In early 1998, Novartis finally relented. It would synthesize and release a few grams of CGP57148, just about enough to run a trial on about a hundred patients. Druker would have a shot—but only one shot. To Novartis, CGP57148, the product of its most ambitious drug-discovery program to date, was already a failure.
I first heard of Druker’s drug in the fall of 2002. I was a medical resident triaging patients in the emergency room at Mass General when an intern called me about a middle-aged man with a history of CML who had come in with a rash. I heard the story almost instinctively, drawing quick conclusions. The patient, I surmised, had been transplanted with foreign bone marrow, and the rash was the first blush of a cataclysm to come. The immune cells in the foreign marrow were attacking his own body—graft-versus-host disease. His prognosis was grim. He would need steroids, immunosuppressives, and immediate admission to the transplant floor.
But I was wrong. Glancing at the chart in the red folder, I saw no mention of a transplant. Under the stark neon light of the examining room when he held out his hand to be examined, the rash was just a few scattered, harmless-looking papules—nothing like the dusky, mottled haze that is often the harbinger of a graft reaction. Searching for an alternative explanation, I quickly ran my eye through his list of medicines. Only one drug was listed: Gleevec, the new name for Druker’s drug, CGP57148.*
The rash was a minor side effect of the drug. The major effect of the drug, though, was less visible but far more dramatic. Smeared under the microscope in the pathology lab on the second floor, his blood cells looked extraordinarily ordinary—“normal red cells, normal platelets, normal white blood cells,” I whispered under my breath as I ran my eyes slowly over the three lineages. It was hard to reconcile this field of blood cells in front of my eyes with the diagnosis; not a single leukemic blast was to be seen. If this man had CML, he was in a remission so deep that the disease had virtually vanished from sight.
By the winter of 1998, Druker, Sawyers, and Talpaz had witnessed dozens of such remissions. Druker’s first patient to be treated with Gleevec was a sixty-year-old retired train conductor from the Oregon coast. The patient had read about the drug in an article about Druker in a local newspaper. He had called Druker immediately and offered to be a “guinea pig.” Druker gave him a small dose of the drug, then stood by his bedside for the rest of the afternoon, nervously awaiting any signs of toxicity. By the end of the day there were no adverse effects; the man was still alive. “It was the first time that the molecule had entered a human body, and it could easily have created havoc, but it didn’t,” Druker recalled. “The sense of relief was incredible.”
Druker edged into higher and higher doses—25, 50, 85, and 140 mg. His cohort of patients grew as well. As the dose was escalated in patients, Gleevec’s effect became even more evident. One patient, a Portland woman, had come to his clinic with a blood count that had risen to nearly thirtyfold the normal number; her blood vessels were engorged with leukemia, her spleen virtually heaving with leukemic cells. After a few doses of the drug, Druker found her counts dropping precipitously, then normalizing within one week. Other patients, treated by Sawyers at UCLA and Talpaz in Houston, responded similarly, with blood counts normalizing within a few weeks.
News of the drug spread quickly. The development of Gleevec paralleled the birth of the patient chat room on the Internet; by 1999, patients were exchanging information about trials online. In many cases, it was patients who informed their doctors about Druker’s drug and then, finding their own doctors poorly informed and incredulous, flew to Oregon or Los Angeles to enroll themselves in the Gleevec trial.
Of the fifty-four patients who received high doses of the drug in the initial phase I study, fifty-three showed a complete response within days of starting Gleevec. Patients continued the medicine for weeks, then months, and the malignant cells did not visibly return in the bone marrow. Left untreated, chronic myeloid leukemia is only “chronic” by the standards of leukemia: as the disease accelerates, the symptoms run on a tighter, faster arc and most patients live only three to five years. Patients on Gleevec experienced a palpable deceleration of their disease. The balance between normal and malignant cells was restored. It was an unsuppuration of blood.
