The role of BRCA1 and BRCA2 in prostate cancer

Abstract

One of the strongest risk factors for prostate cancer is a family history of the disease. Germline mutations in the breast cancer predisposition gene 2 (BRCA2) are the genetic events known to date that confer the highest risk of prostate cancer (8.6-fold in men ≤65 years). Although the role of BRCA2 and BRCA1 in prostate tumorigenesis remains unrevealed, deleterious mutations in both genes have been associated with more aggressive disease and poor clinical outcomes. The increasing incidence of prostate cancer worldwide supports the need for new methods to predict outcome and identify patients with potentially lethal forms of the disease. As we present here, BRCA germline mutations, mainly in the BRCA2 gene, are one of those predictive factors. We will also discuss the implications of these mutations in the management of prostate cancer and hypothesize on the potential for the development of strategies for sporadic cases with similar characteristics.

Introduction

Prostate cancer (PCa) is the second commonest tumour in men worldwide, with about 900 000 new cases diagnosed annually, although there is substantial variation in disease incidence across regions. Australia/New Zealand and the countries from North America and Western and Northern Europe have the highest rates of PCa, partially because PCa screening with prostate specific antigen (PSA) and subsequent biopsy has become common practice (although this is not the sole reason), while countries in South-Central Asia have the lowest incidence.¹ Studies of incidence rates in migrant populations from Asia to America² suggest that lifestyle risk factors are important determinants of PCa risk. To date, however, these lifestyle risk factors have not been identified. PCa accounts for the second commonest cause of male cancer-related deaths in the United States³ and the sixth worldwide, with more than 250 000 deaths a year.¹ Thus, it is essential to identify those patients with potentially lethal forms of PCa at their presentation.

PCa is rarely diagnosed in men younger than 50 years, but its incidence rises rapidly thereafter. Excluding advanced age and African-American ancestries, the strongest risk factor for the disease is a family history of PCa.^{4, 5, 6} PCa is one of the common cancers with a large genetic component, as up to 42% of the risk could be explained by inheritance from studies about twins.⁷ Genome-wide association studies (GWAs) have identified 46 susceptibility loci associated with PCa^{8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20} that individually contribute to a small increase in PCa risk, but which taken together, could explain more than 25% of the excess of PCa familial risk.¹³

Although present in only 1%–2% of sporadic PCa cases^{21, 22, 23} germline mutations in the breast cancer predisposition gene 2 (BRCA2) are the genetic events known to date that confer the highest risk of PCa (8.6-fold in men ≤65 years).²³ BRCA1 has also been associated with an increased risk of sporadic PCa (3.5-fold), even though germline mutations in this gene have only been observed in 0.44% of PCa cases (Leongamonlert et al., in press). Lifetime risk of PCa in BRCA2 mutation carriers has been estimated to be 20%,²⁴ while for BRCA1, the risk is 9.5% by age 65 years (Leongamonlert et al., in press), similar to that in non-carriers.

The genetic aetiology of PCa is complex and poorly understood, with multiple predisposing factors which may also affect presentation, progression and outcome.^{25, 26, 27} While the clinical implications of common genetic variants associated with PCa risk remain unclear,^{28, 29, 30, 31, 32, 33, 34} deleterious germline mutations in both genes, BRCA1 and BRCA2, have been associated with more aggressive disease and poor clinical outcomes.

The increasing incidence of PCa supports the need for new methods to predict outcome, because the factors currently used (TNM stage, tumour grade and PSA at diagnosis) have been proven to be insufficient to identify which men are at risk of developing lethal PCa. As we present here, BRCA germline mutations, mainly those mutations in the BRCA2 gene, are one of those predictive factors. We will also discuss the implications of BRCA mutations in the management of PCa and hypothesize on the potential for the development of strategies for sporadic PCa tumours with similar characteristics.

BRCA1 and BRCA2 genes and the risk of PROSTATE CANCER

Although multiple aetiologies have been proposed to predispose to the development of PCa, the only well-established risk factors are age, ethnic background (men of African descendent have an increased risk of the disease) and a family history of the disease.³⁵

