sensitivity to platinum chemotherapy, better five-year survival, high
grade serous (HGS) histology, and somatic TP53 mutations [5–8]. Our
group and others have recently shown that somatic mutations in
BRCA1/2 and some non-BRCA1/2 HRR genes including RAD51C are also
associated with platinum sensitivity, improved survival, and sensitivity
to poly (ADP-ribose) polymerase (PARP) inhibitors [2,9–11].
Promoter hypermethylation (methylation) is another biologic
mechanism of reducing gene expression and occurs frequently in cancer
[12].BRCA1 promoter methylation is common in ovarian carcinoma
with rates from 8 to 15% and is associated with both reduced RNA and
protein expression [12–19]. In HGS ovarian carcinomas, The Cancer Ge-
nome Atlas (TCGA) identified BRCA1 and RAD51C as the only HRR genes
in which promoter methylation correlated with reduced RNA expres-
sion [1].RAD51C methylation has been described in 1–3% of ovarian car-
cinomas [1,15].
Given the improved outcomes seen with BRCA1 mutation and the
functional impact of BRCA1 promoter methylation (decreased RNA
and protein expression), we hypothesized that BRCA1 and RAD51C
methylation would also be associated with improved outcomes
and similar clinicopathological characteristics. However, there
are conflicting reports of the prognostic value of BRCA1 methylation
in regards to sensitivity to platinum chemotherapy and survival
[1,14,15,20,21]. No previous study except TCGA, which was limited
only to ovarian carcinomas of HGS histology, has considered both
methylation and mutation in many HRR genes in comparing out-
comes [1,14,15,20,21]. We sought to evaluate BRCA1 methylation
and RAD51C methylation in ovarian carcinomas of varying histolo-
gies and compare outcomes and characteristics of methylated versus
mutated cases.
2. Methods
Patients with ovarian, fallopian tube, or peritoneal carcinoma pro-
vided IRB-approved informed consent to enroll in the University of
Washington gynecologic oncology tissue bank at the time of their pri-
mary debulking surgery. Patients diagnosed with carcinoma at the
time of planned prophylactic bilateral salpingo-oophorectomy were ex-
cluded. Germline and neoplastic DNA was sequenced using BROCA, a
targeted capture and massively parallel sequencing platform previously
described [2,22,23].
Carcinomas with a damaging germline or somatic mutation in ATM,
ATR,BARD1,BLM,BRCA1,BRCA2,BRIP1,CDK12,CHEK2,MRE11A,NBN,
PALB2,RAD51C,RAD51D,RBBP8,SLX4,orXRCC2 were classed as having
HRR deficiency.
Methylation of RAD51C and BRCA1 in neoplastic DNA was assessed
by bisulfite conversion using the Zymo Research EZ DNA Methylation-
Direct kit followed by methylation-specific PCR as previously described
[9,17]. Methylation was evaluated in primary carcinomas and, if avail-
able, in corresponding recurrent carcinomas.
Patients were considered to have a “strong family history”of cancer
if they had a relative with OC, a relative with breast cancer before age 50,
or two relatives with breast cancer at any age.
All statistical analyses were pre-planned based on our hypothe-
ses. The Fisher's exact test was used to test the significance of
contingency tables (Tables 1 and 2). Progression-free survival (PFS)
was defined as the time between study enrollment and disease
progression or death. Patients who did not receive chemotherapy
and those for whom chemotherapy information was not available
were excluded from the PFS analysis. Overall survival (OS) was
defined as the time between study enrollment and last follow up
visit or death. Patients with stage I disease were excluded from
survival analyses. Kaplan Meier curves were generated for OS and
PFS and evaluated by the Log Rank test. The study sample size was
not large enough to correct for confounders with multivariate
analysis.
3. Results
Primary carcinomas from a total of 332 patients were evaluated.
Table 1 provides demographic information for all subjects. For 12 pa-
tients, paired recurrent neoplasms were also analyzed for methylation.
