Evaluating genetic variants associated with breast cancer risk in high and moderate-penetrance genes in Asians

Summary

We investigated up to 654 SNPs in 13 genes carrying high or moderate-penetrance mutations in association with breast cancer risk. Our study identified several new risk variants in the BRCA1, BRCA2, CHEK2 and PALB2 genes in relation to breast cancer risk in Asian women.

Abstract

Over the past 20 years, high-penetrance pathogenic mutations in genes BRCA1, BRCA2, TP53, PTEN, STK11 and CDH1 and moderate-penetrance mutations in genes CHEK2, ATM, BRIP1, PALB2, RAD51C, RAD50 and NBN have been identified for breast cancer. In this study, we investigated whether there are additional variants in these 13 genes associated with breast cancer among women of Asian ancestry. We analyzed up to 654 single nucleotide polymorphisms (SNPs) from 6269 cases and 6624 controls of Asian descent included in the Breast Cancer Association Consortium (BCAC), and up to 236 SNPs from 5794 cases and 5529 controls included in the Shanghai Breast Cancer Genetics Study (SBCGS). We found three missense variants with minor allele frequency (MAF) <0.05: rs80358978 (Gly2508Ser), rs80359065 (Lys2729Asn) and rs11571653 (Met784Val) in the BRCA2 gene, showing statistically significant associations with breast cancer risk, with P-values of 1.2 × 10^–4, 1.0 × 10^–3 and 5.0 × 10^–3, respectively. In addition, we found four low-frequency variants (rs8176085, rs799923, rs8176173 and rs8176258) in the BRCA1 gene, one common variant in the CHEK2 gene (rs9620817), and one common variant in the PALB2 gene (rs13330119) associated with breast cancer risk at P < 0.01. Our study identified several new risk variants in BRCA1, BRCA2, CHEK2, and PALB2 genes in relation to breast cancer risk in Asian women. These results provide further insights that, in addition to the high/moderate penetrance mutations, other low-penetrance variants in these genes may also contribute to breast cancer risk.

Introduction

Breast cancer is the most common invasive cancer in females worldwide and in Asian countries (1). Over the past 20 years, high-penetrance pathogenic mutations for breast cancer were identified in the genes BRCA1, BRCA2, TP53, PTEN, STK11 and CDH1 and moderate-penetrance mutations in the genes CHEK2, ATM, BRIP1, PALB2, RAD51C, RAD50 and NBN (2–5). Most of these germline mutations were discovered in women of European ancestry. Pathogenic mutations in the BRCA1 and BRCA2 genes are associated with a 10- to 20-fold increased risk of breast cancer, which corresponds to a cumulative risk of 55–65% of developing breast cancer by age 70 years for BRCA1 mutation carriers and 45–47% for BRCA2 mutation carriers (6,7). Mutations in the TP53 gene are associated with at least a 10-fold increased risk of breast cancer and account for 2–7% of early-onset breast cancer (8,9). It has been recently revealed that female PTEN mutation carriers have an 85% lifetime risk of developing breast cancer with 50% penetrance by age 50 years (10), and these findings were subsequently confirmed by two other groups (11,12). It is estimated that the cumulative risk of breast cancer by age 70 years is approximately 14% for women who carry CHEK2 1100delC A meta-analysis, based on 29 154 cases and 37 064 controls from 25 case–control studies, reported a significant association between CHEK2 1100delC heterozygotes and breast cancer risk with an odds ratio (OR) of 2.75 and a 95% confidence interval (CI) of 2.25–3.36 (13,14). Similarly, the approximate risk of breast cancer is 15% for those who have ATM mutations (15). It is estimated that the eight confirmed high and moderate-penetrance genes (BRCA1, BRCA2, PTEN, TP53, CHEK2, ATM, BRIP1 and PALP2) explain approximately 20% of the familial risk of breast cancer (3).

Most of these studies were conducted in women of European ancestry (3). It is possible that the risk and the proportion of familial risk explained by these genes, as well as the mutation landscape of these genes, differ by study population. Previous studies of high and moderate-penetrance genes conducted in Asian populations were limited by a small sample size, as only a few of the studies had a sample size of over 1000 cases (16–18). In this study, we combined our data from the Shanghai Breast Cancer Genetic Study (SBCGS) and from the Breast Cancer Association Consortium (BCAC) to investigate whether there are additional variants in the 13 high and moderate-penetrance genes associated with breast cancer in women of Asian ancestry.