PCa exhibits a significant familial aggregation in some men, particularly when affected at young age. High-risk PCa familes have been described as those that present with: (i) three or more first-degree relatives affected with PCa; (ii) PCa in three or more generations, or (iii) an early age of PCa diagnosis among at least two first-degree relatives, usually siblings.³⁶ The risk of the disease in first-degree relatives of cases is approximately twice that in the general population. This familial risk is greater among young cases, being more than fourfold for close relatives of cases diagnosed below age 60. Higher risks have been shown for men with two or more affected relatives.^{4, 5, 6, 7, 37, 38, 39, 40} This increase is too great to be explained by non-genetic factors alone. Analyses based on the Nordic twin registries have found higher risks in monozygotic than dizygotic twins, supporting the hypothesis that much of this familial aggregation is due to genetic factors (42%), rather than shared lifestyle factors.⁷ Several linkage studies for PCa have been completed, often with conflicting results,^{26, 41, 42, 43} therefore, attempts to identify high-prevalence genes with a dominant mode of inheritance, such as APC for colon cancer⁴⁴ or BRCA for breast and ovarian cancer^{45, 46} have failed in PCa.

There is a recognized association of breast cancer with PCa in families.^{47, 48, 49} The contribution of the BRCA1 and BRCA2 to male cancer has been extensively studied, since families with these mutations show clustering of cancer in men. These genes have not only been associated with an increased risk of developing breast, ovarian, fallopian tube and peritoneal cancer⁵⁰ in women, but also with male breast cancer and PCa.^{24, 51} For BRCA1 mutation carriers, the Breast Cancer Linkage Consortium (BCLC) reported an increase in PCa risk in men aged <65 years with a relative risk of 1.82, but no risk increase was seen for men aged ≥65 years.⁵² For BRCA2 mutation carriers, the BCLC estimated the relative risk for developing PCa as 4.65, rising to 7.33 for men aged <65 years.²⁴ We have recently concluded two large studies reporting an 3.5-fold and 8.6-fold increase in the risk of PCa for BRCA1 and BRCA2 mutation carriers by age 65, respectively (Leongmonrlert et al., in press).²³

BRCA1 and BRCA2 are tumour suppressor genes and both are inherited in an autosomal dominant fashion with incomplete penetrance. Tumorigenesis in individuals with germline mutations in the BRCA genes requires somatic inactivation of the remaining wild-type allele. Both genes encode large proteins that function in multiple cellular pathways. BRCA1 is a key player in cellular control systems, having been linked to a range of cellular processes, such as DNA damage response and repair, transcriptional regulation and chromatin modelling.^{53, 54} By contrast, BRCA2 function seems to be limited to DNA recombination and repair processes, being of particular importance in the regulation of RAD51 activity.^{53, 54, 55} BRCA1 or BRCA2 function loss is associated with a deficiency in repairing DNA double-strand breaks (DSBs) by the conservative mechanism of homologous recombination (HR). Therefore, cells have to repair these lesions through other non-conservative and potentially mutagenic mechanisms. This genomic instability may underlie the cancer predisposition caused by deleterious mutations in the BRCA genes, although the reason why these mutations are particularly associated with some specific types of cancer, such as breast, ovarian and PCa, remains unknown.⁵⁶

Although some studies have started to elucidate the role of these genes in PCa, the specific role of BRCA1 and BRCA2 in the development and progression of PCa has not been yet elucidated. BRCA1 has not only been identified as one of the coregulators of the androgen receptor (AR)^{57, 58} which mediates a signalling pathway essential in the development and progression of PCa, but has also been suggested to modulate another important pathway in PCa through the regulation of IGF-1R in an AR-dependent manner.⁵⁹ Using animal models, Francis et al.⁶⁰ have proposed that BRCA2 may act as a tumour suppressor in epithelial prostate tissue and its functional loss predisposes to premalignant prostatic lesions. They observed that deletion of Brca2 in murine prostatic epithelia results in focal hyperplasia and low-grade prostate epithelial neoplasia (PIN) in animals over 12 months of age and that the simultaneous deletion of Brca2 and the tumour suppressor tp53 lead to focal hyperplasia and atypical cells at 6 months, and to high-grade PIN in animals at 12 months. Moro et al.⁶¹ have proposed that a functional BRCA2 gene may limit the metastatic potential of neoplastic cells by downregulating matrix metalloproteinase 9 (MMP-9) through inhibition of PI3-kinase/AKT and activation of MAP/ERK, effectively preventing cancer cell migration and invasion. In a cohort of sporadic PCas treated with prostatectomy, Bednarz et al.⁶² found that those tumours with somatic BRCA1 loss in at last one tumour focus, due to a deletion or hypermethylation of the promotor, had a higher probability of advanced tumour stage and a shorter disease-free survival than BRCA wild-type tumours. These results may help to explain the fact that patients with germline deleterious BRCA1 and BRCA2 mutations frequently present with nodal involvement and distant metastasis at diagnosis^{63, 64} and those with local disease, have a shorter disease-free survival than those patients BRCA wild type.^{63, 64, 65}

On the other hand, BRCA functional loss leads to genomic instability and the accumulation of genetic aberrations in cells. As a consequence, one would expect other recurrent driving molecular events to be present in BRCA-mutated tumours. Several translocations have been recently described in sporadic PCa (i.e., TMPRSS2/ETS gene fusions, SLC45A3/BRAF, RAF1-ESRP1), and some of them could have prognostic and therapeutic implications,^{66, 67} although the frequencies of such events have not yet been characterized in BRCA-related PCa.

