Comprehensive molecular comparison of BRCA1 hypermethylated and BRCA1 mutated triple negative breast cancers

Glodzik, Dominik; Bosch, Ana; Hartman, Johan; Aine, Mattias; Vallon-Christersson, Johan; Reuterswärd, Christel; Karlsson, Anna; Mitra, Shamik; Niméus, Emma; Holm, Karolina; Häkkinen, Jari; Hegardt, Cecilia; Saal, Lao H.; Larsson, Christer; Malmberg, Martin; Rydén, Lisa; Ehinger, Anna; Loman, Niklas; Kvist, Anders; Ehrencrona, Hans; Nik-Zainal, Serena; Borg, Åke; Staaf, Johan

doi:10.1038/s41467-020-17537-2

Download PDF

Article
Open access
Published: 27 July 2020

Comprehensive molecular comparison of BRCA1 hypermethylated and BRCA1 mutated triple negative breast cancers

Nature Communications volume 11, Article number: 3747 (2020) Cite this article

9567 Accesses
46 Citations
16 Altmetric
Metrics details

Subjects

Abstract

Homologous recombination deficiency (HRD) is a defining characteristic in BRCA-deficient breast tumors caused by genetic or epigenetic alterations in key pathway genes. We investigated the frequency of BRCA1 promoter hypermethylation in 237 triple-negative breast cancers (TNBCs) from a population-based study using reported whole genome and RNA sequencing data, complemented with analyses of genetic, epigenetic, transcriptomic and immune infiltration phenotypes. We demonstrate that BRCA1 promoter hypermethylation is twice as frequent as BRCA1 pathogenic variants in early-stage TNBC and that hypermethylated and mutated cases have similarly improved prognosis after adjuvant chemotherapy. BRCA1 hypermethylation confers an HRD, immune cell type, genome-wide DNA methylation, and transcriptional phenotype similar to TNBC tumors with BRCA1-inactivating variants, and it can be observed in matched peripheral blood of patients with tumor hypermethylation. Hypermethylation may be an early event in tumor development that progress along a common pathway with BRCA1-mutated disease, representing a promising DNA-based biomarker for early-stage TNBC.

Single-cell multiomics reveals the interplay of clonal evolution and cellular plasticity in hepatoblastoma

Article Open access 08 April 2024

Amélie Roehrig, Theo Z. Hirsch, … Eric Letouzé

A single-cell and spatially resolved atlas of human breast cancers

Article 06 September 2021

Sunny Z. Wu, Ghamdan Al-Eryani, … Alexander Swarbrick

Massively parallel screen uncovers many rare 3′ UTR variants regulating mRNA abundance of cancer driver genes

Article Open access 18 April 2024

Ting Fu, Kofi Amoah, … Xinshu Xiao

Introduction

Triple-negative breast cancer (TNBC) encompasses a subgroup of tumors defined by absence of estrogen and progesterone receptor expression and lack of amplification of the human epidermal receptor growth factor 2/erythroblastic oncogene B (HER2/ERBB2) gene. TNBCs comprise ~10% of all breast cancers and are clinically aggressive with an often poor prognosis, partly due to the lack of targeted therapeutics. Although being classified as a clinical tumor entity, TNBC tumors display a high degree of molecular heterogeneity. TNBC tumors are associated with pathogenic variants in the BRCA1 breast cancer susceptibility gene, with 7–20% of diagnosed patients harboring pathogenic germline or somatic variants^1,2,3,4. The BRCA1 protein has multiple distinct roles in maintaining genome integrity, particularly, through homologous recombination (HR)-mediated double strand break repair⁵. Tumor cells deficient for BRCA1 (or BRCA2) are considered HR-deficient (HRD) and sensitive to cytotoxic agents and poly (ADP-ribose) polymerase (PARP) inhibitors, which cause DNA damage or increased demand for double strand break repair⁶. The HRD phenotype is utilized in promising clinical studies of germline BRCA1/BRCA2-mutated breast or ovarian cancer^3,7,8. In addition, BRCA1/BRCA2-mutated breast cancers have been suggested to be more immunogenic than non-HR defective tumors^9,10,11. The increase in immunogenicity may be related to better responses to checkpoint blockade response, although this remains to be proven.

Between 40% and 70% of TNBC tumors are reported to have a presumed HRD phenotype, which exceeds the number of cases with germline/somatic BRCA1/BRCA2 inactivating variants^3,12,13. This suggests that other mechanisms and/or genes may confer a similar phenotype. DNA promoter hypermethylation could be an alternative mechanism of inactivating BRCA1. BRCA1 promoter hypermethylation has been reported in 16–57% of TNBCs across studies^{14,15,16,17,18,19,20}, superseding the frequency of germline BRCA1 alterations, however, with conflicting reports about association with prognosis (e.g. refs. ^{14,15,16,18,21}). Currently, we lack a detailed multi-layer comparison of BRCA1 hypermethylated versus BRCA1-mutated early stage TNBCs using current state-of-the-art profiling techniques that thoroughly investigates similarities and differences between the two groups.

In the current study, we pursued the hypothesis that BRCA1 promoter hypermethylation confers an omics phenotype identical to that of BRCA1-mutated TNBCs, and that the two entities have equivalent patient outcomes in response to standard of care chemotherapy. To this end, we analyzed a recently reported unselected population-based cohort of 237 early stage TNBC tumors profiled by comprehensive whole-genome sequencing (WGS), RNA sequencing, global DNA methylation analysis²², further complemented with 54 additional BRCA1-mutated tumors from previous studies^23,24. Herein, we sought the frequency of BRCA1 promoter hypermethylation, its tumor phenotype compared to tumors with BRCA1 inactivating genetic variants (somatic or germline, BRCA1-null), and its association with clinicopathological variables, molecular subtypes, and patient outcomes in early-stage TNBC. We demonstrate that BRCA1 hypermethylation is twice as frequent as BRCA1-null tumors in early-stage unselected TNBC and that elevated BRCA1 promoter methylation is detectable in peripheral blood DNA of patients with hypermethylation in the tumor. Moreover, we show that in terms of mutational, epigenetic, transcriptional, and immune infiltration profiles, BRCA1 hypermethylation confers a tumor phenotype practically identical to that of BRCA1-null cases. BRCA1 hypermethylation and BRCA1 mutation are equally associated with better outcome after adjuvant chemotherapy when compared to TNBC patients without BRCA1 inactivation, thus BRCA1 hypermethylation represents a promising DNA-based prognostic marker.

Results

BRCA1 mutations and promoter hypermethylation in TNBC

Table 1, Fig. 1, and Supplementary Fig. 1 outlines patient demographics, selection, and study layout. The original 237 TNBC patients (hereinafter referred to as the SCAN-B cohort) reported by Staaf et al.²² represent 58% of the total number of diagnosed TNBC cases in the studied healthcare region during the inclusion period (September 1 2010 to March 31, 2015), and has been shown to be representative of the total regional population with respect to key clinicopathological variables. Of these patients, 24.1% (57/237) were classified as BRCA1 promoter hypermethylated based on pyrosequencing, while 25 were BRCA1-null cases of which 19 carried germline variants and six somatic variants (Supplementary Data 1). Similar hypermethylation rates were observed across different years of diagnosis in the SCAN-B cohort: 21.0% hypermethylated cases diagnosed 2011, 22.2% in 2012, 26.4% in 2013, and 21.7% in 2014. Pyrosequencing classifications were corroborated by Illumina DNA methylation profiling data for BRCA1 gene associated CpGs (Fig. 2a), and markedly reduced BRCA1 mRNA expression from RNA sequencing for hypermethylated cases (Fig. 2b), similar to previous reports²⁵, that were in line with expression levels for cases with BRCA1 frame shift, nonsense and indel variants (Supplementary Fig. 2). Of the 57 hypermethylated cases, 51 (89.5%) showed concurrent LOH of BRCA1 (tumor cell content by WGS range 23–82%) with the six remaining cases having low estimated tumor cell content (between 11% and 23% by WGS), which possibly interfered with the copy number analysis. In comparison, 84% of BRCA1 germline cases showed LOH of the wild-type allele, again with low associated tumor cell content for cases without LOH (n = 3, 11%, 18%, and 27% by WGS). Furthermore, BRCA1 hypermethylated and BRCA1 germline mutated cases without BRCA1 LOH had similar proportions of rearrangement signature 3 and deletions with microhomology, representing prototypical signatures of BRCA1-deficient cancer²⁴ (Fig. 2c). This suggests that BRCA1 hypermethylation or BRCA1 germline alteration without concurrent LOH is rare in early-stage TNBC specimens and when observed may be due to tissue sampling limitations. Consistent with Fig. 2c, HRD classification from the previous report²² revealed that 98.2% (56/57) of hypermethylated cases were called as HR deficient by two different HRD algorithms^13,26. Presence of nonmalignant cells in tumors may skew analyses of genomic data obtained from bulk tumor analyses (such as RNA sequencing and DNA methylation). In the SCAN-B cohort, there was no statistical difference in tumor cell content estimated by WGS between hypermethylated and BRCA1-null cases (t-test, p = 0.47), suggesting that nonmalignant infiltration should not affect group-level conclusions. Tumor cell content showed a strong linear correlation with BRCA1 CpG allele methylation rate for hypermethylated cases when using WGS specific estimates also accounting for possible subclonality (r² = 0.84, slope = 0.90, Fig. 2d).

