Hereditary Ovarian Carcinoma: Cancer Pathogenesis Looking beyond BRCA1 and BRCA2

Besides BRCA1 and BRCA2, several other inheritable mutations have been identified that increase ovarian cancer risk. Surgical excision of the fallopian tubes and ovaries reduces ovarian cancer risk, but for some non-BRCA hereditary ovarian cancer mutations the benefit of this intervention is unclear. The fallopian tubes of women with hereditary ovarian cancer mutations provide many insights into the early events of carcinogenesis and process of malignant transformation. Here we review cancer pathogenesis in hereditary cases of ovarian cancer, the occurrence of pre-invasive lesions and occult carcinoma in mutation carriers and their clinical management.


Introduction
Familial or hereditary ovarian cancer syndromes (HOC) encompass a number of tumor suppressor genes in which heritable mutations together account for 24% of epithelial ovarian cancer cases [1]. The last decade has seen major clinical advances in the treatment of ovarian cancer including the use of angiogenesis inhibitors [2] and poly-ADP ribose polymerase inhibitors [3][4][5][6] which have dramatically improved overall survival in this disease. Despite these improvements, ovarian carcinoma remains the most lethal gynecologic malignancy with more than 21,000 new cases and 13,000 deaths in the U.S. in 2020 [7]. Among these, hereditary cancer syndromes are a prime target for early detection and prevention. Current guidelines recommend germline genetic testing for all women with ovarian cancer, cascade testing of family members and risk-reducing surgery for those with actionable mutations [8,9]. In this review, we will summarize what is known about hereditary ovarian cancer pathogenesis in women with germline mutations in several DNA repair pathways.
The standard of care for women with BRCA1 and BRCA2 mutations is risk-reducing surgical resection of the fallopian tubes and ovaries (RRBSO) [8]. Mutations in non-BRCA HOC-associated genes are more frequently being detected with increased utilization of germline genetic screening but their optimal management is less clear. For several non-BRCA HOC mutations the National Comprehensive Cancer Network cites unknown risk or insufficient evidence to support increased screening or prophylaxis [8]. The mechanistic similarities and differences between carcinogenesis in BRCA mutation carriers, other HOC mutation carriers and sporadic cases remain unclear. Here, we discuss pre-neoplastic lesions and occult carcinoma occurring in women with BRCA mutations as well as other rare and recently described HOC mutations and describe the early events in carcinogenesis in this population.

