Druker also recruited Charles Sawyers, from the University of California—Los Angeles, to the team. Sawyers was an oncologist with expertise in leukemia. Sawyers and Druker had met on the oncology conference circuit years earlier and were kindred spirits. Both held medical degrees, both were irresistibly drawn to lab research, and both had zeroed in on leukemia as the focus of their work. In fact, Sawyers had cut his scientific teeth in the laboratory of Owen Witte, who, with David Baltimore, had been largely responsible for figuring out the link between bcr/abl and CML.
When Druker showed Sawyers his results from his first cell-line and marrow studies of CGP-57148B in early 1995, Sawyers was immediately on board with Druker’s desire to ask the drug company to run a clinical trial. “Why wouldn’t we?” he said to Druker, awestruck by the data. Sawyers knew that the kinase program was a peripheral project that held little interest for the company. In part, Druker’s request for Sawyers to join the clinical trial planning team was to show the company that, together with John Goldman, they could recruit enough CML patients for a phase I clinical trial.
In designing a clinical trial protocol, the team had to decide whether the compound should be tested only for CML or also other types of cancer; which stages of CML would be included; where the studies would be conducted; how many investigators to include; how many patients to enroll; and on and on. There would be reams of paperwork to prepare, so although the molecule had not yet entered a single animal, there was no time to waste.
BY THE EARLY spring of 1996, the intravenous formulation was ready, the team in Basel having added all the necessary attributes without destroying the compound’s anti-kinase activity. Now the drug—as it had become—could advance to the next stage: animal testing. The rules of the FDA state that of the two species in which a new drug candidate had to be tested, one had to be a non-rodent. The intravenous formulation would be given to rats and dogs. A couple of different toxicology studies were planned. In the first one, the animals would be given the compound as a bolus injection; that is, the entire dose would be inserted into the bloodstream fairly rapidly. In the next study, the dose would be given as an infusion over three hours.
During the months when CGP-57148B was being converted from a drug candidate into an intravenous drug, a man named Peter Graf, a former director of pharmacokinetic studies at Ciba-Geigy, had continued to work on the oral formulation of CGP-57148B. Like Lydon and Matter, Graf knew that a pill version of the drug would be much better than an intravenous version. His additional research contradicted those first findings of which Matter had been so suspicious. Graf showed that the compound could be made into a soluble formulation, one that would remain intact in the body’s watery ecology and be absorbed into the bloodstream upon being swallowed, rather than being digested and eliminated with other bodily waste. What others on the development team had concluded was impossible, Graf had accomplished with certainty. But with the intravenous form already in toxicology testing, the pill was stored away.
In mid-April 1996, Druker learned that the preclinical paper had finally been accepted for publication at Nature Medicine. On April 30, Druker’s forty-first birthday, the report was published. The conclusion was the same one Druker had noted during his ASH presentation: “This compound may be useful in the treatment of bcr-abl-positive leukemia.” It was the stock, neutral language of so many reports of experimental drugs, mild enough to be supported by the data, piquing interest without being overblown. At that moment, though, Druker was entirely uncertain about whether he would ever get a chance to find out what the compound could really do. The process, though moving along, seemed to be taking forever. And although animal testing had begun and study protocols had been discussed, Ciba-Geigy had still not given full, official clearance for an eventual clinical trial if the toxicology work went smoothly. For Druker, confident in the drug and ready to move forward, the pace was excruciating.
On May 13, 1996, Druker received a fax from John Ford, who was the liaison between the investigators on the outside and the company. In two paragraphs, Ford delivered the news for which Druker had been so anxiously waiting. The compound had been discussed at a meeting with Ciba-Geigy’s research board early that month. “Our proposals received a favourable reception and we are reasonably confident that full approval to proceed into clinical trials will be forthcoming,” Ford wrote him. The next step was to complete the protocols for the human studies. Ford forecasted early July for those planning sessions. The FDA had been in touch with the company and was eager for a face-to-face meeting. It was happening.