Table 1 Clinicopathological characteristics of SCAN-B TNBC patients.

Full size table

**Fig. 1: Study scheme, performed analyses, and cohorts used.**

**Fig. 2: BRCA1 hypermethylation, gene expression, and HRD association.**

Overall, BRCA1 hypermethylation was 2.3 times more frequent than BRCA1-null cases, and three times more frequent than BRCA1 germline alterations. In the SCAN-B cohort, BRCA1 hypermethylation was mutually exclusive with BRCA1-null cases and BRCA2 tumors with only one exception. One of the ten BRCA2-null SCAN-B cases displayed BRCA1 hypermethylation. This case, PD35990a, showed a pathogenic biallelic BRCA2 variant (p.Pro3194Asnfs*2 germline mutation and BRCA2 LOH) and distinct BRCA1 hypermethylation (61% CpG allele methylation) together with BRCA1 LOH. The tumor had genomic patterns of Substitution Signature 3, Rearrangement Signatures 2 and 3 and loss of BRCA1 gene expression, all characteristic of BRCA1- but not BRCA2-deficient tumors^24,27 (Fig. 2e).

Elevated BRCA1 promoter hypermethylation in peripheral blood

Forty-six of the 237 SCAN-B patients underwent clinical BRCA1/2 germline screening due to family history and/or young age at diagnosis according to Swedish guidelines. Nine cases showed germline inactivating BRCA1 variants and three had germline inactivating BRCA2 variants. Among the clinically screened patients without a pathogenic germline BRCA1/BRCA2 variant, 16 out of 34 (47.1%) displayed tumor BRCA1 hypermethylation while 18 did not. Patients with BRCA1 hypermethylated tumors were significantly younger at diagnosis compared to the patients negative for both tumor methylation and BRCA1/2 loss of function variants (median age 36 versus 50 years, t-test, p = 8e−5). In fact, age at diagnosis for hypermethylated patients was similar to patients with pathogenic germline BRCA1 variants (median age 36 versus 32 years, t-test, p = 0.82) (Fig. 3a). Based on oncogenetic counseling data, only one of the 16 BRCA1 hypermethylated patients had a first-degree relative with breast/ovarian cancer, indicating that these TNBC patients had likely been referred for screening based on young age rather than family history. SCAN-B patients with BRCA1 hypermethylated tumors (n = 16) also showed higher levels of BRCA1 promoter methylation in matched peripheral blood DNA than the 18 patients with unmethylated tumors and no BRCA1/2 variants (hypermethylated: n = 16, mean CpG allele frequency = 4.03, median = 3.5, sd = 1.53; unmethylated: n = 18, mean = 3.03, median = 3.0, sd = 0.61; t-test p = 0.024, Wilcoxon’s test, p = 0.015, Fig. 3b). Trends also remained significant after transformation of pyrosequencing methylation rates to M-values (t-test, p = 0.02).

**Fig. 3: *BRCA1* methylation in peripheral blood, gene expression subtypes, and prognosis after therapy.**

To confirm these findings, we analyzed peripheral blood DNA from 71 additional SCAN-B cases from our cohort not subjected to prior clinical screening, including 39 cases with tumor BRCA1 hypermethylation and 32 without. Again, higher BRCA1 promoter methylation levels were found in blood DNA from patients with BRCA1 tumor hypermethylation (t-test, p = 0.005, Wilcoxon’s test, p = 0.01). The finding remained significant after transformation of methylation rates to M-values (t-test, p = 0.006, Wilcoxon’s test, p = 0.01). When we combined both clinically screened and unscreened SCAN-B patients (n = 105 in total, reanalyzed using the same instrument settings and reagent lots), we observed that the higher BRCA1 blood DNA methylation levels appeared independent of age (two-way ANOVA interaction model with tumor hypermethylation status and age groups, p = 0.002, Fig. 3c). This finding remained significant after transformation of methylation rates to M-values (two-way ANOVA interaction model with tumor hypermethylation status and age groups, p = 0.002).

Clinical and genomic features of BRCA1 methylated TNBC

BRCA1 promoter hypermethylation and BRCA1-null frequencies in subgroups of SCAN-B TNBC patients defined by clinicopathological variables and molecular subtypes are shown in Table 2. Hypermethylation frequency was especially high in patients under 50 years (46.2% frequency, Fisher’s exact test, p = 8e−5). A trend of lower age at diagnosis for hereditary BRCA1 cases, followed by BRCA1 hypermethylated cases, BRCA1/BRCA2 somatic cases, and non-altered cases was observed in the total SCAN-B cohort (Fig. 3d).

Table 2 BRCA1 promoter hypermethylation and BRCA1-null frequency in TNBC.

Full size table

Several gene expression based subtyping schemes have been proposed in breast cancer (e.g., the general PAM50²⁸, CIT²⁹, and IntClust 10³⁰ as well as TNBC specific subtypes³¹). In all instances, BRCA1 hypermethylation was strongly associated with the proposed basal-like phenotype (PAM50 basal-like p = 2e−5, CIT basal-like p = 1e−6, TNBCtype BL1 p = 0.02, and IntClust 10 cluster 10 p = 4e−6, Fisher’s exact test performed in a 2 × 2 basal-like vs. non-basal context) (Table 2 and Fig. 3e). The subtype proportions for BRCA1-null and non-BRCA1 altered cases are shown in Supplementary Fig. 3, demonstrating similarity between hypermethylated and BRCA1-null cases.

BRCA1 hypermethylation was also observed in clinicopathological and molecular subgroups not commonly associated with BRCA1-deficiency, including old patients, tumors with some ER staining (1–10% by immunohistochemistry), and in non-basal like gene expression subtypes, although at low relative frequencies similar to those observed in BRCA1-null cases (Table 2).

BRCA1 hypermethylation and patient outcome in TNBC

Association of BRCA1 promoter hypermethylation status with outcome after adjuvant chemotherapy (mainly FEC ± taxane therapy) was investigated using invasive disease-free survival (IDFS) as the primary clinical endpoint in 149 eligible SCAN-B patients (Fig. 1). In both univariable Cox regression and Kaplan–Meier analyses, BRCA1 hypermethylation alone or combined with BRCA1-null cases was associated with significantly longer IDFS than non-altered TNBC cases (Fig. 3f, which also shows univariate results for standard prognostic variables in breast cancer for reference and Fig. 3g). The 5-year IDFS was 88% for BRCA1 hypermethylated patients versus 71% for non-methylated TNBC patients, while the 5-year distant relapse-free interval (DRFI) was 92% versus 80%, respectively. No difference in IDFS after adjuvant chemotherapy was observed between BRCA1 hypermethylated and BRCA1-null cases (Fig. 3g, log-rank test, p = 0.29).

To assess the independent prognostic value after adjuvant chemotherapy of BRCA1 hypermethylation alone or in combination with BRCA1-null status, we performed multivariable Cox regression analysis adjusting for tumor size (≤20 mm, >20 mm), patient age (<40, 40–59, and ≥60 years), tumor grade (2, 3), and lymph node status (N0, N+) using IDFS as the clinical endpoint. BRCA1 hypermethylation was significantly associated with improved IDFS alone (HR = 0.33, 95% confidence interval (CI): 0.12–0.88) and when combined with BRCA1-null cases (HR = 0.35, 95% CI: 0.14–0.84), although the formal proportional hazard assumption was not fulfilled in these analyses (proportional hazard test, p < 0.05). When patient age was used as continuous variable in the IDFS analyses the results were borderline nonsignificant (Cox regression p = 0.07 for BRCA1 hypermethylation). Borderline nonsignificant results for BRCA1 hypermethylation versus non-methylated cases in multivariable Cox regression using the above models with stratified age bins (HR = 0.28, 95% CI: 0.08–1.02, p = 0.053) or continuous age (Cox regression p = 0.09) were also observed for DRFI. A significant association was seen for DRFI for the combined hypermethylation/BRCA1-null group, irrespective of whether binned age (HR = 0.23, 95% CI: 0.07–0.72) or continuous age (HR = 0.32, 95% CI: 0.11–0.94) was used. For untreated or neoadjuvantly treated SCAN-B patients the number of BRCA1 hypermethylated cases was too low to allow for robust outcome analysis.

Genetic phenotypes of BRCA1 methylated and BRCA1-null TNBC

To test the hypothesis that BRCA1 hypermethylation confers a similar genetic and genomic tumor phenotype as BRCA1-null tumors we compared an array of readouts from WGS between the two groups using the combined SCAN-B and Nik-Zainal et al.²⁴ cohorts (n = 57 hypermethylated and n = 52 BRCA1-null cases in total, Fig. 1).