Clinical implications

Several epidemiological studies that examined the risk of PCa among BRCA1 and BRCA2 mutation carriers have suggested that protein-truncating BRCA2 mutations confer an increased risk of PCa, while the effect of mutations in the BRCA1 gene seems to be more modest (Leongamonlert et al., in press).^{23, 24, 51, 68, 69, 70, 71, 72, 73} However, one of the challenges of studies analysing the susceptibility of PCa in BRCA mutation carriers is the low incidence of germline mutations in those genes in the general population which is estimated to be between 0.07% and 0.24% in different studies,^{72, 74, 75, 76} and therefore, results have often been inconsistent. The ethnic population that has more extensively been studied is the Ashkenazim Jewish, due to the relative high incidence of three founder mutations (BRCA1 185delA, BRCA1 5382insC and BRCA2 6174delT). It has been estimated that ∼2% of this population carries at least one of these mutations.^{77, 78, 79, 80} Even in this population, the results of these epidemiological studies have been contradictory, but many of them were underpowered, due to small sample size or lack of covariate information.^{22, 81, 82, 83, 84, 85, 86}

Another BRCA2 founder mutation that has been extensively studied is the Icelandic BRCA2 999del5 mutation. Johannesdottir et al.⁸⁷ reported that this mutation was present in 2.7% of unselected Icelandic patients diagnosed with PCa at age ≤65 years. This prevalence is double the prevalence of germline BRCA2 mutations reported by Kote-Jarai et al.²³ and Agalliu et al.²² in two cohorts of PCa patients diagnosed at age ≤65 years (1.2% and 0.78%, respectively). In these latter two studies, patients were screened for any²² mutation, not only the founder ones.

To assess whether BRCA2 999del5 mutations contribute to PCa phenotype and prognosis, Tryggvadottir et al.⁶⁴ analysed a group of 89 PCa patients diagnosed in Iceland between 1955 and 2004. The study included 30 men carrying the Icelandic founder mutation, and for each carrier, two control patients without the mutation. Compared with non-carriers, the mutation carriers were younger at diagnosis (69 years vs. 74 years); presented with more advanced tumour stage (T) (T3–4: 79% vs. 36%) and poorly differentiated tumours (84% vs. 52.7%). Median cause-specific survival (CSS) for carriers was 2.1 years compared with 12.4 years for non-carriers.

After Tryggvadottir et al.⁶⁴ suggested that BRCA mutations could have clinical implications beyond increasing the risk of PCa, another two studies analysed the role of a wide spectrum of mutations in BRCA2. None of them included patients with the Icelandic founder mutation. Edwards et al.⁸⁸ compared overall survival (OS) after PCa diagnosis in a series of 21 BRCA2 mutation carriers and 1587 controls, 263 of which had tested negative for BRCA2 mutations. Carriers had a median OS of only 4.8 years compared with 8.5 years for the non-carriers. More recently, Thorne et al.⁸⁹ reported the analysis of the outcome of PCa in 40 BRCA2 mutation carriers compared with 97 BRCA wild-type patients. Non-carriers belonged to families with a history of breast cancer. No difference in age or PSA at diagnosis between carriers and non-carriers was seen. BRCA2 carriers presented with less differentiated (65.8% vs. 33%) and larger (T≥3) tumours (39.5% vs. 22.6%) than non-carriers. Despite 79% of patients in both groups (carriers and non-carriers), receiving similar local treatment with curative intent, those with BRCA2 mutations had a worse CSS. BRCA2 mutation status was shown to be an independent predictor of CSS (hazard ratio (HR): 4.97; 95% confidence interval (CI): 2.19–11.25). Interestingly, the non-carrier group also had a poorer outcome compared with other series of sporadic PCa, suggesting that a family history of breast cancer might, somehow, affect the prognosis of PCa patients.

