Medicine Gets Personal
By Marc Wortman
THE OLD WAY—One size of drug fits all.
THE NEW WAY—Each medicine will be adapted to the genetic makeup of individuals or subgroups of the population.
THE GOAL—Better treatments, better control of side effects.
Illustrations by Stuart Bradford
In early 1999, Karen Cassidy, a dental hygienist and former high school athletics official, saw ads touting the new LYMErix vaccine for preventing Lyme disease, a bacterial infection passed to people by the bloodsucking bite of the deer tick. "Deer are right at my back door," says Cassidy, who hoped the vaccine would let her rake the yard of her suburban Philadelphia home without fear. But shortly after she completed two of a course of three inoculations in May 1999, Cassidy began experiencing burning pain in her back, numbness in her arms and aches and swelling in an ankle. "Just the thought of walking across the yard hurt me," says Cassidy.
By the following December, the pain was so intense she was considering reconstructive surgery for her ankle and had joined over 100 others in a class action lawsuit charging LYMErix's developer, the pharmaceutical giant SmithKline Beecham, with, among other things, ignoring warnings that a genetically identifiable class of recipients—as many as 30 percent of patients—may develop an incurable autoimmune disorder called treatment-resistant Lyme arthritis from the vaccine.
While Cassidy's diagnosis is in dispute, her legal case is an undeniable indicator of how dramatically medicine is being transformed by a new science known as pharmacogenomics.
Now that the Human Genome Project is largely complete, academics and companies are shifting their emphasis to the study of how genes vary among people. The Human Genome Project involves decoding the sequence of one complete set of human genes, a kind of genetic Everyman. But that template won't describe everyone, since DNA varies slightly from person to person. And those minute variations in the DNA, scientists believe, may determine which patients will benefit most from particular drugs—as well as which subgroups may be harmed by them.
That's the idea of pharmacogenomics, and it is taking the drug industry by storm. The term, coined only four years ago, now encompasses the aspirations of a large number of enterprising biotechnology companies and academic laboratories. This is an era of transition in medicine: from the time of "one size fits all" drugs created for and marketed to all patients, to the emerging epoch of personalized medicines, in which drugs are geared to the specific genetic makeup of groups or individuals. This transition is causing growing pains for some companies. But the ultimate payoff, a decade or more away, should be enormously beneficial to patients, enabling doctors to think about stopping tumors before they begin and heart attacks before they happen.
"This is not a fad," says Gualberto Ruano, CEO of Genaissance Pharmaceuticals in New Haven, Conn., a player in the transition. "It's a major tidal wave changing the entire pattern of health care." And Alan Roses, who leads genetics research at drug giant Glaxo Wellcome in Research Triangle Park, N.C., agrees wholeheartedly that pharmacogenomics "is a disruptive technology, not a technology that sustains what organizations are used to doing. It's going to be a part of everybody's business, and that's what most people don't seem to understand."
The DNA Differential
What will the fully realized pharmacogenomics revolution in medicine bring us? Numerous questions involving health insurance and other nagging ethical, political and social issues must be confronted over the next decade (see sidebar, "Yourgenome.com"). But eventually knowledge of the relevant variations in your own genome—perhaps via a readout in your doctor's office using a biochip—will provide predictions of your potential health problems before they happen. To start with, there will be indications of whether a particular medication might have toxic side effects for you, based on your DNA. Then there will be prescriptions specifying which of the many pharmaceutical and other health care options will be optimal for you. Finally, physicians will be armed with genetically targeted medications, advice for behavior change and other elements of what Nicholas Dracopoli, executive director of the Pharmacogenomics and Human Genetics Group at Bristol-Myers Squibb, calls "disease management packages." This will allow doctors to intervene well in advance of symptoms, so tumors won't form, arteries won't clog, bones won't grow brittle and aging brain cells won't die.
Before we reach this payoff, several key technological challenges must be overcome (see sidebar, "Breakthroughs Ahead"). The first is to identify as many of the small genetic variations between individuals as possible. These variations, known as single nucleotide polymorphisms (SNPs), are simple chemical substitutions of one letter of the DNA alphabet for another in a person's genes; though "minute," these substitutions can make a world of difference in how a person responds to a given medication. Yet not all SNPs actually have much medical significance, and sorting out the important from the unimportant is another crucial technological challenge.
Sidestepping Side Effects
Even before these technological problems are conclusively solved, the first steps toward personalized medicine are already being taken. Indeed, two waves of innovation are likely before the full array of "disease management packages" is in place. The first wave includes genetic tests to predict which patients will suffer "adverse reactions," the sometimes fatal side effects caused by drugs.