Materials and methods

Study populations

Data from the BCAC and the SBCGS were included in this study. The BCAC, a large collaborative study, includes a total of 12 893 Asian women (6269 cases and 6624 controls) from nine studies including samples from the SBCGS. The nine studies were composed of seven hospital-based studies (Asia Cancer Program, Hospital-based Epidemiologic Research Program at Aichi Cancer Center, Malaysian Breast Cancer Genetic Study, Seoul Breast Cancer Study, Singapore Breast Cancer Cohort, IARC – Thai Breast Cancer Study and Taiwanese Breast Cancer Study), and two population-based studies [Los Angeles County Asian-American Breast Cancer Case–Control Study and Shanghai Breast Cancer Genetic Study (SBCGS)]. Details of the study protocol for the BCAC are described elsewhere (19–21). Study participants were recruited following protocols approved by the Institutional Review Board at each institution and all subjects provided written informed consent. All participating BCAC studies are described in detail in Supplementary Table 1, available at Carcinogenesis Online.

The SBCGS includes four population-based studies conducted among Chinese women in urban Shanghai: the Shanghai Breast Cancer Study (22,23), the Shanghai Breast Cancer Survival Study (24), the Shanghai Endometrial Cancer Study (including controls only for the present study) (25) and the Shanghai Women’s Health Study (26). Details of the SBCGS studies have been described previously (22,24). All participants provided written informed consent, and institutional review boards of all institutes in both China and the United States approved the study protocols. In this study, the SBCGS data includes samples from a total of 4866 women (2731 cases and 2135 controls) that were genotyped through genome-wide association studies (GWAS) (referred to in this paper as SBCGS-GWAS) and samples from 11 470 women (5767 cases and 5703 controls) that were genotyped with the Exome array (referred to in this paper as SBCGS-exomechip). Samples that overlapped between studies were removed so that individual genetic data were used only once in the meta-analysis. After eliminating overlapping samples, our final dataset of the Exome array includes 6457 samples (3063 cases and 3394 controls). There were no overlapping samples between BCAC and SBCGS-GWAS samples.

Genotyping and sample quality control

In the BCAC, 12 893 Asian samples were genotyped using the custom Illumina Infinium array (iCOGS) with 211 155 SNPs included. Details of the genotyping and quality control measures for the BCAC are described elsewhere (19,27). In brief, SNPs on this array were primarily selected to replicate the results of the GWAS of five cancer sites, including breast. Genetic variants available through iCOGS genotyping in the BRCA1 (chr17:41,196,311-41,277,500), BRCA2 (chr13:32,889,616-32,973,809), PTEN (chr10:89,623,194-89,728,532), TP53 (chr17:7,571,719-7,590,868), CHEK2 (chr22:29,083,730-29,137,822), ATM (chr11:108,093,558-108,239,826), BRIP1 (chr17:59,756,546-59,940,920), PALB2 (chr16:23,614,482-23,652,678), RAD51C (chr17:56,769,933-56,811,703), STK11 (chr19:1,205,797-1,228,434), CDH1 (chr16:68,771,194-68,869,444), RAD50 (chr5:131,892,615-131,980,313) and NBN (chr8:90,945,563-90,996,899) genes were included. A total of 654 SNPs in these 13 high/moderate penetrance genes were evaluated in the current analysis.

In the SBCGS-GWAS, 4866 samples were genotyped using Affymetrix Genome-Wide Human SNP Array 6.0. Detailed genotyping information has been previously described (22). In the SBCGS-exomechip, 6457 samples were genotyped using the Asian Exomechip, an expanded Illumina HumanExome-12v1_A Beadchip. In order to improve the coverage for the low frequency variants in Asian populations, we added additional customer content variants onto the Illumina HumanExome-12v1_A Beadchip. The original Exome array included 247 870 markers focused on protein-altering variants selected from sequencing data in >12 000 subjects, mostly from European-ancestry populations. In the Asian Exomechip, the additional variants were primarily selected from exome sequencing in 581 Chinese women from SBCS, exome sequencing in 496 Singapore Chinese and Asian data in the 1000 Genomes Project. We added nonsynonymous, splicing and stop-altering variants observed two or more times in any of these datasets or if we observed one in any two of the three datasets. The genotyping protocol and array design using the Asian Exomechip has been described previously (28–30).

