| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Commentary |
Howard Hughes Medical Institute, Oregon Health and Science University Cancer Institute, Portland, Oregon, USA
Correspondence: Brian J. Druker, M.D., Investigator, Howard Hughes Medical Institute, JELD-WEN Chair of Leukemia Research, Oregon Health and Science University Cancer Institute, L592, 3181 SW Sam Jackson Park Road, Portland, Oregon, USA 97239. Telephone: 503-494-5596; Fax: 503-494-3688; e-mail: drukerb{at}ohsu.edu
Key Words. Imatinib • Gefitinib • Chronic myeloid leukemia • Gastrointestinal stromal tumor • Lung cancer
Approximately 3 years ago, an editorial in The Oncologist by Dr. Bruce Chabner concluded by saying that "STI571, or Gleevec, represents a monumental leap forward in cancer chemotherapy. It proves a principle. It justifies an approach. It demonstrates that highly specific, nontoxic therapy is possible. It does not guarantee success of similar efforts, because CML may not be typical of most other malignancies. And we have much to learn about maximizing its value. Congratulations to Novartis, to Brian Druker, and their colleagues for accomplishing the equivalent of the 4-minute mile. To their colleagues in the fight against cancer, are the floodgates finally open [1]?"
In the past 2 months, two important events have stimulated us to revisit this concluding statement and ask whether the floodgates have opened. The first event is no less than the 50th anniversary of Roger Bannister breaking the 4-minute barrier [2]. The second is the discovery that specific mutations in the epidermal growth factor receptor (EGFR) impart sensitivity to an EGFR inhibitor, gefitinib (Iressa®) [3, 4].
The analogy of imatinib (Gleevec®) to the 4-minute mile is quite apt. For years, there was skepticism that the 4-minute barrier could be broken. Many even postulated that there was a physiological barrier that would prevent anyone from running a mile in less than 4 minutes, but a relatively obscure runner proved that it could be done. Similar to this story, in the development of Gleevec® there was enormous skepticism that inhibiting kinases would be a useful strategy [5, 6]. Biochemists argued that high intracellular concentrations of adenosine triphosphate (ATP) would preclude the use of competitive inhibitors of ATP binding. Chemists argued that the high degree of conservation in the ATP binding pocket of kinases would preclude the development of specific inhibitors. Biologists argued that inhibiting kinases, which have critical roles in cell growth, differentiation, and survival, would have numerous untoward effects on normal cells. Despite this skepticism, Gleevec® has shown that specific kinase inhibitors can be developed, competitive ATP analogues work, and inhibition of kinases is an extremely useful approach in a disease that is caused by a mutation in a kinase [7].
Despite the success of Gleevec® in chronic myeloid leukemia (CML), many have argued that CML was a simple, homogeneous disease and that similar results would not be possible in common solid tumors such as breast, lung, colon, and prostate. However, several examples suggest this skepticism is unfounded. Hematologists and oncologists who treat CML know that CML blast crisis is one of the most treatment-refractory leukemias, yet Gleevec® as a single agent yields a response rate in excess of 50%. Responses to Gleevec® in blast crisis are typically rapid and dramatic, but, unfortunately, rarely durable [8]. In addition, the remarkable results of Gleevec® in gastrointestinal stromal tumors (GIST) and a molecular genetic definition of the subset of non-small cell lung cancer (NSCLC) patients with dramatic responses to Iressa® make it clear that this same paradigm can apply to complex solid tumors.
Before describing the results of these clinical trials, the concept of patient selection in these clinical trials will be examined. Figure 1A
shows a hypothetical set of clinical trials in which therapy with a targeted agent results in an 80% clinical response rate, if and only if the target is present in the tumor. In the four examples shown in this figure, the target frequency varies from 90% to 10%, and the corresponding overall response rates vary from 72% to 8%. Without knowledge of the target frequency in a treated population, one might conclude that the trial with a 72% response rate was successful but the trial with an 8% response rate was a failure. However, in both cases the therapy was equally effective in patients who expressed the target: 80% in both cases. The above example is the simplest possibility for clinical trial design, with target expression sufficient to predict responses. However, a more complicated situation is portrayed in Figure 1B, in which target expression is not sufficient to predict responses, rather, some evidence for target activation could be required and only a subset of patients with target expression would respond. The issue of how target activation should be defined will be explored later. In this example, the target expression is 100%, but target activation varies from 90% to 10%. In patients with target expression and activation, the response rate is again 80%. Thus, the overall response varies from 72% to 8%. If the target expression and activation are high, it might be concluded erroneously that expression correlates with response, whereas if expression is high and activation is low, one might conclude correctly that expression does not correlate with response, but conclude erroneously that the agent lacks activity.