A month later, Druker received a second fax from Ford inviting him to a meeting in Washington, DC, later that summer to discuss the next steps in the clinical development of CGP-57148B. The company was meeting with FDA representatives for a preliminary talk about filing an investigational new drug (IND) application. The FDA would tell Ciba-Geigy everything that was needed before a clinical trial could begin. The agency would also go over what clinical data were required for the drug to be considered for approval. Along with Druker, John Goldman, Charles Sawyers from UCLA, and Moshe Talpaz from MD Anderson, were also invited. These were the individuals whom Ciba-Geigy planned to involve in the phase I trial. They would be the principal investigators.
Suddenly the air had cleared. Everything was looking good. The first human study of CGP-57148B had finally come within reach. The company was on board, meeting with the FDA, finding out what would be needed for the road ahead. And Druker would be one of the investigators. The wait was over. The drug was going to be tested in CML patients.
Then, on July 9, 1996, the fax machine rang again with a third letter from Ford. The first toxicology report was in, and the news was not good. In a four-week study, a group of dogs had received daily infusions of CGP-57148B at doses of 6, 20, and 60 milligrams per kilogram. The infusions were given through a catheter into the jugular vein for 28 straight days, each infusion taking about 3 hours. Some of the dogs given the two higher doses had “massive necrotising thrombophlebitis starting at the tip of the catheter and extending into the lungs,” Ford wrote. The drug had crystallized in the blood, resulting in clots. The problem had occurred in the first week of the study, and some of the dogs had died as a result. The toxicologists altered the infusion schedule a bit, but the problem remained.
The clots were puzzling. In the first toxicology study, with the bolus injection, rats and dogs tolerated the compound well. Ford and the toxicology team had no explanation. He suspected some “undiscovered technical problem,” he told Druker. Regardless of the cause, though, the study would have to be repeated, resulting in a six-month delay in the schedule toward opening the first clinical trial.
Druker stared at the page. He could not believe what he was reading: “Given the seriousness of the toxicity and the time penalty which would be incurred in trying to rectify it, we feel inclined to abandon our plans to initiate trials with the intravenous formulation.”
There was a bright side, though. Ford was hopeful that the program could continue with the oral formulation. Graf’s work had now rescued the program. He reminded Druker that the intravenous approach was only used because the pharmaceutical development team had wrongly predicted that the compound would not be absorbed when taken orally. “Abandoning the intravenous trials will make all of our lives much more straightforward,” Ford wrote. However, the oral formulation would now need to enter toxicology studies from the beginning. His revised estimate for when a human study would launch was March 1997. In the meantime, the company had cancelled its meeting with the FDA. The FDA required toxicology data in two species given the oral formulation before it would discuss the IND, even in a preliminary fashion.
Ford persisted in his encouraging tone. “Naturally, we are all very disappointed by these unexpected events but we sincerely believe they represent only a temporary set-back,” Ford wrote, concluding with the cheerful news that CGP-57148B had been promoted to clinical development status by the company’s board of devel
opment.
23
_______
“NOT OVER MY DEAD BODY WILL THIS COMPOUND GO INTO MAN”
At nearly the same time that CGP-57148B was infiltrating the veins of laboratory animals, Ciba-Geigy and its rival from across the Rhine, Sandoz, were merging. The merger created the largest pharmaceutical company in the world, renamed Novartis. For Druker, the news wasn’t without irony. It was Dana-Farber’s agreement with Sandoz that had thwarted his work with Lydon years earlier. Now here he was in the midst of trying to develop the very kind of drug for which Sandoz had once barely mustered the minutest enthusiasm.