Pathogenesis in Hereditary Ovarian Cancer: Insights from Risk-Reducing Surgery Specimens
The fallopian tubes of women with HOC mutations who have undergone RRBSO provide important insights into the process of malignant transformation. Many previous studies have made inferences about ovarian cancer pathogenesis from high grade serous carcinoma cases with concurrent precursor lesions. There is, however, a concern that rapidly expanding invasive carcinoma destroys precursor lesions and fundamentally alters the precancerous environment of the tube [36,37]. What then can we learn from the histopathological and molecular genetic analysis of the fallopian tubes of women with HOC mutations but without cancer?
The distal fallopian tube epithelia (FTE) of BRCA mutations carriers shows several transcriptional differences from tubal epithelia of unaffected women. Monoallelic BRCA1 and BRCA2 mutation carriers have significant modulation of genes associated with transcriptional regulation, cell cycle control and cell adhesion [38]. Multiple pathways in carcinogenesis are implicated including MAPK, adipokine, TGF-B and p53 signaling [39]. Notably, the differentially expressed genes in BRCA mutation carriers closely resemble serous carcinoma, rather than those of the low-risk (non-carrier) fallopian tube. Other studies have shown transcriptional changes in tubal epithelia of BRCA mutation carriers that indicate altered responses to inflammation and cellular stress. Compared to wild-type, tubal epithelia from BRCA1 mutation carriers have increased expression of NAMPT, an enzyme involved in glucose metabolism and cellular stress as well as C/EBP-δ, a transcription factor implicated in inflammation, DNA damage response and tumorigenesis [39,40]. The DNA damage and repair markers P53 and γ-H2A also show high nuclear expression in the FTE and ovarian surface epithelium of BRCA1/2 mutation carriers compared to non-mutation carriers [41]. The physiologic process of ovulation exposes the tubal fimbriae to reactive oxygen species and induces DNA double strand breaks [42] and BRCA mutation carriers appear to be more susceptible to these stresses. The cyclic hormonal context provided by the ovulatory cycle cannot be ignored [43,44].
BRCA1 and BRCA2 heterozygous fallopian tubes collected during the post-ovulatory luteal phase have a global gene expression profile similar to high grade serous carcinoma but distinct from follicular phase mutation carriers and both follicular and luteal phase controls [39]. BRCA1-mutated luteal phase fallopian tubes also demonstrate differential expression of more than 100 inflammation-related and NF-κB signaling genes compared to post-ovulatory non-mutation carriers which likely contributes to tumorigenesis [45]. Despite these observed transcriptional changes, most women with BRCA mutations will never develop ovarian cancer, complicating the identification of the specific transcriptional program that leads to neoplasia. One study has addressed this challenge by using BRCA1 fallopian tubes with STIC (i.e., demonstrated proclivity to neoplasia) and performing microdissection and expression profiling of the histologically normal epithelial regions [46]. The authors compared gene expression in this histologically normal FTE from mutation carriers to non-mutation carriers. Notably, nearly one quarter of the highly differentially expressed genes were also differentially expressed in BRCA1-associated carcinomas. The preneoplastic signature present in BRCA1 mutation carriers included modulation of EFEMP1, involved with cell-cell communication and angiogenesis regulation; E2F3, a cell cycle regulator implicated in malignant transformation; and CSPG5, a growth factor and activator of HER2/neu, all of which have been described in human carcinogenesis from multiple disease sites.
Epigenetic mechanisms also contribute to hereditary carcinogenesis and several studies have addressed this using risk-reducing surgery specimens from BRCA mutation carriers without cancer. FTE from BRCA mutation carriers show increased promoter methylation of tumor suppressor genes compared with non-mutation carriers [47,48]. This increased methylation pattern does not correspond to preneoplastic foci and there is no difference in methylation between p53 signatures and adjacent histologically normal FTE [47]. Disruption of existing epigenetic control is also a relevant mechanism. Analysis of differentially methylated genes in prophylactically removed fallopian tubes from BRCA mutation carriers shows the gene HOXC4 is highly expressed in the fimbriae which is a key trigger for activation-induced deaminase and induces epigenetic reprogramming [48]. Bartlett et al show overexpression of activation induced deaminase in cultured fimbrial cells modifies the methylation pattern to one that closely resembles HGSC and cellular differentiation. Yet another epigenetic mechanism is monoubiquitination of histone H2B which is performed by BRCA1 and regulates gene expression, DNA damage response and chromatin segregation [49]. In non-mutation carriers, there is incremental loss of monoubiquitination of histone H2B from histologically normal fallopian tube, to STIC to HGSC, indicating this is an early and progressive epigenetic step in carcinogenesis [50]. Interestingly, germline BRCA1 mutants do not show loss of monoubiquitination of histone H2B compared to wild type, including after sub-analysis of BRCA1 mutations specifically involving the functional domain responsible for ubiquitination [49]. While these findings need to be validated using larger sample sizes, they suggest one distinct mechanism of carcinogenesis occurring in sporadic but not hereditary cases.
These findings add to the growing evidence of haploinsufficiency as a carcinogenic mechanism in HOC. Contrary to the "two-hit hypothesis" which requires loss-of-function mutations in both alleles of a tumor suppressor gene for tumorigenesis, the studies highlighted above indicate multiple pathogenic tubal changes occur with monoallelic germline BRCA mutation. Increased cancer susceptibility in the fallopian tube appears to begin with loss of a single allele of BRCA1 or BRCA2. There is parallel evidence that other, non-BRCA mutations may also model haploinsufficiency. Lee et al [51] showed both monoallelic and biallelic PALB2 mutated breast tumors have no significant difference in homologous recombination deficiency scores. Tumors with monoallelic germline PALB2 mutations are also more likely to have high risk frameshift mutations than somatic PALB2-null tumors [52]. Additionally, both preneoplastic FTE lesions and primary peritoneal carcinoma have been reported in PALB2 heterozygote patients [53,54]. Taken together, loss of heterozygosity does not appear to be the rate-limiting step in HOC pathogenesis.