Statistical analyses

Differences of study characteristics between cases and controls in each study were compared using a Wilcoxon rank sum test (continuous variables) or χ² test (categorical variables). An association analysis between single variants and breast cancer risk was conducted using ORs and 95% CIs derived from logistic regression models. The association between each variant and breast cancer was established using an additive genetic model estimating the OR of the BC per increase in effect allele. Analyses were adjusted for age, study site and the top two principal components (PCs) for the BCAC (19), and age and the top five PCs for the SBCGS (22,29). Additional adjustment for other potential confounders did not change the results materially. Meta-analyses of single variant results were conducted using the fixed-effect inverse variance method to combine the β estimates and standard errors from each dataset (BCAC, SBCGS-GWAS and SBCGS exomechip). Stratified analysis by estrogen receptor (ER) status was carried out. Cochran’s Q test assessed heterogeneity across studies and ER status. Each SNP was annotated using ANNOVAR (31) and LOFTEE (32). We used the SIFT algorithm (33), PROVEAN (34) and PolyPhen-2 (35) to predict the possible impact of an amino acid substitution on protein function and structure. For SBCGS-GWAS, imputation was performed using SHAPEIT to derive phased genotypes and Minimac2 to perform imputation on the phased data (36,37). The 1000 Genome Project phase 3 was used as the reference data for imputation (http://www.1000genomes.org/), and SNPs with high imputation quality (RSQR ≥ 0.8) were included in the analysis. Dosage data with imputation accuracy taken into consideration from SBCGS-GWAS were used in the present study.

Results

Selected characteristics of the study population are shown in Table 1. Patients included in BCAC were more likely to be ER positive, progesterone receptor positive, and human epidermal growth factor receptor 2 positive than SBCGS patients. Breast cancer patients have a higher BMI, more first-degree relatives with breast cancer, and earlier menarche than controls (Supplementary Table 2, available at Carcinogenesis Online).

Table 1.

Selected characteristics of study participants in the current analysis from the Breast Cancer Association Consortium (BCAC) and Shanghai Breast Cancer Genetics Study (SBCGS)

Characteristics	BCAC_Asian	SBCGS-GWAS	SBCGS-exome chip
	(N = 12 893)	(N = 4866)	(N = 6457)
Total no. of cases/controls	6269/6624	2731/2135	3063/3394
Age (year, mean ± SD)	51.3 ± 10.4	50.7 ± 9.1	53.5 ± 9.5
Age at menarche (year, mean ± SD)	13.9 ± 1.9	14.6 ± 1.7	14.6 ± 1.7
Age at menopause (year, mean ± SD)	49.0 ± 4.8	48.8 ± 4.7	49.0 ± 4.3
Postmenopausal (%)	47.5	41.7	52.0
ER status (%)
Positive	26.1	32.6	29.6
Negative	12.5	15.8	16.0
Othera	61.4	51.6	54.4
Progesterone receptor status (%)
Positive	22.4	30.6	26.5
Negative	13.1	17.7	19.0
Othera	64.6	51.7	54.5
Human epidermal growth factor receptor 2 status (%)
Positive	6.4	7.4	9.4
Negative	8.6	18.7	20.9
Othera	85.1	73.9	69.7

^aOther includes ‘Unknown, not tested, neutral’.

For high-penetrance genes (BRCA1, BRCA2, TP53, PTEN, STK11 and CDH1), 353 variants in the BCAC dataset, 92 variants in SBCGS-GWAS and 46 variants in SBCGS-exomechip were included in the current analysis. We found that common variants (MAF ≥ 0.05), rs9620817 in the CHEK2 gene and rs13330119 in the PALB2 gene, were associated with breast cancer at P < 0.01 in the combined BCAC and SBCGS dataset of 9000 cases and 8758 controls (Table 2). The ORs (95% CI) for those common SNPs were 0.90 (0.83–0.96) for rs9620817, (P = 0.002) and 1.07 (1.02–1.13) for rs13330119, (P = 0.008).

