Lasker~DeBakey C l i n i c a l m e d i c a l COMMENTARY research award

Perspectives on the development of imatinib and the future of

Brian J Druker

Imatinib: a personal history I am incredibly fortunate. To receive an award from the Lasker Foundation, with its preemi- nent jury, is indeed an honor. But the greatest reward is seeing patients every week in clinic who have benefitted from my work. The life expectancy of these patients was originally three to five years. They are now leading active and productive lives, and many have been with me for over ten years. My desire to pursue a career in oncology began when I was a medical student, when I took an elective course on the history of . We learned about the cure of childhood with combination che- motherapy and about the contributions of pioneers such as Sidney Farber and Lasker laureates Emil Frei and Emil Freireich. Despite these breakthroughs, I imagined there had to be a better way to treat cancer, and I wrote in my final essay that only through a molecular understanding of cancer would we be able to more rationally treat the disease. Even though the prognosis for most cancers was quite poor, it seemed to me that there would be great opportunities in cancer research over the next Figure 1 Members of the Roberts lab. The males in Tom’s lab attempted to dress as physicians in honor 20 to 30 years, and I wanted to be a part of of Brian’s birthday in 1988. Brian and Tom are both in the top row; Brian is first from left and Tom is this. second from right. When I started my training in oncology at Dana-Farber Cancer Institute, we spent most of our time trying to extend our patients’, lives, but battling chronic myeloid leukemia (CML), we MD postdoctoral fellow. I entered a lab filled our tools were limited. We had 5-flurouracil for dreaded each visit. We checked blood counts, with incredible postdocs and graduate stu- colon cancer, doxorubicin (Adriamycin) and looking for the inevitable signs of progression dents who became my mentors and friends. one of the first targeted cancer agents, tamox- of the disease to a fatal and untreatable acute This group included Helen Piwnica-Worms, ifen, for breast cancer. For prostate cancer, we leukemia, knowing we would have to deliver David Kaplan, Michael Corbley, Tony West, were testing long-acting androgen antagonists harsh news. With this background, I decided Connie Gee, Harvey Mamom, Kenneth Wood as an alternative to castration. As for patients to pursue a laboratory career and pledged to and David Pallas (Fig. 1). All were extremely continue the work until I was confident that patient and helpful in advising a relatively green Brian J. Druker is an Investigator of the Howard I had something to contribute that was a vast trainee. We also worked closely with the labs Hughes Medical Institute, JELD-WEN Chair of improvement over what was then available. of David Livingston and Charles Stiles and had Leukemia Research, and Director of the Oregon In early 1985, I met with Thomas Roberts an ongoing collaboration with Lewis Cantley. Health & Science University Knight Cancer at Dana-Farber on the advice of George It was an exciting time in Tom’s lab, as many Institute, Portland, Oregon, USA. Khoury1. Despite my limited lab background, insights into kinase signaling pathways were e-mail: [email protected] Tom was willing to take a chance on his first being discovered. My initial project was to work

nature medicine volume 15 | number 10 | october 2009 xv commentary

100 lar. By early 1999, six months after beginning Gleevec the phase 1 study, we reached effective doses, and virtually all of our patients were respond- 80 ing and experiencing few, if any, side effects4. Despite my own concerns about whether Interferon responses would be durable, the patients al (%) 60

viv embraced this new therapy. All they knew is that their blood counts were normal for the 40 first time in years. They felt better than they Chemotherapy erall sur had in a long time and had their hope for the Ov future was restored. Internet chat rooms were 20 a new phenomenon, and patients were describ- ing their experiences with imatinib even before