By June 1999, with many of the original patients still in deep remissions, Gleevec was evidently a success. This success continues; Gleevec has become the standard of care for patients with CML. Oncologists now use the phrases “pre-Gleevec era” and “post-Gleevec era” when discussing this once-fatal disease. Hagop Kantarjian, the leukemia physician at the MD Anderson Cancer Center in Texas, recently summarized the impact of the drug on CML: “Before the year 2000, when we saw patients with chronic myeloid leukemia, we told them that they had a very bad disease, that their course was fatal, their prognosis was poor with a median survival of maybe three to six years, frontline therapy was allogeneic transplant . . . and there was no second-line treatment. . . . Today when I see a patient with CML, I tell them that the disease is an indolent leukemia with an excellent prognosis, that they will usually live their functional life span provided they take an oral medicine, Gleevec, for the rest of their lives.”
CML, as Novartis noted, is hardly a scourge on public health, but cancer is a disease of symbols. Seminal ideas begin in the far peripheries of cancer biology, then ricochet back into more common forms of the disease. And leukemia, of all forms of cancer, is often the seed of new paradigms. This story began with leukemia in Sidney Farber’s clinic in 1948, and it must return to leukemia. If cancer is in our blood, as Varmus reminded us, then it seems only appropriate that we keep returning, in ever-widening circles, to cancer of the blood.
The success of Druker’s drug left a deep impression on the field of oncology. “When I was a youngster in Illinois in the 1950s,” Bruce Chabner wrote in an editorial, “the world of sport was shocked by the feat of Roger Bannister. . . . On May 6, 1954, he broke the four-minute barrier in the mile. While improving upon the world record by only a few seconds, he changed the complexion of distance running in a single afternoon. . . . Track records fell like ripe apples in the late 50s and 60s. Will the same happen in the field of cancer treatment?”
Chabner’s analogy was carefully chosen. Bannister’s mile remains a touchstone in the history of athletics not because Bannister set an unbreachable record—currently, the fastest mile is a good fifteen seconds under Bannister’s. For generations, four minutes was thought to represent an intrinsic physiological limit, as if muscles could inherently not be made to move any faster or lungs breathe any deeper. What Bannister proved was that such notions about intrinsic boundaries are mythical. What he broke permanently was not a limit, but the idea of limits.
So it was with Gleevec. “It proves a principle. It justifies an approach,” Chabner continued. “It demonstrates that highly specific, non-toxic therapy is possible.” Gleevec opened a new door for cancer therapeutics. The rational synthesis of a molecule to kill cancer cells—a drug designed to specifically inactivate an oncogene—validated Ehrlich’s fantasy of “specific affinity.” Targeted molecular therapy for cancer was possible; one only needed to hunt for it by studying the deep biology of cancer cells.
A final note: I said CML was a “rare” disease, and that was true in the era before Gleevec. The incidence of CML remains unchanged from the past: only a few thousand patients are diagnosed with this form of leukemia every year. But the prevalence of CML—the number of patients presently alive with the disease—has dramatically changed with the introduction of Gleevec. As of 2009, CML patients treated with Gleevec survive an average of thirty years after their diagnosis. Based on that survival figure, Hagop Kantarjian estimates that within the next decade, 250,000 people will be living with CML in America, all of them on targeted therapy. Druker’s drug will alter the national physiognomy of cancer, converting a once-rare disease into a relatively common one. (Druker jokes that he has achieved the perfect inversion of the goals of cancer medicine: his drug has inc
reased the prevalence of cancer in the world.) Given that most of our social networks typically extend to about one thousand individuals, each of us, on average, will know one person with this leukemia who is being kept alive by a targeted anticancer drug.
*Abl, too, was first discovered in a virus, and later found to be present in human cells—again recapitulating the story of ras and src. Once more, a retrovirus had “pirated” a human cancer gene and turned into a cancer-causing virus.
* Gleevec, the commercial name, is used here because it is more familiar to patients. The scientific name for CGP57148 is imatinib. The drug was also called STI571.
The Red Queen’s Race
“Well, in our country,” said Alice, still panting a little, “you’d generally get to somewhere else—if you ran very fast for a long time, as we’ve been doing.”
“A slow sort of country!” said the Queen. “Now, here, you see, it takes all the running you can do, to keep in the same place. If you want to get somewhere else, you must run at least twice as fast as that!”