Tables 1 and 2 summarize the methylation and germline mutation
status of the primary carcinomas. Mutation information for all but
three patients was previously published [2]. Sixty-nine (20.8%) carcino-
mas had a germline mutation in one or more of the HRR genes assayed
(1 ATM,2BARD1,38BRCA1,13BRCA2,4BRIP1,3CHEK2,1NBN,1PALB2,
3RAD51C, and 4 RAD51D). Twenty-eight (8.4%) carcinomas had a so-
matic mutation in one or more HRR genes (2 ATM,16BRCA1,5BRCA2,
1BRIP1,4CHEK2,1MRE11A,1RAD51C,1SLX4). Among these mutated
cases, six (1.8%) had mutations in more than one gene. These cases
were a germline BRCA1 mutation with a somatic CHEK2 mutation, a
germline BRCA1 mutation with a somatic ATM mutation, a germline
BRCA1 mutation with a germline BARD1 mutation, a germline CHEK2
mutation with a somatic BRCA2 mutation, a somatic CHEK2 mutation
with a somatic SLX4 mutation, and a case with somatic mutations in
BRCA1,BRIP1,andMRE11A.
A total of 31 primary carcinomas were found to have methylation of
either BRCA1 (22, 6.6%) or RAD51C (9, 2.7%). Two hundred seven
(62.3%) carcinomas had neither methylation nor mutation. No carcino-
ma had both a germline mutation in a HRR gene and methylation of ei-
ther BRCA1 or RAD51C (P= 0.0008). One carcinoma with a somatic
mutation in BRIP1 had methylation of BRCA1.
When available, paired recurrent carcinomas were also evaluated for
somatic mutations and methylation. All paired carcinomas had concor-
dant methylation status: eleven of the recurrent carcinomas were
unmethylated as were their corresponding primary neoplasms, and
one recurrent carcinoma was BRCA1 methylated, as was the primary
neoplasm associated with it.
Table 1 describes the clinical features of the study population by
methylation and mutation status. At time of diagnosis, patients with
BRCA1 methylation of their carcinoma were on average 5.6 years youn-
ger (mean 57.7 years ± 2.5) than patients without germline or somatic
mutations and without methylation (mean 63.3 years ± 0.79) (P=
0.029). Patients with germline or somatic BRCA1 mutations were youn-
ger at diagnosis than patients without mutations and without methyla-
tion by 9.3 years ± 1.7 (mean 54.1 years ± 1.4) (Pb0.0001). The
difference in age at diagnosis was more pronounced when the analysis
was restricted to patients with germline BRCA1 mutations, who were on
average 10.9 years younger (mean 52.4 years ± 1.4) than patients with-
out mutations and without methylation (Pb0.0001). Patients with
germline or somatic mutations in HRR genes other than BRCA1 or
RAD51C were also younger at diagnosis than patients without muta-
tions and without methylation by 5.83 ± 2.04 years (mean 57.5 ±
1.9 years) (P= 0.005).
Germline BRCA1 mutations and BRCA1 methylation were both asso-
ciated with HGS histology when compared to carcinomas without mu-
tation or methylation (95% vs 71%, P= 0.001; 91% vs 71%, P=0.045),
but this was not true for BRCA1 somatic mutations. Of the 16 somatically
mutated BRCA1 carcinomas, eight were HGS, four were undifferentiated,
two were endometrioid, one was clear cell, and one was a carcinosarco-
ma. A strong family history of breast or ovarian carcinoma was more
common in patients with germline HRR mutations including mutations
in BRCA1 and RAD51C (55% vs 17% of patients whose carcinomas were
neither mutated nor methylated, Pb0.0001). Patients with germline
HRR gene mutations had a significantly higher incidence of a personal
history of breast cancer (21.7% vs 3.4% of those with neither mutation
nor methylation, Pb0.0001). Advanced stage disease (stage III or IV),
utilization of neoadjuvant chemotherapy, and achievement of optimal
primary surgical debulking was similar in carcinomas with or without
mutation or methylation. As previously published, germline or somatic
BRCA1 mutations were associated with increased platinum sensitivity
compared to carcinomas without mutation or methylation (80% vs
2S.S. Bernards et al. / Gynecologic Oncology xxx (2017) xxx–xxx
Please cite this article as: S.S. Bernards, et al., Clinical characteristics and outcomes of patients with BRCA1 or RAD51C methylated versus mutated
ovarian carcinoma, Gynecol Oncol (2017), https://doi.org/10.1016/j.ygyno.2017.12.004