Abstract

Genetic variants may influence miRNA–mRNA interaction through modulate binding affinity, creating or destroying miRNA-binding sites. Twenty-four single nucleotide polymorphisms (SNPs) that were predicted to affect the binding affinity of breast cancer-related miRNAs to 3′-untranslated regions (UTR) of known genes were genotyped in 878 breast cancer cases and 900 controls in Chinese women. Three promising SNPs (rs10494836, rs10857748 and rs7963551) were further validated in additional 914 breast cancer cases and 967 controls. The variant allele (C) of rs7963551 at 3′-UTR of RAD52 showed a consistently reduced breast cancer risk in two stages with a combined odds ratio (OR) of 0.84 [95% confidence interval (CI) = 0.75–0.95], which was more prominent among women with early age at first live birth (OR = 0.71, 95% CI = 0.58–0.87). A significant interaction was observed between rs7963551 and age at first live birth on breast cancer risk ( P for interaction = 0.04). Luciferase activity assay showed a higher expression level for rs7963551 C allele as compared with A allele ( P = 5.19×10 –3 for MCF-7 cell lines), which might be due to a reduced inhibition from a weakened binding capacity of miRNA to 3′-UTR of RAD52 harboring C allele. These findings indicate that rs7963551 located at hsa-let-7 binding site may alter expression of RAD52 through modulating miRNA–mRNA interaction and contribute to the development of breast cancer in Chinese women.

Introduction

Breast cancer is one of the most prevalent cancers in both developing and developed countries and caused by a combination of genetic and environmental factors. Familial aggregation and twins studies have indicated the contribution of hereditary factors to breast cancer ( 1 , 2 ). Over the past few years, a series of susceptibility genes have been identified to be implicated with breast cancer risk, including high-penetrance genes ( BRCA1 , BRCA2 , TP53 and PTEN ), moderate-penetrance genes ( CHEK2 , ATM , BRIP1 and PALB2 ) and common low-penetrance variants in multiple genes ( 3 ). However, <20% of familial risk of breast cancer can be explained by high-to-moderate penetrance genes, whereas ~9% have been attributed to low-penetrance variants ( 4 , 5 ). Therefore, there are still important regions harboring genetic variants associated with breast cancer risk to be further identified.

miRNAs are a class of small (~20–22 nt), non-coding RNAs that exert their regulatory effects by incomplete combination to target sequences mainly on 3′-untranslated regions (UTRs) of messenger RNA (mRNA) ( 6 ). A single miRNA can bind to approximately 200 genes and potentially regulate the expression of nearly 30% of human mRNAs ( 7 ). A number of studies have highlighted the importance of miRNA in the development of cancers. miRNA expression levels signify initiation, progression, metastasis of human tumors and even the different neoplasm types ( 8 ). For example, expressions of miR-200 family were found to be lost in invasive breast cancer cell lines with mesenchymal phenotype ( 9 ). Moreover, knockdown of miR-21 increased TIMP3 protein expression and luciferase reporter activity, suggesting miR-21 might promote invasion in breast cancer cells via its regulation of TIMP3 ( 10 ). Genetic variants may modulate the miRNA–mRNA interaction through altering miRNA expression, maturation, creating or destroying the miRNA-binding sites, and result in diverse functional consequences ( 11 , 12 ). As yet, genetic variants in pre-miRNA and mature miRNA regions have been implicated in cancer susceptibility, prognosis and sensitivity to chemotherapy ( 13–16 ). Besides, polymorphisms in miRNA-binding sites may also involve in the process of carcinogenesis ( 11 ). Yu et al . ( 17 ) conducted a genome-wide analysis of genetic variants located in the miRNA-binding sites of 3′-UTR of human genes and found that genetic variants in miRNA-binding sites might affect expression of miRNA targets that were potentially associated with various human cancers. Furthermore, several independent studies have found SNPs in 3′-UTR affect genes expressions via miRNA–mRNA interaction, including MYCL1 , SET8 and KRAS ( 18–20 ).

To systematically investigate the effect of genetic variants in miRNA-binding sites on the development of breast cancer, we defined miRNAs that were reported to be associated with breast cancer and then selected single nucleotide polymorphisms (SNPs) that were predicted to be in the miRNA-binding sites. Furthermore, we performed a two-stage case–control study and functional studies to evaluate the effect of selected SNPs on breast cancer risk in Chinese women.

Materials and methods

Ethics statement

This study was approved by the institutional review board of Nanjing Medical University. The design and performance of this study involving human subjects were clearly described in a research protocol. All participants were voluntary and would complete the informed consent in written before taking part in this research.