Table 2.

Associations of breast cancer with SNPs with MAF ≥0.05 among Asians^a

Gene	SNP (alleles) annotationb	Chr:Positionc	Study	No. of cases/ controls	EAF	OR (95% CI)d	P valuee
CHEK2	rs9620817 (A/T)	chr22:29108556	BCAC	6269/6623	0.90	0.89 (0.83–0.97)	0.006
	Intron		SBCGS-GWAS	2731/2135	0.90	0.91 (0.79–1.04)	0.178
			Meta-analysis	9000/8758		0.90 (0.83–0.96)	0.002
CDH1	rs8063605 (T/C)	chr16:68836665	BCAC	6268/6624	0.77	0.94 (0.88–0.99)	0.026
	Intron		SBCGS-GWAS	2731/2135	0.75	0.95 (0.86–1.05)	0.333
			Meta-analysis	8999/8759		0.94 (0.90–0.99)	0.016
CDH1	rs1801552 (T/C)	chr16:68857441	BCAC	6268/6624	0.37	0.94 (0.89–0.99)	0.011
	Synonymous		SBCGS-GWAS	2731/2135	0.34	0.99 (0.91–1.09)	0.987
			Meta-analysis	8999/8759		0.94 (0.90–0.99)	0.03
CDH1	rs9935563 (T/C)	chr16:68861656	BCAC	6265/6624	0.37	0.94 (0.89–0.99)	0.017
	Intron		SBCGS-GWAS	2731/2135	0.34	0.99 (0.91–1.09)	0.965
			Meta-analysis	8996/8759		0.94 (0.89–0.99)	0.039
CDH1	rs2902185 (T/C)	chr16:68774168	BCAC	6267/6623	0.77	0.94 (0.89–0.99)	0.042
	Intron		SBCGS-GWAS	2731/2135	0.74	0.97 (0.89–1.07)	0.571
			Meta-analysis	8998/8758		0.95 (0.90–1.00)	0.043
CDH1	rs7188750 (A/G)	chr16:68842895	BCAC	6269/6622	0.16	1.08 (1.01–1.16)	0.022
	Intron		SBCGS-GWAS	2731/2135	0.19	1.00 (0.90–1.11)	0.95
			Meta-analysis	9000/8757		1.08 (1.01–1.15)	0.047
NBN	rs1805796 (T/C)	chr8:90993395	BCAC	6265/6624	0.66	0.96 (0.91–1.01)	0.152
	Intron		SBCGS-GWAS	2731/2135	0.64	0.92 (0.85–1.00)	0.056
			Meta-analysis	8996/8759		0.95 (0.90–0.99)	0.026
NBN	rs867185 (A/G)	chr8:90975150	BCAC	6267/6623	0.66	0.96 (0.91–1.01)	0.146
	Intron		SBCGS-GWAS	2731/2135	0.65	0.92 (0.85–1.00)	0.066
			Meta-analysis	8998/8758		0.95 (0.91–0.99)	0.028
PALB2	rs13330119 (T/C)	chr16:23630071	BCAC	6269/6622	0.20	1.08 (1.02–1.15)	0.015
	Intron		SBCGS-GWAS	2731/2135	0.22	1.06 (0.96–1.16)	0.277
			Meta-analysis	9000/8757		1.07 (1.02–1.13)	0.008
PALB2	rs420259 (A/G)	chr16:23634026	BCAC	6268/6622	0.63	0.93 (0.88–0.98)	0.005
	Intron		SBCGS-GWAS	2731/2135	0.62	0.94 (0.87–1.03)	0.175
			SBCGS-Exomechip	3061/3393	0.61	1.00 (0.94–1.08)	0.875
			Meta-analysis	12 060/12 150		0.93 (0.89–0.97)	0.01
PALB2	rs249954 (A/G)	chr16:23640467	BCAC	6268/6623	0.37	1.08 (1.02–1.13)	0.005
	Intron		SBCGS-GWAS	2731/2135	0.38	1.06 (0.97–1.15)	0.181
			SBCGS-Exomechip	3063/3392	0.39	0.99 (0.93–1.07)	0.868
			Meta-analysis	12 062/12 150		1.06 (1.02–1.11)	0.01

^aAll SNPs with meta-analysis P < 0.05.

