Gleevec began life in the late 1980s as a compound known as STI571, synthesized by Ciba Geigy to inhibit tyrosine kinase oncogenes. Brian J. Druker, who is based at the Oregon Health & Science University Cancer Institute, Portland, then put it to work on the tyrosine kinase that was known to be activated in CML. It was Druker’s lab that did the basic work showing that STI571 had potential, and Druker who talked the pharmaceutical giant Novartis into testing the compound on patients. Druker’s impact in this field is now dramatic. His 1996 Nature Medicine article reporting on the performance of STI571 in the laboratory has been cited more than 400 times; his back-to-back papers in the April 5, 2001 New England Journal of Medicine reporting on the clinical trials had each attracted upwards of 300 citations little more than 18 months after publication (see the table on page 4). In all, 14 of Druker’s papers published since 1995 have now been cited more than 100 times each. In addition to two reports in the current Medicine Top Ten (page 5, paper #1 and #4), Druker and colleagues have published six Hot Papers over the last two years, placing Druker on this issue’s page-one list of "hot" scientists.

Druker, 47, obtained his bachelor’s degree from the University of California, San Diego, in 1977 and his medical degree from UCSD four years later. He went on to do an internship and residency in Barnes Hospital at the Washington University School of Medicine and then a three-year fellowship in Medical Oncology at the Dana-Farber Cancer Center. Between 1987 and 1993 he was an instructor at Harvard Medical School, and then moved on to Oregon Health & Sciences University, where he is now a professor of medicine and director of the OHSU Cancer Institute Leukemia Center. Last year, Druker became a Howard Hughes Medical Institute investigator.

Can you start by telling us how you got started working with chronic myelogenous leukemia, and how you came up with the idea of inhibiting the activated tyrosine kinase?

Druker: In 1990, I was working in the laboratory of Thomas Roberts at Dana-Farber. He was investigating tyrosine kinases and their role in regulating cell growth. After gaining some expertise, I decided I really needed to combine my clinical cancer background with my basic science training and work on a human disease. The obvious choice was a disease activated by a tyrosine kinase. It quickly became clear to me that CML would fit perfectly, because it is a human leukemia that is known to be caused by an activated tyrosine kinase, called BCR-ABL. So, in collaboration with Jim Griffin at Dana-Farber, I started a project in 1990 working on the mechanism of transformation by this tyrosine kinase.

  When did the idea of attacking tyrosine kinases emerge as a potential route to controlling cancer?

Druker: Tyrosine kinases were first described in 1979 by Tony Hunter and Jonathan Cooper. By the mid-1980s it was known that a number of oncogenes were tyrosine kinases, and it was obvious to many that they would be an attractive target.

It sounds like there was a catch involved.

Druker: Most people thought it would be extremely unlikely to develop specific tyrosine kinase inhibitors, particularly if you were targeting the ATP-binding pocket. Researchers were thinking, ATP is ATP, and a thousand different kinases bind ATP, so you would never get specificity but you would get a lot of toxicity. Despite that, by the late 1980s it became clear that ATP-binding site inhibitors could be developed with some specificity. As we learned more about the structure of the kinases, it became evident that there is a fair bit of variability even within the ATP-binding pockets of different kinases, so specificity is, in fact, quite possible.

How did you convince a major pharmaceutical company to develop a drug for a disease that strikes a comparatively small group of roughly 5,000 Americans a year?

Druker: That was certainly the hard part, and a lot of things had to fall into place. We faced three major hurdles. One was that it was a drug for a small market. Second, there was some liver toxicity in early animal tests. For a variety of reasons, Ciba-Geigy was adverse to taking a risk with a potentially toxic drug. But for me, I was dealing with cancer patients who were dying, and I was accustomed to giving drugs with significant toxicity. Thus, I was much less concerned about potential risks. Here was a compound that might help patients, and I absolutely wanted it to be given a fair trial in people. If it proved to be toxic in people, then of course it wouldn’t move forward. But if it worked with a therapeutic window, that would be worth developing.

The last hurdle was that Ciba Geigy had just merged with Sandoz to create Novartis. That led to down-sizing and re-evaluations of all the programs. STI571 wasn’t even on anybody’s radar screen. In addition, Nick Lydon, who was the lead investigator of the tyrosine kinase inhibitor program at Ciba Geigy and the main internal proponent for the development of STI571, had left the company.

Ultimately, I was able to convince a few key people at Novartis to move this compound forward to clinical trials. One of the factors that helped was that this compound also inhibited the platelet-derived growth factor receptor (PDGF-R) tyrosine kinase. The PDGF-R is expressed in most cancers, including most of the common ones, such as breast and prostate cancers. Whether its activity is required for the growth and survival of these cancers is unknown, but if STI571 worked in these diseases, this would be the market size that would interest a large pharmeuctical company. So the view was that if it worked in CML, Novartis had an easy way to the market. It could then be tested broadly and might have a very large market potential.