Study population

This study was approved by the institutional review board of Nanjing Medical University. All subjects were genetically unrelated ethnic Han Chinese women from Nanjing and surrounding regions in southeastern China. We collected breast cancer cases from the First Affiliated Hospital of Nanjing Medical University, the Cancer Hospital of Jiangsu Province and the Gulou Hospital, Nanjing, China from January 2004 to April 2010. Controls were randomly selected from >30 000 participants in a community-based screening program for non-infectious diseases conducted from 2004 to 2006 in Jiangsu Province, China. The controls were frequency matched to the cases on age and residential area (urban or rural). Study characteristics have been described previously ( 21 , 22 ). Briefly, 878 cases and 900 controls were randomly selected into stage 1, and the remaining 914 cases and 967 controls formed the stage 2. In total, 1792 breast cancer cases and 1867 cancer-free controls were included in analysis.

SNP selection and genotyping

We carried out a literature search by using the Medline/PubMed database ( http://www.ncbi.nlm.nih.gov/pubmed/ ) to retrieve papers that reported an important role of miRNA in breast cancer development before November 2010. A total of 20 miRNAs [let-7 ( 23 ), miR-9 ( 24 ), miR-10b ( 25 ), miR-17 ( 26 ), miR-20 ( 26 ), miR-21 ( 25 , 27 ), miR-31 ( 28 ), miR-34 ( 29 ), miR-101 ( 30 ), miR-125b ( 25 ), miR-126 ( 31 ), miR-145 ( 25 ), miR-155 ( 25 ), miR-200 ( 9 ), miR-206 ( 32 ), miR-221 ( 33 ), miR-222 ( 33 ), miR-335 ( 31 ), miR-373 ( 34 ) and miR-520 ( 34 )] according to retrieve articles were selected. We further assessed these 20 miRNAs in Patrocles ( 35 ) to predict the miRNA target genes, and chose SNPs located on binding site of 3′-UTR for genes that were reported to be involved in cancers. Linkage disequilibrium value ( r2 < 0.8) and minor allele frequency (≥0.05) in Chinese Han population were further applied to filter SNPs. With one SNP (rs27770) removed for genotyping failure (call rates <95%), a total of 24 SNPs related to 14 miRNAs that may affect the binding affinity of miRNA–mRNA were finally included in this study ( Table I ).

Table I.

Summary of associations between 24 SNPs in miRNA-binding sites and breast cancer risk in stage 1 study

