focus on cancer
news and views
Targeting the missing links for cancer therapy Kornelia Polyak & Judy Garber
© 2011 Nature America, Inc. All rights reserved.
A continuing quest in clinical oncology is to effectively eliminate tumors without major side effects. But drugs rationally tailored against specific tumors and predictive markers for patient selection are very limited, and their identification is challenging. A phase 1 study has provided proof of concept for the use of PARP inhibitors in tumors from individuals carrying BRCA mutations—a remarkable success in rational drug design and translational research. A quote by Paracelsus, the father of toxicology, “it is the dose that determines that a thing is not a poison,” highlights one crucial bottleneck of anticancer drug development: the need to identify compounds that eliminate cancer cells at concentrations harmless to the individual. Most traditional chemotherapeutic agents were identified on the basis of their ability to kill fast-growing cancer cells in in vitro models. Not surprisingly, they also eliminate rapidly dividing normal cells, such as hematopoietic and intestinal epithelial progenitors, leading to serious side effects that limit their effective dosing. But even these imperfect drugs can eradicate a subset of tumors, including testicular cancer and many childhood leukemias. Their efficacy in other tumor types, however, is fairly limited, especially in advanced disease. A major goal of cancer research in the past decades has therefore been to identify cancer-specific molecular alterations as targets for rational drug design and biomarkers to help select those people who will probably benefit from the treatment. Tumorigenesis is an evolutionary process driven by the accumulation of genetic and epigenetic alterations that create a diverse population of cancer cells, of which the most successful are continuously being selected by the tumor microenvironment1. Thus, drugs that selectively eliminate cancer cells can be developed in three general ways. First, we can identify and target mutant or abnormally expressed proteins that are only present in Kornelia Polyak and Judy Garber are at the Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, Massachusetts, USA and the Department of Medicine, Harvard Medical School and Brigham and Women’s Hospital, Boston, Massachusetts, USA. e-mail:
[email protected]
c ancer cells. Successful agents of this type in the clinic include imatinib, targeting BCRABL oncoprotein in chronic myelogenous leukemia, and trastuzumab, targeting HER2 oncoprotein in a subset of breast tumors2. Recent whole-genome sequencing projects aimed to identify additional cancer-specific mutations that can be exploited therapeutically, but thus far there has been little success. Second, the microenvironment could be altered to reduce the fitness of cancer cells by limiting infiltration of leukocytes that secrete cytokines promoting tumor cell survival or angiogenesis with anti-inflammatory and antiangiogenic drugs. The third approach may consist of identifying and targeting genes and pathways necessary for cancer cells because of their abnormal cellular milieu caused by genetic and epigenetic abnormalities—even targeting proteins other than the mutant protein per se. The so-called synthetic lethal strategy belongs to this general approach3, where two genes show a synthetic lethal relationship if the elimination of either one is compatible with cellular survival but the inhibition of both is not. A cancer-specific mutation in one of two genes in synthetic lethality combined with the therapeutic targeting of the other one will selectively kill tumors without harming normal cells. Previous studies showed that cancer cells deficient in either BRCA1 or BRCA2 (the breast-cancer–associated proteins) are specifically sensitive to poly(adenosine diphosphate (ADP)-ribose) polymerase (PARP) inhibition4,5, which is a family of enzymes involved in base-excision repair, a key pathway in DNA single-strand break repair. Inhibition of PARP activity led to accumulation of nuclear RAD51 foci, indicating the presence of double-strand breaks that are normally repaired by homologous recombination6.