|
is directly applicable to the clinical trials of Gleevec® in GIST and Iressa® in NSCLC. In the former, the majority of patients expressed an activating KIT mutation that is Gleevec® sensitive. Accordingly, the overall response rate was quite high [9, 10]. However, closer inspection of the data clearly showed that KIT mutational status correlated with response to Gleevec®. Thus, patients with activating KIT mutations in exon 11 had a partial response rate of close to 80%. In contrast, patients whose tumors expressed wild-type KIT, with no mutation, had a response rate of only 18% [11]. It has also been instructive to determine why 18% of patients with wild-type KIT expression respond to Gleevec®. Examination of tumors from patients with wild-type KIT expression showed that one-third of these tumors had activating mutations of the platelet-derived growth factor receptor alpha gene [12]. These mutations occurred in two different exons. One set of mutations was Gleevec® sensitive and this accounted for responses observed in patients whose tumors expressed wild-type KIT [12]. The results of Iressa®, an EGFR in NSCLC, have been similarly instructive. Single-agent response rates in heavily pretreated patients are in the 12%–18% range [13, 14]. As the majority of these patients express the EGFR, it has been concluded that EGFR expression does not correlate with response. The modest results in these clinical trials have fueled skepticism regarding the utility of molecularly targeted agents. However, it has always been clear that there is a subset of lung cancer patients who responded dramatically to Iressa®, and these were predominantly nonsmoking women with bronchoalveolar tumors [15]. Two recent reports have shown that this subset of patients has EGFR mutations that render the EGFR hypersensitive to ligand and to Iressa® [3, 4]. Thus, this represents an example where the target expression is high but frequency of activation is low. However, this subgroup of patients has an extremely high response rate and demonstrates in this subgroup of patients that their tumors are dependent on the activity of the mutated EGFR. In both the preceding examples, careful evaluations of subsets of responding patients can yield important insights into disease pathogenesis and the mechanism of response to an agent.
The implications of the results with Gleevec® in GIST and Iressa® in patients with EGFR mutations are extremely important. It would be hard to argue that these are simple cancers. GISTs are extremely heterogeneous tumors, often with complex karyotypes, and responses to multiagent chemotherapy are observed in less than 5% of patients. Although it could be argued that tumors arising in nonsmokers may be less complex than those occurring in smokers, all of the patients enrolled in the Iressa® clinical trials had treatment-refractory disease. Thus, the implications of these results are that even advanced solid tumors have Achilles’ heels and that developing agents that modulate these targets is an extremely effective therapeutic strategy. Of course, it could now be said that this represents only 10% of lung cancers. However, the issue then becomes defining similar Achilles’ heels in the other 90% of lung cancers as well as other tumors.
A concern cited about the Iressa® results is that the market size might become too small to develop drugs as the molecular definition of cancer might divide tissue-specific tumors into numerous smaller molecularly defined entities, all of which would be niche diseases. The counter argument to this is that there could be overlap of molecularly defined entities. For example, there could be similar molecular phenotypes of breast or prostate cancer that would increase market potential of a drug. However, the greatest benefit is that drug development costs could be significantly decreased. The clinical trials of a drug are the most expensive part of the drug development cost, and the risk of failure is high. Imagine if the molecular defects in the subset of lung cancer patients were known in advance of the Iressa® clinical trials. Were this the case, a small study could have been performed in this molecularly defined subset. Dramatic responses would have been observed and approval could likely have been obtained rapidly, and at a fraction of the cost of the actual trials that were performed.
