Personalizing Therapy for Ovarian Cancer: BRCAness and Beyond

Although 70% to 80% of epithelial ovarian cancers respond to chemotherapy, less than 30% of patients can be cured with cytoreductive surgery combined with the best currently available drugs. Over the last three decades, the median survival for patients with ovarian cancer has improved significantly, but the plateau of the survival curve has not changed appreciably. Empirical combinations of up to five individually active drugs have proven no more effective in extending disease-free or overall survival than a two-drug combination of carboplatin and paclitaxel. In this setting, many investigators believe that the most effective strategy to improve outcomes is to personalize therapy, targeting those abnormalities that occur in each individual ovarian cancer and avoiding drugs that will cause excessive toxicity in that particular patient. Striking responses to poly(adenosine diphosphate-ribose) polymerase (PARP) inhibitors in patients with inactivating germline mutations of BRCA1 or BRCA2 have provided one of the best examples to date of personalized therapy for ovarian cancer. During transformation of cancer cells in mutation carriers, the second functional allele of BRCA1 or BRCA2 is lost, inhibiting the ability to repair double-strand breaks in DNA through homologous recombination (HR). In this setting, cancer cells are selectively sensitized to inhibitors of other DNA repair pathways, including base excision repair, a process that requires PARP. Data from phase I and II trials of PARP inhibitors in women with germline mutations of BRCA1 and BRCA2 are quite promising, with a 40% objective response rate as single agents in ovarian and breast cancers and acceptable toxicity. Because only 10% to 15% of women who develop ovarian cancer carry germline mutations of BRCA1 or BRCA2, PARP inhibitors were initially expected to benefit only a small, but important, subset of ovarian cancer patients with a strong family history of breast and ovarian cancer. However, recent data suggest that dysfunction of BRCA1 and BRCA2, so-called BRCAness, may be more prevalent than originally assumed. BRCA1 and BRCA2 mutations have been found in up to 20% of unselected ovarian cancers. BRCA1 promoter methylation has been reported to markedly decrease BRCA1 expression in 5% to 31% of sporadic epithelial ovarian cancers. Amplification of EMSY, an inhibitor of BRCA2, has been observed in 17% of sporadic cases. Furthermore, inactivating mutations and promoter methylation can disrupt the function or expression of other genes that contribute to HR. Thus, both genetic and epigenetic mechanisms can create the BRCAness phenotype in at least a third of all epithelial ovarian cancers, with the expectation that these patients would benefit from PARP inhibitors. The challenge remains to develop a convenient and accurate method to identify ovarian cancers with BRCAness and, in particular, to identify patients likely to benefit from PARP inhibitor therapy. The cost of high-throughput sequencing is being halved every 6 months, supporting the expectation that the thousand-dollar genome challenge will be fulfilled in the near future, permitting the cost-effective sequencing of cancers. Coupling next-generation sequencing with other analytic approaches will also make measurement of DNA methylation in multiple promoters practical. At present, judging BRCAness by sequencing is a challenge because multiple genes are involved in the HR pathway, BRCA1 and BRCA2 are large genes, and somatic mutations do not seem limited to particular nucleotides or hot spots. An alternative approach has been to look downstream of the fundamental genetic or epigenetic changes to identify a distinctive signature of BRCAness at the level of mRNAs or proteins, using gene expression arrays or reverse-phase protein arrays (RRPAs). In this issue, Konstantinopoulos et al report a 60-gene signature for BRCAness in familial and sporadic ovarian cancers. Preliminary evidence for the validity of the signature was sought using several different approaches. Initially, they determined whether the BRCAness profile could distinguish platinum-sensitive from platinum-resistant ovarian cancers, profiling biopsies from six patients with germline mutations of BRCA1 or BRCA2. In this and other studies, ovarian cancers from patients with BRCA1 or BRCA2 germline mutations exhibit greater sensitivity to platinum compounds because these agents can induce double-strand breaks in DNA that are less readily repaired by HR in the presence of BRCA dysfunction. In four of the six patients studied in the report by Konstantinopoulos et al, emergence of platinum resistance was associated with additional revertant mutations in BRCA genes that presumably restored normal HR function to the cancer cells. This observation strongly supports platinum sensitivity as a predictor of BRCAness. The BRCA-like signature predicted platinum sensitivity or resistance in eight of 10 biopsies from cancers with mutated or revertant BRCA1 or BRCA2. Although the number of cases is small, platinum resistance correlated with BRCA status. Interestingly, the two discordant results were observed with BRCA2 mutations, raising the question of whether the 60-gene BRCA-like signature is able to detect dysfunction of BRCA1 with greater accuracy. JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L S VOLUME 28 NUMBER 22 AUGUST 1 2010