Chromosome Target gene MiRNA SNP  Allele a  Case b  Control b  Minor allele frequency c  Hardy–Weinberg equilibrium d  Codominant model e  Additive model e 
N = 878  N = 900  (Case/control) Phet Phom Padd 
1q24.2 SLC19A2 miR-101 rs12091844 C/T 638/194/18 673/196/18 0.267/0.239 0.377 0.48 0.86 0.50 
1q32.1 JARID1B miR-125b-2 rs10494836 C/T 456/353/54 520/322/53 0.267/0.239 0.714 0.02 0.55 0.06 
2p22.2 CYP1B1 let-7 rs2855658 G/A 661/188/15 678/204/9 0.126/0.125 0.167 0.59 0.26 0.97 
2q13 PAX8 miR-520e rs874898 G/C 461/350/55 485/350/53 0.266/0.257 0.379 0.72 0.49 0.50 
2q34 ERBB4 miR-200 rs1595066 C/T 376/380/98 396/386/103 0.337/0.335 0.546 0.79 0.89 0.99 
3p25.3 VHL let-7 rs1642742 T/C 508/308/50 545/297/43 0.236/0.216 0.766 0.48 0.58 0.41 
3q13.11 ALCAM miR-221 rs1157 G/A 599/237/27 613/251/27 0.169/0.171 0.814 0.95 0.61 0.74 
5p13.2 GDNF let-7 rs3749692 G/A 359/391/112 358/425/105 0.357/0.357 0.242 0.55 0.78 0.96 
7q22.1 EIF2AK1 miR-520c-5p rs3801030 T/C 778/90/1 808/80/1 0.053/0.046 1.000 0.37 0.86 0.40 
7q31.2 WNT2 miR-10b rs2024233 A/G 222/434/193 220/434/231 0.483/0.506 0.591 0.95 0.16 0.16 
8q24.11 EXT1 miR-373 rs11785084 C/T 453/354/61 494/338/57 0.274/0.254 1.000 0.24 0.45 0.22 
9q33.1 TLR4 let-7 rs7873784 G/C 721/113/13 726/145/14 0.082/0.098 0.053 0.19 0.93 0.31 
10q26.3 CYP2E1 miR-34a rs10857748 T/C 521/315/40 498/335/62 0.225/0.256 0.599 0.34 0.03 0.05 
11p15.5 H19 miR-520a-5p rs2839701 G/C 404/370/84 449/354/84 0.314/0.294 0.258 0.09 0.67 0.23 
12p13.2 ETV6 miR-34b rs1573613 T/C 260/425/179 268/466/154 0.453/0.436 0.048    
12p13.33 RAD52 let-7 rs7963551 A/C 597/241/26 581/270/41 0.169/0.197 0.185 0.40 0.05 0.08 
12q23.1 HAL let-7 rs2230885 C/A 656/197/14 672/205/14 0.130/0.131 0.883 0.92 0.85 0.86 
14q11.2 NDRG2 miR-21 rs10196 T/C 377/393/99 405/382/100 0.340/0.328 0.493 0.37 0.63 0.43 
15q24.1 CSK miR-373 rs2071501 T/G 313/406/130 309/448/129 0.392/0.398 0.108 0.56 0.84 0.72 
15q26.1 FURIN miR-200 rs4702 A/G 232/415/215 243/467/184 0.490/0.467 0.158 0.27 0.27 0.34 
17q23.2 BRIP1 miR-101 rs7213430 A/G 429/357/65 463/342/82 0.286/0.285 0.117 0.14 0.39 0.83 
19q13.2 CEACAM5 let-7 rs7252828 A/G 427/346/76 439/365/82 0.293/0.298 0.631 0.82 0.79 0.75 
19q13.33 KLK2 miR-206 rs198978 G/T 518/307/35 556/277/54 0.219/0.217 0.018    
22q11.21 MAPK1 let-7 rs2276008 G/C 720/122/8 746/134/6 0.081/0.082 1.000 0.72 0.55 0.95 
Chromosome Target gene MiRNA SNP  Allele a  Case b  Control b  Minor allele frequency c  Hardy–Weinberg equilibrium d  Codominant model e  Additive model e 
N = 878  N = 900  (Case/control) Phet Phom Padd 
1q24.2 SLC19A2 miR-101 rs12091844 C/T 638/194/18 673/196/18 0.267/0.239 0.377 0.48 0.86 0.50 
1q32.1 JARID1B miR-125b-2 rs10494836 C/T 456/353/54 520/322/53 0.267/0.239 0.714 0.02 0.55 0.06 
2p22.2 CYP1B1 let-7 rs2855658 G/A 661/188/15 678/204/9 0.126/0.125 0.167 0.59 0.26 0.97 
2q13 PAX8 miR-520e rs874898 G/C 461/350/55 485/350/53 0.266/0.257 0.379 0.72 0.49 0.50 
2q34 ERBB4 miR-200 rs1595066 C/T 376/380/98 396/386/103 0.337/0.335 0.546 0.79 0.89 0.99 
3p25.3 VHL let-7 rs1642742 T/C 508/308/50 545/297/43 0.236/0.216 0.766 0.48 0.58 0.41 
3q13.11 ALCAM miR-221 rs1157 G/A 599/237/27 613/251/27 0.169/0.171 0.814 0.95 0.61 0.74 
5p13.2 GDNF let-7 rs3749692 G/A 359/391/112 358/425/105 0.357/0.357 0.242 0.55 0.78 0.96 
7q22.1 EIF2AK1 miR-520c-5p rs3801030 T/C 778/90/1 808/80/1 0.053/0.046 1.000 0.37 0.86 0.40 
7q31.2 WNT2 miR-10b rs2024233 A/G 222/434/193 220/434/231 0.483/0.506 0.591 0.95 0.16 0.16 
8q24.11 EXT1 miR-373 rs11785084 C/T 453/354/61 494/338/57 0.274/0.254 1.000 0.24 0.45 0.22 
9q33.1 TLR4 let-7 rs7873784 G/C 721/113/13 726/145/14 0.082/0.098 0.053 0.19 0.93 0.31 
10q26.3 CYP2E1 miR-34a rs10857748 T/C 521/315/40 498/335/62 0.225/0.256 0.599 0.34 0.03 0.05 
11p15.5 H19 miR-520a-5p rs2839701 G/C 404/370/84 449/354/84 0.314/0.294 0.258 0.09 0.67 0.23 
12p13.2 ETV6 miR-34b rs1573613 T/C 260/425/179 268/466/154 0.453/0.436 0.048    
12p13.33 RAD52 let-7 rs7963551 A/C 597/241/26 581/270/41 0.169/0.197 0.185 0.40 0.05 0.08 
12q23.1 HAL let-7 rs2230885 C/A 656/197/14 672/205/14 0.130/0.131 0.883 0.92 0.85 0.86 
14q11.2 NDRG2 miR-21 rs10196 T/C 377/393/99 405/382/100 0.340/0.328 0.493 0.37 0.63 0.43 
15q24.1 CSK miR-373 rs2071501 T/G 313/406/130 309/448/129 0.392/0.398 0.108 0.56 0.84 0.72 
15q26.1 FURIN miR-200 rs4702 A/G 232/415/215 243/467/184 0.490/0.467 0.158 0.27 0.27 0.34 
17q23.2 BRIP1 miR-101 rs7213430 A/G 429/357/65 463/342/82 0.286/0.285 0.117 0.14 0.39 0.83 
19q13.2 CEACAM5 let-7 rs7252828 A/G 427/346/76 439/365/82 0.293/0.298 0.631 0.82 0.79 0.75 
19q13.33 KLK2 miR-206 rs198978 G/T 518/307/35 556/277/54 0.219/0.217 0.018    
22q11.21 MAPK1 let-7 rs2276008 G/C 720/122/8 746/134/6 0.081/0.082 1.000 0.72 0.55 0.95 

a Major/minor allele.