The genome-wide landscape of copy number gain and loss appeared highly similar between the two groups (Fig. 4a). A full copy number analysis is available in Supplementary Fig. 4, demonstrating no statistical differences between the groups (FDR adjusted Fisher’s exact test, p > 0.05). Concerning specific breast cancer driver genes/alterations reported by Nik-Zainal et al.²⁴ we observed some small frequency differences (typically < 5–10%, when excluding MYC amplifications and TP53 mutations) between the two groups regarding copy number amplifications (Fig. 4b) and mutations (Fig. 4c). RB1 mutation status appeared to show the largest difference between hypermethylated (n = 1/57, 1.8%) and BRCA1-null (n = 4/52, 7.8%) cases for the investigated specific driver alterations, albeit not significant possibly due to low numbers (Fisher’s exact test, p = 0.19). A high frequency of RB1 alterations, particularly intra-gene rearrangements, has previously been linked with BRCA1-null disease²³. However, when considering all detected RB1 mutations (i.e., not filtered for specific drivers), proportions were similar between hypermethylated and BRCA1-null cases in the SCAN-B cohort specifically (35 and 32%, respectively). Only a small statistical difference in the number of indels (BRCA1-null: median = 410, hypermethylated median = 633) was observed between the two groups for cases sequenced with at least 30-fold sequence depth when comparing absolute numbers of substitutions, indels and rearrangements between the two groups (Fig. 4d). This absolute difference in numbers did, however, not correspond to a difference in proportions of indels at repeats or with microhomology, as previously shown (Fig. 2c). Moreover, these results imply a similar tumor mutational burden for hypermethylated and BRCA1-null cases.

**Fig. 4: Genetic phenotypes of *BRCA1* hypermethylated and *BRCA1*-null TNBC.**

For mutational signatures, BRCA1-null cases had higher proportions of substitution signatures suggested to be associated with age at diagnosis (Signatures 1 and 5), an APOBEC signature (Signature 2), and, to some extent, a substitution signature (Signature 8) reported to be elevated in cases with HRD²⁷ (Fig. 4e). When tested for differential proportions in the SCAN-B cases specifically (25 BRCA1-null versus 57 hypermethylated), only the specific APOBEC signature remained elevated in BRCA1-null cases (Wilcoxon’s test, p = 1e−05), driven by a subset (8/25 cases) of BRCA1-null cases with elevated signature proportions. Strikingly, in hypermethylated cases no substitutions were assigned to the particular APOBEC signature (Signature 2), while many cases showed substitutions assigned to another reported APOBEC signature (Signature 13). For comparison, in non-basal tumors from Nik-Zainal et al.²⁴ the median APOBEC signature (Signature 2) proportions were 0.05–0.08 across PAM50 subtypes, with 7% of cases having no signature exposure (Fig. 4e).

No distinct differences were seen for the six rearrangement signatures reported by Nik-Zainal et al.²⁴, including the two BRCA1/BRCA2 associated signatures (RS3 and RS5) (Fig. 4f). Hierarchical clustering and principal component analyses of substitution and rearrangement signatures, or HRDetect²⁶ components, did not separate hypermethylated from BRCA1-null cases (Fig. 4g–i), supporting similar effects on DNA-repair.

DNA methylation in BRCA1 methylated and BRCA1-null TNBC

To investigate differences in global DNA methylation patterns between BRCA1-null and hypermethylated cases, Illumina MethylationEPIC profiles from 25 BRCA1-null and 57 hypermethylated SCAN-B cases were compared. Preprocessing and filtering left 614,977 informative CpGs sites. Supervised differential methylation analysis between the groups identified 32 significant CpGs (false-discovery rate (FDR) adjusted Wilcoxon’s test, p < 0.05). Strikingly, all 32 CpGs were associated with BRCA1, with 28 CpGs within the canonical promoter region (+500 to −1500 base pairs, chr17:43124984-43126983) and four CpGs further upstream BRCA1 (chr17:43169746–43171745). Cluster analysis based on the 32 CpGs recreated, as expected, the division of BRCA1 hypermethylated versus non-methylated tumors perfectly in the entire cohort of 235 DNA methylation profiled cases (Fig. 5a).

**Fig. 5: DNA methylation and transcriptional characteristics in *BRCA1* hypermethylated and *BRCA1*-null TNBC.**

Gene expression in BRCA1 methylated and BRCA1-null TNBC

Transcriptomic differences between BRCA1-null and BRCA1 hypermethylated cases were investigated in 52 BRCA1-null cases versus 57 hypermethylated cases by combining SCAN-B cases with 27 RNAseq analyzed BRCA1-null cases from Jönsson et al.²³ (Fig. 1). Unsupervised clustering using 7224 highly varying genes did not reveal any apparent subclusters specific for the two sample groups (Fig. 5b). This finding was further substantiated by both: (i) unsupervised consensus clustering using the same gene set that showed similar proportions of BRCA1 hypermethylated and BRCA1-null cases across different cluster solutions (Fig. 5c), and (ii) principal component analysis showing that BRCA1 hypermethylation/BRCA1-null status, cohort (SCAN-B or non-SCAN-B) or PAM50 subtypes did not contribute significantly to explaining the variation in gene expression among these tumors (Fig. 5d).

Differential gene expression analysis using significance analysis of microarrays (SAM)³² between the groups was performed to identify differentially expressed genes using the same 7224 genes as for the cluster analyses. Merely eight genes, BRCA1, UQCRHL, SOX6, HAPLN1, POLR2J3, H2AC20, MUCL1, and HYI, were differentially expressed at a FDR of 1% (Supplementary Data 2). Of these, BRCA1 (downregulated in hypermethylated cases, see Fig. 2b), SOX6 (upregulated in hypermethylated cases), and MUCL1 (downregulated in hypermethylated cases) showed differential expression also in SCAN-B cases (Wilcoxon’s test, p < 0.05), while HAPLN1 was borderline nonsignificant (Wilcoxon’s test, p = 0.051) when analyzed separately (not accounting for multiple testing). Taken together, the low number of differentially expressed genes identified and the unsupervised analyses suggest that there is no strong transcriptional difference between the hypermethylated and BRCA1-null patient subgroups.

Immune infiltration in BRCA1 methylated and BRCA1-null TNBC

In silico estimated immune cell composition in BRCA1 hypermethylated versus BRCA1-null cases was compared using three different bulk tissue de-convolution methods based on either gene expression (CIBERSORTx; 6 cell types³³, xCell; 64 cell types³⁴) or DNA methylation (EpiDish; 9 cell types³⁵) for the 82 hypermethylated or BRCA1-null SCAN-B cases. For neither of the methods was a statistical difference in cell type proportions observed between hypermethylated and BRCA1-null tumors for any cell population (t-test, p > 0.05, data provided in Supplementary Data 1). Supporting these observations, the expression of 102 immune cell type associated genes was analyzed by unsupervised clustering and supervised differential gene expression in the 82 SCAN-B cases. While groups of tumors displayed distinct immune infiltration (e.g., consistent with the immune modulatory TNBCtype subtype) these were not defined by BRCA1-status or PD-L1 immunohistochemistry status, nor was any gene statistically different in expression between sample groups (FDR adjusted Wilcoxon’s test p > 0.05) (Supplementary Fig. 5).

Next, PD-L1 immunohistochemistry scores (using the Roche SP-142 antibody) were examined in 53 BRCA1 hypermethylated and 25 BRCA1-null SCAN-B patients, finding no significant difference in PD-L1 classification (≥1% staining in immune cells), but a borderline nonsignificant trend of higher PD-L1 scores in BRCA1-null cases (Wilcoxon’s test, p = 0.051, Fig. 6a). Presence of TILs were evaluated on available whole slide H&E sections for 52 hypermethylated and 24 BRCA1-null SCAN-B cases, revealing no significant difference between the groups (Wilcoxon’s test, p = 0.70) (Fig. 6b).

**Fig. 6: Immune cell infiltration phenotypes in *BRCA1*-null and hypermethylated cases.**

To investigate whether PD-L1 differences between BRCA1-null and hypermethylated cases were related to neoantigen burden, we calculated number of expressed neoantigens from somatic substitutions by integrating WGS-based HLA-typing, neoantigen prediction, and mRNA expression for 232 of 237 SCAN-B cases^36,37. The Pearson correlation between the total substitution burden and predicted number of expressed neoantigens from substitutions was 0.91. PD-L1 positive cases (n = 120) had a higher number of expressed neoantigens than PD-L1 negative cases (n = 93) (Wilcoxon’s test, p = 0.02). In BRCA1-null and hypermethylated cases there was however no statistical difference (Wilcoxon’s test, p > 0.05) between the groups in total, or when stratified by PD-L1 status, irrespective of whether all predicted neoantigens were assessed (Fig. 6c, d) or weak or strong binders were assessed separately³⁶.