nature medicine volume 17 | number 3 | MARCH 2011
Cells with defective homologous recombination, such as those lacking BRCA1 or BRCA2, are therefore killed by PARP inhibitors, whereas normal cells are relatively unaffected (Fig. 1). PARP inhibitors can then be used as a different therapeutic strategy for the treatment of tumors that lack BRCA function. The availability of a highly effective and well-tolerated oral PARP inhibitor, olaparib, developed by Kudos and acquired by AstraZeneca (AZD2281), allowed the rapid translation of these findings into a clinical trial7. One of the first studies showing the potential of synthetic lethal strategy in clinical trials in people with cancer was published in the New England Journal of Medicine by Fong et al.7. They designed a phase 1 clinical trial to test the pharmacokinetic and pharmacodynamic profiles of olaparib in a cohort of individuals enriched for BRCA1 or BRCA2 mutation carriers7 independent of their type of tumor. The authors tested six different treatment regimens using a range of drug concentrations (100–600 mg) and compared continuous daily drug administration with drug administration in two of every 3 weeks (2 weeks of daily treatment, 1 week with no treatment)7. On the basis of dose-limiting toxicity, they established 600 mg and 400 mg as the maximum administered and tolerated doses, respectively. Among the 60 patients enrolled in the study, only the 22 BRCA1 or BRCA2 mutation carriers showed an objective antitumor response, and 12 out of 19 mutation carriers who had breast, ovarian or prostate tumors, had a clinical benefit from PARP inhibitor treatment, with remarkable responses being achieved in a few cases7. The side effects, mostly nausea and vomiting, were relatively mild at doses where effective antitumor responses (decreases in tumor size) were observed, although myelosuppression was also detected. 283
news and views
BRCA1/2 mutation carriers (BRCA1/2+/–)
Breast tumor
Normal cell BRCA1/2+/– SSB
PARP inhibitor
DSB
Tumor cell BRCA1/2+/–
LOH
Active BRCA1/2
Active BRCA1/2
Katie Vicari
Resistant tumor cell
PARP inhibitor Synthetic lethality
© 2011 Nature America, Inc. All rights reserved.
Cell survival
Sensitive tumor cell
Tumor cell BRCA1/2–/–
Figure 1 Synthetic lethality in tumors from BRCA1 and BRCA2 mutation carriers treated with PARP inhibitors. DNA repair pathways are categorized into single and double-strand break–specific mechanisms, both of which are further divided into subgroups. PARP inhibitors block the repair of single-strand breaks (SSBs), which if left unrepaired are converted to double-strand breaks (DSBs) during replication. In normal cells of BRCA1 and BRCA2 mutation carriers (BRCA1/2+/–), these DSB lesions are repaired by homologous recombination because one copy of BRCA1 or BRCA2 is sufficient for repair proficiency, and the cells remain viable. However, in cells with defective homologous recombination, such as tumor cells in BRCA mutation carriers that lost the wild-type copy of BRCA by loss of heterozygosity (LOH), double-strand breaks cannot be efficiently repaired, leading to cancer cell death and elimination of the tumor. Resistance may arise because of the presence of tumor cells that retained a wild-type copy of BRCA or mutations in BRCA or other genes that restore repair proficiency or overcome PARP inhibition by other mechanisms.
