The development of poly(ADP-ribose) polymerase (PARP) inhibitors as a new class of anticancer agents has created a tremendous amount of hope in the ovarian cancer community, especially in the high-risk, difficult-to-screen, hereditary ovarian cancer population. With the resurrection of the concept of synthetic lethality in recent years, focus has shifted to the development of PARP inhibitors in many different cancers—especially in BRCA-deficient ovarian cancer, both as a single agent and in combination with chemotherapy, radiation therapy, and other targeted therapy. Early results from these trials seem to have left more questions than answers. The article by Zorn et al in this issue of ONCOLOGY focuses on the data so far and provides a balanced perspective into the many challenges involved in the clinical development of PARP inhibitors as a treatment for epithelial ovarian cancer (EOC).
Ovarian cancer remains an important public health concern in the world; it accounts for 140,000 deaths per year worldwide and is the leading cause of gynecological cancer–related mortalities both in the United States and the United Kingdom.[1,2] The landscape of this disease has been plagued by the knowledge that there is no reliable screening technique available that has been shown to improve mortality in this disease. Hence most women present in advanced stages and with 5-year survival rates of 30% to 40%. Approximately 90% of ovarian cancers are epithelial in origin, with subtypes of papillary serous (75%), mucinous (10%); and endometrioid tumors (10%), clear-cell tumors, Brenner (transitional-cell) tumors; and undifferentiated carcinomas. Of these, patients with clear-cell and mucinous types appear to have a worse prognosis than those with other types. In recent years, the pathologic grade of the tumor has been shown to be an important prognostic feature, and high-grade and low-grade tumors have distinct molecular profiles with different mutational characteristics. Approximately 10% of all ovarian cancers have a hereditary basis, and 90% of hereditary ovarian cancers are caused by mutations in BRCA1 and BRCA2. These two genes play an important role in double-stranded DNA repair and thereby in genome integrity through the homologous recombination (HR) pathway (which in turn becomes defective in cells harboring these mutations). This has led to selectively targeting the BRCA-deficient population for the PARP inhibitor trials. PARP is an enzyme in the base excision repair pathway for repairing single-stranded DNA breaks, and inhibition of this pathway for DNA repair in HR-defective cells leads to collapsed replication forks and cell death. This concept of synthetic lethality also explains the higher sensitivity of this group to platinum-based and other DNA-damaging chemotherapy in preclinical studies.[4-6]
What Is Known So Far?
With the initial excitement brought on by the positive iniparib phase II study in triple-negative breast cancer (TNBC), at least half a dozen PARP inhibitors have been actively developed at various phases, as noted in Table 1 and Table 2 of Zorn's article. The initial excitement was dampened by the report that iniparib does not inhibit PARP, and by the subsequent negative results reported from its phase III study in TNBC. However, olaparib had reported single-agent activity in BRCA mutation ovarian and breast cancer patients.[10-13] As the study authors have pointed out, even though the activity of PARP inhibitors was seen in a very specific group of hereditary ovarian cancer patients, these agents may have application to a broader range of cancer types that harbor germline or acquired somatic defects in the HR pathway, a phenomenon more recently described as “BRCAness.” The first proof of this concept came when Gelmon et al showed that women with sporadic recurrent high-grade serous EOC had an objective response rate of 24% to single-agent olaparib.
What About Combination Trials?
With exciting preliminary single-agent activity reported for a few of the PARP inhibitors in BRCA-mutation patients, combination studies of chemotherapy regimens with PARP inhibitors have mushroomed in selected and unselected patient populations, as outlined in Zorn's article.[15-19] However, it remains to be proven whether adding a PARP inhibitor to chemotherapy has an additive or synergistic effect clinically. Many questions need to be addressed in combination trials. Questions about the sequence (concurrent or sequential), the dose, and the duration of PARP inhibitor therapy have not been adequately answered either preclinically or clinically. Preclinical data suggest that chemotherapy followed by PARP inhibition may be more effective.[20,21] The optimal chemotherapy combinations are also not clear. Trials with platinum agents are ongoing, but other agents, such as temozolomide(Drug information on temozolomide) (Temodar), capecitabine(Drug information on capecitabine) (Xeloda), topotecan(Drug information on topotecan), metronomic cyclophosphamide(Drug information on cyclophosphamide), and irinotecan(Drug information on irinotecan), are also being investigated. Is a reduced dose of chemotherapy in combination with a PARP inhibitor as effective as, or more effective than, the chemotherapy alone? Olaparib and veliparib had significant myelosuppression when added to various chemotherapy regimens. Another important question is whether these agents should be given until progression, similar to other targeted agents like bevacizumab(Drug information on bevacizumab) (Avastin), or in predetermined cycles like conventional chemotherapy. In addition, would a continuous or maintenance approach result in a longer progression-free period or in the growth of more resistant tumor populations with compensatory DNA repair mechanisms? What about prior therapy and its effect on response to PARP inhibition? Limited data from the trial by Fong et al indicate that platinum-refractory patients do not appear to respond to olaparib. If that is the case, should this group be excluded from future PARP inhibitor trials? Lastly, although in the olaparib trials higher doses of olaparib were associated with higher response rates, there was no correlation with the degree of PARP inhibition.