b Major homozygote/heterozygote/rare homozygote between cases and controls.

c Minor allele frequency.

d Hardy–Weinberg equilibrium test among controls.

e Logistic regression with adjustment for age, age at menarche and menopausal status were used to test associations in codominant ( Phet : heterozygote versus major homozygote; Phom : minor homozygote versus major homozygote) and additive ( Padd : minor homozygote versus heterozygote versus major homozygote) models.

All 24 SNPs were genotyped in stage 1 samples using TaqMan OpenArray Genotyping System (Applied Biosystems). Promising SNPs were further validated in stage 2 samples using ABI PRISM 7900 HT platform (Applied Biosystems). The genotyping methods and quality control have been described previously ( 21 ). The overall call rates of 24 SNPs in stage 1 were from 96.2 to 99.2%. About 5% of samples in stage 1 were randomly selected to be repeated on ABI PRISM 7900 HT platform and the concordance rates were 97.9–100%.

RAD52 3′-UTR luciferase reporter plasmid

The 3′-UTR region of RAD52 containing the putative recognition site rs7963551 was amplified from a DNA sample carrying AA genotype. The primers were 5′-GTAGACGCGTTGGTCAAGGTGGGTTACATC-3′ (sense) and 5′-TCCAAAGCTTTACTGGTGTGAGCCACTGTG-3′ (antisense). PCR products were separated in agarose gel, extracted, purified and then cloned into the pMIR-REPORT™ (Applied Biosystems) vector with Mlu I and Hind III digestions. Plasmid containing the rs7963551 C allele was generated using site-specific mutagenesis method. All the constructs used in this study were restriction mapped and sequenced to confirm the authenticity.

Transient transfections and luciferase assays

The 293T, Hela and MCF-7 cells were maintained in Dulbecco’s modified Eagle’s medium, respectively, supplemented with 10% heat-inactivated fetal bovine serum (Gibco) and 50 µg/ml streptomycin (Gibco) at a 37ºC incubator supplemented with 5% CO 2 . Transfections were performed with cells using Lipofectamine2000 according to manufacturer’s introduction (Invitrogen) after 24h. The RAD52 3′-UTR luciferase plasmids (different alleles) and chemically synthesized mature let-7 were cotransfected into three different cells, respectively. The pRL-SV40 plasmid (Promega) was also cotransfected as a normalizing control. Six replicates for each group and the experiment were repeated at least twice. After 24h of incubation, cells were collected and analyzed for luciferase activity with Dual-Luciferase Reporter Assay System (Promega).

Statistical analyses

Student’s t -test (for continuous variables) and Chi-squared test (for categorical variables) were used to examine the differences in the distributions of demographic characteristics, selected variables and genotype frequencies of 24 SNPs between cases and controls. The associations between genotype and breast cancer risk were estimated by calculating odds ratios (ORs) and 95% confidence intervals (CIs) through logistic regression model with adjustment for age, age at menarche and menopausal status. Differences in the expression levels of RAD52 among subgroups were examined using t -test. All of the statistical analyses were performed with Statistical Analysis System software (9.1.3; SAS Institute, Cary, NC).

Result

The information of 24 SNPs was described in Table I . Two SNPs (rs1573613 and rs198978) showed departures from Hardy–Weinberg equilibrium among controls ( P < 0.05) and were excluded in the further analysis. The association results of 22 SNPs in condominant and additive models in stage 1 showed that three SNPs ( JARID1B rs10494836, CYP2E1 rs10857748 and RAD52 rs7963551) were suggestively associated with breast cancer risk, which are located in the target binding sites of miR-125b-2, miR-34a and let-7, respectively.

To validate the results of stage 1, we further genotyped these three promising SNPs in stage 2 samples ( Table II ). The results showed that RAD52 rs7963551 was consistently associated with breast cancer risk in two stages (stage 1: OR = 0.85, 95% CI = 0.71–1.07; stage 2: OR = 0.81, 95% CI = 0.68–0.97) for additive model. After combining these two stages, we found that variant allele of rs7963551 was significantly associated with a decreased risk of breast cancer (additive model: OR = 0.84, 95% CI = 0.75–0.95).

Table II.

