On Target
A new generation of drugs offers customized cures
The two women, on opposite coasts, have led very different lives. Ginger Empey of Bakersfield, Calif., can no longer work. Gloria Caruso is an airline reservation agent in Tampa. Caruso, 63, is married; Empey, 58, divorced. The women have also struggled with two different illnesses. Empey has battled breast cancer that spread to her liver, while Caruso fought lung cancer that spawned tumors in her neck and back.
But the two also share some good fortune, including joy in their grandchildren. Empey was amazed over the Christmas holiday to see that Liam, 7, had shot up to over 5 feet tall. And Matthew, 5, delighted Caruso by appearing as a reindeer in a play. "You really haven't lived until you've seen little kids put on a musical, yelling, `Hey, Mommy' in the middle of a song," she says. The women also share in a medical revolution that has made such moments possible for them, a revolution known as personalized medicine.
Unlike the shotgun strategy of most drugs on the market today, new drugs can be tailored to fit an individual patient with a particular disease. Empey is taking Herceptin, a drug that homes in on tumor cell proteins that flourish only in patients who have her specific subtype of breast cancer. Just last month, a study showed that the medicine kept the disease at bay for an unprecedented 14 months. For Empey, it has already added nearly eight years to her life. Iressa, the drug that has allowed Caruso to hear her reindeer sing, focuses on another protein that has been found in abundance in certain lung cancers.
"This is a revolution in drug use," says Edward Sausville, associate director of developmental therapeutics at the National Cancer Institute. "Now we can use the biology of the disease, the genetics, to see a target in a patient. This is a lot different from the old mind-set." Doctors used to assault an amorphous disease with a blunt instrument, like a chemical selected for its ability to kill lots of cells in a lab dish. That's a far cry from using the individuality of genetics.
Matchmaking. The revolution extends beyond cancer and beyond the development of new drugs. It includes new tests that can better match patients with existing drugs, improving the chance of good results and avoiding toxic side effects. Gene tests can now reveal whether an organ transplant patient is likely to have a bad reaction to a particular transplant drug, forewarning doctors about the risk. A commercially available gene test can identify patients in danger of severe bleeding if they take the anticoagulant warfarin. And tests under development can pick out people who have both good and poor responses to drugs for high blood pressure, which should allow doctors to match patients to the most suitable drugs in advance. Taken together, these discoveries lead Dennis Slamon, who pushed the development of Herceptin while an oncologist at the University of California-Los Angeles, to this pronouncement: "The one-size-fits-all approach doesn't work anymore. Targeted medicine is the future."
But the new one-size-fits-one approach doesn't work perfectly yet, either. Iressa, for instance, helped only 10 percent of patients in one trial, and in the newest, largest set of trials it didn't improve survival, leading scientists to wonder if they are really seeing the target clearly. And several of the diagnostic tests aren't being used in standard practice, for a couple of reasons: First, doctors unfamiliar with genetics don't know about them or don't trust them, and second, insurance companies won't pay for them. They're too new. "Until the issue of reimbursement is settled, these tests are going to have it tough," says Robert Roberts, an expert in the genetics of heart disease and chief of cardiology at Baylor College of Medicine in Houston.
Despite such obstacles, forays into personalized medicine are now possible because of new technologies that are capable of rapidly analyzing the DNA of diseased and normal tissue. Genes come in different flavors, or variants, and those variants can be as small as a change in a single molecule in the massive genetic chain. High-speed computers and DNA filters called microarrays have made it possible to quickly identify these varieties in a gene.
The first step in targeting a treatment is to link a genetic variation to a specific disease. Then the genetic aberration can be matched with a drug that somehow interferes with it. Such is the case with Herceptin and breast cancer. Malignant cells have deranged DNA to begin with. In the case of about 30 percent of breast cancer patients, the derangement takes the form of an overabundance of a gene called Her2. "It's like the Xerox machine inside the cell gets stuck," says Slamon. "It keeps on churning out copies. Once you get over a half-million, you're off to the races."
