|Home | About | Journals | Submit | Contact Us | Français|
Ovarian cancer is a heterogeneous disease that encompasses a number of different cellular subtypes, the most common of which is high-grade serous ovarian cancer (HGSOC). Still today, ovarian cancer is primarily treated with chemotherapy and surgery. Recent advances in the hereditary understanding of this disease have shown a significant role for the BRCA gene. While only a minority of patients with HGSOC will have a germline BRCA mutation, many others may have tumor genetic aberrations within BRCA or other homologous recombination proteins. Genetic screening for these BRCA mutations has allowed improved preventative measures and therapeutic development. This review focuses on the understanding of BRCA mutations and their relationship with ovarian cancer development, as well as future therapeutic targets.
Ovarian cancer is the most lethal of all gynecologic malignancies in the United States. In 2016 it was estimated that approximately 14,240 patients with ovarian cancer would succumb to their disease.1 Despite advances in care, for newly diagnosed patients the overall survival at 5 years has only marginally improved to 46% during the last 20 years. The two most important factors for the lack of improvement are a high rate of advanced disease at diagnosis and lack of new therapeutic options. Epithelial ovarian cancer (EOC), which accounts for a majority of diagnoses, is further subdivided into various cell types, grades, and anatomic locations. The most common form is high-grade serous ovarian cancer (HGSOC), which accounts for approximately 70% of all EOC.2 Historically, the treatment of ovarian cancer has been surgical cytoreduction followed by adjuvant chemotherapy. The concept of improved outcomes in patients for optimal or complete surgical cytoreduction has been explored extensively over the years, with a consistent benefit seen for upfront surgical management.3–5 The cornerstone of chemotherapy for ovarian cancer is platinum and taxane-based treatment. Advances in the route of administration, including intraperitoneal chemotherapy, have helped to delay progression and increase survival.6,7 While some more recent gains have been seen from newer targeted therapies (i.e. bevacizumab), moving the field forward in the future is going to depend on a greater understanding of the genetic basis of the disease to identify new targets.
Ovarian carcinoma, especially HGSOC, is a highly mutated cancer. In 2011, a comprehensive analysis performed by The Cancer Genome Atlas (TCGA) found a number of genes to be significantly mutated in ovarian carcinoma; most notably p53, which was mutated in nearly 96% of HGSOC.8 From this work, it was also found that BRCA1/2 genes play a role in many HGSOC, irrespective of germline status. Further analysis on pathways found nearly half of all tumors tested had a mutation in one gene related to homologous recombination function. These findings point to an important role, as well as therapeutic potential to exploit, for tumors displaying deficiency in homologous recombination.
Hereditary ovarian cancer was first identified by Pierre Paul Broca in 1866 with his documentation of breast and ovary cancer within his wife’s family.9 Nearly 130 years passed until molecular confirmation of this hereditary cancer syndrome was announced. Mary Claire-King and colleagues first published a linkage analysis of families with early-onset breast cancer and identified the gene locus of BRCA1 at 17q21.10 The gene was cloned in 1994, which allowed reproducible testing.11 Shortly thereafter, the BRCA2 gene was identified and cloned as well on chromosome 13.12 Over the last 20 years, research has expanded to improve the understanding of BRCA-related ovarian cancers, specifically how they respond to treatment as well as the expected clinical course. Better characterization of alterations in these genes may enable development of new, targeted therapies, or broadening the clinical application of current therapies.
The process of repairing DNA damage from external or internal sources of derangement is an essential task of the genome in order to prevent cell death. One of the most significant alterations to DNA can occur through a double strand break (DSB), and if left unchecked it is lethal to a cell.13 DSBs are disruptions in both reading frames of the DNA, often caused by external insults such as ionizing radiation. These breaks are more difficult for DNA repair because there is a lack of a normal reading frame to repair nucleotides to, and for this reason are prone to error. Two main mechanisms allow a cell to repair a DSB: non-homologous end joining (NHEJ) and homologous recombination (HR). NHEJ causes open ends of the DNA to attach binding proteins to stabilize and ultimately reconnect the sides of the DNA, but without regard for the reading frame.14,15 This induces errors into the DNA. HR allows for repairing an unaltered reading frame. From the open ends, a single strand 3′ opening is created. This allows a series of proteins (including RAD51/BRCA2) to populate to begin searching for a compatible sequence with which to invade and create a D-loop. This process allows both sides to faithfully reconstruct the reading frame.15 BRCA 1/2 each play multiple, unique roles in HR repair. BRCA1 is thought to be part of a larger complex molecule that helps to survey the DNA for DSB damage.16 The role of BRCA2 is less clear, but it likely has a more direct role in repair by assisting the RAD51 complex in attaching to the repair site.17 Both genes serve as important pieces in a large framework of repair molecules.
