Starting in June 1998, each site would enroll one patient per month, for about ten months, for a total of about thirty patients. Druker, Sawyers, and Talpaz knew that the low starting dose—25 milligrams—would not have any effect. But starting low was essential to preventing dangerous toxicities. Month by month for up to a year, they would raise the dose higher and higher, looking for a change in blood counts and keeping vigilant for side effects.
The pills were small and orange, the whitish powder that Jürg Zimmermann had created held inside a shell made of gelatin and dyes. Upon entering the body, the shell dissolved, releasing crystals of STI-571. Until now, those molecules had been released only into the bodies of rats, mice, dogs, rabbits, and monkeys. Every clinical trial investigator knows that animal studies are often poor predictors for what a drug will do to a human being and how a human body will handle a drug. So regardless of the years of toxicity testing and the endless fights with Novartis to test the drug in people, Druker, Sawyers, and Talpaz still proceeded with trepidation. They stood at the crossroads of the theoretical power of hitting a single kinase and the unknown reality of what would happen when this foreign chemical was unleashed in the body. Would it really block just the one Bcr/Abl kinase and leave the person unharmed? Or was there a chance it would shut down multiple kinases, an event that, it was not melodramatic to think, could kill someone. “What if it blocks ATP binding of every single kinase in the body?” said Sawyers, recalling the concern at that time. “That’s going to be one sick person.”
In Portland, Bud Romine barely hesitated before swallowing his 25 milligrams. The same for the minister from Bakersfield, California, who was Charles Sawyers’s first enrollee. They would remain in the clinic for eight hours, under careful surveillance for any signs of potentially dangerous side effects, and then return the following morning. In Houston, it was the same.
Enrollment required a commitment. The phase I patients had to live near the study site for about three months. In Portland, Los Angeles, and Houston, patients, often accompanied by their spouses, rented homes, temporarily leaving their lives behind for a chance to try this new drug. They also had to agree to close medical scrutiny, including blood tests three times a week, periodic bone marrow biopsies, and other monitoring. For each patient, the investigators had to record white blood cell counts, red blood cell counts, liver enzyme levels, kidney function, weight, temperature, blood pressure, and any other measures that might reveal a problem the drug was causing. Check-ups were at least once a week.
The investigators explained to the trial patients the various ways in which their response to the drug would be measured. First was the hematologic response rate, which showed changes in the number of red and white blood cells based on samples usually drawn from the vein at the crook of the elbow. A decrease in the number of white blood cells would indicate something positive. The number of platelets, the part of the blood responsible for clotting, can rise or fall as CML progresses, so a count moving toward the normal range was a good sign. A hematologic response was defined as a decrease in the white blood cell count by half. A complete hematologic response meant a patient had normal counts of white blood cells, platelets, and red blood cells, and no blast cells in the bloodstream.
Responses to STI-571 would also be measured in terms of a cytogenetic response. Cytogenetics examined the connection between genes and disease, and so a cytogenetic response meant a change at the genetic level as a result of the treatment. Using bone marrow biopsies (the same chromosome analysis tests that Janet Rowley had used in the 1970s), and FISH to a lesser extent, the investigators could examine how the drug affected the prevalence of the Philadelphia chromosome. Were there fewer cells containing this genetic abnormality following treatment with STI-571? When the bone marrow was biopsied at the time of diagnosis, usually 100 percent of the cells examined had this abnormality. A so-called major cytogenetic response meant the Philadelphia chromosome was now present in just 35 percent or less of the biopsied cells. A complete cytogenetic response meant the Philadelphia chromosome was entirely absent.
Druker knew that the starting dose of an experimental drug is always based on toxicity observed in animal studies. The first human dose would be one tenth of whatever dose caused animals to be sick in some way. But Druker also knew the drug level—technically different from drug dose—that was needed for the cancer to be killed. Concentration is measured in molars, a unit representing the amount of material held in a solution. Based on his studies in CML cells, Druker predicted that a human bloodstream would need to contain a minimum of one micromolar of STI-571 for the drug to have any anticancer activity. At less than that, the drug would be too dilute to do any good.
However, the dosage needed to achieve those concentrations—one micromolar, ten micromolars, or anywhere in between—was a complete mystery. There was no way to estimate how many milligrams of STI-571 a patient would have to swallow to bring the concentration of the drug in the blood to one micromolar or higher. That measurement could be deduced only when the drug entered a human bloodstream.
Almost more important than the predicted minimum drug level, though, was the maximum drug level on which the planning team had agreed. Deciding in advance on a measure at which one can safely conclude a drug doesn’t work is common in clinical trial planning; investigators and industry representatives want to know when they can say when. “We are always [asking], ‘What are the endpoints that will kill this project?’” said Druker. For STI-571, that endpoint was set at ten micromolars, a far higher concentration than the minimum required for activity. “If we got to ten times our prediction and saw nothing,” said Druker, “we would stop the trial.”