0 we had presented clinical data or published our 0 6 12 18 24 30 36 42 48 54 60 66 72 results—and the results seemed too good to be Time (months) after beginning of treatment true. Now, with five years of follow-up, survival for people with CML treated with imatinib is Figure 2 Survival of patients with CML. Curves show overall survival for patients with CML treated with 89%, compared to less than 50% with previous imatinib (Gleevec)5, interferon or conventional chemotherapy24. Conventional chemotherapy delivers 5 only a minimal effect on survival compared with no treatment.Thanks to C. Lynm for preparing the therapies (Figs. 2 and 3). In addition, the risk figure. of relapse has decreased over time, such that no subject randomized to imatinib in a compara- on polyoma middle-T antigen and its mecha- nate to be recruited by Grover Bagby at Oregon tive study of imatinib versus interferon plus nism of transformation, including its inter- Health & Science University, who shared my cytarabine has relapsed in years 6 and 7 after action with tyrosine kinases. Soon, Deborah vision for targeted therapies. When I moved to starting therapy6. Morrison joined the lab, and, with her expertise Oregon, I had one goal: to find a drug company and guidance in monoclonal antibody produc- that had a BCR-ABL kinase inhibitor and to Translating the success of imatinib to tion, I took on a project to develop an antibody get it into clinic. I was lucky enough that after other malignancies to phosphotyrosine. This eventually resulted in only one phone call—to Nick Lydon—I had It was an understanding of the target that led the production of 4G10 (ref. 2). My involve- found a willing collaborator with candidate to the success of imatinib. What distinguishes ment with this antibody led to numerous col- compounds who shared my passion to move imatinib from chemotherapy is that we know laborations with investigators working on a drug into the clinic. which patients will benefit from this drug and tyrosine kinase signaling, and I rapidly gained Over the course of the next several years, why. This was made possible by numerous expertise in this field. When Nick Lydon and things moved reasonably smoothly. Of the scientific discoveries that allowed a precise Alex Matter at Ciba-Geigy began working with compounds Nick sent me for testing, I iden- understanding of the molecular pathogenesis Tom’s lab on their kinase inhibitor project, Nick tified CGP 57148B (also known as STI571, of CML7. To achieve quantum improvements and I formed a natural partnership because of Gleevec, Glivec or imatinib mesylate) as the in cancer outcomes, it is necessary to have a my development of the 4G10 antibody, which most effective at killing CML cells without thorough and comprehensive understanding was an essential reagent in their screening pro- harming normal cells3. Nick and I worked of what distinguishes cancer cells from normal gram for kinase inhibitors. closely during that time, sharing data, devising cells, including interactions with surrounding By 1990, I was at a crossroads in my career. experiments and timelines, outlining clinical stromal cells. The field is transitioning from The polyoma project was moving slowly. With development plans and discussing challenges. an era of empirically based cancer therapy to my background in oncology and my emerging Despite the promising preclinical data, there one based on a precise understanding of the expertise in tyrosine kinase signaling, it seemed were still hurdles to overcome before clinical molecular defects in cancer. In the not-too- to me that I should be working on a human trials commenced. These included concerns distant future, clinicians will be able to thor- disease caused by a tyrosine kinase. At the time, about toxicity, whether targeting a single oughly analyze individuals’ tumors for molecu- the BCR-ABL tyrosine kinase was the best- kinase would be an effective anticancer strat- lar defects and match each person with specific, credentialed tyrosine kinase target in oncology. egy, and whether Ciba-Geigy (now Novartis) effective therapies that will yield a durable A couple of years earlier, I had advised Nick that would realize a return on their investment for response with minimal toxicity. CML, driven by BCR-ABL kinase activity, was a small-market disease such as CML. In addi- To speed this transition, there is a clear need likely to be the first and best cancer in which tion, Nick had left Novartis to form his own for more funding for cancer research. But to validate the paradigm of kinase inhibition. biotech company. At the time, I was caring for researchers also need to reevaluate how funds With this in mind, I approached James Griffin, patients in my clinic with CML who had no are spent and to better coordinate their efforts. with whom I had already been collaborating treatment options remaining. This connection They need to examine the entire process from and, with his guidance, began to develop model was significant. I became their voice, lobbying target identification to a drug becoming the systems to study BCR-ABL signaling. my remaining contacts at Novartis, includ- standard of care and determine, for example, In 1993, I reached another crossroads. I had ing Alex Matter and Elisabeth Buchdunger, where knowledge deficits exist but technology established various BCR-ABL–driven model to move this project forward. Ultimately, we is available to fill in these gaps. These would systems and had even written a grant with Jim prevailed. be areas where large investments would yield Griffin outlining how to characterize BCR-ABL As Nick writes in his accompanying com- considerable payoffs. There will be other areas kinase inhibitors. I was ready to put my knowl- mentary (pages xix–xxiii), the clinical results where technology has not yet been developed edge to work in my own lab, and I was fortu- with imatinib were nothing short of spectacu- to circumvent the hurdles. In these areas, sub-