Association results for three promising SNPs in two stages

SNP Stage  Case a  Control a  Minor allele frequency b (case/control)   OR het (95% CI) c  OR hom (95% CI) c  OR add (95% CI) c 
rs10494836 
C>T 456/353/54 520/322/53 0.267/0.239 1.27(1.04,1.55) 1.13(0.75,1.71) 1.16(0.99,1.36) 
511/332/58 550/336/67 0.249/0.247 1.07(0.87,1.31) 0.93(0.63,1.38) 1.01(0.87,1.18) 
Combined 967/685/112 1070/658/120 0.258/0.243 1.17(1.01,1.35) 1.02(0.77,1.36) 1.08(0.97,1.21) 
rs10857748 
T>C 521/315/40 498/335/62 0.225/0.256 0.91(0.74,1.11) 0.63(0.41,0.96) 0.85(0.73,0.99) 
517/329/52 545/360/52 0.241/0.242 0.96(0.78,1.17) 1.09(0.71,1.67) 0.99(0.85,1.17) 
Combined 1038/644/92 1043/695/114 0.233/0.249 0.94(0.81,1.08) 0.83(0.62,1.08) 0.93(0.83,1.04) 
rs7963551 
A>C 597/241/26 581/270/41 0.169/0.197 0.91(0.74,1.13) 0.59(0.35,0.99) 0.85(0.72,1.01) 
587/201/82 603/249/96 0.177/0.202 0.82(0.66,1.01) 0.67(0.39,1.13) 0.81(0.68,0.97) 
Combined 1188/443/108 1185/519/137 0.173/0.200 0.87(0.75,1.01) 0.65(0.45,0.94) 0.84(0.75,0.95) 
SNP Stage  Case a  Control a  Minor allele frequency b (case/control)   OR het (95% CI) c  OR hom (95% CI) c  OR add (95% CI) c 
rs10494836 
C>T 456/353/54 520/322/53 0.267/0.239 1.27(1.04,1.55) 1.13(0.75,1.71) 1.16(0.99,1.36) 
511/332/58 550/336/67 0.249/0.247 1.07(0.87,1.31) 0.93(0.63,1.38) 1.01(0.87,1.18) 
Combined 967/685/112 1070/658/120 0.258/0.243 1.17(1.01,1.35) 1.02(0.77,1.36) 1.08(0.97,1.21) 
rs10857748 
T>C 521/315/40 498/335/62 0.225/0.256 0.91(0.74,1.11) 0.63(0.41,0.96) 0.85(0.73,0.99) 
517/329/52 545/360/52 0.241/0.242 0.96(0.78,1.17) 1.09(0.71,1.67) 0.99(0.85,1.17) 
Combined 1038/644/92 1043/695/114 0.233/0.249 0.94(0.81,1.08) 0.83(0.62,1.08) 0.93(0.83,1.04) 
rs7963551 
A>C 597/241/26 581/270/41 0.169/0.197 0.91(0.74,1.13) 0.59(0.35,0.99) 0.85(0.72,1.01) 
587/201/82 603/249/96 0.177/0.202 0.82(0.66,1.01) 0.67(0.39,1.13) 0.81(0.68,0.97) 
Combined 1188/443/108 1185/519/137 0.173/0.200 0.87(0.75,1.01) 0.65(0.45,0.94) 0.84(0.75,0.95) 

a Major homozygote/heterozygote/rare homozygote between cases and controls.

b Minor allele frequency between cases and controls.

c Logistic regression with adjustment for age, age at menarche and menopausal status were used to test associations in codominant (OR het : heterozygote versus major homozygote; OR hom : minor homozygote versus major homozygote) and additive (OR add : minor homozygote versus heterozygote versus major homozygote) models.

As shown in Table III , stratification analysis showed that the protective effect of rs7963551 was more prominent in women with early age at first live birth (OR = 0.71, 95% CI = 0.58–0.87, P for heterogeneity = 0.02). Furthermore, a significant interaction between rs7963551 and age at first live birth was detected on the risk of breast cancer ( P for interaction = 0.04) ( Table IV ). Among women with early age at first live birth (<25 years old), those with CC genotype had a 62% reduced risk of breast cancer (OR = 0.38, 95% CI = 0.24–0.59).

Table III.