Patients who have germline mutations in either BRCA1/2 are at a higher risk for certain cancers compared to the general public. In rational terms, this would mean many tissues would be at a higher risk of tumor development. However, the majority of cancers developing from BRCA mutations are of either breast or ovarian origin. Some research suggests that menstrual cycle oxidative stress may play a role in ovarian tumorigenesis.18 Also, hormone regulation, especially estrogen, appears to increase DSB, which may explain tissue specificity.19
Germline mutations in BRCA1/2 have been extensively studied in the population to ascribe a risk associated to carriers for the development of breast and ovarian carcinoma. In a seminal paper analyzing over 8000 unselected cases of breast or ovarian cancer, the average cumulative risk of developing ovarian cancer with a BRCA1/2 mutation was 39% and 11% respectively.20 The authors also found convincing evidence of an age discrepancy for onset of disease between BRCA1/2, with BRCA1 patients having an increased risk after age 40 and BRCA2 patients after age 50. This becomes important when counseling patients regarding options for risk reduction. Of all patients who are diagnosed with serous ovarian carcinoma, over 15% will have a germline BRCA mutation (gBRCAmut) present.21 Particularly noteworthy is that these patients are the incident case in the family over 40% of the time.22 Ethnic minorities, in some instances, are affected with BRCA mutation more frequently. Ashkenazi Jewish descendants have a 1–2% chance of harboring a BRCA1/2 mutation compared to the general public, which has a rate of 1/400.23
Female and male relatives may harbor germline mutations in the BRCA genes. This point is crucial to understand for an adequate assessment of genetic history. The risk of breast cancer (up to 80% lifetime) in BRCA 1/2, along with risks for ovarian cancer, usually dominates the discussion. It is important when counseling patients, however, to note the increased risk of pancreatic cancer, melanoma, as well as breast and prostate cancers in men.24–27 Given the implications for treatment and cancer risk determination, there is widespread agreement among professional organizations like the Society of Gynecologic Oncology (SGO) and the National Comprehensive Cancer Network (NCCN) that all women diagnosed with epithelial ovarian, fallopian tube, and/or peritoneal cancers should be offered cancer genetic counseling and testing for germline BRCA1/2 mutations. Genetic counseling should include the collection of a three-generation pedigree and involves comprehensive risk assessment based on the patient’s personal and family histories.28
The US Supreme Court’s 2013 ruling to dismiss gene patents29 was a landmark case in the field of molecular genetics. The case allowed for a competitive availability of next-generation sequencing technology for the genetic screening of BRCA mutations. In addition to single-gene/syndrome testing, multi-gene panel tests, from various distributors, have been adopted by clinicians as affordable and efficient alternatives. Clinical use of multi-gene panels is not without controversy, particularly when less-studied moderate-risk susceptibility genes are included. The NCCN advises that in patients who have a personal and/or family history suggestive of more than one potential hereditary cancer syndrome, it may be appropriate to consider multi-gene panel testing but does not provide guidance for deciding which of the many available testing options should be offered in specific clinical situations. Fortunately, several studies have described the spectrum of gene mutations identified in patients with EOC. Norquist and colleagues recently reported multi-gene panel testing outcomes from 1915 unselected patients with EOC and found that 3.3% of patients had mutations in genes other than BRCA1, BRCA2, or the Lynch syndrome-associated mismatch repair genes. Derangements in genes like BRIP1, PALB2, RAD51C, RAD51D, and BARD1 made up 20% of the mutations identified in this study, and each confer an estimated 5–15% lifetime risk of ovarian cancer.30 It is important to note that these genes are not often referred to as ‘moderate-risk genes’. The most recent version of the NCCN Guidelines for Genetic/Familial High-Risk Assessment includes interventions for individuals with mutations in these genes, with the exception of BARD1, perhaps making multi-gene panel testing for women with ovarian cancer less controversial. It should also be noted that given the added complexities inherent to multi-gene panel testing it is generally recommended that these tests be ordered by providers with specific cancer genetics expertise. Expertise is required to give the essential elements of informed consent necessary for any cancer genetic test. During the consent process, special attention should be paid to the potential limitations of result interpretation, the application to clinical management, as well as the possibility of receiving an uncertain test result.31
Genetic testing in clinical practice should always begin with the affected individual. This allows for accurate interpretation of the results, either positive for mutation or negative. Cascade testing begins with an affected relative, and then progresses to unaffected patients. Only in situations of patients who meet high-risk criteria, and have no known living affected relative should testing occur in an unaffected individual first. Testing in these situations can lead to confusion of results, and may require broadening the scope of screening (i.e. multi-gene panel testing).