One test for such differences is already saving lives: a genetic screen developed by St. Jude Children's Research Hospital in Memphis, Tenn., for mutations in the thiopurine S-methyltransferase (TPMT) gene. TPMT regulates how the liver breaks down certain drugs, including 6-mercaptopurine, a chemical that's a lifesaver for victims of acute lymphoblastic leukemia, a deadly form of cancer that afflicts approximately 2,400 American children and adolescents each year, according to the National Cancer Institute. Between 10 and 15 percent of children metabolize the drug either too quickly or too slowly. The former don't gain a benefit from a standard dose, while the latter can accumulate lethal levels of the drug.
Although the TPMT test is now being used to choose the right drug dosage, few treatments have such tests. For instance, there's no routine testing before giving chemotherapy for most other cancers. Patricia Rodriguez, a hematology oncologist at Arlington-Fairfax Hematology-Oncology Group in Arlington, Va., says, "You know the drugs are going to be toxic, but sometimes there are idiosyncratic side effects, and you can't predict who is going to have them. When it happens, it's terrible."
Although developing tests that predict side effects seems like a commonsense proposition, in fact it runs counter to almost every conventional notion of how drugs are developed and marketed. The pharmaceutical industry wants single drugs that work safely for everybody (see "Pharma's Blockbuster Habit," TR July/ August 2000). Of course, most drugs don't. And now that genetics is providing some insights as to why, the "one size fits all" attitude is becoming more and more out of tune with reality—sometimes dangerously so.
Despite safety study and clinical trial protocols that typically involve many thousands of patients and five or more years of testing, drugs causing serious, unpredicted side effects do make it through to FDA approval and the marketplace. According to a 1998 study published in Journal of the American Medical Association, an estimated 2.2 million patients had adverse reactions to drugs in 1994, and 106,000 patients died. Adverse drug reactions are by some estimates the fourth leading cause of death in America.
According to Dale Pfost, president and CEO of the Princeton, N.J.-based Orchid BioSciences, a company developing SNP technologies, the pharmaceuticals industry has an unpleasant secret: "More money is spent on adverse drug response than on drug development."
Safety First
The pharmacogenomics revolution has a chance to change this picture. Within five years, genetic tests identifying individuals at risk for an adverse reaction will very likely be a more routine part of how new drugs are developed; after that, such tests may actually be co-marketed with new drugs.
Taking the lead are companies like Glaxo Wellcome, which will soon merge with SmithKline Beecham. Glaxo signaled its commitment to the area by hiring pharmacogenomics pioneer Roses in 1997 to lead its gene research efforts. As director of the Center for Human Genetics at Duke University, Roses had a hand in tracking down the gene that causes amyotrophic lateral sclerosis (Lou Gehrig's disease) and led the team that identified apolipoprotein E (apoE), a major genetic factor in Alzheimer's disease.
Roses has set up a widespread network of medical-center-based programs to speed the search for disease genes. One of the first successes was creating a map of the SNPs present in the apoE gene; this development has opened up an entire new frontier in searching out genetically targeted treatments for Alzheimer's, which has proved notoriously hard to treat.
While developing a more effective Alzheimer's drug could take a decade, Roses expects in the next two to five years to submit for FDA approval a pharmacogenetic test for the safety of Glaxo's anti-HIV drug Ziagen. Around 5 percent of AIDS patients have a predisposition to develop dangerous and potentially fatal hypersensitivity reactions to this medication, a rate found with other AIDS drugs as well. If the FDA approves the test, then the right 5 percent of the population will know not to take Ziagen.
That test, Roses says, will be the "proof of principle" that the genetic revolution is both smart business and good for people. He contends it will "have a revolutionary effect on the pharmaceutical industry," dramatically changing the way the industry operates. "It will be inescapable. You cannot deny evidence-based safety testing to people taking a drug."
Many of Roses' peers worry that pharmacogenomics may fragment their markets in ways that could be lethal to the corporate bottom line. Roses believes otherwise, arguing that the new Ziagen test will drive far more people to Glaxo's drug. Armed with pharmacogenetic tests for safety, he says, Glaxo's products will have "...a huge competitive advantage....Would you pay more for a pill that is a thousand times safer?" And once that first test for Ziagen demonstrates its market, "the dominoes will start to fall," he says. "Consumers will want it. We're a regulated industry. Regulators will demand it."
Targeted Medicine
If the first wave of the pharmacogenomics revolution is already under way—using genetic data to avoid adverse reactions—a second wave is ready to break right behind it: using genetic information to find out which medicines will work optimally for which patients. A Sweden-based firm, Gemini Genomics AB, offers an assay approved for use in Sweden that identifies a subgroup (30 percent) of hypertensive patients who should be treated with angiotensin-converting enzyme (ACE) inhibitors. ACE inhibitors are the most widely used of several different types of hypertension treatments. When they're effective, they save lives. They don't work for many people, though, and cost a great deal, so knowing who would benefit would be a significant advantage to doctors.