Stratification analysis of the association between rs7963551 of RAD52 and risk of breast cancer

Characteristics Case Control  OR (95%CI) a Pheterogeneity 
AA (%) AC (%) CC (%) AA (%) AC (%) CC (%) 
Age 
 <50 593(68.32) 245(28.23) 30(3.46) 623(63.57) 313(31.94) 44(4.49) 0.84(0.71,1.00) 0.85 
 ≥50 591(68.40) 250(28.94) 23(2.66) 554(65.10) 262(30.79) 35(4.11) 0.86(0.72,1.03) 
Menopausal status 
 Premenopausal 563(68.08) 240(29.02) 24(2.90) 618(64.44) 299(31.18) 42(4.38) 0.86(0.72,1.03) 1.00 
 Postmenopausal b 504(68.20) 210(28.42) 25(3.38) 533(64.76) 255(30.98) 35(4.25) 0.86(0.72,1.03) 
Age at menarche 
 <16 687(69.32) 273(27.55) 31(3.13) 463(64.04) 226(31.26) 34(4.70) 0.83(0.69,0.99) 0.64 
 ≥16 476(67.23) 210(29.66) 22(3.11) 709(64.51) 345(31.39) 45(4.10) 0.88(0.74,1.04) 
Age at first live birth 
 <25 423(70.62) 160(26.71) 16(2.67) 564(62.60) 298(33.07) 39(4.33) 0.71(0.58,0.87) 0.02 
 ≥25 693(67.35) 302(29.35) 34(3.30) 585(66.63) 254(28.93) 39(4.44) 0.96(0.81,1.13) 
Estrogen receptor status 
 ER+ 548(71.08) 201(26.07) 22(2.85)    0.76(0.65,0.90) 0.22 
 ER– 414(66.56) 190(30.55) 18(2.89)    0.88(0.74,1.04) 
Progesterone receptor status 
 PR+ 549(70.03) 214(27.30) 21(2.68)    0.80(0.68,0.94) 0.68 
 PR– 413(67.93) 177(29.11) 18(2.96)    0.84(0.71,0.99) 
Characteristics Case Control  OR (95%CI) a Pheterogeneity 
AA (%) AC (%) CC (%) AA (%) AC (%) CC (%) 
Age 
 <50 593(68.32) 245(28.23) 30(3.46) 623(63.57) 313(31.94) 44(4.49) 0.84(0.71,1.00) 0.85 
 ≥50 591(68.40) 250(28.94) 23(2.66) 554(65.10) 262(30.79) 35(4.11) 0.86(0.72,1.03) 
Menopausal status 
 Premenopausal 563(68.08) 240(29.02) 24(2.90) 618(64.44) 299(31.18) 42(4.38) 0.86(0.72,1.03) 1.00 
 Postmenopausal b 504(68.20) 210(28.42) 25(3.38) 533(64.76) 255(30.98) 35(4.25) 0.86(0.72,1.03) 
Age at menarche 
 <16 687(69.32) 273(27.55) 31(3.13) 463(64.04) 226(31.26) 34(4.70) 0.83(0.69,0.99) 0.64 
 ≥16 476(67.23) 210(29.66) 22(3.11) 709(64.51) 345(31.39) 45(4.10) 0.88(0.74,1.04) 
Age at first live birth 
 <25 423(70.62) 160(26.71) 16(2.67) 564(62.60) 298(33.07) 39(4.33) 0.71(0.58,0.87) 0.02 
 ≥25 693(67.35) 302(29.35) 34(3.30) 585(66.63) 254(28.93) 39(4.44) 0.96(0.81,1.13) 
Estrogen receptor status 
 ER+ 548(71.08) 201(26.07) 22(2.85)    0.76(0.65,0.90) 0.22 
 ER– 414(66.56) 190(30.55) 18(2.89)    0.88(0.74,1.04) 
Progesterone receptor status 
 PR+ 549(70.03) 214(27.30) 21(2.68)    0.80(0.68,0.94) 0.68 
 PR– 413(67.93) 177(29.11) 18(2.96)    0.84(0.71,0.99) 

a Derived from additive model with adjustment for age, age at menarche and menopausal status where appropriate.

b Postmenopausal status for natural menopause.

Table IV.

Interaction between age at first live birth and rs7963551 of RAD52 on breast cancer risk

Age at first live birth Genotype Case Control  OR (95% CI) a 
≥25 AA 696 587 1.00 
≥25 AC 269 231 1.00 (0.81,1.25) 
≥25 CC 69 63 0.93 (0.64,1.35) 
<25 AA 424 570 0.66 (0.55,0.78) 
<25 AC 146 269 0.48 (0.38,0.61) 
<25 CC 31 69 0.38 (0.24,0.59) 
P for multiplicative interaction  0.04 
Age at first live birth Genotype Case Control  OR (95% CI) a 
≥25 AA 696 587 1.00 
≥25 AC 269 231 1.00 (0.81,1.25) 
≥25 CC 69 63 0.93 (0.64,1.35) 
<25 AA 424 570 0.66 (0.55,0.78) 
<25 AC 146 269 0.48 (0.38,0.61) 
<25 CC 31 69 0.38 (0.24,0.59) 
P for multiplicative interaction  0.04 

a Adjusted by age, age at menarche and menopausal status.