Discussion

DNA hypermethylation of promoter CpG islands is associated with loss of gene expression and constitutes a long-known mechanism of functional inactivation of tumor suppressor genes in cancer cells. In the current study, we have comprehensively analyzed the occurrence of BRCA1 promoter hypermethylation in a population-based early-stage TNBC cohort, its readout on the tumor genome and immune microenvironment, and its implications on patient outcomes after adjuvant standard chemotherapy in the context of BRCA1-null tumors.

An important feature of our study is its population-based nature with integrated tissue sampling in conjunction with routine diagnostics³⁸, exemplified by the similar methylation rates observed across individual years of patient enrollment. This lends support to the reproducibility and generalizability of our results in the context of the studied patient demographics for both molecular and patient outcome findings, despite limited sample numbers for some comparisons. In the total SCAN-B cohort, 24.1% of TNBC patients were BRCA1 hypermethylated, a proportion in the lower to mid-range of previous reports^{14,15,16,17,18,19,20}. However, in comparison to the observed BRCA1-null rate of 10.5% among all SCAN-B patients, BRCA1 hypermethylation is more than twice as frequent. Our analyses demonstrate a high frequency of BRCA1 hypermethylation in young patients (46.2% of TNBC patients less than 50 years). These women are often referred to clinical germline BRCA-screening, which is increasingly being considered for guiding the surgical procedure and post-operative treatment³⁹. Indeed, almost half of the women who had undergone clinical BRCA1/2 sequencing without findings of pathogenic germline variants had a tumor with BRCA1 hypermethylation. These women had no clear family history of breast/ovarian cancer but had an age of diagnosis similar to patients with germline BRCA1 alterations. Moreover, they had a low but elevated level of BRCA1 promoter methylation also in peripheral blood DNA, an observation confirmed in additional clinically unscreened SCAN-B patients of varying age. Although these results must be interpreted with caution as the hypermethylation levels are at the limit of detection by pyrosequencing (and thus not suitable for prediction of somatic hypermethylation), they are intriguing in the context of both possible circulating tumor DNA and findings of mosaic constitutional BRCA1 hypermethylation. Regarding the latter, mosaic constitutional BRCA1 hypermethylation has been reported in 4–7% of newborn females^40,41 as well as promoter methylation of BRCA1 or other cancer-related genes in peripheral blood in women who developed TNBC or high grade serous ovarian cancer^42,43,44,45 (see also Tang et al.⁴⁶ for review). Unfortunately, a lack of corresponding fresh normal tissue hindered us from analyzing whether a mosaic methylation pattern was present also in non-malignant breast tissue. If elevated BRCA1 hypermethylation levels in blood cells constitute a risk factor for TNBC (and ovarian cancer) development and are present prior to tumorigenesis, these observations raise important questions regarding potential screening, prevention and early detection, requiring development of more sensitive assays for patient classification. Irrespectively, our results implicate BRCA1 hypermethylation as the likely causative factor underlying a significant proportion of women with TNBC, particularly in those without a family history of disease.

In agreement with previous studies^19,23,24,47, BRCA1 hypermethylation was strongly associated with classical features of BRCA1-null and basal-like disease. Despite this, BRCA1 promoter hypermethylation was also present in subgroups such as old patients and in molecular subtypes not commonly associated with BRCA1 deficiency in TNBC, albeit at lower frequency. The latter indicates that current gene expression phenotypes are not a perfect surrogate for identification of BRCA1 deficiency. No case of combined BRCA1 germline/somatic mutation and hypermethylation was observed, suggesting that these are separate routes for gene inactivation and that loss of heterozygosity or deletion of the wild type BRCA1 copy is the dominant second hit in both hypermethylated and mutated cases. Moreover, the close relationship between BRCA1 promoter allele methylation levels and tumor cell content support a view that the hypermethylation is already present in the main clone, as opposed to subclonal, and therefore early in evolution of the tumor (Fig. 2d). While BRCA1 hypermethylation and germline/somatic mutations represent separate mechanisms for gene inactivation, an objective of this study was to explicitly and in detail test the hypothesis that these alterations result in similar genomic phenotypes. Based on comprehensive global genetic, gene expression, and DNA methylation analyses we found support for this hypothesis. BRCA1 hypermethylated cases share, with few exceptions, similar frequencies of copy number alterations, proportions of reported driver genes in breast cancer, tumor mutational burden, mutational signatures, rearrangement signatures, gene expression patterns, immune cell infiltration, and DNA methylation patterns with BRCA1-null cases.

There are conflicting reports in the literature about whether BRCA1 hypermethylation is associated with a better or worse prognosis in early-stage disease (e.g., refs. ^{14,15,16,18,19,21}). In metastatic TNBC, BRCA1 hypermethylation has been reported not to be associated with a better treatment response to carboplatin compared with docetaxel⁴⁸ (although hypermethylation in this study was measured in the primary tumor tissue, not the actual metastatic tissue) or to single-agent carboplatin effect⁴⁹. This has led to a hypothesis of a soft/plastic BRCAness phenotype for hypermethylated patients compared with germline BRCA1-altered cases⁴⁸, which remains to be proven by analysis of matched tissue from primary and relapsed tumors. Our survival analyses show that BRCA1 promoter methylation appears as a marker for favorable response to current conventional adjuvant chemotherapy in early-stage TNBC, with a prognosis similar to that of BRCA1-null cases. When considering the observed survival curves for hypermethylated cases and the typical early relapse pattern of TNBC, the borderline nonsignificant statistical findings for BRCA1 hypermethylation are likely due to a relatively short follow-up time in the SCAN-B cohort (59% of cases have ≤5 years follow-up) and the limited sample size of the study. Similarly, we saw significant or borderline nonsignificant patterns also for other clinical endpoints like overall survival (OS) and DRFI. Moreover, the observation of BRCA1 hypermethylation also in old TNBC patients, often excluded from adjuvant therapy and clinical trials, may warrant reconsideration in the clinical management of such patients. The explanation for the improved prognosis of BRCA1 hypermethylated versus non-BRCA1 altered cases may lie in the exclusive association with an HRD phenotype as demonstrated in our prior study²². A growing body of evidence suggests that TNBCs with an HRD phenotype (including non-BRCA1/2 mutated cases with HRD inferred by other alterations) respond better to DNA damaging agents^{3,13,22,50,51}. In this context, our observed high incidence of BRCA1 hypermethylation implies that BRCA1 hypermethylation likely constitutes the most common underlying cause of HRD in unselected early-stage TNBC. Recent randomized clinical trial data in ovarian cancer has showed that HRD-positive tumors without BRCA1/2 alterations respond favorably to PARP-inhibitors as first-line maintenance therapy⁸. Similarly, advanced breast cancer patients with HRDetect-high tumors have been shown to be associated with clinical improvement on platinum-based chemotherapy⁵⁰, suggesting that also early stage patients with HRD-positive hypermethylated tumors may benefit from tailored therapies. If so, the early stage patient cohort that could be considered for tailored treatment would increase considerably, although the full treatment effect remains to be determined in clinical trials.

TNBC tumors have been associated with high TIL-levels (signaling immunogenicity) which are also prognostically favorable^52,53. A recent study has suggested a higher frequency of TIL-positive tumors in BRCA1/2-mutated patients than in wild-type cases⁵⁴. In the SCAN-B cohort, TIL-levels did not differ between hypermethylated and BRCA1-null cases, which is consistent with the similar prognosis in these groups after adjuvant chemotherapy. PD-L1 expression is also frequent in TNBC, including BRCA1 hypermethylated tumors¹⁹, but studies have not found that levels of PD-L1-positive TILs in TNBC cancers are driven by a high mutation rate or by BRCA1 mutation status^52,54. This lack of association may now be appreciated through our delineation of the BRCA1/2-wt subgroup by BRCA1 hypermethylation, forming a major subgroup of TNBC with similar genetic and immune cell phenotypes as BRCA1-null cases. Taken together, on a group level these results suggest that early-stage hypermethylated cases may have similar response to immune checkpoint inhibitors as the BRCA1-null group. Both groups are, however, clearly heterogeneous and include both inflamed and cold tumors with and without PD-L1 expression. Consequently, the trend of higher PD-L1 levels in BRCA1-null tumors compared with hypermethylated cases warrants further investigation, as does the intersection between HRD-positivity and a variable immunogenicity.