Hereditary cancer syndromes, such as the one associated with BRCA mutation, provide an opportunity to screen family members earlier, and in some cases carry out preventative measures to greatly reduce the risk of developing cancer. The NCCN has compiled a list of guidelines for providers to use when counseling for genetic risk evaluation. The SGO and American College of Obstetrician and Gynecologists have joined in a consensus statement regarding genetic counseling as well.32 Notably, all patients with EOC are recommended to receive genetic counseling and testing for BRCA mutations. In addition, patients who have been diagnosed with early-onset breast cancer (age <45) and patients diagnosed with triple negative breast cancer prior to age 60.
Breast cancer screening, with mammograms, remains one of the most effective tools to date in reducing the risk of breast cancer mortality. The recommendations for breast cancer screening in a high-risk population have been extensively studied and covered in various publications.33–35 Unfortunately, ovarian cancer screening has a long history of relatively poor outcomes with regard to early detection or prevention. Ovarian cancer screening on trial has been performed with pelvic ultrasonography and CA-125 measurement. The relative short interval from early disease to advanced-stage disease in ovarian cancer makes surveillance particularly difficult. The large US-based Prostate, Lung, Colorectal and Ovarian (PLCO) cancer trial, which evaluated screening for ovarian cancer among other diseases, failed to show reduction in ovarian cancer mortality among a non-selected population.36 These results were further reiterated in the randomized UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS).37 Unfortunately, even in high-risk populations, no screening modality has been shown to be effective at reducing mortality or detecting early disease.38,39 NCCN guidelines state that although the data on screening have been inconclusive, it is reasonable to consider in patients who are unwilling to go through a risk-reducing salpingo-oophorectomy (RRSO) at a young age (<35 years old). Table 1 outlines the recommendations from national organizations for ovarian cancer surveillance and risk-reducing management for the high-risk population.
Prophylactic surgery in patients with gBRCAmut has been shown to be beneficial in prevention of ovarian cancer. RRSO was analyzed in two separate studies that were released in 2002. Kauff and colleagues showed that in patients that chose RRSO, who were at high risk for a breast or a BRCA-related gynecologic malignancy, had a 75% decreased risk of developing ovarian cancer following surgery.41 Overall, patients who underwent an RRSO had <1% chance of developing a primary fallopian tube or ovarian malignancy. Rebbeck and colleagues showed essentially the same results, but over a longer follow-up period. In that study, RRSO conferred a 96% reduction in BRCA-related gynecologic cancer.42 A large meta-analysis confirmed the significant reduction in ovarian cancer risk among patients with BRCA who undergo an RRSO (HR = 0.21). The authors reported a continued small risk of developing a primary peritoneal cancer after RRSO. In addition, a statistically significant decrease in the risk of breast cancer in patients undergoing RRSO (HR = 0.47) was also seen.43 The ages recommended for consideration of RRSO are also based on these studies. Currently, NCCN and SGO recommend consideration of RRSO following completion of childbearing and after age 35. This is based on the relative increase in risk of a gynecologic malignancy in a BRCA1 carrier after age 40. It is reasonable to consider a delay of RRSO in patients who have a BRCA2 mutation, since their age-adjusted risk for ovarian cancer does not start to increase until age 50.41
There are some early data regarding the role of prophylactic salpingectomy only in younger patients who do not desire or are unwilling to pursue oophorectomy. The risk reduction for ovarian cancer is 35–50% with a salpingectomy alone in a non-selected population.44 This hypothesis is based on the theory that a majority of epithelial ovarian carcinoma (serous in particular) actually originates in the fallopian tube. The serous tubal intraepithelial carcinoma (STIC) theory originated in the late 1990s when pathologists noted occult lesions on the fallopian tubes of women with BRCA1/2 mutations following prophylactic surgery.45 The concept is that these serous carcinomas more closely resembled the cells of the fallopian tube fimbria, and STICs are found in high numbers of patients with HGSOC.46–48 While the data are thought-provoking, it is still too early to formally recommend this for all patients at high risk of BRCA-related gynecologic malignancy.