Two studies have analysed the implications of both BRCA1 and BRCA2 mutations. Narod et al.⁹⁰ reported a series of 67 BRCA2 and 37 BRCA1 mutation carriers or obligate carriers of different BRCA mutations. The study analysed the prognosis of BRCA1 vs. BRCA2 mutation carriers and did not include a BRCA wild-type group for comparison. No information was available on pathology or clinical features of the PCa cases. They observed that the 5-year OS for all causes of death was shorter for BRCA2 than for BRCA1 (42% vs. 64%, respectively) with a median OS of 15 years for BRCA1 mutation carriers and 5 years for BRCA2 mutation carriers. Gallagher et al.⁶⁵ reported a series that included 26 PCa cases carrying the Ashkenazi founder mutations BRCA2 6174delT (20 men) and BRCA1 185delAG (6 men), and 806 non-carriers. All patients presented with early-stage PCa were screened for the two mutations. Patients carrying BRCA2 6174delT presented with poorly differentiated tumours more frequently than non-carriers (85% vs. 57%). No association with the tumour phenotype was observed for BRCA1 185delAG. Both mutations conferred a higher risk of biochemical relapse and metastasis. CSS was 19.1 years for non-carriers compared with 13 and 12.5 years for BRCA1 185delAG and BRCA2 6174delT, respectively. Although these results should be assessed cautiously due to the sample size, both mutations were found to be predictors of poorer CSS (HR: 5.16; 95% CI: 1.09–24.53; and HR: 5.48; 95% CI: 2.03–14.79, for BRCA1 and BRCA2 mutation carriers, respectively). In general, these previous series were limited when analysing patient outcomes, due to the small number of BRCA mutation carriers with little clinical information, or to the lack of a comprehensive multivariate analysis. Thus, the real independent prognostic contribution of BRCA mutations might be overestimated, compared with other classical prognostic factors for PCa outcome. In an attempt to better understand the implications of BRCA1 and BRCA2 germline mutations, we have recently reported the analysis of 2019 PCa patients, 18 and 61 of them harbouring BRCA1 and BRCA2 germline mutations, respectively.⁶³ This is the largest study to date investigating the clinical characteristics and outcome of PCa patients with and without BRCA mutations. Our results reveal that a wide spectrum of pathogenic mutations in the BRCA1 and BRCA2 genes confers a more aggressive PCa phenotype and these tumours are more frequently associated with lymph node involvement and distant metastasis at diagnosis than PCa in non-carriers. We have not seen differences in age at diagnosis as previously reported by others.^{65, 84, 89, 91} This contradicts Tryggvadottir's observations,⁶⁴ but our study⁶³ did not include any patients with the Icelandic founder mutation BRCA2 999del5, and we cannot exclude that some mutations are associated with younger age at diagnosis. BRCA1 and BRCA2 carriers had a higher incidence of poorly differentiated tumours, presented with larger tumours and a higher incidence of node involvement and distant metastasis. We have also demonstrated that BRCA germline mutations in PCa, particularly in the BRCA2 gene, are associated with poor OS and CSS independently of other classical prognostic factors, including stage, Gleason score and PSA.

Current management of BRCA-mutated prostate tumours

While PCa screening in unselected men remains controversial,^{92, 93, 94} targeting screening to higher-risk groups may have a greater impact in men with a higher risk of aggressive and fatal PCa. BRCA1 and BRCA2 mutation carriers may constitute one of such populations. Early diagnosis in these patients may be crucial, and currently, the IMPACT study is evaluating the utility of PSA-based PCa screening in asymptomatic BRCA1 and BRCA2 mutation carriers.⁹⁵

At this time, there is no evidence as to which is the most appropriate radical treatment for BRCA mutation carriers with local PCa. Although clinical trials to assess the most adequate management of these cases are still needed, radical treatment with either surgery or radiotherapy seems to be preferable to active surveillance, even for low-risk cases. The studies previously presented also support clinical trials evaluating the role of adjuvant treatment in BRCA mutation carriers as they have a higher early nodal spread and metastasis rate.

One of the key clinical issues to be addressed in the treatment of BRCA-related tumours is the choice of chemotherapy in both, the adjuvant and the palliative settings. Platinum, taxanes and more recently poly(ADP-ribose) polymerase (PARP) inhibitors have been demonstrated to be active in breast and ovarian tumours harbouring germline mutations in either of the two BRCA genes, although their roles in PCa needs further investigation.