In summary, our analyses show that the genomic characteristics of BRCA1 hypermethylated and BRCA1-null TNBCs are highly similar, thus representing genomic phenocopies. This suggests that BRCA1 hypermethylation is an early event in tumor development and that tumor progression proceeds along pathways common with BRCA1 germline mutated cases. TNBC patients with BRCA1 hypermethylated tumors share a similar beneficial outcome after standard of care adjuvant chemotherapy as BRCA1-null patients, suggesting that BRCA1 hypermethylation may represent a DNA based prognostic biomarker that is detectable also in low-cellularity tumor tissue specimens. Whether BRCA1 hypermethylated early-stage TNBC patients respond similar as BRCA1-null patients to platinum-based chemotherapy regimens, PARP inhibitors, or immune checkpoint inhibitors merits investigation. A key component in utilizing findings from this study is to include methylation testing in routine clinical diagnostics, requiring robust assays applicable to archival tissue to be developed, and to develop sensitive assays for further investigation of hypermethylation patterns in peripheral blood and their potential association with disease risk.

Methods

Ethics approval and consent to participate

Patients were enrolled in the Sweden Cancerome Analysis Network—Breast (SCAN-B) study (ClinicalTrials.gov ID NCT02306096)^38,55,56, approved by the Regional Ethical Review Board in Lund, Sweden (Registration numbers 2009/658, 2010/383, 2012/58, 2016/742, 2018/267, and 2019/01252). All patients provided written informed consent prior to enrollment. All analyses were performed in accordance with patient consent and ethical regulations and decisions.

SCAN-B patient cohort

From our²² recently reported Swedish population representative TNBC cohort (SCAN-B) of 237 patients we identified 57 BRCA1 hypermethylated cases based on pyrosequencing and 25 BRCA1-null cases by WGS. All identified cases had complete WGS, RNA sequencing, global DNA methylation profiles (performed for this study), and extensive clinical follow-up data. A CONSORT diagram is provided in Supplementary Fig. 1. Clinicopathological characteristics for the 82 SCAN-B patients with genetic or epigenetic BRCA1 alterations are summarized in Table 1 and detailed in Supplementary Data 1. Of the 237 SCAN-B patients, 46 patients underwent clinical screening for germline BRCA1 or BRCA2 alterations based on national guidelines at the time.

BRCA1 promoter hypermethylation analysis

BRCA1 pyrosequencing promoter hypermethylation status was available from a prior study²² based on a protocol reported by Saal et al.⁵⁷. Briefly, the pyrosequencing assay involved analysis of two CpG-dense amplicons containing five and four CpGs, respectively, with internal controls for bisulfite treatment (nucleotide positions are relative to the +1 adenine of the ATG translational start codon on the annotated Ensembl v36 BRCA1 genomic sequence): BRCA1 region 1, −1291 to −1267 (containing 5 CpGs); and region 2, −1332 to −1310 (4 CpGs). The following primers were used: BRCA1_Region1/2_PCR_F, 5′-NNTATTTTGAGAGGTTGTTGTT TAG-3′; BRCA1_Region1/2_PCR_R, 5′-biotin-TAA AAAACCCCACAACCTATCC-3′; BRCA1_Region1 _Seq, 5′-GAATTATAGATAAATTAAAA-3′; and BRCA1_Region2_Seq, 5′-GGTAGTTTTTTGGTTT T-3′. CpGs covered by the assay include cg16630982, cg15419295, cg16963062, cg20187250, cg04658354, cg04110421, and cg21253966. Methylated and unmethylated controls were included in each bisulfite conversion and pyrosequencing run. For the methylated control samples, across ten pyrosequencing tumor runs, the mean and standard deviation for the percentage of BRCA1 allele methylation was 95.48 ± 1.22 for primer set 1 and 96.16 ± 1.32 for primer set 2. For the unmethylated control, the corresponding values were 2.68 ± 0.75 for primer set 1 and 1.55 ± 0.90 for primer set 2. The combined methylation estimates for unmethylated samples (2.26 ± SD 0.45; max 3.59%) are therefore in line with the estimates for the negative conversion controls. Taken together, this demonstrates that the bisulfite conversion efficiency is sufficiently high to ensure to avoid false-positive methylation results. A cut-off of 7% in allele methylation was used to call samples as hypermethylated or not. The chosen cut-off is between the highest methylation level of 3.6% for patients classified as unmethylated (mean 2.3% and standard deviation 0.48%), and the lowest observed methylation level of 13.6% for hypermethylated patients. Importantly, the 7% cut-off is relevant for pyrosequencing data performed according to the specified protocol. It may therefore not be suitable for calling BRCA1 promoter hypermethylation in Illumina Infinium Methylation beadchips.

In addition, pyrosequencing based BRCA1 hypermethylation analysis of blood DNA for constitutional methylation was first performed for 34 of the 46 patients that underwent clinical screening without germline BRCA1/2 findings. DNA from matched blood samples were extracted using the Qiagen Midiprep kit according to manufacturer’s instructions at the Labmedicin Skåne Biobank, Lund, Sweden. Experiments were performed as a single batch experiment including bisulfite conversion and pyrosequencing (Supplementary Methods). In a second analysis, the original 34 patients were rerun together with 71 additional patients from the SCAN-B cohort (n = 105 in total) covering in total 55 of the 57 included cases with BRCA1 tumor hypermethylation. This second run was performed using the exact same instrument settings, reagent lots, and PCR product amounts for all cases, but using a separate bisulfite conversion for the 71 new cases.

Global epigenetic profiling

Global epigenetic profiling of BRCA1 hypermethylated and BRCA1-null cases were performed using Illumina Infinium MethylationEPIC beadchips (interrogating ~800,000 CpGs) according to manufacturer’s instructions at the Center for Translational Genomics, Lund University and Clinical Genomics Lund, SciLifeLab. Input DNA was the same as for WGS, extracted from tumor tissue preserved in RNAlater (Qiagen, Hilden, Germany)^38,55. Beadchip data analysis and immune cell deconvolution based on DNA methylation data using EpiDISH with Robust Partial Correlations (RPC)³⁵ were performed as detailed I Supplementary Methods. M-values were defined as log2(Beta/(1-Beta)) in line with⁵⁸.

Gene-expression analyses

Gene-expression profiling was performed using RNA sequencing (RNAseq) as outlined⁵⁵, and data has been reported elsewhere as Fragments Per Kilobase per Million reads (FPKM)^59,60. Supervised and unsupervised analyses were performed as outlined in Supplementary Methods. Molecular subtype classification according to PAM50⁶¹, IntClust 10⁶², CIT²⁹, and reported TNBC subtypes (TNBCtype)^31,63 were obtained from²². To increase the number of BRCA1-null tumors in gene expression analyses for greater statistical power we generated new RNA sequencing data (according to the same experimental and data analysis protocol as for SCAN-B cases) from 27 ER- and PR-negative BRCA1 germline cases reported originally by Jönsson et al.²³. These were added to the SCAN-B cases (n = 52 BRCA1-null cases in total). Immune cell deconvolution was performed for RNAseq data using xCell³⁴ and CIBERSORTx³³ as outlined in Supplementary Methods.

Whole-genome sequencing analyses

Whole-genome sequencing data for BRCA1 hypermethylated and BRCA1-null SCAN-B cases, including mutational calls, mutational and rearrangement signatures, copy number profiles, and HRD classifications (HRDetect²⁶ and genomic scars, HRD-score¹³) were obtained from ref. ²² (Supplementary Methods). Eighty-four percent of BRCA1-null and 82.5% of BRCA1 hypermethylated cases were sequenced with a 30X target depth, while the remaining cases were sequenced with a 15X target depth.

Expanded BRCA1-null whole genome sequencing cohort

To expand the BRCA1-null cohort for analyses based on WGS data we used the reported cohort of 560 WGS analyzed breast cancers by Nik-Zainal et al.²⁴. In this cohort, 27 TNBC tumors were identified as germline BRCA1-null and added to the SCAN-B cases (n = 52 BRCA1-null cases in total).

PD-L1 immunohistochemistry and TILs

PD-L1 immunohistochemistry using the SP-142 antibody (Roche) was performed on a tissue microarray including 53 BRCA1 hypermethylated and 25 BRCA1-null SCAN-B tumors (two 1 mm cores/tumor) on a Ventana instrument (Roche) according to manufacturer’s recommendations. The SP-142 antibody is the intended PD-L1 test for atezolizumab checkpoint inhibitor used in the Impassion 130 trial⁶⁴. PD-L1 assessment was performed according to antibody instructions on immune cells by a board certified breast cancer pathologist, using a ≥1% cut-off for positivity. Negative cases in both TMA cores were set to score 0.

TILs were scored on available whole section formalin-fixed paraffine embedded haemotoxylin and eosin (H&E) slides by a board certified breast cancer pathologist and summarized as a percentage per sample. Scoring was performed according to the international consensus scoring recommendations of the International Immuno-Oncology Biomarker Working Group on Breast Cancer (www.tilsinbreastcancer.org). Scores were averaged when multiple slides were available per patient. All pathology scorings were performed blinded to clinicopathological and BRCA1 classifications.