Patients who are diagnosed at a younger age with a BRCA mutation often question what can be done to reduce their risk for ovarian cancer prior to definitive surgical intervention. Oral contraceptive pills (OCPs) have been studied as a type of ‘chemoprophylaxis’ for ovarian carcinoma. Narod and colleagues found that patients who had taken OCPs for any length of time saw a reduction in risk of ovarian cancer by about 50%.49 Further, as the timing of OCP use extended past 6 years, the risk reduction was up to 60%. A meta-analysis confirmed this finding, and showed that the benefit of OCP use among patients with a BRCA mutation may be similar or better than the general population.50 However, the use of OCPs has to be weighed against the theoretical risk of or impact on breast cancer and the impact of hormonal manipulation. Whether OCP use increases the risk of breast cancer in BRCA mutation carriers is conflicting.51,52 Based on the current data disparity, BRCA carriers should be counseled on the potential benefits and perils of OCP use and it should be considered a cautious choice when seeking an alternative way to reduce their risk of ovarian cancer.
The advent of multi-gene panel testing for hereditary breast and ovarian cancer has increased the finding of germline genetic mutations in genes associated with an increased risk of these cancers. Walsh and colleagues found that approximately 6% of patients with ovarian cancer had a mutation that was a non-BRCA loss of function.53 BRCA mutations are highly penetrant, while other genes have variable penetrance. RAD51C and RAD51D mutant carriers have been shown to have a relative risk of ovarian cancer of 5.88 and 6.30 respectively.54,55 Furthermore, a more recent case-control study of over 3000 patients with ovarian cancer found a significantly higher proportion of patients with mutations in RAD51C and RAD51D compared to controls. The researchers noted that by age 70, the risk of ovarian cancer for RAD51C and RAD51D was 5.2% and 12% respectively.56 Though other mutations have yet to yield compelling evidence for pre-emptive surgical management, with careful counseling regarding the early nature of the research, patients carrying alterations in the RAD51 genes, in context with family history, can be considered for RRSO.57
The presence of a germline BRCA mutation (gBRCAmut) in a patient with HGSOC confers a survival benefit when compared to patients without the mutation. In 1996, the first study analyzing outcomes among patients with a BRCA mutation showed that BRCA mutant patients lived longer than non-BRCA patients (77 versus 29 months).58 Further studies have confirmed that these patients have a better response to platinum therapy compared to patients without BRCA mutations.22,59,60 gBRCAmut carriers appear to also be more sensitive to the benefits of intraperitoneal chemotherapy.61 In a large pooled analysis of 26 observational studies, BRCA1/2 germline mutations were shown to have a definitive improvement in overall survival compared to patients without a mutation. For BRCA2 mutation carriers, the mean 5-year overall survival was 52% compared to 36% in non-carriers.62 BRCA2 mutations, in particular, carry a higher survival rate. This may be due to its mechanism of action; BRCA2 protein more closely regulates the process by which crosslink damage repair occurs, thus making these patients more sensitive to DNA-damaging chemotherapy.63 Unfortunately, when analyzing survival out to 10 years, the protective effect of a BRCA mutation seems to diminish.64
Poly (ADP-ribose) polymerase (PARP) was first discovered as a molecule in 1963.65 PARP is a member of the collection of proteins that aides in the HR repair of DSBs. The first inhibitor of PARP was discovered in 1980 and was originally designed for possible use in chemotherapy sensitization.66 Originally, these molecules were not thought to be a single-agent therapy choice for patients with cancer due to its mechanism of action, which was thought to only slow down cancer cell growth, but not induce lethality. In 2005, two published reports showed that combining a PARP inhibitor with cells that were deficient in BRCA1 caused significant cellular death compared to cells with BRCA intact.67,68 Independent researchers had identified a new term called ‘synthetic lethality’, where either endogenous or exogenous depletion of two molecules in a DNA repair pathway becomes lethal to a cell. This exploitative function for PARP inhibitors became especially noteworthy for cancers such as breast and ovary related to germline BRCA mutations.