xvi volume 15 | number 10 | october 2009 nature medicine commentary stantial and broad investments need to be made that will allow exploration and inno- vation to find solutions. Researchers already know that there are procedural barriers that slow their progress, and they must work to fix these. Finally, they need to put in place the infrastructure required for a future where per- sonalized cancer therapy is a reality. Workers in the field must have a sense of urgency to their mission, never forgetting the 1.5 million individuals diagnosed and the 500,000 who die of cancer each year and are in desperate need of better therapies. One area where an immediate large invest- ment would yield a major payoff would be the sequencing of 25,000 pairs of tumor and normal genomes (estimated cost of $1 billion, at $20,000 per genome). This would need to be followed with a similarly massive effort to carry out functional characterization and validation of the targets that might emerge Individuals with CML and their families participating in a Leukemia and Lymphoma Society from this effort. With this, the way clinicians Figure 3 event. Many of the patients in the photo were treated in the phase 1 and early phase 2 clinical trials of understand, classify and treat cancer would imatinib. change forever. Sequencing cancer genomes and validating targets will not be the panacea for cancer. Some vulnerabilities to cancer will As part of their efforts to accelerate progress, cancer will be too small for companies to see be epigenetically based, others will be based on investigators must find ways to encourage aca- potential market benefit. Early genomic efforts metabolic susceptibilities, and still others will demic and industrial collaborations. The devel- have suggested that, although large numbers of be based on interactions between tumors and opment of imatinib is an outstanding example mutated cancer genes may be identified from surrounding stroma. Finally, some genomic of how academic investigators and industry can these screens, the numbers of pathways impli- changes will be in genes that will be difficult form productive collaborations. We each did cated in cancer may be finite11–13. Regardless, to target. Understanding each of these issues what we do well. Academics had identified and a new paradigm needs to be developed for will require substantial effort while providing validated a target, and I had established all the cancer drug development that allows for rapid key opportunities for investigator-initiated model systems to evaluate compounds. Nick’s approval of drugs that achieve high response studies. group had developed compounds and was rates in a well-defined patient population with The field will continue to confront molecu- willing to share their compounds for testing. an unmet medical need. This would keep drug lar defects that are difficult to target, includ- Then, academics carried out the clinical trials development costs down but would require ing well-validated targets such as KRAS, RB1 in collaboration with industry. Unfortunately, additional changes in postapproval safety and TP53. Two basic-research strategies may these types of relationships are hampered by a monitoring, reimbursement policies for off- assist in overcoming this type of hurdle. One culture that has evolved in both academia and label use, and a much more nimble clinical- is an understanding of signaling mechanisms industry. An example of this is that negotiations trials system. Much of this was achieved with affected by tumor-causing genes. For example, go on for months over material transfer agree- imatinib in CML and gastrointestinal stromal in lung cancer with mutated KRAS, it has been ments, which only slows progress. Academic tumors, but, even for these trials, the approval suggested that targeting the phosphatidylinos- institutions want to share in the profits, and process could have been faster. itol-3-kinase and RAF kinase pathways would drug companies want to preserve their ability The clinical trials process is cumbersome be equivalent to KRAS inhibition8. This strat- to recoup an investment. What is desperately and needs revamping. Thirty percent of clinical egy may be tumor specific, as our lab has seen needed is a uniform material transfer agree- trials at most academic medical centers do not substantial synergy in acute myeloid leukemia ment so that inordinate amounts of time and enroll a single subject14, and many others are cell lines with these combinations, much more effort are not wasted on paperwork. It took me not published15. The time from protocol con- so than in pancreatic cancer cell lines with six weeks to get imatinib into my lab. At the ception to enrollment of the first subject can be KRAS mutations. A second approach has used time, it was already patented against any tumor as long as two-and-a-half years16. Funding for a synthetic lethal screen to identify two kinases in a warm-blooded mammal. Had my insti- cooperative group studies from the National downstream from RAS that may allow an effec- tution negotiated for royalty rights, it would Cancer Institute does not come close to cov- tive drug development strategy. The synthetic have delayed progress at best; at worst, this ering costs. What is needed is a much more lethal strategy has already shown promise with compound would never have made it to my efficient system that focuses on fewer studies the finding that inhibition of poly(ADP-ribose) lab or into the clinic. that are well funded, with a quick determina- polymerase is an effective treatment strategy As medicine move into an era of personal- tion of whether additional investments are for cancers arising in carriers of a BRCA1 or ized cancer therapeutics, where cancers are worthwhile. BRCA2 mutation or tumors that share features defined by molecular targets instead of site Since the first clinical results of imatinib with BRCA1-related cancers, such as triple- of origin, one concern is that the number of were presented ten years ago, hundreds of negative breast cancer9,10. individuals with a specific molecularly defined targeted drugs have entered clinical trials,

