Selective Testing

Does Big Pharma stand to gain the most from new genetic tests for drug therapy?

Illustration by Neil Doshi

It’s a mild June day in New York, and Mark Milano, a fit fifty-two-year-old HIV health educator with the AIDS Community Research Initiative of America, is dressed for comfort in running shoes, grey chinos, and a striped short-sleeved shirt. Sitting in his standard-issue cubicle, which is sparingly decorated with a colourful abstract painting, he tells me why he doesnt take his good health for granted. Milano became HIV positive in the early 1980s, when infection was equivalent to a death sentence. Against the odds, he remained healthy enough to avoid antiretroviral drug therapy for twenty-five years. Then, last fall, he was diagnosed with anal cancer, and his oncologist insisted he drive down the amount of the virus in his system before beginning chemotherapy.

Milanos HIV specialist determined that a combination drug called Epzicom which contains lamivudine and abacavir was indicated, but she was reluctant to prescribe it right away. Abacavir is known to cause life-threatening reactions ranging from fever to respiratory distress in some 5 to 8 percent of patients. There is, however, a way to mitigate the risk: a test for abacavir hypersensitivity. Milano, whod recently had an adverse reaction to another drug, was already familiar with the technology, and says he would have refused Epzicom if he hadnt been offered the test. His doctor took a blood sample and sent it off to a lab equipped to screen for HLA-B*5701, an allele one of two alternate forms of a gene on a chromosome closely correlated with abacavir hypersensitivity. In about a week, Milano would find out whether or not he was a candidate for a medication that could save his life.

Welcome to the frontier of pharmocogenetics, a field that can be traced back to nineteenth century English physician Archibald Garrod, who first postulated that many previously mysterious diseases were caused by genetic variations. The term pharmocogenetics wasnt coined, however, until the 1950s, when the German geneticist Friedrich Vogel defined it as the study of genetics role in drug response. The field exploded with the decoding of the human genome in 2000, and since then rapid advances in genomic science and computer technology have led to the development of tests for everything from antidepressants to anticancer drugs.

Adverse reactions to medications like these result in a significant number of emergency room visits in North America and abroad every year, and pharmacogenetic testing holds the tantalizing promise of reducing some of the harm doctors unwittingly cause when they prescribe drug therapy. But the potential usefulness of pharmocogenetics extends beyond safety. Most drugs on the market are, by some accounts, only effective for 50 percent of those who take them. Researchers envision a time when youll be able to walk into your doctors office and have him or her conduct a simple test to determine whether this or that drug will work best for you, given the presence (or absence) of specific genetic variations the age of designer drugs. Combine that with genetic testing that reveals predispositions toward particular diseases, and you have a future in which medical science could circumvent potential problems by intervening even before symptoms appear. But were not there yet.

Currently, all of the major pharmaceutical companies including UK-based GlaxoSmithKline and the American heavyweight Pfizer are either acquiring or entering into partnerships with firms that develop pharmacogenetic tests, or building up their own in-house genetic testing divisions. The United States Food and Drug Administration has approved numerous tests, some of which indicate sensitivity to a particular medication (e.g., the HLA-B*5701 test, developed by GlaxoSmithKline and the Laboratory Corporation of America), while others probe for genes that may regulate the metabolism of many different medicines (such devices as the Roche Ampli-Chip, or Third Wave Technologies Invader Molecular Assay). The regulatory body has even stipulated that the labels of some twenty medications contain information on pharmacogenetic testing.

Despite the availability of such tests in the US, however, most prospective drug takers are unlikely to encounter them at this point. They can cost upwards of $500, insurers are reluctant to cover them, and many doctors are either unaware of the new technology or skeptical of its value. Milanos cancer specialist, for example, wasnt interested in ordering the pharmacogenetic test for the chemotherapy drug fluorouracil and referred him back to his HIV doctor.

Another problem is that certain tests are difficult to interpret. The test for warfarin, an anticoagulant considered the poster child for pharmocogenetics because it is both widely prescribed (more than 200,000 Canadians currently take it) and potentially lethal at the wrong dosage, identifies genotypes associated with different metabolic profiles; that is, the presence of one or another allele suggests how quickly the patient will metabolize the drug. Doctors must then undertake the tricky business of translating this information into dosing recommendations.