The first trials in PARP inhibitors for patients with solid tumors with a gBRCAmut were published in 2009. The population studied was enriched with patients who had a known mutation in BRCA and included patients with ovarian tumors. Other tumors included were breast, colon, melanoma, prostate, and sarcomas. In patients with known BRCA1/2 mutations, single-agent treatment with olaparib showed a 63% clinical benefit (including disease stabilization).69 These results were followed up with an expansion cohort (phase IB) looking at recurrent ovarian/fallopian tube/primary peritoneal cancer patients only. The expansion included only known germline BRCA mutation carriers and heavy pre-treatment. The results showed an overall response rate of 40%, with a subanalysis showing a 62% response rate in patients who had been platinum sensitive with their last platinum treatment.70 Following a trial by Kaye and colleagues [ClinicalTrials.gov identifier: NCT00628251] in which olaparib showed comparable efficacy to a standard of care option (pegylated liposomal doxorubicin),71 a larger phase II randomized study was opened with olaparib. This trial (study 19) was the first to enroll patients with recurrent disease who may or may not have a gBRCAmut. This trial was specifically studying whether patients who display a BRCA-like phenotype respond similarly to those with an actual BRCA mutation. Inclusion in the study did not require BRCA 1/2 mutation status to be known. Patients were required to be platinum sensitive to most recent platinum chemotherapy, showing an objective response. Olaparib or placebo was provided in a maintenance setting. Overall, the progression free survival (PFS) with olaparib was 8.4 months versus 4.8 months compared to placebo.72 In a second paper updating overall survival, the authors presented pre-planned subanalysis on BRCA status. The two arms were well balanced with over 50% of patients having either a germline or a tumor somatic mutation of BRCA (the majority were gBRCAm). In this population, the PFS was 11.2 versus 4.3 months (HR 0.18; p < 0.0001) comparing olaparib and placebo.73 The results of this study led to the European Medicines Agency granting approval for olaparib in the maintenance setting for patients with recurrent HGSOC who are platinum sensitive.
A second pivotal study, by Kaufman and colleagues (study 42), was a multicenter phase II trial that enrolled patients with a BRCA1/2 mutation who had recurrent cancer. The majority of the tumor types were ovarian; however, other solid tumors such as pancreatic and prostate were enrolled. In this trial, patients who had ovarian cancer had to be resistant to platinum therapy. All patients received olaparib 400 mg twice daily. The primary outcome was tumor response rate (TRR). The overall TRR was 26%; however, for patients with ovarian cancer the TRR was 31%. In addition, ovarian cancer patients showed a stable disease rate of 40%. The PFS and overall survival (OS) for ovarian cancer patients were 7 months and 16.6 months respectively, with over 64% of patients alive at 12 months.74 On the basis of this trial, the Food and Drug Administration (FDA) approved use of olaparib as monotherapy in patients with a gBRCAm who have received 3 lines of chemotherapy in the United States.