^bEffect allele/reference allele based on forward strand.

^cChromosome position (bp) based on NCBI Human Genome Build 37.

^dOR (95% CI) was adjusted for study site (BCAC only), age and PCs.

^e P value in each study obtained from logistic regression analysis. Meta-analysis P-value derived from a weighted z statistic-based meta-analysis.

For low-frequency SNPs with MAF < 0.05 in the combined BCAC and SBCGS dataset, three missense SNPs in the BRCA2 gene (rs80358978, rs80359065 and rs11571653) were associated with breast cancer risk at P < 0.01 (Table 3). We found a novel missense variant, rs80359065 (Lys2729Asn) in the BRCA2 gene, that was significantly associated with breast cancer risk with OR (95% CI) of 1.35 (1.13–1.63) and P = 0.001. Another missense variant, rs80358978 (Gly2508Ser) in the BRCA2 gene, was associated with breast cancer risk with OR (95% CI) of 3.02 (1.69–5.39) and P = 1.23 × 10^–4. In addition, we found one more missense variant, rs11571653 (Met784Val) in the BRCA2 gene, that was significantly associated with breast cancer risk with OR (95% CI) of 0.77 (0.65–0.92), (P = 0.005). None of the tests for heterogeneity gave statistically significant results (P for heterogeneity > 0.05).

Table 3.

Associations of breast cancer with SNPs with MAF<0.05 among Asians^a

Gene	SNP (alleles) annotationb	Chr:Positionc	Study	No. of cases/controlsd	EAF (no. of case/control)e	ORf	Lower 95% CI	Upper 95% CI	P valueg
BRCA1	rs8176085 (T/C)	chr17:41274789	BCAC	26/1	0.0011 (26/1)	26.98	12.65	57.55	3.33E−06
	Intron
BRCA1	rs799923 (A/G)	chr17:41251931	BCAC	92/77	0.0074 (95/85)	1.82	1.34	2.47	9.65E−05
	Intron
BRCA1	rs8176173 (T/C)	chr17:41238619	BCAC	9/1	0.0004 (9/1)	14.92	4.07	54.69	0.001
	Intron
BRCA1	rs8176258 (A/G)	chr17:41216021	BCAC	9/1	0.0004 (9/1)	14.88	4.06	54.53	0.001
	Intron
BRCA1	rs8176305 (C/T)	chr17:41201364	BCAC	21/17	0.0015 (21/17)	2.08	1.05	4.00	0.03
	Intron
BRCA1	rs273898669 (C/T)	chr17:41276135	BCAC	21/9	0.0012 (21/9)	2.22	1.09	4.55	0.038
	Intron
BRCA2	rs80358978 (A/G)	chr13:32930651	BCAC	31/12	0.0017 (31/12)	2.62	1.44	4.78	0.003
	Missense		SBCGS-Exomechip	10/2	0.0009 (10/2)	5.59	1.80	17.38	0.012
			Meta-analysis			3.02	1.69	5.39	1.23E−04
BRCA2	rs80359065 (T/G)	chr13:32937526	BCAC	162/128	0.0114 (164/128)	1.43	1.13	1.80	0.003
	Missense		SBCGS-Exomechip	93/82	0.0137 (93/83)	1.24	0.92	1.67	0.157
			Meta-analysis			1.35	1.13	1.63	0.001
BRCA2	rs11571653 (G/A)	chr13:32910842	BCAC	221/323	0.023 (226/343)	0.78	0.66	0.92	0.004
	Missense		SBCGS-Exomechip	2/4	0.0005 (2/4)	0.55	0.11	2.70	0.481
			Meta-analysis			0.77	0.65	0.92	0.005
BRCA2	rs28897700 (G/A)	chr13:32893344	BCAC	10/6	0.0006 (10/6)	3.23	1.14	9.09	0.019
	Synonymous
BRCA2	rs11571789 (A/C)	chr13:32959240	BCAC	37/32	0.0027 (37/31)	1.68	1.02	2.76	0.037
	Intron
PTEN	rs1234218 (T/C)	chr10:89650759	BCAC	69/104	0.0068 (69/105)	0.70	0.52	0.94	0.02
	Intron
PTEN	rs3781195 (G/A)	chr10:89642609	BCAC	182/258	0.0177 (184/264)	0.82	0.68	0.99	0.043
	Intron
RAD51C	rs17175543 (G/T)	chr17:56803598	BCAC	45/50	0.9962 (6250/6618)	0.65	0.43	1.00	0.042
	Intron

^aAll SNPs with meta-analysis P-value < 0.05 or P-value from BCAC < 0.05.