Types of Preneoplastic Fallopian Tube Lesions and Significance
Several preneoplastic histopathological entities associated with HGSC have been identified in FTE including the p53-signature [30,31], secretory cell outgrowths (SCOUTs) [55], serous tubal intraepithelial lesions (STILs) and serous tubal intraepithelial carcinoma (STIC) [56] and are depicted in Figure 2. Secretory cell outgrowths (SCOUTs) are arrays of >30 cells found in either the proximal or distal tube with absent ciliated differentiation, loss of PAX2 and intact p53 [55]. It remains unclear if SCOUTs progress to other pre-neoplastic lesions or serous carcinoma. The p53 signature is a closely related entity but characterized by loss of p53 with accumulation of the deficient protein. The p53 signature occurs in the distal fallopian tube, involves an array of >12 benign secretory cells, exhibits strong p53 immunostaining, low proliferative index and co-localizes with gamma-H2AX [31,57]. In contrast, serous tubal intraepithelial carcinoma (STIC) is a noninvasive lesion characterized by nuclear enlargement and hyperchromasia, chromatin aggregation and strongly positive p53 immunostaining [58]. Additionally, they have loss of polarity, epithelial tufting, and high proliferative index [59]. Intermediate lesions with aberrant p53 expression but lacking these diagnostic features of STIC and low proliferation are termed serous tubal intraepithelial lesions (STIL) [60]. Although the clinical significance of some of these lesions remains unclear, there is growing evidence that they represent a spectrum of "early serous proliferations" that coexist with or precede the development of HGSC [61,62]. Evolutionary analyses show that p53 signatures and STICs are the precursors to ovarian carcinoma and ultimately metastases [63]. However, the paradigm of transformation of a single preneoplastic lesion into HGSC has been challenged by whole-exome mutational analysis of micro-dissected p53 signatures and STIC. Using this approach, Wu et al found multiple unique p53 mutations present in multiple STIC lesions isolated from a single patient. In patients with both STIC and occult carcinoma, phylogenetic analysis showed STIC was in some cases clonally related and in others, not [36]. This suggests multiple STIC may develop in tandem and divergently before one expands and becomes HGSC.

Incidence of Preneoplastic Lesions and Occult Carcinoma in HOC Mutation Carriers
Reports on the incidence of STICs in patients with increased hereditary ovarian cancer risk vary widely [56]. The diagnosis of these rare lesions has poor inter/intra-observer reproducibility and is dependent on micro-sectioning technique [25,64]. For BRCA1 and BRCA2 mutation carriers, the incidence of isolated STIC was thought to be 3% based on pooled data from more than 2000 RRBSOs from nine studies [65]. Recent data indicate that the incidence is higher. A multicenter retrospective study of more than 450 BRCA1/2 mutation carriers undergoing RRBSO showed a 3.5% incidence of STIC but when a subset of these samples underwent detailed pathologic review, 12.4% were found to have either STIC, STIL or invasive carcinoma [66]. A prospective study which used the SEE-FIM protocol for all pathologic analyses found STIC lesions in 9.9% of 500 RRBSO specimens from patients with BRCA1/2 mutations [67]. There is a paucity of clinical data available on the prevalence of precursor lesions in asymptomatic women with other, non-BRCA mutations implicated in HOC due to their rarity. The largest study of this population to date was performed by Rush et al and included 273 patients with non-BRCA HOC mutations or high-risk family history who underwent RRBSO. Isolated STIC was found in one PALB2 mutation carrier and none were found in the high-risk family history group or among 41 patients with defects in other homologous recombination genes or mismatch repair genes [54]. One other study has identified intraepithelial neoplasia without invasive carcinoma in a PALB2 mutation carrier, although only preliminary results are available [68]. A small retrospective study of 18 patients showed no serous intraepithelial precursor lesions among patients with RAD51C, BARD1, BRIP1, PALB2 or CHEK2 mutations who underwent RRBSO [11].
Multiple studies have reported on occult carcinoma detected at RRBSO, with estimates between 1-12% for BRCA mutations carriers [69]. A recent systematic review suggests the incidence may be at the lower end of published estimates [70]. Using 27 studies including over 6000 high risk women undergoing RRBSO, Piedimonte et al found a pooled prevalence of occult carcinoma of only 1.2%. This study was limited by excluding non-BRCA mutations and by combined analysis of BRCA mutations carriers with patients with high-risk family history, which could underestimate the true prevalence. A summary of the incidence of preneoplastic lesions and occult carcinoma identified following riskreducing surgery including non-BRCA mutations is provided in Table 1. Though limited to a few studies with small sample sizes, these data can inform preoperative counselling and planning. They indicate there is minimal occult disease at time of risk-reducing surgery for most non-BRCA HOC mutations. Pathologic findings from larger cohorts of women with these rare mutations may support or refute the timing of risk reducing surgery in this population. The unique observation of STIC in patients with PALB2 mutations warrants further investigation.