However, accessibility is only a matter of time, and all signs point toward rapid growth of the industry. The market research firm Kalorama Information projects that sales of pharmacogenetic tests in the US will triple, to nearly $35 billion, by 2013. Canada is slightly behind the curve: Health Canada has yet to approve any of the pharmacogenetic tests currently on the market. The government wont comment on the issue, but according to Don Husereau of the government-funded Canadian Agency for Drugs and Technologies in Health, the guardians of our public health care system can be quite reactionary in their attitudes toward novel treatments. For now, Canadian doctors cannot order such tests for their own patients.

Meanwhile, Canadian companies, such as Quebecs Warnex Medical Laboratories, can and do provide pharmacogenetic testing services to American drug companies. And Genome Canada, a not-for-profit organization with a mandate from the federal government to encourage genomics research, is pumping tens of millions of dollars 83 percent of which comes from the public sector into pharmacogenetic projects across the country. With so much public and even more private money swirling around, Canada is actually home to some of the most cutting-edge work in the field worldwide.

Located in one of several nondescript office blocks in an industrial park in Saint Laurent, Quebec, Genizon BioSciences shows few outward signs of being a major player in the pharmacogenetic revolution. Yet it is one of a handful of companies involved in producing the richly detailed genetic maps required to make new, more effective pharmacogenetic tests. These maps not only identify particular genes; they illustrate how those genes contribute to the complex biological pathways that can ultimately lead to schizophrenia, diabetes, or adverse drug reactions. The company has already attracted $130 million in venture capital (it was denied funding three times by Génome Québec, the local provincial affiliate of Genome Canada) and has licensed its GeneMaps to Pfizer and Genentech.

Genizon bases its maps on the DNA of the Quebec founder population, a community of six million French-Canadians descended from the 2,600 pioneers who settled the province. The QFP is one of the largest genetically homogeneous founder populations in the world, and includes anyone who can lay claim to a complete set of Québécois grandparents. Genizon collects blood samples from members across Quebec and processes them in its sprawling lab, storing the extracted DNA in banks of freezers locked away in a giant vault. The genetic material is eventually scanned by a series of eight compact genotypers labelled with the French names of Snow White and the Seven Dwarfs (Timide, Atchoum, etc.). Using proprietary software algorithms, Genizons scientists are able to identify population-wide genetic variations that can trigger the biochemical pathways that eventually lead to illness and lethal drug allergies. With fewer variations per gene than one would find in a larger, more heterogeneous group, the signal-to-noise ratio is higher, and the most significant genes stand out.

But the narrow ethnic composition of the QFP, so useful in singling out troublesome genes, also points toward a major concern among critics of pharmocogenetics: namely, that tests will target only the most commercially attractive gene pools, like Caucasians and Asians, further widening the gap between those who have access to the best medical care and those who do not. The test for abacavir hypersensitivity, for example, was born out of a study that correlated the HLA-B*7501 allele with abacavir hypersensitivity in a population consisting predominantly of Caucasians of northern European descent. In a later study of Hispanics with abacavir hypersensitivity, only one-fifth of participants tested positive. Obviously, that population isnt going to benefit equally from HLA-B*7501 screening. And there are other pharmacogenetic tests on the market that suffer from a similar ethnically blinkered focus.

Genizon CEO John Hooper, a lanky Englishman with a background in drug development and analysis, believes fears about the drug markets potential fragmentation along racial lines are misplaced. Only the prevalence of the occasional gene can differ substantially between populations, he says, but the science is not yet developed enough to make predictions about this. Mark Milano acknowledges that ethnic bias is a problem, and that it needs to be resolved through more tests, [developed through research] on more diverse populations. Whether Big Pharma will expend the required resources remains to be seen. Even Hooper anticipates that the pharmaceutical industry will concentrate research efforts on the worlds most lucrative drug markets, spreading the net just wide enough to catch the biggest fish, e.g., all extant Caucasian populations.

A more fundamental criticism of pharmacogenetic testing has to do with the growing authority of genomic models of illness, which threatens to obscure other, equally important, risk factors. In the case of adverse drug reactions, its true that some demonstrate an overwhelming genetic component: if you have the allele and you take the drug, youll almost certainly get sick. Abacavir hypersensitivity exemplifies this, and the associated pharmacogenetic test identifies individuals at risk with incredible accuracy, according to Allen Roses, a former VP at GlaxoSmithKline. For the first time in the history of medicine, he declares, you can predict an allergy to a drug before the patient has to take it.