The mid-1990s was an era of pharmaceutical mergers that brought together other giants, including Glaxo and Wellcome, Pharmacia and Upjohn, and Roche and Boehringer Mannheim, among others. This spate of deals, which continued for several years, was the result of an industry-wide decrease in productivity in research and development and an onslaught of generic drugs. The Hatch-Waxman Act, passed into federal law in 1984, loosened restrictions for generic drug makers, allowing their products to enter the market on the basis of bioequivalence rather than clinical data. That is, as long as the manufacturer demonstrates that the generic drug has the same properties as the original, there is no need for costly clinical trials to test for safety and efficacy to earn FDA approval. The change made developing generics much cheaper and bringing them to market much easier. As patents expired, generics, usually exact replicas of the once-protected molecules, flooded the market, and revenues for brand-name manufacturers decreased. The moment a generic drug was approved, sales of the original faded fast, a moment known in the industry as the patent cliff.
The mergers were a response to these combined forces, a way for pharmaceutical companies to keep their profits up and their pipelines full. (Eventually, tactics to prevent such revenue loss would also include paying off generic drug makers to delay the introduction of their product onto the market and creating new indications for branded drugs to extend patent time. These practices, dubbed “evergreening,” are generally considered shady, operating through legal loopholes, but can bring a company millions of dollars in additional revenue, even when the result is just a few extra months of brand-name exclusivity.)
Yet the promise of ramped-up research and development that inspired the merger of Ciba-Geigy and Sandoz did not extend to the kinase inhibitor program. When Novartis was launched, “everything ground to a halt,” said Lydon. The initial toxicology report from the intravenous formulation of CGP-57148B had dampened the company’s interest in the drug, and although other animal tests were now in progress, all of the clinical programs were being reexamined as a routine part of the merger. It was the usual big-company problem, Lydon said. “No one was a champion for the compound within Novartis.” The company had little motive to push ahead a drug that would ultimately be given to very few people.
Frustrated by the slow pace at which this pharmaceutical behemoth was now moving, Lydon resigned shortly after the merger. He struck out on his own to start a small biotech company called Kinetix, which would be bought by Amgen a few years later, resulting in the clinical development of several kinase inhibitors. Elisabeth Buchdunger took over as the head of the biology department for Novartis’s cancer research program.
Alongside Matter, Lydon had been the force behind moving CGP-57148B—renamed STI-571 under Novartis—toward clinical trials. Without his greatest industry ally, and in light of the first toxicology report, Druker was worried that Novartis would lose interest in the drug.
The Nature Medicine paper of April 1996 had garnered little attention. That spring, the headlines were focused on the promise of angiogenesis inhibitors. Pioneered largely by a scientist named Judah Folkman, angiogenesis inhibitors were supposed to kill tumors by cutting off their blood supply. James Watson (who, with Francis Crick, Maurice Wilkins, and Rosalind Franklin determined the helical structure of DNA) famously pronounced that cancer would be cured in just a few years via this new approach. Kinase inhibition barely scored a footnote in discussions of exciting treatments on the horizon. Druker received some calls to talk about his work, and inevitably the interviewers asked whether the drug would be moving into a clinical trial. “That was the logical question,” said Druker. But for him, it was the toughest one. He didn’t want to acknowledge that he didn’t know if the drug would ever move forward, and he didn’t want to risk antagonizing the company and thereby jeopardizing the compound by pointing fingers.
One journalist visited Druker for an in-person interview. Alexandra Hardy, a writer from Houston who’d recently relocated to Portland for her husband’s work, had been assigned the story by the Associated Press. Hardy was lukewarm about doing the article. She’d covered medical breakthroughs before and knew they rarely lived up to the hype. “I did not think the drug was going anywhere,” she said. Something else did catch her interest, though. “The way he interacted with his patients struck me more [than the research],” she said. Hardy, not generally fond of doctors, noticed immediately how respectful Druker was to those in his care. It left a lasting impression, even if the reason for her visit didn’t.
Hardy’s story was picked up by the Oregonian, and then the attention quickly faded—almost. A man named Bud Romine, who was suffering from CML, sent Druker a note saying that if the drug ever went into a clinical trial, he wanted to be the first patient. The article also brought more people with CML to Druker, now a more widely recognized expert, for treatment.