^bEffect allele/reference allele based on forward strand.

^cChromosome position (bp) based on NCBI Human Genome Build 37.

^dNumber of samples carrying heterozygous variant.

^eEffect allele frequency and number of cases and controls carrying effect allele are listed.

^fOR (95% CI) was adjusted for study site (BCAC only), age, and PCs.

^g P-value in each study obtained from logistic regression analysis. Meta-analysis p-value derived from a weighted z statistic-based meta-analysis.

Among BCAC study participants, we identified 10 low-frequency SNPs (MAF < 0.02) in three high-penetrance genes (BRCA1, BRCA2 and PTEN) and one moderate-penetrance gene (RAD51C) at P < 0.05 (Table 3). Results included six SNPs in the BRCA1 gene (rs8176085, rs799923, rs8176173, rs8176258, rs8176305 and rs273898669), two SNPs in the BRCA2 gene (rs28897700 and rs11571789), two SNPs in the PTEN gene (rs1234218 and rs3781195) and one SNP in the RAD51C gene (rs17175543). These SNPs were associated (P < 0.05) with breast cancer risk; the SNP rs8176085 in the BRCA1 gene showed the strongest association (P = 3.33 × 10^–6). This SNP was observed in 26 cases and in only one control. These rare SNPs were either intronic or synonymous SNPs, and they were not genotyped in the SBCGS data. No significant associations were observed for the other five high/moderate penetrance genes (TP53, ATM, BRIP1, RAD50 and STK11).

When stratified by ER status, 5 SNPs (rs9620817, rs420259, rs249954, rs8176085, rs11571653) showed a stronger association for ER-positive breast cancer than ER-negative breast cancer according to association P value (P < 0.05), and the difference was statistically significant (P for heterogeneity < 0.05) (Supplementary Table 3, available at Carcinogenesis Online). In addition, two SNPs (rs80358978, rs80359065) showed a stronger association for ER-negative breast cancer than ER-positive breast cancer according to association p-value (P < 0.05). The difference was statistically significant (P for heterogeneity < 0.05) for these two SNPs (Supplementary Table 3, available at Carcinogenesis Online).

We investigated the association of our identified SNPs with breast cancer risk in women of European ancestry using data from the DRIVE GAME-ON Consortium (38), consisting of 122 977 cases and 105 974 controls from 11 breast cancer GWAS. SNP rs9620817 with MAF > 0.05 showed a significant association with breast cancer risk in women of European ancestry at P =8.08 × 10^–12. The OR (95% CI) for the association was 0.93 (0.91–0.95), consistent with the association observed in the East Asian population with OR (95% CI) being 0.90 (0.83–0.96).

Discussion

In this study, we investigated associations between genetic variants in 13 high or moderate-penetrance genes (BRCA1, BRCA2, TP53, PTEN, STK11, CDH1, CHEK2, ATM, BRIP1, PALB2, RAD51C, RAD50 and NBN) and breast cancer risk in 6269 cases and 6624 controls from BCAC Asian women (up to 654 SNPs) and 5794 cases and 5529 controls from SBCGS Chinese women (up to 236 SNPs). Our study identified several new risk variants in BRCA1, BRCA2, CHEK2 and PALB2 genes in relation to breast cancer risk in Asian women at P < 0.01. Six low-frequency variants (MAF < 0.01) in the BRCA1 gene showed associations with breast cancer risk among the BCAC data, although these six BRCA1 SNPs were not genotyped in SBCGS data. These low-frequency variants are not in linkage disequilibrium with each other (r² < 0.1). SNP rs8176085 showed the strongest evidence of association with an approximately 27-fold increased risk of breast cancer with P = 3.33 × 10^–6. These six low-frequency variants are located in intron regions with no obvious functional effect, so further research is needed to determine whether the association with these variants may be due to their correlation with high risk variants in the coding region. For combined BCAC and SBCGS data, we found three missense variants in the BRCA2 gene which showed significant associations with breast cancer risk: rs80359065 (Lys2729Asn), rs80358978 (Gly2508Ser) and rs11571653 (Met784Val).

