36,000 potential hits (despite there being only 25,000 genes in the human genome).
Modern research shows us that, for most human traits, there is huge genotypic variation behind phenotypic variation, he said, with each individual gene having only a small effect on the overall phenotype. To illustrate his point, Professor Jones used the example of human height. Scientists have long been trying to understand the genetics of human height and genomic searches have lead to the discovery of a number of associated genetic variants. Nevertheless, different studies have identified different variants and it is likely that the total number of genes identified so far only accounts for approximately 3 per cent of variation.
The 1948 Framingham Heart Study provides another example. It was a project that sought to identify the common genetic factors contributing to cardiovascular disease by following its development over a lengthy period in a large group of participants. Over the years, this study has lead to a significant number of risk variants attributable to cardiovascular disease being identified. But Professor Jones questioned the value of these discoveries when the likely interventions, such as quitting smoking, exercise and healthy diet, are of benefit to everyone, regardless of risk. He suggested that a better approach might be to target many low risk groups of people, rather than a few high risk individuals.
Professor Mark McCarthy, from the Oxford Centre for Diabetes, Endocrinology and Metabolism, cast further doubt on the usefulness of genetic tests for diabetes. He told the audience that, at present, it would be easier to identify a predisposition to the condition by reference to a patient's family history and lifestyle choices, rather than genetics. Yet, he acknowledged that the predictive value of genetics is therapeutically important where there is limited biological understanding. Although, at present, we do not all have the genetic information needed to accurately predict most disorders, we will have new ways of treating and preventing disease in the future.
Realising this goal requires embracing competing methodologies. Professor McCarthy questioned GWA studies, which restrict research to certain common gene variants, but it is important to acknowledge that rare gene variants are also making a significant contribution. To date, 40 genetic variants have been found that influence diabetes, but these account for perhaps less than 10 per cent of the genetic contribution to the disease. Furthermore, we understand little of the biology of these gene variants, meaning there is little scope for clinical translation. Furthermore, the majority have been identified within populations of European origin meaning further studies will be needed to establish whether they have predictive value among other non-European populations.
In conclusion, Professor McCarthy stressed that what we know about the genetics of diabetes was really just the tip of the iceberg. He predicted that further research into understanding the biology of certain risk variants and identifying risk factors of intermediate predictive power and prevalence, would help provide a more complete view of the condition and lead to improved treatment and prevention.
David Melzer, Professor of Epidemiology and Public Health at the Peninsula Medical School, expanded the debate by introducing questions about the evaluation of genetic tests for both scientists and consumers. For example, the (now insolvent) company deCODE genetics launched a genetic test which it claimed could predict 21 per cent of cases of diabetes. But, although proof of association may well be prerequisite to the marketing of the test, it does not mean it is sufficient to accurately identify personal predictors in a particular individual, he warned.
One exception to this is genetic variants for 'age-related macular degeneration' (AMD), the most common form of visual impairment in the developed world. 50 per cent of the heritability of AMD may be explained by known genetic markers, meaning that genetic tests with very good predictive power are possible. But the drawback is there is currently no preventative treatment for those identified as being at high risk, highlighting that genetic tests are only useful if an appropriate intervention exists.
Moving on to the issue of regulation, Professor Melzer highlighted the deficiencies of European law in this area, which tends to take a reactive approach as opposed to the proactive approach taken by the United States. In Europe, for example, there is no requirement for pre-market approval and companies offering genetic tests must only comply with self-certification requirements under the In Vitro Diagnostic Directive (Directive 98/79/EC). Crucially, in Europe, the clinical evidence to support genetic tests is only a requirement if the test asserts medical qualities. Furthermore, secrecy in the presentation of evidence for approval hampers real scrutiny of the data. We should arguably be attempting to construct a permissive yet protective regulatory environment to ensure the fluid development of genetic testing and to engender public confidence, rather than slapping down restrictions in response to public concerns, he said.
Looking forward, Professor John Burn, Head of the Institute of Human Genetics at Newcastle University, discussed individuals taking control of their own healthcare management, but predicted this will emerge with the support of healthcare professionals as part of targeted treatment programmes, rather than through consumers dealing with genetics companies such as '23andMe' directly.
He predicted that 'personalised medicine' would be the first form of genetic medicine with real clinical benefits. For example, research has identified approximately 61 mutations in the BRCA1 gene among women in Norway. A new breast cancer drug, known as PARP inhibitors, is soon to be available, but it doesn't work among women who have one of these 61 mutations and is extremely expensive to produce. With the availability of cheap genetic testing, this drug could be targeted to those women who will benefit most, saving money and potentially lives.
Professor Burns said this area of science is in flux to some extent, with deregulated countries like the US marketing tests that don't necessarily work and regulated countries, like the UK, unable to afford the scientifically robust tests that are available. While this situation is currently hampering the translation of personalised healthcare into the clinic, he predicted that clinical applications in this field would increase markedly over the next five years.
The discussion that followed covered the idea that science is in flux in this area, but that some 'horizon scanning' is necessary to ensure that it moves forward in the right way. There is little doubt that genes will have a crucial impact on the delivery of future healthcare, but this impact must be in the right context. On the whole, the seminar emphasised that realism must prevail - only a tiny number of tests are currently being sold to the public and companies offering genetic tests are struggling financially - but with the development of scientific accuracy, the utility of tests will improve and this may catalyse the market. The regulatory environment needs to be prepared for this by being in a position to protect the interests of the individual consumer whilst not imposing onerous regulatory hurdles that may push the market outside of Europe, and hence, arguably towards more lax regulatory regimes.