Meanwhile the toxicology studies for the pill formulation were continuing at Novartis. Rats were given a low, middle, or high dose of STI-571 for two studies of thirteen weeks each. Some of the animals experienced kidney problems in the first round, but in the second thirteen-week stretch, the problems disappeared. In some animals, sperm production slowed down. Rats given the highest dose had bloody or dark urine, swollen muzzles, and increased salivation. The animals were euthanized at the end of twenty-six weeks and their organs were weighed. In many of them, testicular weight dropped. Some liver problems arose but were not life threatening. Two of the rats given a high dose of the drug died as a result, but none of the rats in the low- or middle-dose groups died.
In some of the female rats, the drug caused problems with the development of follicles in the ovaries. Blood samples revealed that the drug accumulated in their bodies at a faster pace than in the male rats. In pregnant rats and rabbits given the drug, the fetuses were damaged. In lactating rats, the chemical moved from the blood into the mother’s milk supply.
The studies continued for more than a year, far longer than the few months John Ford had predicted in his last fax to Druker, and much to the consternation of Alex Matter, who could only watch in silence, aghast. To investigate how STI-571 might affect the central nervous system, mice were given a single dose and observed for side effects such as tremors and impeded motor function. No problems occurred. To study cardiovascular toxicity, a group of rats was anesthetized and given a shot of the drug. The animals did not experience any heart trouble except a short-lived decrease in arterial blood pressure.
To study the gastrointestinal system, a group of mice was given varying doses of STI-571. Two hours later, the animals were fed a liquid containing a small amount of charcoal, which cannot be digested, to see if mice given a higher dose of the drug would eliminate the charcoal more slowly than those given a lower dose. If the drug was affecting the intestines, the charcoal would take longer to be excreted. The charcoal was passed by all the mice in about the same amount of time.
A group of beagles were given an oral dose of 60 milligrams per day for thirteen weeks to investigate any potential side effects. Some of the animals had severe diarrhea that resolved.
As the studies went on, Druker didn’t hear much from Novartis. It wasn’t that he was being ignored; the radio silence was just part of the standard procedure. The data were being recorded and would be assembled into a final toxicology report. And the studies took time. Each test took about three months, wi
th a further three months needed for analysis, which would include sacrificing the animals so that a pathologist could conduct a full exam of the internal organs. Until the analysis was completed, the findings would not be distributed outside the company. Druker kept himself busy with research.
Finally, someone from the company contacted him with results. Once again, the news was not good. In a study of dogs given doses up to 600 milligrams per day, high doses of the drug had caused liver failure. In rats, even low doses had resulted in some liver damage. The drug was considered too dangerous for humans. One of the toxicology experts told Alex Matter, “Not over my dead body will this compound go into man.”
To examine liver toxicity, the drug had been given in increasing doses for up to thirteen weeks. As the weeks passed, cells inside the animals’ livers began dying off, and inside the bile duct, cells began growing at a pace that was faster than normal, often a precursor to cancer. At regular, frequent intervals, the toxicologists checked the level of enzymes in the liver. Heightened levels of enzymes indicated liver damage. But when they noticed that liver enzymes were elevated in the animals, they continued to give the drugs. By the end of the study, two or three months later, the dogs had liver failure. Four weeks after the study, some of the dogs still showed signs of liver trouble.
Druker took issue with the rationale behind their approach. “If you know the dogs’ liver enzymes were elevated, wouldn’t you stop?” he explained. “If I was giving this to people, I’d stop the drug.” Unschooled as he was in running a clinical trial, Druker knew that administering experimental cancer drugs followed the same principles as giving highly toxic chemotherapy. When a dose is too high for a patient to tolerate, it’s lowered. When the side effects are too much to bear, the treatment is halted. Cancer patients aren’t force-fed medications until their livers fail. Rather, the functioning of all organs is carefully monitored throughout treatment. “We know what to look for, we know when to hold the drugs,” Druker said. In a clinical trial, patients are monitored even more frequently than in a regular clinic.
The Philadelphia Chromosome Page 17