nature medicine volume 15 | number 10 | october 2009 xvii commentary with numerous examples showing potential, diagnostic technologies, improved outcomes the balance they provide to my life. I am extremely even in patients with advanced solid tumors. will also be possible, even with single-agent appreciative of the assistance from A. Hardy on the writing of this manuscript. This award was made For example, the response rate to imatinib therapy. possible by the hundreds of people who have assisted for individuals with metastatic gastrointes- Imatinib has created considerable excite- me in my career. This includes all of my current and tinal stromal tumors is close to 60%, with a ment about targeted therapies in both aca- former laboratory staff, my colleagues and mentors durability of over two years17. Similar results demia and industry. There is still much work at Oregon Health & Science University, the clinical are seen with imatinib in the small percentage to be done in defining molecular patho- faculty, and our nurses and data managers. I would like to thank my mentors from the Dana-Farber of patients with melanoma who harbor KIT genetic events in cancer that will allow for Cancer Institute and the dedicated scientific and mutations18. Rapid and profound responses rational combinations of targeted therapies, clinical staff at Novartis, who shepherded imatinib to epidermal growth factor receptor inhibitors resulting in improved outcomes for patients. through clinical trials. I would also like to thank the have also been observed in patients with lung Targeted therapies are important but will not numerous investigators who enrolled subjects on our 19,20 clinical trials. During this time, I have been supported cancers driven by EGFR mutations . At the become the exclusive approach to cancer. by various funding agencies, including the National annual meeting of the American Society of As clinicians have learned from combating Cancer Institute, the Leukemia and Lymphoma Clinical Oncology in 2009, there was a pre- infectious diseases over the past century, they Society, the Burroughs Wellcome Fund, the T.J. Martell sentation on an ALK tyrosine kinase inhibitor cannot adopt a one-size-fits-all approach. In Foundation, the Doris Duke Charitable Foundation, targeting an ALK fusion protein present in the 1940s, there was enormous excitement and the American Cancer Society and the Howard Hughes Medical Institute. I am grateful for this people with lung cancer that elicited rapid over antibiotics as a cure-all for infections. support. Lastly, I would like to thank my patients who and dramatic responses21. Similarly, a potent But there were several events that led to have gone on this incredible journey with me. inhibitor of mutant B-RAF showed consider- the treatability or eradication of infections. 1. Livingston, D.M. Cell 49, 577 (1987). able activity in individuals with melanoma Antibiotics were an obvious breakthrough; 2. Druker, B.J., Mamon, H.J. & Roberts, T.M. N. Engl. J. and BRAF mutations22. vaccinations, another. Public-health mea- Med. 321, 1383–1391 (1989). In these examples of advanced-phase sures such as chlorination of water and pas- 3. Druker, B.J. et al. Nat. Med. 2, 561–566 (1996). 