‘BRCAness and its implications for platinum action in gynecologic cancer

Abstract

Gynecologic cancers are major therapeutic targets of platinum-containing regimens. They may be particularly susceptible to these agents if their origins are related to hereditary BRCA-mutations; this implicates defective DNA repair secondary to inherited alterations in BRCA function. The concept of ‘BRCAness’ was introduced by Ashworth and colleagues in order to identify phenotypic changes in sporadic cancers that would lead to analogous treatment susceptibility. In fact, recent analyses of the genetic alterations in ovarian cancer have led to further extending this concept to all women with high-grade serous cancers –the predominant form of ovarian cancer arising in association with hereditary mutations in BRCA genes. Presumably, most serous cancers of gynecologic origin share to some extent BRCA dysfunction rendering these cancers susceptible to platinums and to other DNA damaging agents, and justifying general inclusion of this histology in trials of new drugs and therapeutic strategies that have shown activity against hereditary cancers. More recently, however, differences in outcome between BRCA mutation carriers vis-á-vis those with no mutations or those with epigenetic or acquired forms of BRCA genes (somatic mutations) in their respective tumors have been identified. These findings raise additional questions on modifiers of ‘BRCAness’ and other pathways that appear to contribute to the effects of platinums and other DNA damaging agents in ovarian cancer. The Cancer Genome Atlas analyses delineate the complexity of genomic alterations in ovarian cancer and other malignancies of mullerian epithelial origin promising further refinements of the ‘BRCAness’ concept.

Introduction

Platinum drugs play an essential role in gynecologic cancer treatment. Cisplatin or carboplatin[1-3] are coupled with surgery as part of the initial treatment in more than 90% of epithelial ovarian cancer patients. Upon recurrence all patients except those labeled as ‘platinum resistant’ usually receive multiple courses of carboplatin[4]. Platinum-based chemotherapy also has emerged in the last decade as the prevailing strategy (over radiation) for adjuvant treatment of endometrial cancer following identification of some high-risk features at hysterectomy, and forms part of the systemic treatment of patients with metastatic disease beyond rare cases of well-differentiated tumors that metastasize[5-13]. Finally, in uterine cervix cancer, the use of either cisplatin or carboplatin in combination with other drugs or use of cisplatin as a radiosensitizer have yielded improvements in outcome for patients with locally advanced or metastatic presentations[14-18].

The Cancer Genome Atlas (TCGA) provides some insight into the effectiveness of platinum compounds either alone or in combination with other drugs in the treatment of ovarian and endometrial cancers[19]. Specifically, TCGA has shown some common abnormalities among high-grade ovarian cancers with poorly differentiated endometrioid, high grade serous of the endometrium, and basal-like carcinomas of breast origin (also updated on line at cancergenome.nih.gov). All of these are characterized by high genomic instability and BRCA mutations and/or silencing through epigenetic changes. Such genomic changes did not readily provide identifiable ‘driver mutations’ that may be targeted; however, they have reinforced the concept of ‘BRCAness’ introduced by Ashworth and colleagues to identify phenotypic changes in sporadic cancers that would imply similar treatment susceptibility to DNA-damaging agents[20].We elaborate on the evolving clinical implications behind this concept in the main portion of this manuscript.

BRCA function and DNA repair

The past decade has witnessed impressive advances in our understanding of the various cellular components required for maintenance of the genome under the onslaught of DNA damage. Among these, the sensors of DNA damage, ataxia telangiectasia and Rad3-related protein (ATR) and ataxia telangiectasia mutated (ATM) are central to turning on DNA repair machinery [21-30]. In recent years, major roles have also been assigned to BRCA1 and BRCA2 functions in leading high-fidelity repair by ‘homologous recombination’ (HR)[31-41]. Cellular systems have served to probe and identify defects in such repair functions by verifying the formation of foci when BRCA-mediated repair is intact. Lack of foci formation implies defects in the HR pathway, among others. BRCA2 and a number of other Fanconi-related genes also have additional key functions in non-HR repair[20, 42-56].Selective forms of DNA damage such as nucleotide (NER) or base excision repair (BER) rely on other repair pathways, in which poly-adenosyldiphosphate-ribose polymerase (PARP) plays a major role, particularly in the absence of HR.

Interest in tumor repair pathways has led to the recognition of their importance in determining cellular sensitivity to chemotherapeutic agents. Attention was naturally drawn to epithelial ovarian cancer and its known sensitivity to the classical DNA damaging drugs -alkylating agents (from 1950-1980) and subsequently the more effective platinums. In particular, cisplatin’s clinical and analog development was greatly stimulated by its remarkable activity in the common advanced presentations of ovarian cancer.