It's the cell, actually, that begins racing out of control. Those overabundant genes instruct the cell to make huge numbers of receptors on its surface, which in turn snag signaling molecules called growth factors. Growth factors are essential for a cell's normal, healthy functioning. But when a cell grabs a few hundred thousand or more of them, it begins to grow wildly. It divides and divides, rapidly turning into an aggressive tumor that quickly spreads to other parts of the body. Patients with Her2 have a gloomy outlook.
But in the 1990s, Slamon and his colleagues discovered Herceptin, a drug that fit into the aberrant receptor like a hand in a glove. It leaves no room in that glove for circulating growth factors. What makes Herceptin an especially good drug is that it only hits cells covered with this specific receptor, ignoring healthy cells and thus avoiding the shotgun approach of chemotherapy--including the well-known side effects of nausea, hair loss, exhaustion, and cognitive deficits.
Once it was known how the drug works on the cellular level, the next step was to identify patients overwhelmed by the receptors. Fortunately, there is a test that measures the receptor-making Her2 genes. "Without the test, we would have had nothing," Slamon concedes. Herceptin was approved in 1998 for metastatic or recurrent breast cancer in patients with the receptor. At the time, it looked as if the drug shrank tumors or held them at bay for seven months. Today, given in combination with other therapies, that figure is up to 17 months.
Reprieve. Ginger Empey started the drug in 1995, as part of a clinical trial. Her cancer had spread to her liver, chemo had failed, and her oncologist told her to get her affairs in order. "So I fired him!" Empey says. "It felt like getting hit with a sledgehammer. You can't help but hate a person who delivers a message that way." But she'd heard about Herceptin and banged on doors at UCLA until they tested her and found she was Her2 positive. She started the drug, and within months her tumors had shrunk to next to nothing. Today Herceptin still drips in, every three weeks, through a tiny tube implanted in her chest. "This is my 385th week on the drug. But who's counting? Look, people with my kind of cancer aren't supposed to live very long. I've already passed that. I really hope I'm the one who's going to make it out alive."
Not everyone does, of course. The drug seems to help about 40 percent to 50 percent of those who take it. "So clearly, we're not at the end of the road with it," says Mark Pegram, a UCLA oncologist. "We have to find new approaches for the nonresponders."
Those same sentiments are voiced by researchers who work on Iressa, the lung cancer drug taken by Gloria Caruso. "I've seen some remarkable things with patients on this drug," says Alan Sandler, an oncologist at Vanderbilt University Medical Center who worked on some of the early Iressa tests. "One woman came in a wheelchair. She couldn't even get dressed because her body was filled with fluid from the cancer. I would have given her a few weeks. But in a few weeks, on the drug, she was up and around. MRI images confirmed that her lesions shrank." Then there's Caruso, on the drug for 23 months--it's one pill, once a day--and still going strong. "My neck tumor was supposed to swell and cut off my air," she says. "Now, it's disappeared."
Fuzzy target. But Caruso is in the fortunate, small minority who get such positive results from Iressa. Unlike Herceptin, Iressa has no test that identifies patients most likely to benefit from it. Indeed, many scientists argue that it was a mistake to push Iressa ahead without such a test, because it's basically a targeted drug without a clear patient to target.
The drug's maker has a sharp retort: "That's insane!" says Mary Lynn Carver, a spokesperson for AstraZeneca. "That's saying, `Let's wait until we have everything tied up with a nice neat bow before we save somebody's life.' In lung cancer, people usually die within a year of diagnosis. You can't wait if you want to save people like Gloria Caruso." The gene analyzers are busily churning through tumor samples, searching for a difference between responders and nonresponders that could refine the target. Meanwhile, the drug may be approved next month as a last-ditch therapy.
Given the difficulties of developing new drugs, some of targeted medicine's pioneers have turned to diagnostic tests instead (box, Page 56). For example, a drug that's often used to prevent organ rejection after a transplant is broken down by an enzyme called TPMT. This prevents the drug from reaching toxic levels in the body. However, Bill Evans of St. Jude Children's Research Hospital in Memphis has found three genetic variants in TPMT that essentially cripple it in 1 of every 300 people. People who have these variants must take much lower doses of the transplant drug, because the normal doses could be fatal. A genetic test to identify those people has become standard in the transplant field. "It's made its way all the way from the lab to the bedside," Evans says.