Clinical Management of Isolated Pre-Invasive Lesions in Hereditary Ovarian Cancer Mutation Carriers
The optimal clinical management of HOC mutation carriers with isolated pre-invasive FTE lesions detected following prophylactic surgery remains unknown. A systematic review of 14 studies showed that management practices vary widely: 36% of patients received subsequent hysterectomy, 21% received lymph node assessment and 14% received adjuvant chemotherapy. Among women with BRCA mutations with STIC identified after RRBSO, 7.5% of patients later developed primary peritoneal carcinoma [71]. Management is thus directed at risk reduction for and early detection of primary peritoneal carcinoma. Surveillance with physical exam, CA-125 and ultrasound and CT imaging can be considered but may be low yield given that carcinoma develops several years after RRBSO [71,72]. Limited evidence argues against comprehensive staging surgery for this group. In one systematic review, zero patients with BRCA1 or BRCA2 mutations and isolated STIC who subsequently underwent staging surgery had invasive cancer [72]. Similarly, Wethington et al reported on seven patients with BRCA mutations and isolated STICs who underwent staging and none had other disease [73]. Adjuvant chemotherapy may not be warranted in cases of isolated STIC but data on use and development of primary peritoneal carcinoma are lacking [72].

Incidence of Ovarian Carcinoma with Rare HOC Susceptibility Mutations
Pathogenic variants in PALB2, ATM and BARD1 are recognized by the NCCN as involved in hereditary breast and ovarian cancer syndrome but have insufficient evidence to determine their ovarian cancer risk [74].

BARD1
Massive parallel sequencing studies have shown BARD1 loss-of-function mutation occurs in <1% of women with ovarian cancer [1]. Among women with BARD1 mutations one study appeared to show an increased ovarian cancer risk (OR 4.2 95% CI 1.4-12.5) compared to population estimates from large, publicly available exome sequencing datasets [10]. Other U.S. cohorts compared to the same controls have not found BARD1 to be significantly associated with ovarian carcinoma [75].

PALB2
PALB2 was initially reported to occur at similar frequency in ovarian cancer to controls in large European cohorts [76]. However, the largest study to-date of patients with PALB2 mutations including more than 500 families found a relative risk of ovarian cancer of 2.91 (95% CI 1.40-6.04) compared to country-specific population incidence [77]. Primary peritoneal carcinoma has also been reported in a patient with PALB2 mutation following RRBSO [53].

ATM
For ATM, the prevalence of this mutation was 0.6% out of more than 1500 women with ovarian cancer reported to SEER registries. ATM was associated with a relative risk between 2.25 and 2.97% and has been estimated from clinical trial populations [10] using Exome Aggregation Consortium controls [78] and local controls [79].
Recent evidence suggests PALB2 and ATM are associated with increased risk and expanded screening and risk-reducing surgery should be investigated for these groups. Studies of BARD1 mutation and ovarian cancer risk are contradictory and the link remains unclear.
The MRN complex (MRE11, RAD50, and NBN) is implicated in nonhomologous end joining DNA repair and has rarely been reported in association with ovarian cancer.

NBN
One biallelic NBN germline mutation was identified from 354 Russian women with ovarian carcinoma [80]. Other reports from the same geographic region have identified the same founder mutation in 1.7% of women with ovarian cancer [81]. However, in another study comparing more than 3000 women with epithelial ovarian cancer to healthy controls, there was no difference in NBN mutation frequency [76].