4. Druker, B.J. et al. N. Engl. J. Med. 344, 1031–1037 malignancies, relapses and resistance, as teurization of milk were major contributions. (2001). described in Charles Sawyers’s accompa- Recast for cancer, these three measures are 5. Druker, B.J. et al. N. Engl. J. Med. 355, 2408–2417 (2006). nying commentary (pages xxiv–xxvii), has targeted therapy, immune modulation and 6. O’Brien, S.G. et al. Blood 112 (ASH Annual Meeting been common. For now, combinations of preventative measures. Targeted therapies Abstracts), 186 (2008). targeted agents with chemotherapy will be will have a key role, but these will need to 7. Druker, B.J. Blood 112, 4808–4817 (2008). 8. Engelman, J.A. et al. Nat. Med. 14, 1351–1356 used. As our understanding of the molecu- be broadly directed to genetic or epigenetic (2008). lar pathogenesis of cancer improves, it will changes in tumors, tumor metabolism, stem 9. Fong, P.C. et al. N. Engl. J. Med. 361, 123–134 (2009). be possible to rationally combine targeted cells and tumor-stroma interactions. To effec- 10. O’Shaughnessy, J. et al. J. Clin. Oncol. 27(Suppl), 3 therapies. For example, addition of both tively manage cancer, clinicians also need (2009). trastuzumab to the aromatase inhibitor innovations in immune therapy and early 11. Cancer Genome Atlas Research Network. Nature 455, 1061–1068 (2008). anastrozole and the tyrosine kinase inhibi- diagnostics. Ultimately, some of the genetic 12. Jones, S. et al. Science 321, 1801–1806 (2008). tor lapatinib to letrozole, have markedly and epigenetic changes that lead to cancer 13. Parsons, D.W. et al. Science 321, 1807–1812 (2008). 14. Dilts, D.M. et al. J. Clin. Oncol. 26(Suppl), 6543 improved progression-free survival in indi- will have been programmed at birth, and an (2008). viduals with advanced breast cancer positive understanding of these changes will eventu- 15. Ramsey, S. & Scoggins, J. Oncologist 13, 925–929 for estrogen receptor and HER2 (ref. 23). The ally allow molecularly targeted preventative (2008). 16. Dilts, D.M. et al. J. Clin. Oncol. 24, 4553–4557 challenges will be to determine what combi- therapies. (2006). nations of targeted therapies will work best, I am truly fortunate to have witnessed 17. Blanke, C.D. et al. J. Clin. Oncol. 26, 620–625 (2008). what resistance mechanisms will need to be the advances in cancer therapy over the past 18. Hodi, F.S. et al. J. Clin. Oncol. 26, 2046–2051 circumvented, how to serially analyze solid decade. Patients and physicians are far more (2008). tumors for molecular defects using noninva- optimistic about our ability to treat cancer. 19. Lynch, T.J. et al. N. Engl. J. Med. 350, 2129–2139 (2004). sive technologies and how to serially analyze Clinicians and scientists now have the tools 20. Paez, J.G. et al. Science 304, 1497–1500 (2004). for modulation of targets in vivo. As learned at hand to turn cancer into a manageable 21. Kwak, E.L. et al. J. Clin. Oncol. 27(Suppl), 3509 from the imatinib example, earlier treatment chronic disease. I am encouraged by what will (2009). 22. Flaherty, K. et al. J. Clin. Oncol. 27(Suppl), 9000 in the course of the disease, when there is be accomplished in the coming decades. (2009). less tumor heterogeneity and less chance for 23. Johnston, S.R. Clin. Breast Cancer 9, S28–S36 Acknowledgments (2009). resistant mutations to be present, will pro- I am indebted to my wife Alexandra and children, 24. The Italian Cooperative Study Group On Chronic Myeloid duce better results. So, with improved early Holden, Julia and Claire for their love and support and Leukemia. N. Engl. J. Med. 330, 820–825 (1994).

xviii volume 15 | number 10 | october 2009 nature medicine