Why did CML constitute an easy way to test and market a drug?

Druker: CML has several advantages. In the chronic phase of the disease, it is likely driven by BCR-ABL as the sole oncogenic abnormality. In addition, it is a relatively homogeneous disease, with virtually all cases being driven by the BCR-ABL kinase. Thus, in the clinical trials with STI571, an ABL inhibitor, you have the capacity to test patients with a disease that is driven by an activated ABL kinase. If the drug hits the target and shuts it down, it should work. If you hit the target and shut it down and it doesn’t affect the disease, then that means it’s not a very good target. Either way, you learn that very quickly. It either works or it doesn’t. The only way that you lose is if the drug is too toxic.

How quickly did you see an effect in your clinical trials?

Druker: In a phase I study, which is a dose-finding study, the primary endpoint is the safety and tolerability of the drug. This means you start with relatively low doses that you hope might have benefit but are certainly likely to be safe. With STI571, we started at 25 milligrams. We treated three patients and saw no benefits but also no toxicity. A month later we enrolled three more patients at 50 milligrams and so on. By the fourth month, we were seeing some hints of activity. By the time we got to six months, everybody was responding, with all patients having their blood counts returning to normal. That was virtually unheard of in a phase I clinical trial. Usually in a phase I trial, if you see a 20% response rate, that’s remarkable. We had a drug that was extremely well-tolerated and had a 100% response rate. It was absolutely incredible to see this unfold.

Did the results hold up, and for how long?

Druker: That was the big issue. If we treated patients and their blood counts returned to normal for only a month, nobody would really care. So we had to wait see whether our results would be durable. With six months of follow up, everyone treated still had normal blood counts. Even more remarkably, we saw that a significant number of patients were achieving the disappearance of leukemia cells. Thus, we knew we were doing more than just controlling blood counts. We were actually getting rid of the leukemia. As that story began to unfold, it was clear that this was a breakthrough treatment. But what captured the attention of the cancer research community was that this validated the paradigm of targeting molecular pathogenetic events in cancer, with the hope that this paradigm would be applicable to all cancers.

What does it tell you about treating other cancers and about a second-generation Gleevec?

Druker: With patients who relapse while on treatment, we’re finding that they relapse because of mutations in the BCR-ABL tyrosine kinase. The task now is to develop a drug or a set of drugs that will inhibit these mutated kinases, as well as to identify patients who harbor mutations and would be candidates for treatment with Gleevec derivatives. The other task is to identify targets like BCR-ABL in other cancers. For example, we already know that Gleevec is incredibly effective in treating gastrointestinal stromal tumors (GIST), a tumor that is notoriously refractory to chemotherapy. This is because the majority of GISTs are driven by mutations in the KIT tyrosine kinase, another kinase inhibited by Gleevec. In addition, we have been working on another tyrosine kinase, FLT3. This kinase is activated by mutation in about one-third of all cases of acute myeloid leukemia. No less than four drug companies are currently running clinical trials with inhibitors of that particular tyrosine kinase.

How about cancers other than leukemia?

Druker: In most cancers, the problem is that we have much less of an understanding of the critical molecular abnormalities that are driving the growth and survival of the cancer.

How about giving us a 10-year prediction for the future of cancer therapy?

Druker: How about a 50-year prediction? The analogy I like best is the analogy to infectious diseases. If this were the year 1900, we would be talking about how a few years earlier Louis Pasteur had developed a rabies vaccine. We would be excited about the identification of the tuberculosis bacteria by Koch and about other investigators identifying bacteria that caused infections. The pessimist would say, "Okay, that’s great, but how does that help you do anything about it? " And I’d say, maybe some day we’ll develop drugs to specifically attack those bacteria. Then 30 years later we would have sulfa drugs, in 40 years we’d have penicillin, and in 50 years we’d have good vaccines for polio and smallpox. If you look at the landscape of infectious disease in the year 2003, we still have challenges like HIV, but many of the common infections are easily treatable or have been eradicated. That’s been accomplished through public health measures, specific therapies, and vaccines. All of this arose from the knowledge and understanding of infectious disorders.

I believe that we’re in a very similar situation with cancer in the year 2003. A couple months back we heard about a vaccine for cervical cancer. Now we have a drug like Gleevec, which is a specific therapy for cancer. And we continually hear about genes being identified that predispose to various cancers. As we gain a greater understanding of cancer, we will develop better therapies; we have a better understanding of how to harness the power of the immune system to make vaccines or immune therapies. So in the next 10 to 50 years, our ability to treat cancer will become significantly better than it is today, and it will be better through an understanding of cancer at a molecular level.