Anticlotting. The next test to travel this route may be one for reactions to the anticlotting drug warfarin. Warfarin is one of the most frequently prescribed medications for heart patients, but at too high a dose it can cause uncontrolled bleeding, requiring hospitalization and emergency transfusions. "So doctors monitor patients who take the drug pretty carefully," says David Veenstra, a drug outcomes researcher in the pharmacy department at the University of Washington in Seattle. Given that kind of care, Veenstra was surprised to find one group of warfarin patients with twice the bleeding risk of others. The high-risk patients turned out to have variations in a gene called 2C9, which codes for a liver enzyme that removes warfarin from the body. As Veenstra reported last year in the Journal of the American Medical Association, these variants make for a slower-acting enzyme, which in turn makes for more warfarin and more bleeding. The test is fairly straightforward. "I'd encourage patients to get this screen, and I'd certainly use the data," says Baylor's Roberts. "But most cardiologists haven't grown up with genetics, and they're not familiar with it."
They'll probably get to know it well in the next couple of years, many in the field predict, along with another test for sensitivity to blood pressure medications. Three months ago, Julie Johnson of the University of Florida presented data at an American Heart Association meeting indicating that people have different forms of a receptor that takes up drugs called beta blockers. Some forms produce no response to the drugs, and some produce a good response. "Those with the no-response variant obviously should try a different drug," she says. She's trying to replicate the finding in a larger study. If the receptor test does indeed have value, "we should be able to get away from trial and error and predict good responses to drugs using genetics."
But leaving errors behind and hitting the targets more precisely may call for more than larger clinical trials. It may require changes outside the lab and the clinic. A lot of those changes involve how medical care is paid for. If you're an insurance company, asks Roberts, would you pay for a test predicting someone's response to a heart drug? The way people switch health plans so often, it's likely that the prediction will benefit some other insurer--as well as the patient, of course. Plus there are the often-discussed issues of genes and privacy. What's to prevent your insurance company from boosting your premiums if your genes put you at risk for hypertension? After all, they would boost the charges if you smoke. "This country has to come to grips with the implications of this science," Roberts says.
The science, for its part, is moving ahead, clarifying the roles of the myriad genes involved in disease. "It's like when Galileo made the high-powered telescope," says the National Cancer Institute's Sausville. "All of a sudden people could see things around planets, like moons, that hadn't been thought about." But thinking about them opened up a whole universe of ideas.
Stopping cancer inside and out
Think of the cell as a fort. Many deadly cancers do their damage when normal cells go bad and take in too much growth factor, causing uncontrollable proliferation. Personalized drugs use different armaments to stop this unwelcome invasion.
THE INVASION
Human growth factor is a protein that's essential to life. However, cells can mutate and produce more than the normal number of receptors that "grab" this protein. The result is a very strong signal traveling into the cell to its nucleus, telling it to divide and divide. That's the signature of cancer.
[Drawing is not available]
[labels]
Growth factor
Receptor
Signaling molecules
Growing tumor
Nucleus
PERIMETER DEFENSE
The drug Herceptin aims to stop this invading protein by establishing a molecular moat outside a cell's walls, blocking a particular growth factor receptor. With the drug blocking the receptor, the growth factor can't attach. Therefore, it can't send its "grow and divide" signal into the cell, and the cell remains stable.
[Drawing is not available]
[labels]
Growth factor
Receptor
Signaling molecules
Herceptin
Nucleus
INTERIOR DEFENSE
The pill Iressa, by contrast, goes deeper. It homes in on the receptor but dives beyond it, inside the cell wall. Here it lies in wait, blocking messenger molecules within the cell that ordinarily would carry the growth command to the cell nucleus. Again, the goal is to stabilize the cell and keep it from growing out of control.
[Drawing is not available]
[labels]
Growth factor
Receptor
Iressa
Signaling molecules
Nucleus
Sources: Genentech, AstraZeneca; Rod Little
This story appears in the January 20, 2003 print edition of U.S. News & World Report.