At each study site, the patients came through the first twenty-four hours unscathed. The relief among the three investigators and at Novartis headquarters was palpable across the hundreds or thousands of miles separating the sites. In Portland, Druker continued to give Romine his daily capsule for the rest of June. Repeated blood tests showed no change in the number of white blood cells or of cells containing the Philadelphia chromosome. Druker knew that, whether quickly or slowly, Romine’s white blood cell count would continue to climb. But as much as he wished to give this brave man a higher dose of STI-571, the study protocol forbade it. Plus there was no telling whether the drug would elicit side effects later on. Delayed problems could arise in a few weeks or months. For now, Romine had to stay at 25 milligrams.
A husky man with thick glasses who favored brown suspenders that flanked a respectable but not outlandishly large gut, Romine accepted his fate in the name of helping science. “My feeling always was if it didn’t help me, maybe it would help someone in the future,” Romine would later write in the STI Gazette, a newsletter started by Druker’s phase I patients. When Romine and the other first-month patients made it to the end of June with no side effects, the investigators knew it was safe to start the next dose level. But as Romine’s white blood cell counts began to escalate, he had to be taken out of the study. It was hard to not see the irony in Romine’s letter to Druker years earlier asking to be the first person Druker would call if the compound he’d read about in the Oregonian ever went into clinical trials. Enrolling as patient 1 had almost guaranteed that he wouldn’t benefit from the drug. His time on the phase I study having reached its abrupt end, Romine started treatment with hydroxyurea, or Hydrea, a chemotherapy drug that was the only other option for patients who had stopped responding to interferon. This drug might keep his counts in check for a little while, but it was just a temporary fix. Eventually it would stop working.
In July, a new round of patients entered the trial at the 50-milligram dose level. The outcomes were the same: no side effects but also no effect on the cancer. Blood samples from these first few groups of patients were revealing in other ways, though. Looking at the concentration of the drug in the bloodstream, Druker could now see that the dose would need to reach about 200 milligrams before the first response would be seen. But if they didn’
t see any sign of activity at 200 milligrams, he would start to worry. No activity at 200 milligrams meant that the drug probably wasn’t going to work.
DRUKER WAS IN urgent need of a nurse who would help him care for patients and record the profusion of data that were about to be generated. It was only the first month of the study, but so much was needed to prepare for each subsequent month. Patients needed to be walked through clinical trial disclosure forms. Among those who had stopped responding to interferon, blood had to be drawn to ensure the white counts were climbing; the protocol required that they be at least 20,000 and rising for the patient to be eligible for the study because that was the only way to see if the drug was working. Medical records needed to be obtained, and that task alone could require hours on the phone.
Carolyn Blasdel was working as a clinical trial nurse in a Texas cancer center when she decided to relocate to Portland in the summer of 1998. Working on Druker’s trial was a perfect match for her. She had the skills, and the work was inspiring. She’d seen listless patients dying of leukemia, seen how they bled out, how their bodies would shut down from the absence of platelets, how they were overwhelmed by infections after their immune system had been shredded by chemotherapy. She knew about the challenges with interferon, and she knew that patients who took a break from it to recover often refused to resume. “I had several people tell me they’d rather be dead than go back on interferon,” Blasdel recalled. She knew that CML was an inevitably fatal disease except for the rare few who were cured by a bone marrow transplant. If someone was working on a medicine that could change this sorry state, she was game to help.
Druker met Blasdel on a Saturday when she was house hunting in Portland. They talked in the OHSU hospital lobby for an hour, and he offered her the job. She’d be splitting her time between his trial and another one with a different investigator. She knew her job with Druker would probably last about six months or so, since most phase I trials don’t succeed. In early August, Blasdel was ready to begin.
Her first patient was patent number 2, at the 50-milligram dose level. As the study coordinator, Blasdel was drawing the blood, sending the samples for testing, and receiving the results. So she was the first one to see any noteworthy changes in the cancer. Right away, she could see the drug was having no effect on the patient’s blood counts. A few days later, patient number 3 entered the trial. This time, the dose was increased to 85 milligrams per day.
Like everyone in the study, patient number 3 had not taken any CML medication for a week or so before starting STI-571, and the white blood cell counts were rising steadily. For the first couple of weeks at the 85-milligram dose, the counts continued to climb. Then one day, Blasdel got the patient’s blood test results and saw that the counts had not risen. “Could I have made a mistake?” she asked herself. She thought back over every step she’d taken when she’d drawn the patient’s blood. Could she have sent the wrong tube of blood? There was no possible way for her to have done something like that. There was no mistake.
She told Druker, who promptly found out that Sawyers and Talpaz had seen the same thing with their 85-milligram patients.
At 140 milligrams, one patient at each site had a decrease in the number of white blood cells. In Druker’s patient, the white blood cell counts dropped to fewer than 10,000 per microliter, a normal level.
At 200 milligrams, the drop in white blood cell counts became even more prominent. All of the patients were having a hematologic response. The abnormal white blood cells, the very definition of leukemia, were starting to disappear. Repeated blood tests showed the counts were down.