—Lewis Carroll,
Through the Looking-Glass
In August 2000, Jerry Mayfield, a forty-one-year-old Louisiana policeman diagnosed with CML, began treatment with Gleevec. Mayfield’s cancer responded briskly at first. The fraction of leukemic cells in his bone marrow dropped over six months. His blood count normalized and his symptoms improved; he felt rejuvenated—“like a new man [on] a wonderful drug.” But the response was short-lived. In the winter of 2003, Mayfield’s CML stopped responding. Moshe Talpaz, the oncologist treating Mayfield in Houston, increased the dose of Gleevec, then increased it again, hoping to outpace the leukemia. But by October of that year, there was no response. Leukemia cells had fully recolonized his bone marrow and blood and invaded his spleen. Mayfield’s cancer had become resistant to targeted therapy.
Now in the fifth year of their Gleevec trial, Talpaz and Sawyers had seen several cases like Mayfield’s. They were rare. The vast proportion of CML patients maintained deep, striking remissions on the drug, requiring no other therapy. But occasionally, a patient’s leukemia stopped responding to Gleevec, and Gleevec-resistant leukemia cells grew back. Sawyers, having just entered the world of targeted therapy, swiftly entered a molecular world beyond targeted therapy: how might a cancer cell become resistant to a drug that directly inhibits its driving oncogene?
In the era of nontargeted drugs, cancer cells were known to become drug-resistant through a variety of ingenious mechanisms. Some cells acquire mutations that activate molecular pumps. In normal cells, these pumps extrude natural poisons and waste products from a cell’s interior. In cancer cells, these activated pumps push chemotherapy drugs out from the interior of the cell. Spared by chemotherapy, the drug-resistant cells outgrow other cancer cells. Other cancer cells activate proteins that destroy or neutralize drugs. Yet other cancers escape drugs by migrating into reservoirs of the body where drugs cannot penetrate—as in lymphoblastic leukemia relapsing in the brain.
CML cells, Sawyers discovered, become Gleevec-resistant through an even wilier mechanism: the cells acquire mutations that specifically alter the structure of Bcr-abl, creating a protein still able to drive the growth of the leukemia but no longer capable of binding to the drug. Normally, Gleevec slips into a narrow, wedgelike cleft in the center of Bcr-abl—like “an arrow pierced through the center of the protein’s heart,” as one chemist described it. Gleevec-resistant mutations in Bcr-abl change the molecular “heart” of the Bcr-abl protein so that the drug can no longer access the critical cleft in the protein, thus rendering the drug ineffective. In Mayfield’s case, a single alteration in the Bcr-abl protein had rendered it fully resistant to Gleevec, resulting in the sudden relapse of leukemia. To escape targeted therapy, cancer had changed the target.
To Sawyers, these observations suggested that overcoming Gleevec resistance with a second-generation drug would require a very different kind of attack. Increasing the dose of Gleevec, or inventing closely related molecular variants of the drug, would be useless. Since the mutations changed the structure of Bcr-abl, a second-generation drug would need to block the protein through an independent mechanism, perhaps by gaining another entry point into its crucial central cleft.
In 2005, working with chemists at Bristol-Myers Squibb, Sawyers’s team generated another kinase inhibitor to target Gleevec-resistant Bcr-abl. As predicted, this new drug, dasatinib, was not a simple structural analogue of Gleevec; it accessed Bcr-abl’s “heart” through a separate molecular crevice on the protein’s surface. When Sawyers and Talpaz tested dasatinib on Gleevec-resistant patients, the effect was remarkable: the leukemia cells involuted again. Mayfield’s leukemia, fully resistant to Gleevec, was forced back into remission in 2005. His blood count normalized again. Leukemia cells dissipated out of his bone marrow gradually. In 2009, Mayfield still remains in remission, now on dasatinib.
Even targeted therapy, then, was a cat-and-mouse game. One could direct endless arrows at the Achilles’ heel of cancer, but the disease might simply shift its foot, switching one vulnerability for another. We were locked in a perpetual battle with a volatile combatant. When CML cells kicked Gleevec away, only a different molecular variant would drive them down, and when they outgrew that drug, then we would need the next-generation drug. If the vigilance was dropped, even for a moment, then the weight of the battle would shift. In Lewis Carroll’s Through the Looking-Glass, the Red Queen tells Alice that the world keeps shifting so quickly under her feet that she has to keep running just to keep her position. This is our predicament with cancer: we are forced to keep running merely to keep still.