Roses statement reflects the tremendous sense of possibility that surrounds pharmacogenetic testing, but the fact is that genes play a limited role in many adverse reactions. Even among individuals who possess the two genes so far tied to warfarin sensitivity (and, yes, they are especially common in white people), the genes themselves account for only 35 to 40 percent of individual risk. Variables such as age, weight, gender, disease, and interactions with other medications can also factor in. No evidence exists to date that genetic tests will actually decrease the rate of fatal warfarin reactions, and there wont be any until the tests are in widespread use.

Alan Cassels, a drug policy researcher at the University of Victoria and coauthor of Selling Sickness: How Drug Companies Are Turning Us All into Patients, calls pharmocogenetics a largely evidence-free zone, and, in the best-case scenario, a high-tech hunting expedition that will benefit a relatively small proportion of health care consumers. He argues that whatever government money will ultimately be put toward providing these new tests would be better spent promoting basic public health initiatives, like encouraging people to eat better, exercise more, and quit smoking. As pharmacogenetic tests proliferate and more evidence accrues regarding their benefits, Canadian regulators will be forced to confront precisely these concerns as they decide whether or not to include the tests in provincial drug formularies.

Still, there is probably no stopping the advance of pharmocogenetics on either side of the border. The economics of the drug business leave Big Pharma little choice. It can now cost upwards of $1 billion to bring a new drug to market, and drug companies simply cant afford to throw money away on products that fail due to unforeseen side effects. (Think of the embarrassing and costly recall of Vioxx, the popular anti-inflammatory eventually linked to an increased risk of heart disease and stroke.) The industry is at a tipping point, and without genomics, Hooper says, there would be fewer new drugs.

Beyond individual patient testing, pharmacogenetic techniques can be used to develop commercially successful drugs, weeding out dangerous medications early in the development process. Early kills are a great thing in the drug industry, says Keith Johnson, a director at Pfizer. Over 95 percent of drugs in development fail, 43 percent in Phase III trials, he adds, referring to the final stage of drug testing that determines whether a compound is sufficiently safe and effective to merit regulatory approval. By then, youve invested hundreds of millions of dollars.

Pharmocogenetic testing can also be used to find new uses for existing drugs. Five years ago, GlaxoSmithKline began testing the diabetes drug rosiglitazone for possible use as a treatment for Alzheimers disease. Initial trials indicated no significant benefit to the study population, a result that didnt surprise Roses and his team, who had suspected that only patients with a certain genetic makeup would respond. Through their research, they were able to pinpoint the allele involved, and develop a pharmacogenetic test to sort responders from non-responders the first genuine efficacy test. If the Phase III trials currently under way yield positive results, Roses says, from the companys point of view, they will have pharmacogenetic proof, likely accepted by the FDA, that their drug works for 50 percent of subjects.

Roses himself manifests the dual, sometimes conflicting, values that determine virtually all progress in medical technology: health and finance. Appointed director of the new Deane Drug Discovery Institute at Duke University School of Medicine after leaving GSK, he is a medical doctor with a genuine interest in saving lives. But he is also a business-savvy entrepreneur whose consultancy, Cabernet Pharmaceuticals, provides pharmacogenetic advice to pharmaceutical and biotechnical companies. He understands that the pharmaceutical industry will only pursue this shiny new tool if it leads to greater profits, and in whatever way will lead to the greatest profits. That reality is what leads critics to question whether pharmocogenetics will be more of a boon to drug companies than to patients.

Mark Milano and others who have undergone pharmacogenetic testing arent as concerned about whether Big Pharma has ulterior motives. Results of the test for abacavir hypersensitivity showed that Milano does not have the HLA-B*5701 allele, and he started on Epzicom immediately. Everything went according to plan: he tolerated the treatment well, his viral load has been undetectable, and his cancer is now under control. A staunch advocate of genetic testing as a means of making drug therapy safer, Milano calls the abacavir hypersensitivity test, in particular, the standard of care for HIV patients who need the drug. If youre not giving this test, he says, its malpractice.

Alexander Gelfand