What Is the Patient Population? What Are the Biomarkers?
It has become increasingly clear that, as with all other cancers, the answers to most of the questions posed above lie not in conducting innumerable trials in an unselected population but in identifying the selected group for which these agents may be effective, underscoring the need for a reliable biomarker. The work recently published on high-grade serous ovarian cancer by The Cancer Genome Atlas Research Network, involving analysis of messenger RNA and microRNA expression, promoter methylation, DNA copy number, and exome sequences in 489 high-grade serous ovarian adenocarcinomas, has shed some light on the complex biology of this heterogeneous disease. About 50% of the high-grade serous ovarian cancers harbor deficits in the HR DNA repair pathway. In addition, genomic alterations in other HR pathway genes detected could be important in identifying a predictive biomarker signature profile. These include amplification or mutation of EMSY (also known as C11orf30) (8%), focal deletion or mutation of PTEN (7%), hypermethylation of RAD51C (3%), mutation of ATM or ATR (2%), and mutation of Fanconi anemia genes (5%). It is interesting to note that loss of BRCA function as a result of epigenetic changes was also detected, although survival analysis showed numbers similar to those of the wild-type BRCA population rather than the longer survival that has been demonstrated previously in the germline-mutated population. Should patients be tested for defects in the HR pathway? Which defects should be tested? How and what should be tested? Functional assays of DNA repair–like measurement of levels of nuclear RAD51 or γH2AX have been investigated in the past but are technically difficult. The identification of other genomic alterations as described above by The Cancer Genome Atlas group might be useful in this regard.
What Are the Known Resistance Mechanisms?
As with most oncologic therapeutic agents, resistance to PARP inhibitors has been reported. Recent reports suggest that prior exposure and resistance to platinum leads to resistance to PARP inhibitors. One explanation in the BRCA-mutation population may be the development of a secondary mutation that restores the BRCA function.[25-29] It is noteworthy that patients who failed to respond to olaparib may subsequently responded to platinum chemotherapy.
Are All PARP Inhibitors the Same?
Lastly, given the number of PARP inhibitors in clinical development, is there a difference among the various PARP inhibitors? Although attempts at pharmacodynamic comparisons of the different PARP inhibitors have been made, to date there have been no direct or indirect comparisons in the clinical arena. Each drug may have off-target effects that may alter the results of a large clinical trial, as suggested by the olaparib trials. Although PARP was inhibited to a significant degree at both the 100-mg and 400-mg dose levels, a difference in activity was seen, suggesting that other mechanisms may be at work. With more than one PARP inhibitor showing activity in the BRCA-mutation population, should all of them be developed for use in this small population of patients? Also, what standard should be used to compare the activity of these agents, given that there are no historic data to provide guidance?
While consideration of chemoprevention with PARP inhibitors is on the horizon, many knowledge gaps exist regarding these agents. Although the trials in EOC provided some answers regarding the activity of PARP inhibitors, they raised many other questions. These questions may actually complicate the picture as newer agents in this drug class make their way to the clinical arena. A collaborative approach among the researchers is needed to systematically answer these questions so we are better equipped to provide effective treatment to the BRCA-deficient patients. It is noteworthy that The Cancer Genome Atlas Group analysis did reveal other commonly deregulated pathways in this disease—such as RB, RAS/PI3K, FOXM1, and NOTCH—that might provide future opportunities for therapeutic targeting while the story of the PARP inhibitors continues to unfold.
Financial Disclosure: The authors have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article. This article is declared a work of the US Government and is not subject to copyright protection in the United States.