Neoantigen prediction

NeoPredPipe³⁶ was used to predict putative neoantigens with substitution mutation calls provided by CaVEMan (https://cancerit.github.io/CaVEMan/) and HLA typing done with Polysolver³⁷ as input. hg19 was used as the human reference genome throughout all analysis for the neoantigen predictions. NeoPredPipe was run with default parameters except that options -c 1 2 -m where set. Polysolver was run on WGS data from blood DNA with options: (i) unknown ethnicity, (ii) use population-level allele frequencies as priors, and (iii) do not use empirical insert size distribution. Only variants with a PASS flag in the variant call file from CaVEMan was used as input to NeoPredPipe. Integration with RNAseq expression was done as outlined for NeoPredPipe, and only neoantigens with an expression >0.1 was kept for final analysis.

Statistical analyses

Survival analyses were performed in R (ver. 3.6.1) using the survival package with OS, IDFS, or DRFI as endpoints defined according to the STEEP guidelines⁶⁵ (see Supplementary Methods for endpoint definitions and analysis exclusion criteria). Survival curves were compared using Kaplan–Meier estimates and the log-rank test. Hazard ratios were calculated through univariable or multivariable Cox regression. Proportional hazard assumptions were tested using the cox.zph function in the R survival package (ver. 3.1-12). Survival analyses were performed using the 149 eligible cases (62.8%) from the total SCAN-B cohort treated with standard-of-care adjuvant chemotherapy according to national guidelines (in 96% of cases a FEC-based [combination of five fluorouracil, epirubicin, and cyclophosphamide] treatment ± a taxane) to contrast BRCA1-deficient groups versus non-altered patients. Full details on the exclusion criteria for outcome analysis and individual patient treatment are available in Staaf et al.²². All p values reported from statistical tests are two-sided if not otherwise specified. Box-plot elements corresponds to: (i) center line = median, (ii) box limits = upper and lower quartiles, (iii) whiskers = 1.5× interquartile range.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

For BRCA1 hypermethylated and BRCA1-null SCAN-B cases mapping of clinical and molecular classifications to gene expression data GSE96058 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE96058] is available in Supplementary Data 1. Illumina DNA methylation data for BRCA1 hypermethylated and BRCA1-null SCAN-B cases is available as GSE148748 [https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE148748] at Gene Expression Omnibus. FPKM RNAseq data for the additional 27 BRCA1-mutated cases from Jönsson et al.²³ is available in an online repository associated with this study [https://doi.org/10.17632/2dbh285999.1]. Source data are provided with this paper. The source data underlying Figs. 2a–d, 3a–d, 4a, 4d–g, 5a–d, and 6a, b are provided as a Source Data file. Source data are provided with this paper.