Clinical trials by the Gynecologic Oncology Group (GOG) and others eventually identified a consistent pattern of greater sensitivity of high-grade serous cancers to cisplatin, and relative resistance of other epithelial types such as mucinous, clear cell, and low grade serous to this agent[57-68]. Additional studies in patients after platinum-based treatment suggested losing mismatch repair function and increasing tolerance of platinum-DNA adducts as a mechanism of resistance. The incidental finding of invasive and in situ high-grade serous carcinomas in Fallopian tube fimbriae of BRCA mutation carriers subjected to risk reducing surgeries spurred focusing on BRCA function as a determinant of the known platinum sensitivity of this histologic type. In addition, as more reports accumulated[69], it became evident that ovarian adenocarcinomas, arising in BRCA 1 and BRCA2 mutation carriers, had better overall outcomes and greater response-rates to platinum compounds as well as other drugs in common use for ovarian cancer recurrences [e.g. pegylated liposomal doxorubicin (PLD), gemcitabine, topotecan][70]. Such results have strengthened the association between absence of BRCA function and platinum sensitivity, as well as a similar association with sensitivity to other DNA-damaging agents that were known to be active in ovarian cancer.

PARP inhibitors and synthetic lethality

Research groups in Newcastle and the Institute for Cancer Research (led by Curtin and Ashworth, respectively) independently reported in 2005 the remarkable in vitro findings of 'synthetic lethality’ in BRCA −/− cells, i.e. enhanced lethality of DNA damaging agents (including radiation) and PARP, respectively in knockout versus wild-type cells –that is, when HR and PARP-related repair were either defective or blocked. These finding have rekindled the clinical development of PARP inhibitors that had begun at Newcastle under the leadership of A. Hilary Calvert and culminated in a trial of AGO14699 (now known as rucaparib) in a trial mostly consisting of patients with melanoma conducted by Ruth Plummer (summarized in references 71-74).

The subsequent clinical development of PARP inhibitors has not been without challenges. Although these agents are well tolerated by themselves, when administered in combination with other drugs, their doses had to be generally attenuated. Although iniparib was an exception to the above, this agent was subsequently proven as unlikely to function as a PARP1 inhibitor. Disappointingly, also the initial lead identified in phase I showing single-agent efficacy of olaparib against ovarian cancer in mutation carriers , was not pursued as vigorously as many gynecologic oncologists would have wished for their patients. Nevertheless, veliparib (GOG270, unpublished) and niratinib (J DeBono, June 1^st, poster presentation at ASCO 2013) have also shown single-agent activity in BRCA mutation carriers with ovarian cancer . Furthermore, retrospective analyses have also suggested that these patients do particularly well when maintained with oliparib after platinum-induced complete responses, a finding that is less obvious when epigenetic BRCA function is silenced, or if an unidentifiable mutation is present[3, 34, 36, 71-78] (and further documented in the oral presentation by J Ledermann, ASCO June 2^nd, 2013).

Prospective studies of PARP inhibitors have utilized the concept of BRCAness to enrich the population under study beyond those with known hereditary cancers that were shown to benefit during the phase I study of olaparib –as documented through imaging as far back as April 2006 by Fong et al (reference 75). Several studies have shown that olaparib also had some clinical benefit in ovarian cancer patients that are not BRCA mutation carriers, and thus this PARP inhibitor may be used to treat a larger subset of patients with epithelial ovarian cancer -but not patients with low grade tumors, mucinous, or clear cell adenocarcinomas. Some patients with high-grade endometrioid adenocarcinomas do share the sensitivity to platinums and point to the risk of over relying on a histologic diagnosis by itself to identify those tumors that may be extremely sensitive to platinums and PARP inhibitors[79, 80].

Refining the definition of ‘BRCAness’

As noted above, when 'BRCAness' was introduced in 2004, it was hoped that the ‘hallmarks’ of breast and ovarian cancer susceptibility to the known inherited BRCA1 and BRCA2 mutations would be identified in otherwise sporadic cancers. Specifically, the postulate was that ‘the existence of a significant proportion of sporadic breast, ovarian, and other cancers with BRCA-like functional abnormalities raises the possibility of a wider application of treatment regimens designed for familial-BRCA tumors’[20]. Moreover, the authors pointed out the need to seek for phenotypic changes that would allow such BRCAness assignment. Subsequent publications (now 30 in number under pubmed) have proceeded separately in the breast and ovarian cancer literature –not unreasonable, since the phenotypic expression of BRCA1 mutations in breast cancer relates primarily to triple negative with basal cell features, whereas BRCA2 mutations have more variable phenotypic features. Some of the breast cancer literature has sought correlations beyond the anticipated enhanced benefit from DNA-damaging drugs but also with lack of responsiveness to taxanes[81, 82].

The terminology itself, though widely used, is potentially open to ambiguous interpretation. For example, it has been pointed out that ‘BRCAness’ actually refers to ‘BRCAlessness’ because it is the deficiency of BRCA function that defines this phenotype [83]. Tan et al[84] upon comparing chemoresponsiveness in BRCA mutation carriers to ‘non-hereditary’ patients characterized the BRCAness syndrome in ovarian cancer by the following: 1) high response rates to first-line platinum base treatment; 2) high response rates to subsequent therapies including platinums; 3) long treatment-free intervals beyond relapse, 4) improved overall survival; and 5) tumors that are usually, but not exclusively of serous histology. Thus, this study refers to hereditary BRCA mutation carriers, albeit in the less studied non-Ashkenazi Jewish population, under the term ‘BRCAness’. Notwithstanding potential ambiguities, the term has caught on, and although ‘there is not standardized method to detect BRCAness’ [83] the original intent of Turner et al[20] to come up with more robust indications of extending the therapeutic implications beyond BRCA mutations remains viable and awaits further development.