MRE11
The first report of MRE11-associated ovarian carcinoma was from a cohort of 151 women with hereditary breast or ovarian cancer, of whom one was found to have a germline MRE11 mutation [82]. Parallel genomic sequencing of 360 women with pelvic serous carcinoma also identified one patient with this mutation [1]. Conversely, this patient had a concurrent BRCA2 mutation and tumor analysis showed loss of heterozygosity for BRCA2 only. Another study did not find an increased incidence of MRE11 mutation in women with ovarian cancer [10].

RAD50
RAD50 mutations were not identified in a prospective series of 62 patients with ovarian carcinoma [83] but were found in another unselected ovarian cancer cohort with one affected patient out of 360 [1].
The studies of MRN complex (RAD50, MRE11, NBN) mutations and ovarian cancer summarized above are contradictory and are insufficient to determine ovarian cancer risk. Further investigation of the impact of pathogenic variants in these genes is needed.

Strategies for Risk-Reduction in HOC Mutation Carriers
Ovarian cancer screening is ineffective for women of average risk but may have a role for those at high risk due to HOC mutations or family history. Screening with a combination of serial CA-125 and pelvic ultrasound is commonly used for women in this high-risk group. NCCN guidelines note this strategy can be considered but that it has not been shown to decrease ovarian cancer mortality [8,84]. Studies specific to high ovarian cancer risk populations have shown that screening with pelvic ultrasound and CA-125 results in diagnosis at earlier stage [85,86] and may improve survival [86], although an ovarian cancer-specific survival benefit has not been demonstrated. In the absence of data on screening for women with non-BRCA HOC mutations, it is reasonable to consider this approach for this group.
Surgical risk-reduction with bilateral salpingo-oophorectomy is the mainstay of management for women with HOC mutations. In BRCA1 and BRCA2 mutation carriers, RRBSO reduces ovarian, fallopian tube and peritoneal cancer risk by 80% and all-cause mortality by 68% [87]. BRCA1 mutation carriers are recommended to undergo a RRBSO between 35 and 40 years of age, while BRCA2 mutation carriers may delay the procedure until age 40 and 45 because of later presentation of disease [88]. There are currently no prospective data available on RRBSO for other non-BRCA HOC mutation carriers. The mutation-specific lifetime risk of ovarian cancer compared to BRCA-negative/family history positive women (2.6%) has been used as a threshold to guide management [8,89]. On this basis, current guidelines recommend RRBSO or consideration of RRBSO for patients with BRIP1, MSH2, MLH1, EPCAM, RAD51C and RAD51D mutations [8] (Table 2). Family history and other independent risk factors for ovarian cancer should also be considered in timing RRBSO. However, there remain several genes associated with HOC but with insufficient evidence to determine their ovarian cancer risk or to recommend surgical prophylaxis (Figure 1).
While the indication for RRBSO is reduction of ovarian cancer risk, this intervention may have a secondary effect of reducing breast cancer risk. In a study of 876 women with BRCA1 and BRCA2 mutations, Choi et al found a decrease in breast cancer incidence in the first five years following RRBSO and a weaker but significant decrease in long-term risk for BRCA1 mutation carriers only [90]. A study by Stjepanovic et al also found that premenopausal RRSBO decreased short term risk of breast cancer in BRCA1 mutation carriers, though longer follow-up and larger sample size is needed to determine the potential benefit in BRCA2 carriers [91]. For women with other HOC mutations associated with significant breast cancer risk such as CHEK2, PALB2 and ATM, the impact of RRBSO on breast cancer risk remains unknown. Chemoprevention with oral contraceptives can reduce cancer risk for women with HOC who desire contraception but have not completed childbearing or undergone RRBSO. For women with BRCA1 and BRCA2 mutations, oral contraceptive use reduces the risk of ovarian cancer by 50% with a further 36% risk reduction for each additional 10 years of use [92]. There are mixed data [93] on the associated breast cancer risk for women with BRCA mutations receiving oral contraceptives with no significant risk in some reports [92,94] and mild to moderate increased risk in others [95,96]. Patients with HOC mutations should be informed of the possible increase in breast cancer risk when initiating oral contraceptives.
Two studies have shown aspirin may decrease ovarian cancer risk in the general population and this approach may prove useful for women with HOC mutations. A pooled analysis of the Ovarian Cancer Association Consortium found that aspirin use was modestly associated with reduced ovarian cancer risk (OR 0.91 95% CI 0.84-0.99) [97]. A Danish case-control study found no difference in risk for anytime aspirin users but continuous, long-term aspirin use was associated with decreased epithelial ovarian cancer risk (OR 0.56 95% CI 0.32-0.97) [98]. The randomized phase II STICs and Stones trial (NCT03480776) is currently underway to evaluate aspirin use in women with BRCA1 and BRCA2 mutations prior to RRBSO.