References

Gluz, O. et al. Triple-negative breast cancer–current status and future directions. Ann. Oncol. 20, 1913–1927 (2009).
Article CAS PubMed Google Scholar
Foulkes, W. D., Smith, I. E. & Reis-Filho, J. S. Triple-negative breast cancer. N. Engl. J. Med. 363, 1938–1948 (2010).
Article CAS PubMed Google Scholar
Sharma, P. Biology and Management of Patients With Triple-Negative Breast Cancer. Oncologist 21, 1050–1062 (2016).
Article PubMed PubMed Central Google Scholar
Winter, C. et al. Targeted sequencing of BRCA1 and BRCA2 across a large unselected breast cancer cohort suggests that one-third of mutations are somatic. Ann. Oncol. 27, 1532–1538 (2016).
Article CAS PubMed PubMed Central Google Scholar
Lord, C. J. & Ashworth, A. The DNA damage response and cancer therapy. Nature 481, 287–294 (2012).
Article ADS CAS PubMed Google Scholar
Farmer, H. et al. Targeting the DNA repair defect in BRCA mutant cells as a therapeutic strategy. Nature 434, 917–921 (2005).
Article ADS CAS PubMed Google Scholar
Brok, W. D. d. et al. Homologous recombination deficiency in breast cancer: a clinical review. JCO Precis. Oncol. 1, 1–13 (2017).
Google Scholar
Ray-Coquard, I. et al. Olaparib plus bevacizumab as first-line maintenance in ovarian cancer. N. Engl. J. Med. 381, 2416–2428 (2019).
Article CAS PubMed Google Scholar
van Verschuer, V. M. et al. Tumor-associated inflammation as a potential prognostic tool in BRCA1/2-associated breast cancer. Hum. Pathol. 46, 182–190 (2015).
Article PubMed CAS Google Scholar
Jiang, T. et al. Predictors of chemosensitivity in triple negative breast cancer: an integrated genomic analysis. PLoS Med. 13, e1002193 (2016).
Article PubMed PubMed Central CAS Google Scholar
Nolan, E. et al. Combined immune checkpoint blockade as a therapeutic strategy for BRCA1-mutated breast cancer. Sci. Transl. Med. 9, eaal4922 (2017).
Article PubMed PubMed Central CAS Google Scholar
Akashi-Tanaka, S. et al. BRCAness predicts resistance to taxane-containing regimens in triple negative breast cancer during neoadjuvant chemotherapy. Clin. Breast Cancer 15, 80–85 (2015).
Article CAS PubMed Google Scholar
Telli, M. L. et al. Homologous recombination deficiency (HRD) score predicts response to platinum-containing neoadjuvant chemotherapy in patients with triple-negative breast cancer. Clin. Cancer Res. 22, 3764–3773 (2016).
Article CAS PubMed PubMed Central Google Scholar
Zhu, X. et al. Hypermethylation of BRCA1 gene: implication for prognostic biomarker and therapeutic target in sporadic primary triple-negative breast cancer. Breast Cancer Res. Treat. 150, 479–486 (2015).
Article CAS PubMed Google Scholar
Yamashita, N. et al. Epigenetic inactivation of BRCA1 through promoter hypermethylation and its clinical importance in triple-negative breast cancer. Clin. Breast Cancer 15, 498–504 (2015).
Article CAS PubMed Google Scholar
Sharma, P. et al. The prognostic value of BRCA1 promoter methylation in early stage triple negative breast cancer. J. Cancer Ther. Res. 3, 1–11 (2014).
Article ADS PubMed PubMed Central CAS Google Scholar
Xu, Y. et al. Promoter methylation of BRCA1 in triple-negative breast cancer predicts sensitivity to adjuvant chemotherapy. Ann. Oncol. 24, 1498–1505 (2013).
Article CAS PubMed Google Scholar
Sharma, P. et al. Impact of homologous recombination deficiency biomarkers on outcomes in patients with triple-negative breast cancer treated with doxorubicin-based adjuvant chemotherapy (SWOG S9313). Ann. Oncol. 29, 654–660 (2017).
Article PubMed Central Google Scholar
Jacot, W. et al. BRCA1 promoter hypermethylation is associated with good prognosis and chemosensitivity in triple-negative breast cancer. Cancers 12, 828 (2020).
Article PubMed Central CAS Google Scholar
Brianese, R. C. et al. BRCA1 deficiency is a recurrent event in early-onset triple-negative breast cancer: a comprehensive analysis of germline mutations and somatic promoter methylation. Breast Cancer Res. Treat. 167, 803–814 (2018).
Article CAS PubMed Google Scholar
Xie, Y., Gou, Q., Wang, Q., Zhong, X. & Zheng, H. The role of BRCA status on prognosis in patients with triple-negative breast cancer. Oncotarget 8, 87151–87162 (2017).
Article PubMed PubMed Central Google Scholar
Staaf, J. et al. Whole-genome sequencing of triple-negative breast cancers in a population-based clinical study. Nat. Med. 25, 1526–1533 (2019).
Article CAS PubMed PubMed Central Google Scholar
Jonsson, G. et al. The retinoblastoma gene undergoes rearrangements in BRCA1-deficient basal-like breast cancer. Cancer Res. 72, 4028–4036 (2012).
Article PubMed CAS Google Scholar
Nik-Zainal, S. et al. Landscape of somatic mutations in 560 breast cancer whole-genome sequences. Nature 534, 47–54 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Yang, D. et al. Association of BRCA1 and BRCA2 mutations with survival, chemotherapy sensitivity, and gene mutator phenotype in patients with ovarian cancer. J. Am. Med. Assoc. 306, 1557–1565 (2011).
Article CAS Google Scholar
Davies, H. et al. HRDetect is a predictor of BRCA1 and BRCA2 deficiency based on mutational signatures. Nat. Med. 23, 517–525 (2017).
Article CAS PubMed PubMed Central Google Scholar
Nik-Zainal, S. & Morganella, S. Mutational signatures in breast cancer: the problem at the DNA level. Clin. Cancer Res. 23, 2617–2629 (2017).
Article PubMed PubMed Central Google Scholar
Parker, J. S. et al. Supervised risk predictor of breast cancer based on intrinsic subtypes. J. Clin. Oncol. 27, 1160–1167 (2009).
Article PubMed PubMed Central Google Scholar
Guedj, M. et al. A refined molecular taxonomy of breast cancer. Oncogene 31, 1196–1206 (2012).
Article CAS PubMed Google Scholar
Curtis, C. et al. The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups. Nature 486, 346–352 (2012).
Article CAS PubMed PubMed Central Google Scholar
Lehmann, B. D. et al. Identification of human triple-negative breast cancer subtypes and preclinical models for selection of targeted therapies. J. Clin. Investig. 121, 2750–2767 (2011).
Article CAS PubMed PubMed Central Google Scholar
Tusher, V. G., Tibshirani, R. & Chu, G. Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl Acad. Sci. USA 98, 5116–5121 (2001).
Article ADS CAS PubMed MATH PubMed Central Google Scholar
Newman, A. M. et al. Determining cell type abundance and expression from bulk tissues with digital cytometry. Nat. Biotechnol. 37, 773–782 (2019).
Article CAS PubMed PubMed Central Google Scholar
Aran, D., Hu, Z. & Butte, A. J. xCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol. 18, 220 (2017).
Article PubMed PubMed Central CAS Google Scholar
Teschendorff, A. E., Breeze, C. E., Zheng, S. C. & Beck, S. A comparison of reference-based algorithms for correcting cell-type heterogeneity in epigenome-wide association studies. BMC Bioinform. 18, 105 (2017).
Article CAS Google Scholar
Schenck, R. O., Lakatos, E., Gatenbee, C., Graham, T. A. & Anderson, A. R. A. NeoPredPipe: high-throughput neoantigen prediction and recognition potential pipeline. BMC Bioinform. 20, 264 (2019).
Article Google Scholar
Shukla, S. A. et al. Comprehensive analysis of cancer-associated somatic mutations in class I HLA genes. Nat. Biotechnol. 33, 1152–1158 (2015).
Article CAS PubMed PubMed Central Google Scholar
Ryden, L. et al. Minimizing inequality in access to precision medicine in breast cancer by real-time population-based molecular analysis in the SCAN-B initiative. Br. J. Surg. 105, e158–e168 (2018).
Article CAS PubMed PubMed Central Google Scholar
Runowicz, C. D. et al. American Cancer Society/American Society of Clinical Oncology Breast Cancer Survivorship Care Guideline. J. Clin. Oncol. 34, 611–635 (2016).
Article CAS PubMed Google Scholar
Lönning, P., Eikesdal, H., Löes, I. M. & Knappskog, S. Consitutional Mosaic Epimutations—a hidden cause of cancer? Cell Stress 3, 118–135 (2019).
Article PubMed PubMed Central CAS Google Scholar
Al-Moghrabi, N. et al. Methylation of BRCA1 and MGMT genes in white blood cells are transmitted from mothers to daughters. Clin. Epigenet. 10, 99 (2018).
Article CAS Google Scholar
Chen, J. et al. High-resolution bisulfite-sequencing of peripheral blood DNA methylation in early-onset and familial risk breast cancer patients. Clin. Cancer Res. 25, 5301–5314 (2019).
Article CAS PubMed PubMed Central Google Scholar
Lonning, P. E. et al. White blood cell BRCA1 promoter methylation status and ovarian cancer risk. Ann. Intern. Med. 168, 326–334 (2018).
Article PubMed Google Scholar
Prajzendanc, K. et al. BRCA1 promoter methylation in peripheral blood is associated with the risk of triple-negative breast cancer. Int. J. Cancer 146, 1293–1298 (2020).
Article CAS PubMed Google Scholar
Azzollini, J. et al. Constitutive BRCA1 promoter hypermethylation can be a predisposing event in isolated early-onset breast cancer. Cancers 11, 58 (2019).
Article CAS PubMed Central Google Scholar
Tang, Q., Cheng, J., Cao, X., Surowy, H. & Burwinkel, B. Blood-based DNA methylation as biomarker for breast cancer: a systematic review. Clin. Epigenet. 8, 115 (2016).
Article CAS Google Scholar
Polak, P. et al. A mutational signature reveals alterations underlying deficient homologous recombination repair in breast cancer. Nat. Genet. 49, 1476–1486 (2017).
Article CAS PubMed PubMed Central Google Scholar
Tutt, A. et al. Carboplatin in BRCA1/2-mutated and triple-negative breast cancer BRCAness subgroups: the TNT Trial. Nat. Med. 24, 628–637 (2018).
Article CAS PubMed PubMed Central Google Scholar
Isakoff, S. J. et al. TBCRC009: a multicenter phase II clinical trial of platinum monotherapy with biomarker assessment in metastatic triple-negative breast cancer. J. Clin. Oncol. 33, 1902–1909 (2015).
Article CAS PubMed PubMed Central Google Scholar
Zhao, E. Y. et al. Homologous recombination deficiency and platinum-based therapy outcomes in advanced breast cancer. Clin. Cancer Res 23, 7521–7530 (2017).
Article CAS PubMed Google Scholar
Telli, M. L. et al. Homologous recombination deficiency (HRD) status predicts response to standard neoadjuvant chemotherapy in patients with triple-negative or BRCA1/2 mutation-associated breast cancer. Breast Cancer Res. Treat. 168, 625–630 (2018).
Article CAS PubMed Google Scholar
Sobral-Leite, M. et al. Assessment of PD-L1 expression across breast cancer molecular subtypes, in relation to mutation rate, BRCA1-like status, tumor-infiltrating immune cells and survival. Oncoimmunology 7, e1509820 (2018).
Article PubMed PubMed Central Google Scholar
Loi, S. et al. Tumor-infiltrating lymphocytes and prognosis: a pooled individual patient analysis of early-stage triple-negative breast cancers. J. Clin. Oncol. 37, 559–569 (2019).
Article PubMed PubMed Central Google Scholar
Solinas, C. et al. BRCA gene mutations do not shape the extent and organization of tumor infiltrating lymphocytes in triple negative breast cancer. Cancer Lett. 450, 88–97 (2019).
Article CAS PubMed Google Scholar
Saal, L. H. et al. The Sweden Cancerome Analysis Network—Breast (SCAN-B) Initiative: a large-scale multicenter infrastructure towards implementation of breast cancer genomic analyses in the clinical routine. Genome Med. 7, 20 (2015).
Article PubMed PubMed Central CAS Google Scholar
SCAN-B. www.med.lu.se/klinvetlund/canceromics/konsortium/scan_b (2020).
Saal, L. H. et al. Recurrent gross mutations of the PTEN tumor suppressor gene in breast cancers with deficient DSB repair. Nat. Genet. 40, 102–107 (2008).
Article CAS PubMed Google Scholar
Du, P. et al. Comparison of beta-value and M-value methods for quantifying methylation levels by microarray analysis. BMC Bioinform. 11, 587 (2010).
Article CAS Google Scholar
Gene Expression Omnibus. http://www.ncbi.nlm.nih.gov/geo/ (2020).
Brueffer, C. et al. Clinical Value of RNA Sequencing–Based Classifiers for Prediction of the Five Conventional Breast Cancer Biomarkers: A Report From the Population-Based Multicenter Sweden Cancerome Analysis Network—Breast Initiative. JCO Precis. Oncol. 2, 1–18 (2018).
Google Scholar
Paquet, E. R. & Hallett, M. T. Absolute assignment of breast cancer intrinsic molecular subtype. J. Natl Cancer Inst. 107, 357 (2015).
Article PubMed CAS Google Scholar
Ali, H. R. et al. Genome-driven integrated classification of breast cancer validated in over 7,500 samples. Genome Biol. 15, 431 (2014).
Article PubMed PubMed Central Google Scholar
Chen, X. et al. TNBCtype: a subtyping tool for triple-negative breast cancer. Cancer Inf. 11, 147–156 (2012).
Google Scholar
Schmid, P. et al. Atezolizumab and nab-paclitaxel in advanced triple-negative breast cancer. N. Engl. J. Med. 379, 2108–2121 (2018).
Article CAS PubMed Google Scholar
Hudis, C. A. et al. Proposal for standardized definitions for efficacy end points in adjuvant breast cancer trials: the STEEP system. J. Clin. Oncol. 25, 2127–2132 (2007).
Article PubMed Google Scholar

