Must we give our dreadful receptionist a reference?
In the third article in our series on genetics, Professor Steve Humphries looks at the current state and future prospects of testing in this disease area
on genetics, Professor Steve Humphries looks at the current state and future prospects of testing in this disease area
In recent years our understanding of genetics and 'molecular medicine' has advanced dramatically. Newspapers are filled with stories of great discoveries: the gene for breast cancer, a gene for schizophrenia, etc, and last year the announcement of the entire human genome sequence.
This 'treasure trove' of information will take many years to understand and exploit to benefit human health. The question is how soon can we expect this genetic explosion to be of benefit for identifying high-risk subjects and assessing risk in the clinic, in the field of coronary heart disease?
The problem is that CHD is 'multifactorial', meaning that in the general population there are many different genes involved and they interact with many different environmental factors. Any individual will only have inherited some of these 'high risk' genetic variants and will only adopt certain 'high risk' environments.
The environmental factors that influence the development of CHD are well-known and include diet and smoking and the presence of other disorders such as obesity, diabetes and hypertension (where both environmental and genetic factors are also involved).
Nonetheless, premature CHD clearly does have a strong genetic component and several studies have demonstrated that a family history of CHD is an independent predictive risk factor for the development of this disorder.
So the use of genetic tools, together with conventional clinical and biochemical assessment, may prove very useful indeed.
Familial hypercholesterolaemia a single gene defect
There are some types of CHD that are caused by a single gene defect which has been identified, and in these cases a genetic test would be extremely useful for testing in families and identifying at-risk relatives. Currently the best example of this is familial hypercholesterolaemia (FH). FH is inherited as an 'autosomal dominant' disorder, in which a mutation is passed from a parent to on average half the children, which leads to elevated plasma cholesterol levels. The overall standardised mortality ratio for these patients is at least nine-fold higher than normal, but some individuals develop CHD under 35 years of age, while others (notably women) may remain symptom-free until over 70.
Happily, the use of lipid-lowering drugs such as statins reduces CHD mortality in FH patients very effectively. The estimated prevalence of FH is 1/500 in the general population, suggesting that in the UK there are at least 110,000 carriers, with a typical GP health centre of 8,000 patients likely to have four or five affected families. Unfortunately, most of these at-risk individuals remain undiagnosed and are thus denied the potential benefit of treatment.
Surely we don't need a DNA test to make the diagnosis? Just measure the cholesterol! Although this is true for many families this approach has its problems.
Plasma lipid levels in those with an FH-causing mutation overlap with those in the general population, and this is especially true for children, where cholesterol levels may also rise as they grow older, leading to an unacceptably high 'false-negative' rate of at least 10-15 per cent.
As shown in figure 1 on page 64, the availability of routine DNA-based tests for mutations causing FH solves this problem and allows effective testing in families where a patient has been identified. The utility of such testing is now being examined with research funding as part of the London IDEAS Genetics Knowledge Parks, with a clinical genetic diagnostic service for FH.
Future use of genetic tools
to advise smokers
Of the many common environmental factors shown in figure 2 and known to be associated with risk of CHD, smoking makes a major contribution.
Smoking is known to roughly double lifetime risk of CHD and increases cardiovascular risk by directly damaging vascular endothelium, leading to increased secretion of adhesion molecules, and by perturbing lipoprotein metabolism and increasing insulin resistance and lipid intolerance.
In addition, smoking-induced lung and endothelial wall damage will lead to an interleukin-6-mediated inflammatory response, causing hepatic up-regulation of fibrinogen and CRP expression and increased risk of thrombosis. Because of dietary habits, smokers also have lower levels of antioxidants such as ascorbate and tocopherol and thus smoking would favour the oxidation of LDL and increase the risk of atherosclerosis.
Although the mechanisms by which smoking increases CHD risk have been explored, the genetic predisposition to (or protection from) smoking is currently unclear. A better understanding of this would be useful for identifying those at elevated risk, for instance, who would show the greatest benefit from smoking cessation, as well as pointing to possible novel therapeutic approaches.