Such exciting results have led to the use of other PARP inhibitors in clinical trials alone or in combination with other chemotherapeutic drugs, such as carboplatin and cisplatin, gemcitabine, irinotecan and temazolamide, in diverse patient populations: individuals with BRCA1 and/or BRCA2 mutations, triple-negative (negative for estrogen receptor, progesterone receptor, and HER2) breast cancer, serous ovarian cancer, glioblastomas and pancreatic cancer, among others. A recent study showed impressive clinical responses in people with triple-negative breast cancer treated with a combination of the intravenous PARP inhibitor iniparib and carboplatin and gemcitabine8,9. Unfortunately, whereas the synthetic lethality concept was relatively easy to show in BRCA1 and BRCA2 mutation carriers, finding effective and nontoxic combinations, bioavailable formulations and delivery schedules, and molecular profiles of sensitive tumors remains more empiric than would be ideal. Despite the remarkable success of PARP inhibitors in early trials, numerous questions remain unanswered. Not every tumor in BRCA mutation carriers responds to treatment, suggesting that additional factors besides the loss of these genes define cellular
284
response. Loss of the gene encoding the phosphatase and tensin homolog (PTEN) is one of these candidates, as it seems to confer sensitivity to PARP inhibitors even in wild-type BRCA1 and BRCA2 tumor cells, at least in in vitro models10. In addition, PARP inhibitors may be used to treat non-BRCA1 or non-BRCA2 mutated tumors if similar or related defects in DNA damage repair response can be identified or if DNA damage with specific cytotoxic chemo therapeutic agents can be induced when the cell’s ability to repair DNA using the PARP inhibitor is compromised. Which agents would be combined with PARP inhibitors with the best therapeutic index (that is, the most effective and least toxic) and which biomarkers would allow selection of those individuals who would benefit from this treatment remain to be determined. Most tumors eventually escape from PARP inhibition, indicating preexisting or induced resistance11. A study showed that cells from BRCA2 mutation carriers resistant to cisplatin and PARP inhibitors have an interesting mechanism of resistance that restores wildtype BRCA2 function by correcting the defect in homologous recombination, leading to loss of PARP inhibitor sensitivity12.
Although several PARP inhibitors have been developed with specificities for inhibition of different PARPs and nonidentical mechanisms of action, specificity to PARP inhibition does not always correlate with the degree of clinical response, suggesting potential off-target effects. For example, in contrast to olaparib and most other PARP inhibitors, iniparib does not inhibit PARP’s enzymatic activity but rather blocks the interaction of PARP with DNA13. The in vitro potency of iniparib is also much lower (micromolar range) compared to olaparib (low nanomolar range), although presumably a more active metabolite must exist, as the serum half-life of iniparib is very short (~4 min). Because this metabolite has not been conclusively identified, the in vivo targets of this drug are even less clear, although no other targets aside from PARP have been reported. There is a concern for potential long-term toxicities from exposure of healthy tissues to agents inhibiting DNA repair, such as PARP inhibitors, including the induction of secondary tumors or leukemias. These issues can be crucial, as the agents are developed to cure early-stage tumors—through adjuvant and neoadjuvant therapies—and even to reduce cancer risk in BRCA1 and BRCA2 mutation carriers, who have substantial risk of developing lethal ovarian, breast and pancreatic malignancies. In light of the large number of clinical trials ongoing or planned with various PARP inhibitors, the answers to many of these questions will begin to emerge in the near future. COMPETING FINANCIAL INTERESTS The authors declare competing financial interests: details accompany the full-text HTML version of the paper at http://www.nature.com/naturemedicine/. 1. Merlo, L.M., Pepper, J.W., Reid, B.J. & Maley, C.C. Nat. Rev. Cancer 6, 924–935 (2006). 2. Murdoch, D. & Sager, J. Curr. Opin. Oncol. 20, 104–111 (2008). 3. Kaelin, W.G. Jr. Nat. Rev. Cancer 5, 689–698 (2005). 4. Farmer, H. et al. Nature 434, 917–921 (2005). 5. Bryant, H.E. et al. Nature 434, 913–917 (2005). 6. Schultz, N., Lopez, E., Saleh-Gohari, N. & Helleday, T. Nucleic Acids Res. 31, 4959–4964 (2003). 7. Fong, P.C. et al. N. Engl. J. Med. 361, 123–134 (2009). 8. Carey, L.A. & Sharpless, N.E. N. Engl. J. Med. 364, 277–279 (2011). 9. O’Shaughnessy, J. et al. N. Engl. J. Med. 364, 205–214 (2011). 10. Dedes, K.J. et al. Sci. Transl. Med. 2, 53ra75 (2010). 11. Ashworth, A. Cancer Res. 68, 10021–10023 (2008). 12. Edwards, S.L. et al. Nature 451, 1111–1115 (2008). 13. Ferraris, D.V. J. Med. Chem. 53, 4561–4584 (2010).
volume 17 | number 3 | MARCH 2011 nature medicine
