In time, patients' genetic profiles may be checked to ensure new treatments go to those who will benefit most, writes Jenny Bryan

New genetic tests under development at all major pharmaceutical companies cou ld make it much easier to decide which patients should receive expensive new drugs - such as Rebif for multiple sclerosis or Rilutek for motor neurone disease. Out would go prescribing by postcode, in would come treatment by gene code - or pharmacogenetics to give it its scientific title.

Prescribers will soon get a glimpse of the potential of pharmacogenetics with a new drug for breast cancer called Herceptin, recently licensed in the US and under consideration in the EU. It is specially tailored to aggressive tumours which overexpress a protein called HER2, and women will be chosen for treatment according to their HER2 status. About 25-30 per cent of women with metastatic breast cancer have this type of tumour and they appear most likely to respond to Herceptin - a monoclonal antibody which binds to HER2 to inhibit growth of tumour cells.

This early move into the world of pharmacogenetics focuses on the gene product - the HER2 protein - rather than the faulty gene itself. But future developments could see patients having their genetic profile checked before being prescribed certain types of drugs.

'The pharmaceutical industry used to take the view that, like tights, one drug fits all. But they have now realised that it may be better to target smaller numbers of people with drugs that are highly effective and do not cause side-effects, ' explains Dr Neil Barnes, consultant physician at the London Chest Hospital.

His own specialist field of asthma is a strong candidate for pharmacogenetics. Glaxo Wellcome is enrolling 700 families with at least one person with asthma in a genetics network. Blood samples for DNA analysis are being collected from both asthmatic and non-asthmatic family members to try to link rogue genetic sequences with breathing problems.

'This will be by far the biggest coordinated sample collection and gene screening effort ever in asthma, ' says Allen Roses, director of genetics at Glaxo Wellcome. 'We hope that within four years the network will have identified regions of the genome that will eventually tell us more about what causes asthma, about inherited predispositions to developing asthma and how we at Glaxo Wellcome can find new therapies to treat asthma.'

Asthma specialists are already aware that some patients respond better to the new leukotriene antagonists than others, and research suggests that tiny changes in the genetic switch for the enzyme which controls leukotriene production may hold the key. Asthma patients also vary in their responses to inhaled corticosteroids and some become resistant to treatment. If a genetic basis can be identified, poor responders could be spared both a lack of efficacy and potential sideeffects from corticosteroids.

In hypertension, the story is the same. Professor Peter Sever from St Mary's Hospital, London, explains that, while one in five people in the UK has high blood pressure, each of the six main groups of antihypertensive agents work in only about half of those who try them. With so many causes of hyper tension , it is hard to know wh ich mechanism is faulty in any individual patient.

'To have a genetic or biochemical guide to which drug is most likely to work would be enormously helpful, ' says Professor Sever.

Significant progress has already been made on genetic profiling for patients with Alzheimer's disease who are most likely to respond to the new cholinesterase inhibitors, such as Aricept.

In the last few years, research has shown that people who are prone to Alzheimer's disease in their seventies have a pattern of inheritance of the susceptibility gene, apolipoprotein E (APOE), which is different from that of people who develop the disease in extreme old age.

This research has also shown that, unfortunately, patients prone to earlier onset Alzheimer's disease may also be least likely to respond to the new therapies. But, as Dr Simon Lovestone, consultant in old age psychiatry at the Institute of Psychiatry, London, explains, it is not an 'all or nothing' phenomenon.

'The research is very exciting but it also introduces a dilemma. APOE doesn't determine whether or not a patient will respond, it only says whether it is more or less likely. So who will decide where you put the threshold? Some people will be denied treatment not because they definitely won't respond, just because they are less likely to, ' he explains.

Telling people their APOE status can affect the whole family, adds Dr Lovestone. Children whose parents have the worst APOE pattern will inherit the same poor prognosis and will have to live with the knowledge that they stand a strong chance of succumbing to the disease at a relatively early age.

Another important aspect of improving the response rate to treatment for common conditions such as asthma, hypertension and Alzheimer's disease centres on the different ways in which individuals break down (metabolise) drugs in the body. This can affect both the effectiveness of a drug and the severity of its side-effects.

For example, a family of liver enzymes called the cytochrome P450 enzymes metabolise the majority of currently available drugs. Four of these enzymes show considerable genetically determined variability (polymorphism), and this can affect how well a drug is broken down.

Drugs may reach dangerous levels and cause serious side-effects in people who are poor metabolisers, while rapid metabolisers may need a higher than normal dose to gain any effect.

This kind of variability is such bad news for drug companies that they tend to discard experimental drugs that are metabolised by P450 enzymes with the greatest genetic variation.