Download references

Acknowledgements

The authors would like to acknowledge the Center for Translational Genomics, Lund University and Clinical Genomics Lund, SciLifeLab for support with DNA methylation analyses, patients and clinicians participating in the SCAN-B study, personnel at the central SCAN-B laboratory at the Division of Oncology and Pathology, Lund University, the Swedish National Breast Cancer Quality Registry (NKBC), Regional Cancer Center South, RBC Syd, and the South Sweden Breast Cancer Group (SSBCG). Financial support: The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the paper. Ana Bosch: The Governmental Funding for Young Clinical Researchers within the National Health Service (ALF) 2017–2019. Mrs. Berta Kamprad Foundation FBKS 2017-34-199. Shamik Mitra: Shamik Mitra is financially supported by the funding received from the European Community’s Horizon 2020 Framework Program for Research and Innovation (H2020-MSCA-ITN-2014) under Grant Agreement no. 247634. Mattias Aine: Gunnar Nilsson Foundation grant (GN-2018-5). Niklas Loman. The Swedish research Council 2019–2021. The Governmental Funding within the National Health Service (ALF). Mrs. Berta Kamprad Foundation. Åke Borg: The Swedish Cancer Society CAN 2016/659. Mrs. Berta Kamprad Foundation FBKS 2018-3-166. Mats Paulsson Foundation IACD 2017. Lund-Lausanne L2-Bridge/Biltema Foundation (F 2016/1330). Johan Staaf: The Governmental Funding of Clinical Research within the National Health Service (ALF) 2018/40612. The Swedish Cancer Society (CAN 2018/685) and a 2018 Senior Investigator Award (SIA190013). Mrs. Berta Kamprad Foundation FBKS-2018-4-146. The Crafoord Foundation 20180543. The Gustav V:s Jubilee Foundation (174271 and 187041). The research foundation at Department of Oncology in Lund. Serena Nik-Zainal: Wellcome Trust Intermediate Clinical Fellowship WT100183MA. CRUK Advanced Clinician Scientist Award (C60100/A23916). CRUK Grand Challenge Award (C60100/A25274). Open access funding provided by Lund University. Funding: Financial support for this study was provided by the Swedish Cancer Society (CAN 2016/659 and CAN 2018/685 and a 2018 Senior Investigator Award [JS: SIA190013]), the Mrs Berta Kamprad Foundation (FBKS 2018-3-166 and FBKS-2018-4-146), the Lund-Lausanne L2-Bridge/Biltema Foundation (F 2016/1330), the Mats Paulsson Foundation (IACD 2017), the Crafoord Foundation (grant 20180543), the National Society of Breast Cancer Associations in Sweden, the Swedish Breast Cancer Group (SweBCG), the Swedish Research Council, BioCARE a Strategic Research Program at Lund University, the Gustav V:s Jubilee Foundation (174271), the research foundation at Department of Oncology in Lund, Region Skåne Regional funds, and Swedish governmental funding (ALF, grant 2018/40612). Whole genome sequencing and analysis was funded by a Wellcome Trust Intermediate Clinical Fellowship WT100183MA and a CRUK Advanced Clinician Scientist Award (C60100/A23916) and a CRUK Grand Challenge Award (C60100/A25274).

Author information

These authors contributed equally: Dominik Glodzik, Ana Bosch, Serena Nik-Zainal, Åke Borg, Johan Staaf.

Authors and Affiliations

Division of Oncology, Department of Clinical Sciences Lund, Lund University, Medicon Village, SE-22381, Lund, Sweden
Dominik Glodzik, Ana Bosch, Mattias Aine, Johan Vallon-Christersson, Christel Reuterswärd, Anna Karlsson, Shamik Mitra, Emma Niméus, Karolina Holm, Jari Häkkinen, Cecilia Hegardt, Lao H. Saal, Anna Ehinger, Niklas Loman, Anders Kvist, Åke Borg & Johan Staaf
Wellcome Sanger Institute, Wellcome Genome Campus, CB10 1SA, Cambridge, UK
Dominik Glodzik
Department of Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, New York, NY, USA
Dominik Glodzik
Department of Oncology, Skåne University Hospital, SE-22184, Lund, Sweden
Ana Bosch, Martin Malmberg & Niklas Loman
Department of Oncology and Pathology, Karolinska Institute, SE-17177, Stockholm, Sweden
Johan Hartman
Division of Molecular Hematology, Department of Laboratory Medicine, Lund University, SE-22184, Lund, Sweden
Mattias Aine
Division of Surgery, Department of Clinical Sciences, Lund University, SE-22184, Lund, Sweden
Emma Niméus & Lisa Rydén
Division of Translational Cancer Research, Department of Laboratory Medicine, Lund University, Medicon Village, SE-22381, Lund, Sweden
Christer Larsson
Department of Genetics and Pathology, Laboratory Medicine, Region Skåne, SE-22184, Lund, Sweden
Anna Ehinger & Hans Ehrencrona
Division of Clinical Genetics, Department of Laboratory Medicine, Lund University, SE-22184, Lund, Sweden
Hans Ehrencrona
Academic Department of Medical Genetics, The Clinical School University of Cambridge, Cambridge Biomedical Research Campus, CB2 0QQ, Cambridge, UK
Serena Nik-Zainal

Authors

Dominik Glodzik
View author publications
You can also search for this author in PubMed Google Scholar
Ana Bosch
View author publications
You can also search for this author in PubMed Google Scholar
Johan Hartman
View author publications
You can also search for this author in PubMed Google Scholar
Mattias Aine
View author publications
You can also search for this author in PubMed Google Scholar
Johan Vallon-Christersson
View author publications
You can also search for this author in PubMed Google Scholar
Christel Reuterswärd
View author publications
You can also search for this author in PubMed Google Scholar
Anna Karlsson
View author publications
You can also search for this author in PubMed Google Scholar
Shamik Mitra
View author publications
You can also search for this author in PubMed Google Scholar
Emma Niméus
View author publications
You can also search for this author in PubMed Google Scholar
Karolina Holm
View author publications
You can also search for this author in PubMed Google Scholar
Jari Häkkinen
View author publications
You can also search for this author in PubMed Google Scholar
Cecilia Hegardt
View author publications
You can also search for this author in PubMed Google Scholar
Lao H. Saal
View author publications
You can also search for this author in PubMed Google Scholar
Christer Larsson
View author publications
You can also search for this author in PubMed Google Scholar
Martin Malmberg
View author publications
You can also search for this author in PubMed Google Scholar
Lisa Rydén
View author publications
You can also search for this author in PubMed Google Scholar
Anna Ehinger
View author publications
You can also search for this author in PubMed Google Scholar
Niklas Loman
View author publications
You can also search for this author in PubMed Google Scholar
Anders Kvist
View author publications
You can also search for this author in PubMed Google Scholar
Hans Ehrencrona
View author publications
You can also search for this author in PubMed Google Scholar
Serena Nik-Zainal
View author publications
You can also search for this author in PubMed Google Scholar
Åke Borg
View author publications
You can also search for this author in PubMed Google Scholar
Johan Staaf
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

Conception and design: J.S. Collection and assembly of data: J.S., S.N.-Z., D.G., A.B., A.K., C.R., K.H., J.V.-C., J.H., C.H. Provision of study material or patients: C.L., N.L., L.R., Å.B., M.M., A.E., L.H.S., C.H., H.E., A.K., E.N. Data analysis and interpretation: J.S., D.G., M.A., S.M., J.H. (IHC evaluations), S.N.-Z. Financial support: S.N.-Z., Å.B., J.S. Administrative support: J.S., S.N.-Z., D.G., J.V.-C., C.H., J.H. Paper writing: All authors. Final approval of the paper: All authors. Agree to be accountable for all aspects of the work: All authors.

Corresponding author

Correspondence to Johan Staaf.

Ethics declarations

Competing interests

J.H. has received speakers honoraria and travel support from Roche, advisory board fees from MSD, Novartis and Roche, and institutional research grants from Cepheid and Novartis. Anna Ehinger has received speakers honoraria from Novartis, Amgen, Roche, and advisory board fees from Roche. Ana Bosch has participated in advisory boards for Novartis and Pfizer, and has received travel support from Roche. The remaining authors declare no competing interests.

Additional information

Peer review information Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Glodzik, D., Bosch, A., Hartman, J. et al. Comprehensive molecular comparison of BRCA1 hypermethylated and BRCA1 mutated triple negative breast cancers. Nat Commun 11, 3747 (2020). https://doi.org/10.1038/s41467-020-17537-2

Download citation

Received: 20 February 2020
Accepted: 02 July 2020
Published: 27 July 2020
DOI: https://doi.org/10.1038/s41467-020-17537-2

This article is cited by

Sex-specific epigenetics drive low GPER expression in gastrointestinal smooth muscles in type 2 diabetic mice
- Aliyu Muhammad
- Juanita C. Hixon
- Sunila Mahavadi
Scientific Reports (2024)
Long-term outcomes of young, node-negative, chemotherapy-naïve, triple-negative breast cancer patients according to BRCA1 status
- Yuwei Wang
- Gwen M. H. E. Dackus
- Marjanka K. Schmidt
BMC Medicine (2024)
Molecular characteristics of breast tumors in patients screened for germline predisposition from a population-based observational study
- Deborah F. Nacer
- Johan Vallon-Christersson
- Johan Staaf
Genome Medicine (2023)
Contribution of constitutional BRCA1 promoter methylation to early-onset and familial breast cancer patients from Pakistan
- Noor Muhammad
- Ayesha Azeem
- Muhammad Usman Rashid
Breast Cancer Research and Treatment (2023)
Efficacy of different neoadjuvant treatment regimens in BRCA-mutated triple negative breast cancer: a systematic review and meta-analysis
- Olga Caramelo
- Cristina Silva
- Teresa Almeida-Santos
Hereditary Cancer in Clinical Practice (2022)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.