Other pharmacogenetic tests for heart disease will soon reach the U.S. market. University of Cincinnati scientist Steve Liggett, a leading pharmacogenomics researcher, has found SNP markers to identify heart patients most likely to experience rapid decline in cardiac function. Once identified, they can be moved up the list for more radical interventions and prioritized for heart transplant. Given that about 4.6 million people in the United States have suffered congestive heart failure according to the American Heart Association, "the ramifications" of such a test, says Liggett, "are enormous."
In her own practice, oncologist Rodriguez is already seeing how the first wave of testing for tumor genes for certain forms of breast cancer and lymphoma has helped her patients. "More targeted therapies are becoming available. That's the direction for the future in which you'll choose more-selective drugs based on genetic material of the tumor cells."
Although the first two waves of pharmacogenomics are already making an impact, it will be a while—a decade or more—before the full payoff comes. One reason is that genes don't provide all the answers for treating disease. Rather, our genes work in concert with behavioral choices, such as cigarette smoking, exercise, diet and environmental exposure, to determine which illnesses we get. Sorting all of that out will be immensely complex. "Drug metabolism is pretty straightforward," says Rick Sheridan, a vice president of PPGx, a pharmacogenomics testing firm in Morrisville, N.C. "These areas are far more subtle. When you get into things like disease predisposition, the frequency of SNP markers may be quite high, but you have to account for environmental triggers that affect almost everything about disease."
To gather the information needed to unravel the puzzle, governments and enterprising companies have begun combing through epidemiologic and public health information collected over the years. For instance, the startup Framingham Genomic Medicine in Framingham, Mass., was recently formed to take advantage of more than 50 years and thousands of subjects' worth of data collected from the studies first begun as the famous Framingham Heart Study—in which smoking was first linked to heart disease and the notion of "risk factors" was developed. "We're looking for gene-environment interaction," says Fred Ledley, chief scientific officer for Framingham Genomic Medicine. "Good genetic information has to be linked to good clinical data."
Karen Cassidy's case is a good example of how complicated the interactions between our genes, our environment and the drugs we take can be. Was her illness caused by a tick bite, the LYMErix vaccine, some underlying problem in her immune system—or a combination of all three? Today, it is up to the lawyers to sort it all out. But sometime soon clear scientific answers may be possible, not just to resolve health mysteries like Cassidy's, but to prevent them altogether.
When those answers are available, they will do far more than just prevent adverse reactions. They will make it possible to practice medicine in a whole new way. At Genaissance, Ruano's scientific staff is studying patients taking one of the many cholesterol-lowering drugs currently on the market, such as Lipitor, a popular prescription drug that rakes in nearly $4 billion a year in sales. Clinical data showing whose cholesterol drops and whose doesn't, and who has a bad reaction to a drug, is correlated with DNA samples collected from the patients and decoded in the bank of sequencers. The hoped-for outcome: genetic markers that allow the optimal matching of patient and drug.
Armed with the data, a small army of Genaissance software developers are busy writing code for what Ruano hopes will be the "operating system" for the new era of personalized health care—a future in which a doctor seeing a high cholesterol reading, rather than writing a prescription solely on the basis of her accumulated experience, will check your DNA against an online gene database to find the right drug to prescribe.
But isn't that a big change in the doctor's role? Yes, says Herbert Chase, Yale Medical School's deputy dean for education, and the reason is the explosion of medical information. In the future, says Chase, "it is likely that we will know from a drop of blood that a patient has 14 of 19 genes for high blood pressure, and we have 172 drugs that will interact with that. Only a computer will be able to organize this information." Your doctor will become a middleman, Chase suggests, mediating between you, various genomic and information technology systems that will be the backbone of the health care system, and the pharmaceutical treatments that the computer prescribes.
By the time such systems arrive, the current dominant notion that "one size fits all" will likely be a distant memory, having given way to a nuanced, personalized strategy in which health care is focused on finding the right drug for smaller, genetically differentiated segments of the population—even single individuals. For the pharmaceutical industry, it will be a big change. "It's a different mentality," says Genaissance's Ruano. "You need to develop drugs on a smart basis for a targeted market and create a portfolio of drugs that add up to a blockbuster. There will be many more products, and the dynamics of drug development, submission for approval and marketing will have to change." For all who suffer from disease—and sooner or later that's all of us—the changes could be even bigger and more fruitful.
Freelance writer Marc Wortman is a contributing editor to Yale Medicine Magazine, and creator the Yale Children’s Health Newsletter. He lives in New Haven, where he writes about health and medicine, outdoor life, and other subjects. |