There was more. Bone marrow samples from the October group of patients—140 milligrams—revealed an even deeper response to the drug: a decreased number of cells containing the Philadelphia chromosome. They were having cytogenetic responses. In other words, the abnormality responsible for the cancer was disappearing.
plate 11
Changes in blood counts were important because it was these off-kilter numbers that made people feel sick. Too few red blood cells and platelets and too many white blood cells were all dangerous conditions, and as the numbers returned to normal, patients began to feel noticeably better. But a change at the genetic level signified a deeper response, one that bespoke causes, not just symptoms. A decrease in the number of cells containing the Philadelphia chromosome was evidence that STI-571 was truly a targeted therapy. By hitting the haywire kinase, the drug was eradicating the underlying genetic mutation. This drug, rationally designed against a known target, was doing exactly what it had been created to do. No other drug in history had hit cancer at its genetic roots.
Similar breakthroughs were being made elsewhere. In September 1998, the FDA approved Herceptin, a biologic drug aimed against a protein called HER2, which is overexpressed in some breast cancers. The drug prolonged life and was an undeniable advance for breast cancer sufferers and for cancer medicine as a whole. But Herceptin had to be given with chemotherapy, and it worked only for patients with the HER2-positive variety of breast cancer. Plus the survival benefit had a foreseeable endpoint. In clinical trials, patients with advanced breast cancer who’d taken Herceptin plus chemotherapy lived an average of five months longer than those who’d taken chemotherapy alone. STI-571 took the promise of targeted therapy further. This was a single drug given on its own as a once-daily pill targeting a mutation that occurred in nearly every patient with CML. Would it do for this disease what Herceptin had done for breast cancer? Would it do even more? The original purpose of the phase I trial of STI-571 was to prove the principle behind kinase inhibition: Kill the kinase and you kill the cancer. Just months after launching the trial, it was much too soon to think about the broader implications. And yet it was impossible not to.
Druker, Sawyers, and Talpaz held weekly conference calls, joined by John Ford, who was monitoring the trial from Basel. When the counts began dropping at 200 milligrams, even Talpaz, the seasoned investigator who’d witnessed many new agents rapidly elicit promising responses that soon faded, was impressed.
Hesitation lingered, though, as everyone watched for side effects, bracing themselves for disaster. Nearly every single cancer treatment available at that time did something terrible to the body as it killed the malignancy. The theory that a drug aimed against one single kinase would not cause such widespread damage was plausible, but it was hard for the investigators to truly believe it because everything they’d experienced said otherwise.
By a few months into the study, none of the patients had developed any severe side effects. Many had puffiness under the eyes, a result of fluid retention, and in the waiting room patients swapped home remedies to reduce the bagginess, including tea bags and hemorrhoid cream. But the problem was more of a curiosity than a concern. Compared with the side effects of chemotherapy or of interferon, which most of the phase I patients had taken, puffy eyes seemed a small thing.
There were some problems. Patients pushed through dramatic leg pain at the start of therapy as bones swollen from the profusion of immature white blood cells returned to normal. Cramping occurred because trial patients had to take the pill on an empty stomach to prevent food interactions. There was some tiredness. But the debilitating side effects that had made cancer treatment so feared were absent. A patient on chemotherapy was certain to experience some harrowing combination of a number of them: numbness, kidney damage, liver damage, nausea, vomiting, hair loss, peeling skin, rashes, diarrhea, constipation, headaches, changes in blood pressure, dry mouth, insomnia, fever, fatigue, bloody urine or stools, anemia, memory loss, pain, swelling, problems with urination. Drugs could ameliorate some of the problems, but no one taking chemo came through unharmed.
The STI-571 experience wasn’t normal. This wasn’t how cancer treatment worked. Patients were supposed to be doubled over in agony, vomiting into waste bins at the side of their bed, too weak to take care of themselves. These patients were carrying on with their lives as if nothing was wrong. Chronic flu symp
toms and depression had plagued many of them during their stretches on interferon. In light of those dark months, STI-571 was a dream.
Still, Druker barely allowed himself a modicum of excitement. He knew that what he was seeing with the phase I patients could disappear in an instant. STI-571 might be working now, but in another week all those blood counts could escalate right back to where they started. Some late side effect could appear that would render the drug useless. Throughout the ninety-hour workweeks he put in during the study, Druker never stopped to think about the fact that this drug was a success. He couldn’t. Not yet.
As if vindicating his caution, one of the patients on the 200-milligram dose developed liver damage. Some people at the company panicked. They wanted the investigators to go back to 140 milligrams. The investigators vehemently disagreed. “Don’t go back, enroll more,” they insisted. The only way to know whether the toxicity was a one-off situation or a real problem was to continue giving the drug to people at the current dose. They had already seen some activity with the drug, and the only thing to do was keep moving forward. Novartis agreed to continue at 200 milligrams, allowing the investigators to enroll two patients per month starting in January 1999 instead of raising the dose until the problem was resolved.
The Philadelphia Chromosome Page 20