A novel missense variant, rs80359065 (Lys2729Asn), was associated with a 35% increase in breast cancer risk (95% CI: 1.13–1.63; P = 0.001). It was located 35 kb upstream from the GWAS SNP rs11571833 (Lys3326Ter) in the BRCA2 gene that introduces a premature stop codon (19). Several BRCA2 polymorphisms have been studied for their association with breast cancer risk; however, until now, no studies have found an association with rs80359065 (Lys2729Asn) in the BRCA2 gene (28,39–43). This variant was observed in 93 heterozygous breast cancer cases and 82 heterozygous controls in the SBCGS, and 162 heterozygous breast cancer cases and 128 heterozygous controls in the BCAC. Although the clinical importance of this variant for breast cancer is unknown, this SNP may be functionally important since it is predicted to be ‘probably damaging’ based on the Polyphen-2 score (0.9) and ‘probably damaging’ based on the SIFT score (0.001).

A missense variant, rs80358978 (Gly2508Ser), was previously reported by our group to be associated with increased risk of breast cancer in Chinese women (28). As our sample size increased dramatically in the current study compared with the previous report (28), we have more power to detect this SNP, which is located 42 kb upstream from the GWAS SNP rs11571833 in the BRCA2 gene (19). This variant was observed in 10 heterozygous breast cancer cases and two heterozygous controls in the SBCGS, and 31 heterozygous breast cancer cases and 12 heterozygous controls in the BCAC. A sensitivity analysis for this SNP by excluding data presented in the previous analyses (28) showed a consistent significant association. This variant was predicted to be functionally important since the Polyphen-2 score was 1.0 (‘probably damaging’) and the SIFT score was 0 (‘damaging’). However, the PROVEAN prediction score (−1.80) predicted this variant as ‘neutral’.

A missense variant, rs11571653 (Met784Val), was previously associated with an approximately 2-fold increased risk of breast cancer among a Japanese population [OR (95% CI) =2.03 (1.07–3.87)] with heterozygotes in 21% of cases and 13% of controls (39). In our study, we found that this SNP was associated with a 23% decreased risk of breast cancer in women of Asian ancestry. The inconsistent directions in association in these two studies could be due to sample size differences. Ishitobi et al. (39). used 149 cases and 154 controls, and our study included 5794 cases and 5529 controls from the SBCGS and an additional 6269 cases and 6624 controls from the BCAC. This variant was predicted to be functionally not important with a PROVEAN score of 1.09, a Polyphen-2 score of 0, and a SIFT score of 1.0. Although the exact functional significance of this SNP is unknown, several biological studies have suggested that this SNP is associated with breast cancer risk in populations of different ethnic origins (41,44,45). Thus, our findings are biologically plausible and provide functional implications.

Among the 11 common variants that showed an association with breast cancer risk in the combined datasets, we found two variants, rs9620817 in the CHEK2 gene, and rs13330119 in the PALB2 gene, with P < 0.01. These variants are not in linkage disequilibrium with each other and other GWAS-identified breast cancer risk variants (r² < 0.1). SNP rs13330119 was previously identified as a breast cancer risk-associated SNP in a female Chinese population by Chen et al. (46).

We also investigated the genetic relationship between the individuals via pair-wise identity-by-descent (IBD) and did not find any cryptic relatedness among carriers of the rare variants identified from the present study. We also checked the NCBI clinvar database (https://www.ncbi.nlm.nih.gov/clinvar/). We did not find any pathological mutation linked to the rare variants identified in the present study.