An important conclusion from the Gleevec® and Iressa® clinical trials is the crucial role of the target in defining response. As noted in Figure 1B, expression may not be sufficient to predict responses. Thus, despite the preclinical evidence that targeting the EGFR would be a useful strategy, expression of the EGFR and inhibition of signaling from this receptor were not sufficient to inhibit tumor growth in patients who expressed the EGFR without mutations. This is likely due to the fact that tumors have multiple, redundant pathways that can contribute to their growth. Similarly, agents that target late molecular events in tumorigenesis may be somewhat less effective due to tumor heterogeneity or lack of critical dependence of a tumor on a late event [6]. This does not imply that similar agents will always lack activity or that combinations with other agents might be quite useful. It simply implies that agents targeting proteins expressed in tumors must be distinguished from agents that target early, molecular pathogenetic events in tumors.
The critical dependence of a tumor on a genetically defined pathway, such as BCR-ABL in CML, KIT mutations in GIST, and EGFR mutations in NSCLC, is why Gleevec® and Iressa® have worked so well. This implies that throughout the evolution of the cancer, the tumor remains dependent on these events for its growth and survival and that targeting these molecular abnormalities is an extremely successful approach. This leads to the most important conclusion, that all malignancies will depend on this type of a genetic defect and that it is simply a matter of defining the genetic lesion in each malignancy. In addition, if we could diagnose patients earlier in the course of the disease when tumors are less heterogeneous, it is likely that agents that target these critical molecular defects would be even more effective.
As anyone who runs knows, breaking the 4-minute mile still remains a formidable task. It requires years of arduous training, dedication, and talent and is accomplished by only a handful of elite athletes. Breaking the 4-minute barrier did not make it easier, it just proved it was possible. Similarly, developing a successful agent that targets a causal, molecular event in cancer is not easy. It requires knowledge of the molecular genetic events in a tumor, drug design, and testing. All of this requires years of work from dedicated and talented scientists. The success of Gleevec® will not have made it easier, but it has proven that the concept of targeting specific molecular genetic events in cancer can result in remarkably effective therapies.
Roger Bannister did not achieve his groundbreaking accomplishment alone. He actually was paced by two of his friends for three of the four laps. Similarly, a successful drug requires contributions from many individuals. For Iressa®, this included the discovery of the EGFR by Stanley Cohen, the synthesis and development of Iressa®, and the understanding of which patients derive significant benefit. Regardless, this joins the ranks of the oncologic equivalent of a 4-minute mile. Iressa® shows that targeting specific molecular genetic events, even in advanced solid tumors can result in remarkably effective therapies. Although the floodgates may not be open, it is clear that it is only a matter of time.
 |
REFERENCES |
|---|
 Top References |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. Benesch, M. Windelberg, W. Sauseng, V. Witt, G. Fleischhack, H. Lackner, H. Gadner, U. Bode, and C. Urban Compassionate use of bevacizumab (Avastin(R)) in children and young adults with refractory or recurrent solid tumors Ann. Onc., April 1, 2008; 19(4): 807 - 813. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Chen, X. Zhang, and A. Fernandez Molecular basis for specificity in the druggable kinome: sequence-based analysis Bioinformatics, March 1, 2007; 23(5): 563 - 572. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. S. Kirschner Emerging Treatment Strategies for Adrenocortical Carcinoma: A New Hope J. Clin. Endocrinol. Metab., January 1, 2006; 91(1): 14 - 21. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. A. Carter, L. M. Wodicka, N. P. Shah, A. M. Velasco, M. A. Fabian, D. K. Treiber, Z. V. Milanov, C. E. Atteridge, W. H. Biggs III, P. T. Edeen, et al. Inhibition of drug-resistant mutants of ABL, KIT, and EGF receptor kinases PNAS, August 2, 2005; 102(31): 11011 - 11016. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. WORKMAN Drugging the Cancer Kinome: Progress and Challenges in Developing Personalized Molecular Cancer Therapeutics Cold Spring Harb Symp Quant Biol, January 1, 2005; 70(0): 499 - 515. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | CONTACT US | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| THE ONCOLOGIST | STEM CELLS | CME | ALPHAMED PRESS JOURNALS |