As predicted using RNAhybrid ( 36 ), hsa-let-7 has a lower minimum free energy (MFE) with C allele (|MFE|=10.4 kcal/mol) of rs7963551 in RAD52 than that with A allele (|MFE|=11.1 kcal/mol). Thus, we hypothesized the C allele might lead to an increased expression of RAD52 resulted from reduced miRNA repression. To test this hypothesis, we constructed the plasmids containing rs7963551 A allele or C allele to determine whether this polymorphism could affect gene expression. The transcription activity of reporter gene with rs7963551 C allele was significantly enhanced as compared with A allele when we cotransfected chemically synthesized mature hsa-let-7 into MCF-7 cell line (0.912 versus 0.714, P = 5.19×10 –3 ). Similar effects were also observed in 293T (0.989 versus 0.911, P = 2.95×10 –2 ) and Hela (0.986 versus 0.935, P = 2.21×10 –2 ) cell lines ( Figure 1 ).

Fig. 1.

In vitro target binding assays for rs7963551 A/C in 293T, Hela and MCF-7 cell lines. Each transfection was performed with pRL-SV40 plasmids as normalized controls. The mean fold change ± SD for plasmid with different alleles are shown after normalized by control plasmid in parallel transfection experiments. RAD52 3′-UTR luciferase reporter plasmids (A allele or C allele) were cotransfected with chemically synthesized mature hsa-let-7 in 293T, Hela and MCF-7 cell lines. The P values are 2.95×10 –2 , 2.21×10 –2 and 5.19×10 –3 for 293T, Hela and MCF-7 cell lines, respectively.

Fig. 1.

In vitro target binding assays for rs7963551 A/C in 293T, Hela and MCF-7 cell lines. Each transfection was performed with pRL-SV40 plasmids as normalized controls. The mean fold change ± SD for plasmid with different alleles are shown after normalized by control plasmid in parallel transfection experiments. RAD52 3′-UTR luciferase reporter plasmids (A allele or C allele) were cotransfected with chemically synthesized mature hsa-let-7 in 293T, Hela and MCF-7 cell lines. The P values are 2.95×10 –2 , 2.21×10 –2 and 5.19×10 –3 for 293T, Hela and MCF-7 cell lines, respectively.

Discussion

miRNAs function as translational repressors of protein-coding genes through binding to target sites in the 3′-UTRs of mRNAs. Emerging evidence has suggested that genetic variants in the sequence of target sites may affect the miRNA regulation to target gene expression and consequently modify the development of diseases. For instance, the variant C allele of ESR1 rs2747648 enhanced the binding ability of miR-453 and leaded to an increased miRNA-mediated ESR1 repression ( 37 ). A SNP (rs16917496) in miR-502-binding site altered the SET8 expression and contributed to the early age of breast cancer onset ( 19 ). Allelic variation of rs1434536 might destroy the capacity of miR-125b to regulate BMPR1B expression and this variant was in linkage disequilibrium (LD) with two breast cancer risk-related SNPs (rs1970801 and rs11097457) ( 38 ). In this study, we systematically screened SNPs in target sites of breast cancer-related miRNAs and identified a new variant (rs7963551) associated with breast cancer risk in Chinese women. The variant allele (C) of rs7963551 may increase the expression of RAD52 through weakening the binding capacity of hsa-let-7 to target site in the 3′-UTR of RAD52 .

RAD52 protein binds to the single-stranded DNA ends and participates in the intermediate stage following the formation of double-strand breaks during early steps of recombination. It is essential for the DNA double-stranded break repair and meiotic and mitotic recombinations that are the crucial processes for the maintenance of the genetic integrity of living ( 39 ). Genetic variations in DNA repair genes might change the DNA repair capacity that consequently influences the genomic stability even cancer risk. For example, Danoy et al . ( 40 ) found one SNP (rs11571424) in RAD52 gene might contribute to both lung and head and neck cancer. Lee et al . ( 41 ) showed RAD52 C2259T (rs11226) and ERCC1 C354T (rs11615) were associated with the ER/PR negative breast cancer cases. Another study detected the loss of heterozygosity in the region harboring RAD52 in 16% breast tumors, which might support the importance of RAD52 in the development of breast cancer ( 42 ).

Human let-7 family members were reported to be downregulated in several cancers acting as a tumor suppressor by regulating multiple oncogenes, such as MYC , H-RAS and HMGA2 ( 22 , 43–45 ). According to in silico analysis using RNAhybrid database, let-7 is predicted to strongly bind with the target site of RAD52 harboring A allele of rs7963551. Luciferase assay indicated that the transcription activity of reporter gene with rs7963551 C allele was significantly increased than that with A allele. The upregulated level of RAD52 might enhance the DNA repair capacity and result in an inhibition to tumorigenesis. This result was consistent with the association results that the C allele of rs7963551 was associated with a decreased risk of breast cancer. Further functional studies are required to uncover the exact mechanism of this variant.