It will be ideal to do deep sequencing of subjects carrying the most interesting variants. This is one of the limitations of our study that we do not have full coverage of coding regions for rare-variant analysis. Our study has a total of 12 063 cases and 12 153 controls. To our knowledge, this is the largest Asian study to examine these 13 high or moderate-penetrance genes and breast cancer risk. However, statistical power is still very limited for rare variants. In this study, a total of 654 variants were included. Bonferroni correction will result in an association significance level of 7.7 × 10^–5 (=0.05/654). After adjusting for multiple comparisons, most of our results became insignificant except for SNP rs8176085 in BRCA1 gene. However, among the 654 variants, 411 were rare variants with MAF < 0.01, and 359 were extremely rare with MAF < 0.0005. As discussed in the article (47), one of the major challenges in rare variants study is very low statistical power. Even among a relative large study like ours, there is only 3.1% power to detect a variant with MAF = 0.01 and OR = 1.2 at a significance level of P = 7.7 × 10^–5. Therefore, in this study, a majority of the rare SNPs cannot be fully investigated due to very limited power. Further independent studies with larger sample size will be helpful to validate the findings from this study.

In conclusion, our study identified several risk variants in BRCA1, BRCA2, CHEK2 and PALB2 genes in relation to breast cancer risk in Asian women. Most importantly, we have shown an increased risk of breast cancer for Asian women with a newly identified missense variant rs80359065 (Lys2729Asn) in the BRCA2 gene. Results from this study provide additional insights into the effects of genes that carry high/moderate penetrance mutations on breast cancer risk in Asian women.

Supplementary material

Supplementary data are available at Carcinogenesis Online.

Funding

This work was supported by multiple grants. The SBCGS was supported primarily by National Institutes of Health (R01CA64277, R01CA148667, R37CA70867, and UM1CA182910 to W.Z.). Biological sample preparation was conducted the Survey and Biospecimen Shared Resource, which is supported by National Institutes of Health (P30 CA68485). The scientific development and funding of this project were, in part, supported by the Genetic Associations and Mechanisms in Oncology (GAME-ON) Network, funded by the National Institutes of Health (U19 CA148065). The ACP study is funded by the Breast Cancer Research Trust, UK. The HERPACC was supported by a Grant-in-Aid for Scientific Research on Priority Areas from the Ministry of Education, Science, Sports, Culture and Technology of Japan, by the Practical Research for Innovative Cancer Control (15ck0106177h0001) from Japan Agency for Medical Research and Development (AMED), by a Grant-in-Aid for the Third Term Comprehensive 10-year Strategy for Cancer Control from Ministry Health, Labour and Welfare of Japan, by Health and Labour Sciences Research Grants for Research on Applying Health Technology from Ministry Health, Labour and Welfare of Japan, National Cancer Center Research and Development Fund. The LAABC is supported by grants (1RB-0287, 3PB-0102, 5PB-0018, 10PB-0098) from the California Breast Cancer Research Program. Incident breast cancer cases were collected by the USC Cancer Surveillance Program (CSP) which is supported under subcontract by the California Department of Health. The CSP is also part of the National Cancer Institute’s Division of Cancer Prevention and Control Surveillance, Epidemiology, and End Results Program, under contract number N01CN25403. The MYBRCA is funded by research grants from the Malaysian Ministry of Science, Technology and Innovation (MOSTI), Malaysian Ministry of Higher Education (UM.C/HlR/MOHE/06) and Cancer Research Initiatives Foundation (CARIF). Additional controls were recruited by the Singapore Eye Research Institute, which was supported by a grant from the Biomedical Research Council (BMRC08/1/35/19/550), Singapore and the National medical Research Council, Singapore (NMRC/CG/SERI/2010). The SEBCS was supported by the BRL (Basic Research Laboratory) program through the National Research Foundation of Korea funded by the Ministry of Education, Science and Technology (2012-0000347). The SGBCC is funded by the NUS start-up Grant, National University Cancer Institute Singapore (NCIS) Centre Grant and the NMRC Clinician Scientist Award. Additional controls were recruited by the Singapore Consortium of Cohort Studies-Multi-ethnic cohort (SCCS-MEC), which was funded by the Biomedical Research Council, grant number: 05/1/21/19/425. The TBCS was funded by The National Cancer Institute Thailand. The TWBCS is supported by the Taiwan Biobank project of the Institute of Biomedical Sciences, Academia Sinica, Taiwan.

Abbreviations

BCAC

Breast Cancer Association Consortium

ER

estrogen receptor

MAF

minor allele frequency

SNP

single nucleotide polymorphism

SBCGS

Shanghai Breast Cancer Genetics Study