But if checking a patient's P450 or other enzyme status in the future became as routine as testing for antibiotic resistance today, this could improve the choice of treatment for many diseases.

Gene code prescribing will probably be the short term pay-off from pharmacogenetics, but in the longer term it could yield entirely new approaches to treatment. As Professor Martin Bobrow, head of clinical genetics at Addenbrooke's Hospital, Cambridge, says: 'In five years, we may be able to predict who will do well on a particular drug according to their genetic background.

'But in the longer term, finding out why certain genes are important in par t icu lar d iseases will g ive us a handle for developing new treatments.

A single gene could be a very small player but still give us a critical insight into a disease so that a new class of drugs can be developed.'

The good news for health authorities is that the massive investment which pharmaceutical companies are making in pharmacogenetics should not be reflected in a big rise in the cost of the new treatments.

Dr Andrew Rut, director of human genetics at SmithKline Beecham, explains that while the cost of developing any new drug is prohibitive, it shouldn't be any greater just because it has taken the pharmacogenetic route.

'We are simply swapping new tools for old tools, and the cost of genotyping is tr iv ia l compared to the overall cost of clinical trials. We have to invest in the whole process, whether we use genetics or not, and we still have to identify the right targets and show that they are safe and effective, ' he says.

He is putting his money - or at least SmithKline Beecham's - on new drugs for common physical conditions, such as osteoporosis and diabetes, to be first out of the pharmacogenetics pipeline. Disorders with symptoms which are harder to quantify, such as mood disturbances, will take longer to succumb to the new genetics, he believes.

Whether it's people with asthma or Alzheimer's, depression or diabetes who are first to benefit from tailormade treatment, there is little doubt of the importance of pharmacogenetics.

As Professor Bobrow points out: 'It's inconceivable that there won't be advances. We just don't know how quickly they will come or which disease will benefit first.'

Pooling ideas: the Iceland connection No company is yet admitting to testing new drugs on patients according to their genetic profiles (genotypes), but they are collecting DNA samples from people taking part in clinical trials. These will be used to see if people who respond particularly well (or badly) to the treatment on trial have any genetic similarities. Samples will also be used to find out more about the genetic basis of common disorders and hopefully find targets for novel drugs.

Dr Andrew Rut, director of human genetics at SmithKline Beecham, explains that the availability of huge DNA databases, large accessible populations and high-throughput testing systems is pushing pharmacogenetics to the forefront of research and development of new drugs.

'Genetic analysis is becoming routine practice across a number of biotechnology and pharmaceutical activities because it is cheap, quick and easy to use, ' he explains. The ultimate aim, he adds, is to get the right drug to the right patient at the right time.

In addition to taking DNA samples from patients in clinical trials, companies are also buying into DNA databases or, in the case of Glaxo Wellcome, setting up its own clinical networks. One of the most unusual sources of DNA available to pharmaceutical companies is the population of Iceland.

Thanks to its relative isolation over many generations, Iceland's gene pool may be more homogeneous and easier to study than that of most populations.

Roche has a five-year deal with biotech company deCODE, to analyse DNA from large sections of the population in an effort to find new treatments for 12 major diseases.

Getting to grips with the terminology Pharmacogenetics The study of how genetic variability alters people's responses to drug therapy. It uses the variations in human genetic make-up to predict how individuals will respond to and metabolise drugs. As with the new breast cancer drug Herceptin, it is also being used to target new drugs at the gene products - proteins, enzymes and other components of biochemical pathways - implicated in disease processes.

Pharmacogenomics Sometimes used interchangeably with pharmacogenetics, this is a broader term for the use of genetic tools in drug development. It is concerned more with the methodology of understanding responses to treatment.

Genomics The collection, analysis, and interpretation of sequences of DNA.

Human genome The complete map of human DNA, believed to contain 50,000-100,000 genes.

Bioinformatics Information systems for the analysis of biological, especially genomic, data. It helps to link genes with the proteins for which they code, thus providing new targets for drugs.

Genetic linkage Research which sifts through the human genome to find a gene for a particular disease. Such studies have had some success with rare disorders linked to a single gene, such as cystic fibrosis, but are less useful for the majority of human disorders, such as heart disease and asthma, which are likely to be caused by interactions between several genes.

Genetic association Much more popular now than genetic linkage, these studies form the basis of much pharmacogenetics. Instead of looking for an unknown gene in people with a particular disease, it searches the genes of large populations to see if a particular, known sequence of DNA is any more common in one disease than another.

Genechips Genetic probes consisting of glass and silicon wafers containing known DNA sequences. When a sample of unidentified DNA is added to a wafer, matching sequences bind together so that a genetic profile can be constructed.

Genotyping Testing an individual's pattern of genes, or genetic profile.

Polymorphisms Multiple variations in a particular gene.