Konstantinopoulos et al. developed a ‘gene expression profile of BRCAness that correlates with responsiveness to chemotherapy and with outcome in patients with epithelial ovarian cancer’ [85]. Their ‘optimal classifier’ was a 60-gene diagonal linear discriminant predictor and then applied to 35 clinical samples that were sequenced to ensure that BRCA1 and BRCA2 were wild type for classification as BRCA-like (BL) to non-BRCA-like (nBL), and this ‘optimal classifier’ was similarly applied to another 35 clinical samples that were non-sequenced. This BRCAness profile was shown to correlate with responsiveness to platinum and to PARP inhibitors and had an independent prognostic value on multivariate analysis. The 70 patients were stage III (80%), grade 3 (86%) with mostly serous histology (93%). This study supports the notion that genes other than BRCA1 or BRCA2 are responsible for BRCAness in sporadic disease; one cannot exclude, however, that sporadic mutations or epigenetic alterations in the BRCA genes themselves account for their BL classifier. An accompanying editorial by Bast and Mills stressed platinum sensitivity as a reliable predictor of BRCAness[86]: the gene expression profile signature of BRCAness correctly identified 8 of 10 BRCA mutation carriers as responders to platinums –the two exceptions had BRCA2 mutations, suggesting that the signature is better at detecting BRCA1 than BRCA2 dysfunction. These authors go on to emphasize that identifying genomic signatures associated with BRCA dysfunction may directly impact the extent of platinum sensitivity and could have substantial impact on clinical outcomes. In fact, Lesnock and Krivak’s group and our retrospective experience within phase II studies of intraperitoneal platinums [87,88] suggest that BRCA status (by immunohistochemistry in the GOG172 study or by presence of known BRCA mutation carrier state) predicts an especially favorable outcome after intraperitoneal (IP) therapy. Bast and Mills further suggest that reverse-phase protein arrays have the potential to add a new dimension in predictive assays of sensitivity to treatment beyond BRCA status[86].

BRCAness in ovarian cancer: implications for platinums and other drugs

As noted in the preceding section, a decade after introducing ‘BRCAness’, refining our definitions of BRCAness has become a central theme in the treatment of epithelial ovarian cancer. The most dramatic demonstration of its impact may be in the benefit conferred by cisplatin when given by the IP route. Whether this impressive gain in sensitivity carries over to outcomes from PARP inhibitors and other drugs is not known. However, our studies suggest that BRCA mutation carrier status also confers greater sensitivity to PLD[69], and to drugs such as gemcitabine; not enough experience is available for topotecan [70]. Since TGCA did not identify ‘driver mutations’ and the predominant theme remains sensitivity to platinums, future studies into BRCAness and correlations with outcome remain a high priority for study. Additionally, manipulations to enhance platinum sensitivity such as by increasing uptake, --through the use of bortezomib and carboplatin given by IP administration, as an example, --hold high interest. Another area that needs further research is identification of genes that affect BRCA function, such as EMSY amplification and overexpression capable of inhibiting BRCA2 transcriptional activity[20, 89, 90]. Factors resulting in the regaining (e.g., revertant) activity of BRCA even in the presence of deleterious mutations need to be identified, including the suggestion that prior exposure to anticancer agents for breast cancer enhances the likelihood of such revertant activity. Table 1 lists BRCAness-related genetic and epigenetic alterations.

Table 1.

Genetic and other characteristics associated with BRCAness in ovarian cancer^*

Abnormal features and genetic abnormalities	% in ovarian Ca	Reference
BRCA 1 & 2 germline mutation	10-15	96
BRCA 1 & 2 somatic mutation	5-10	97
BRCA promoter methylation	5-30	98
EMSY amplification	20	99
Fanconi anemia complex defects	21	42
PTEN focal deletion/mutation	7	100
Rad 51 hypermethylation	3	101
ATM/ATR mutation	2	102
Serous, pseudo-endometrioid, transitional-cell like, ↑TIL	majority	103

^*Modified from reference 90

In conclusion, the concept of BRCAness as originally introduced has proven useful in emphasizing the central role of the BRCA genes both in breast and in ovarian cancer biology and treatment. Therapeutic implications emanating from this concept appear to differ in ovarian cancer vis-a-vis breast cancer, further reinforcing the importance of the context in which BRCA and related genes function in these malignancies. The remarkable effects of the platinum compounds may hopefully be extended further by studying to what extent the population with BRCAness attains similar outcomes to those with germline BRCA mutations.