The results of the aforementioned studies have led to an increase in the number of phase III trials for multiple PARP inhibitors. This includes trials in the upfront setting with primary therapy as well as trials in maintenance therapy following initial adjuvant treatment [GOG-9923 (ClinicalTrials.gov identifier: NCT02470585) and SOLO-1 (ClinicalTrials.gov identifier: NCT1844986)]. Rucaparib, a PARP-1/2 inhibitor, has shown promise in a similar population of recurrent HGSOC patients. In a trial, now closed to accrual [ARIEL-2 (ClinicalTrials.gov identifier: NCT01891344)], rucaparib was tested in a population with recurrent HGSOC. The trial was conducted in two parts. ARIEL2 Part 1 looked at patients with recurrent, platinum-sensitive disease, who had at least one prior platinum-based therapy. Patients could enroll as known gBRCAm; however, all tissue was tested for BRCA mutations, and confirmed as germline with blood testing. In addition, a molecular signal from the tumor is also being studied to determine if high levels of HRD (homologous recombination deficiency) is present. Testing focused on determining the amount of genomic scarring that is present in the cancer genome, which was quantified by analyzing loss of heterozygosity (LOH) in the tumor. A high amount of LOH indicates genomic instability.75 Recently published results show that in patients with either a germline BRCA mutation or a BRCA wildtype with high LOH on tumor testing, the response is significantly greater than patients with BRCA wildtype and a low LOH. Specifically, the PFS was 12.8 months versus 5.2 months comparing BRCA mutation and BRCA wildtype low-LOH score.76 This trial highlights the understanding that somatic BRCA mutations are not only present in high percentages, but also can be exploited with PARP inhibitors. Based on the results of ARIEL2, as well as a smaller phase I/II by Kristeleit and colleagues (study 10) in Europe,77 rucaparib was recently approved for use in germline or somatic BRCA mutation patients who have had 2 lines of therapy.78 This marked the first PARP inhibitor to receive approval for use in ovary cancers with somatic BRCA mutations in the United States.
Niraparib is a third PARP inhibitor that involves inhibition of PARP-1, PARP-2, and PARP-3. Recently a phase III trial (NOVA/ENGOT-OV16 trial) of recurrent HGSOC patients receiving niraparib versus placebo was published. All patients were platinum sensitive and the trial was conducted with two main objectives: (1) efficacy of the drug over placebo; and (2) identify a biologic marker for HRD (through MyChoice MyriadTM testing). The results demonstrated that the PFS for patients with a gBRCAm was 21 versus 5.5 months. The authors also found there was a significant improvement in PFS in patients with a high HRD score who did not have a gBRCAm. Perhaps most surprising, though, was that patients who had a low HRD score and did not have a germline/somatic BRCA mutation also had a significant PFS advantage.79 The FDA, as a result of this data, just recently approved niraparib to be used in the maintenance setting for platinum-sensitive HGSOC, regardless of BRCA status. PARP inhibitors are also being explored with other targeted therapies as well. A phase II trial looking at combining olaparib and cediranib (VEGF inhibitor) found a significant improvement in PFS compared to olaparib alone.80 The study population in this trial combined those with and without gBRCAmut. PARP inhibitor use will continue to expand in ovarian cancer with the promising results seen so far. Defining optimal patient populations as well as optimal timing for these therapies are key questions going forward in PARP inhibitor development. Table 2 provides a succinct review of the major clinical trials in PARP inhibitors for the treatment of ovarian cancer.
Understanding the role BRCA mutations play in the development, treatment response, and prognosis is an exciting and developing area in the treatment of ovarian cancer. Since its discovery in 1990, research has led to understanding the role of BRCA in tumorigenesis and, more recently, as a therapeutic potential. Identification of a BRCA mutation may not only help the afflicted patient, but also allows for genetic testing to be performed on relatives, allowing for the potential to prevent ovarian cancer. Furthermore, PARP inhibition has an opportunity to significantly improve outcomes in women who harbor germline or somatic BRCA mutations, as well as tumors that display a high degree of HRD. As we continue to advance our understanding of BRCA and its role in the development and outcomes of ovarian cancer, there is great potential to not only prevent many cases through improved access to genetic screening, but also revolutionize the long-term treatment of patients with this insidious disease.
Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement: The authors declare that there is no conflict of interest.
Robert T. Neff, Department of Obstetrics and Gynecology, Division of Gynecologic Oncology, The Ohio State University Wexner Medical Center, Columbus, OH, USA.
Leigha Senter, Department of Internal Medicine, Division of Human Genetics, The Ohio State University Wexner Medical Center Columbus, OH, USA.
Ritu Salani, Ohio State University Wexner Medical Center – James Comprehensive Cancer Center, 320 West 10th Avenue, M210 Starling-Loving Hall, Columbus, OH 43210, USA.