Platinum agents induce DNA crosslinks that are substrates for HR DNA repair, which is deficient in BRCA-mutated cells. Therefore, these tumours show high sensitivity to platinum-based chemotherapy, both in vitro^{96, 97, 98} and in vivo.^{99, 100, 101} Previous studies in ovarian cancer suggested that mutations in both genes were associated with similar responses to platinum-based chemotherapy.¹⁰¹ Recently, Yang et al.¹⁰² have reported a series of 316 ovarian cancer patients treated with surgery and adjuvant platinum-based chemotherapy in which BRCA2 mutations were associated with improved outcomes, while BRCA1-mutated or -hypermethylated tumours were not significantly different from BRCA wild-type cases. BRCA2 mutations were also associated with an increased rate of response to primary platinum chemotherapy. Interestingly, an increase in genomic instability was found in BRCA2-related tumours, but not in those with BRCA1 loss, due to mutation or hypermethylation. This could indicate that BRCA2 lesions cause more substantial HR defects than BRCA1.

Gallagher et al.¹⁰³ have recently suggested that those patients with metastatic disease and BRCA mutations do not respond less to the standard chemotherapy regimen with docetaxel plus prednisolone than PCa which are BRCA wild type. We observed that BRCA2 mutations are associated with worse outcome for all survival endpoints, except for survival from metastasis. No difference was seen between mutation carriers and non-carriers, which may be due to a more favourable response to chemotherapy treatments.⁶³

PARP is an enzyme that produces large chains of poly(ADP-ribose) from NAD⁺.¹⁰⁴ PARP is involved in the repair of single-strand DNA breaks and its inhibition produces the accumulation of these lesions which may end in the arrest of the replication fork and the formation of DSBs.¹⁰⁵ These DSB are only proficiently repaired by HR. In the absence of HR, as occurs when BRCA mutations are present, these DSBs are repaired by error-prone forms of DSB, such as non-homologous en-joining potentially resulting in accumulation of gene aberrations and loss of cell viability.^{106, 107} Significant efficacy has been established for PARP inhibitors in BRCA-mutated breast and ovarian tumours,^{108, 109, 110} although the evidence for PCa is still limited. Clinical trials are currently ongoing to assess the role of PARP inhibitors in PCa BRCA-related, as well as in sporadic tumours. The most effective scheme is still to be established. At the moment, clinical trials are investigating the efficacy of different combinations of chemotherapy and PARP inhibitors in metastatic patients, but there are no data of its role in the adjuvant setting, either in monotherapy or in concomitance with other agents. There is concern that it may potentiate the effects of some modalities such as radiotherapy.

‘BRCAness' in PROSTATE CANCER

Several studies in breast and ovarian cancer have identified sporadic cancers with molecular and clinical characteristics similar to BRCA-related tumours,⁵⁶ defining a group of tumours with BRCAness characteristics that benefits from the same therapeutic strategies as those for BRCA mutant tumours.^{101, 108, 109, 110}

The aggressive phenotype and poor prognosis associated with PCa in patients with germline BRCA mutations is very likely to be associated with a different cancer biology from BRCA wild-type PCa. The identification of these differential molecular characteristics could be essential in tailoring treatment for this population, but it may also be of great importance for the management of some sporadic PCa cases which may have similar clinical characteristics and aggressive behaviour.

Gene promoter hypermethylation is a mechanism of BRCA function loss in sporadic breast and ovarian cancer that triggers tumours with similar characteristics to those with germline mutations in these genes, although this has still not been analysed in PCa. At this time, the study by Bednarz et al.⁶² is the only one that has analysed a series of sporadic PCas, finding that those with BRCA1 somatic loss presented more frequently with node involvement and shorter disease-free survival than those with functional BRCA1.

Conclusion

Germline mutations in the BRCA genes, mainly in BRCA2, not only increase the risk of developing PCa, but also have implications in the prognosis and management of the disease. BRCA-related PCa is usually aggressive, and radical treatments are preferred to surveillance, even for low-risk cases. Further studies are needed to design a tailored management for these patients. An ongoing study, IMPACT, will clarify the benefits of PCa screening in this higher-risk population. Promising clinical trials are evaluating the role of PARP inhibitors in the metastatic setting, but more studies are needed to establish the role of adjuvant treatment, with PARP inhibitors and/or conventional chemotherapy. The role of chemoprophylaxis in patients with high risk of aggressive forms of PCa also needs to be addressed. A better characterization of BRCA-related prostate tumours would help to identify sporadic cases with potential lethal forms of the disease that might benefit from the therapeutic strategies designed for BRCA-mutated tumours.

The authors declare no competing financial interests.