ApoE, smoking interaction
and CHD risk
Of the candidate genes involved in the determination of lipid levels and CHD risk, apoE is the most comprehensively studied. Variation in the apoE gene, coding for the three common isoforms E2, E3, E4, is known to have a strong and consistent influence on plasma lipid levels and on risk of CHD.
Recently we have examined the interaction of smoking with apoE genotype in the UK-based Northwick Park Heart Study of more than 3,000 men, followed prospectively for CHD events for more than six years.
As expected the relative risk of smoking alone on CHD risk (expressed as a hazard ratio) was 1.94 (95%CI 1.25-3.01). As expected, apoE genotype was associated with modest effects on cholesterol and apoB levels, with apoE4 carriers having the highest and apoE2 carriers the lowest levels, irrespective of smoking status.
Compared with all the never smokers, where the hazard ratio was set at 1.00, in men who smoked, those with the genotype apoE3/apoE3, had a hazard ratio of 1.68 (95%CI 1.01-2.83) compared with 1.18 (95%CI 0.46-3.03) for apoE2 carriers and 3.17 (95%CI 1.82-5.51) in apoE4 carriers.
The interaction between smoking status and apoE genotype on CHD risk was significant (p=0.007) and was independent of BMI, blood pressure, lipid levels and markers of inflammation.
As shown in figure 3, the apoE4 men who had quit smoking before the study began showed a lower risk, which did not differ from never smokers, emphasising the benefit of smoking cessation.
There is good evidence that apoE genotype influences lesion formation independently of its effects on fasting plasma lipid levels. The interaction of apoE4 and smoking suggests that smoking exacerbates this.
Future for risk prediction:
the vascular disease clinic
Risk algorithms are already being used at UCL in the vascular disease clinic to assess risk for individual patients, and to give them advice on the relative impact of different lifestyle changes such as weight loss or cessation of smoking, and therapeutic strategies, such as use of statins or ACE inhibitors, on reducing their risk.
Currently the available algorithms include information about a selection of biochemical intermediate phenotypes, plus characteristics such as age,
gender, height and weight, blood pressure, smoking habit, presence of disease etc.
Some also include 'general' genetic information such as family history of premature CHD, and some 'specific' genetic information such as apoE isoform and levels of Lp(a), but no other specific mutation information. For all of these there is a 'weighting factor' or a coefficient, in order to combine the data to estimate the subject's absolute or relative risk of developing CHD over the next five or 10 years or more.
The importance of this 'risk profile' approach has been emphasised by the latest European guidelines for lipid-lowering therapy, which have proposed that only subjects with a relative risk of MI of >20 per cent over 10 years are eligible for therapy.
In the future, if such guidelines are generally applied, the role of specific genetic risk information in assessing relative risk may well be of major impact.
However, as described above, accurate use of genetic information requires knowledge of both the genetic make-up and the environmental risk profile of the patient. It is easy to imagine how this genetic information, such as for the apoE gene and smoking status, once validated by replication, can be combined into these computer-based algorithms.
Other information can also be included as further new risk factors emerge such as plasma levels of homocysteine or C-reactive protein and as their genetic determinants are identified for example the gene coding for the enzyme methyltetrahydrofolate reductase (MTHFR) or the gene for the key inflammatory cytokine interleukin-6.
Currently the cost of the reagents for a DNA test (at £1.50) is high relative to a cholesterol measure or a fibrinogen measure. The usual source of DNA is from the white blood cells in a whole blood sample in EDTA, but purification of DNA from the massive amounts of haemoglobin and other proteins is time-consuming and labour intensive, and this adds additional cost.
Buccal cells obtained from mouthwash are easier to handle and can also survive postal transport. However, over the next few years, cost will come down significantly and throughput will increase dramatically, to a point where testing a DNA sample for 20-100 different mutations will be in the range of £10-£20.
So in the future a patient could be booked for a clinic visit in two-three weeks and sent a mouthwash tube to be returned to the laboratory immediately, so that genetic information would be available for discussion.
Steve Humphries is British Heart Foundation professor of cardiovascular genetics at Royal Free and University College Medical School, UCL, and is the CEO
of the London IDEAS Genetics Knowledge Park