In the decade since the discovery of Gleevec, twenty-four novel drugs have been listed by the National Cancer Institute as cancer-targeted therapies. Dozens more are in development. The twenty-four drugs have been shown to be effective against lung, breast, colon, and prostate cancers, sarcomas, lymphomas, and leukemias. Some, such as dasatinib, directly inactivate oncogenes. Others target oncogene-activated pathways—the “hallmarks of cancer” codified by Weinberg. The drug Avastin interrupts tumor angiogenesis by attacking the capacity of cancer cells to incite blood-vessel growth. Bortezomib, or Velcade, blocks an internal waste-dispensing mechanism for proteins that is particularly hyperactive in cancer cells.
More than nearly any other form of cancer, multiple myeloma, a cancer of immune-system cells, epitomizes the impact of these newly discovered targeted therapies. In the 1980s, multiple myeloma was treated by high doses of standard chemotherapy—old, hard-bitten drugs that typically ended up decimating patients about as quickly as they decimated the cancer. Over a decade, three novel targeted therapies have emerged for myeloma—Velcade, thalidomide, and Revlimid—all of which interrupt activated pathways in myeloma cells. Treatment of multiple myeloma today involves mixing and matching these drugs with standard chemotherapies, switching drugs when the tumor relapses, and switching again when the tumor relapses again. No single drug or treatment cures myeloma outright; myeloma is still a fatal disease. But as with CML, the cat-and-mouse game with cancer has extended the survival of myeloma patients—strikingly in some cases. In 1971, about half the patients diagnosed with multiple myeloma died within twenty-four months of diagnosis; the other half died by the tenth year. In 2008, about half of all myeloma patients treated with the shifting armamentarium of new drugs will still be alive at five years. If the survival trends continue, the other half will continue to be alive well beyond ten years.
In 2005, a man diagnosed with multiple myeloma asked me if he would be alive to watch his daughter graduate from high school in a few months. In 2009, bound to a wheelchair, he watched his daughter graduate from college. The wheelchair had nothing to do with his cancer. The man had fallen down while coaching his youngest son’s baseball team.
In a broader sense, the Red Queen syndrome—moving incessantly just to keep in place—applies equally to every aspect of the battle against cancer, including cancer screening and cancer pr
evention. In the early winter of 2007, I traveled to Framingham in Massachusetts to visit a study site that will likely alter the way we imagine cancer prevention. A small, nondescript Northeastern town bound by a chain of frozen lakes in midwinter, Framingham is nonetheless an iconic place writ large in the history of medicine. In 1948, epidemiologists identified a cohort of about five thousand men and women living in Framingham. The behavior of this cohort, its habits, its interrelationships, and its illnesses, has been documented year after year in exquisite detail, creating an invaluable longitudinal corpus of data for hundreds of epidemiological studies. The English mystery writer Agatha Christie often used a fictional village, St. Mary Mead, as a microcosm of all mankind. Framingham is the American epidemiologist’s English village. Under sharp statistical lenses, its captive cohort has lived, reproduced, aged, and died, affording a rare glimpse of the natural history of life, disease, and death.
The Framingham data set has spawned a host of studies on risk and illness. The link between cholesterol and heart attacks was formally established here, as was the association of stroke and high blood pressure. But recently, a conceptual transformation in epidemiological thinking has also been spearheaded here. Epidemiologists typically measure the risk factors for chronic, noninfectious illnesses by studying the behavior of individuals. But recently, they have asked a very different question: what if the real locus of risk lies not in the behaviors of individual actors, but in social networks?
In May 2008, two Harvard epidemiologists, Nicholas Christakis and James Fowler, used this notion to examine the dynamics of cigarette smoking. First, Fowler and Christakis plotted a diagram of all known relationships in Framingham—friends, neighbors, and relatives, siblings, ex-wives, uncles, aunts—as a densely interconnected web. Viewed abstractly, the network began to assume familiar and intuitive patterns. A few men and women (call them “socializers”) stood at the epicenter of these networks, densely connected to each other through multiple ties. In contrast, others lingered on the outskirts of the social web—“loners”—with few and fleeting contacts.