Ongoing Clinical Trials
Several clinical trials are currently exploring novel risk-reducing interventions for women with high-risk hereditary ovarian cancer mutations and are detailed in Table 3. Although effective in decreasing ovarian cancer risk, RRBSO has many negative sequelae including increased risk of stroke and cardiovascular disease, dementia, and osteoporosis [99]. Besides this, women undergoing oophorectomy are frequently premenopausal, with surgery prompting abrupt genitourinary syndrome of menopause, sexual dysfunction, vasomotor symptoms and mood disturbances which negatively impact quality of life [100]. Given the evidence that the fimbriae of the fallopian tube, rather than the ovary is the etiological site of high-grade serous carcinoma, RRBSO may not be the only preventative strategy, particularly for patients with HOC mutations that impart lower ovarian cancer risk. Two ongoing clinical trials investigate ovarian conservation in women with HOC mutations. SOROCk (NCT04251052) will investigate the efficacy of risk reduction following salpingectomy only and is limited to patients with BRCA1 mutations. In the Radical Fimbriectomy trial (NCT01608074) patients with BRCA1/2 mutations will undergo laparoscopic resection of the fallopian tube and fimbrio-ovarian junction with ovarian conservation. The preliminary data from this trial presented in 2018 demonstrated that of 121 patients who underwent this procedure, 1.7% had STIC and 0.8% had invasive carcinoma with an overall favorable safety profile [101].
Alternately, a staged procedure with salpingectomy followed by interval oophorectomy could temporarily preserve ovarian function while still offering risk reduction. TUBA-WISP-II (NCT04294927) seeks to address the safety and efficacy of this staged approach. TUBA-WISP-II will examine the primary outcome of high grade serous ovarian cancer incidence and secondary outcomes of peri-operative morbidity and completion of oophorectomy for women with BRCA and other HOC mutations. Both trials include women with BRCA and other HOC mutations. Lastly, the PROTECTOR (ISRCTN25173360) and WISP (NCT02760849) trials will focus on quality of life outcomes following staged RRBSO including sexual function, psychological well-being, and satisfaction/regret [102]. Preliminary data from WISP presented in 2019 showed no patients had developed carcinoma and women who underwent a staged procedure had fewer menopausal symptoms and less decision regret, even when compared to women who received hormone replacement therapy [68].

Conclusions
Ovarian cancer pathogenesis in hereditary cases begins with altered responses to inflammation and cellular stress in the fallopian tube which occur in the context of cyclic hormonal control and ovulation. Epigenetic reprogramming and loss of monoubiquitination of histone H2B also play roles in tumor initiation. Multiple pathogenic changes are observed in the tubal fimbria of women with BRCA and other monoallelic tumor suppressor gene mutations, indicating cancer susceptibility is not dependent on loss of heterozygosity. Numerous studies have focused on the cancer risks of BRCA1 and BRCA2, but increased attention is needed to the multitude of non-BRCA mutations that impart ovarian cancer risk. So far, tubal preneoplastic lesions have been identified in PALB2 mutation carriers but not in those with other homologous recombination or mismatch repair gene mutations. Use of germline genetic screening is expanding, and we can expect to see greater numbers of patients with these rare mutations prior to cancer diagnosis. Delineation of mutation-specific cancer risks and the benefits of preventative interventions will allow more personalized management of ovarian cancer risk that also minimizes harm.