Reproductive factors including age at first live birth have been established as breast cancer risk-related factors due to different exposure of endogenous hormones ( 46 ). First full-term pregnancy can change long-term hormonal levels, such as prolactin reduction, estrogen reduction and sex hormone-binding globulin elevation, providing a further protection against breast cancer ( 47 ). Moreover, early age at first pregnancy may strengthen the protective effect and induce a refractory state of mammary gland against carcinogenesis in humans ( 48 , 49 ). During pregnancy, breast epithelial cells rapidly proliferated accompanying a mass of DNA replication response to the estrogen surges. Base mispairing spontaneously emerge in this process more frequently and thus required stronger DNA repair capacity. In our study, a significant interaction was observed between rs7963551 and age at first live birth, suggesting that genetic variations in DNA repair genes may modify the process of breast carcinogenesis mediated by endogenous hormone.

In summary, this is the first systemic evaluation for association of the polymorphisms in miRNA-binding sites with breast cancer risk. Our results indicate that the variant rs7963551 in the hsa-let-7 binding site may modify breast cancer risk through regulating the expression of RAD52 . The findings further highlight that genetic variations in miRNA-binding sites may play an important role in the development of breast cancer.

Funding

National Natural Science Foundation of China (81071715; 81102179); Key Project of the National Natural Science Foundation of China (81230067); Key Grant of Natural Science Foundation of Jiangsu Higher Education Institutions (09KJA330001); Key Project of the Jiangsu Natural Science Foundation (BK2011028); The Program for Changjiang Scholars and Innovative Research Team in University (IRT0631), The Young Talents Support Program from the Organization Department of the CPC Central Committee.

Acknowledgements

We thank the study participants and research staff for their contributions and commitment to this study.

Conflict of Interest Statement: None declared.

References

1.
Lichtenstein
P
et al
.
2000
Environmental and heritable factors in the causation of cancer–analyses of cohorts of twins from Sweden, Denmark, and Finland
.
N. Engl. J. Med.
  ,
343
,
78
85
.
2.
Peto
J
et al
.
2000
High constant incidence in twins and other relatives of women with breast cancer
Nat. Genet.
  ,
26
,
411
414
.
3.
Zhang
B
et al
.
2011
Genetic variants associated with breast-cancer risk: comprehensive research synopsis, meta-analysis, and epidemiological evidence
Lancet Oncol.
  ,
12
,
477
488
.
4.
Varghese
J.S
et al
.
2010
Genome-wide association studies in common cancers–what have we learnt?
Curr. Opin. Genet. Dev.
  ,
20
,
201
209
.
5.
Ghoussaini
M
et al
.
2012
Genome-wide association analysis identifies three new breast cancer susceptibility loci
Nat. Genet.
  ,
44
,
312
318
.
6.
Meister
G
et al
.
2004
Mechanisms of gene silencing by double-stranded RNA
Nature
  ,
431
,
343
349
.
7.
Esquela-Kerscher
A
et al
.
2006
Oncomirs - microRNAs with a role in cancer
Nat. Rev. Cancer
  ,
6
,
259
269
.
8.
Spizzo
R
et al
.
2009
SnapShot: microRNAs in cancer
Cell
  ,
137
,
586
e1
.
9.
Gregory
P.A
et al
.
2008
The miR-200 family and miR-205 regulate epithelial to mesenchymal transition by targeting ZEB1 and SIP1
Nat. Cell Biol.
  ,
10
,
593
601
.
10.
Song
B
et al
.
2010
MicroRNA-21 regulates breast cancer invasion partly by targeting tissue inhibitor of metalloproteinase 3 expression
J. Exp. Clin. Cancer Res.
  ,
29
,
29
.
11.
Ryan
B.M
et al
.
2010
Genetic variation in microRNA networks: the implications or cancer research
Nat. Rev. Cancer
  ,
10
,
389
402
.
12.
Duan
R
et al
.
2007
Single nucleotide polymorphism associated with mature miR-125a alters the processing of pri-miRNA
Hum. Mol. Genet.
  ,
16
,
1124
1131
.
13.
Hu
Z
et al
.
2008
Genetic variants of miRNA sequences and non-small cell lung cancer survival
J. Clin. Invest.
  ,
118
,
2600
2608
.
14.
Hu
Z
et al
.
2009
Common genetic variants in pre-microRNAs were associated with increased risk of breast cancer in Chinese women
Hum. Mutat.
  ,
30
,
79
84
.
15.
Xu
J
et al
.
2009
Functional variant in microRNA-196a2 contributes to the susceptibility of congenital heart disease in a Chinese population
Hum. Mutat.
  ,
30
,
1231
1236
.
16.
Xing
J
et al
.
2012
Genetic polymorphisms in pre-microRNA genes as prognostic markers of colorectal cancer
Cancer Epidemiol. Biomarkers Prev.
  ,
21
,
217
227
.
17.
Yu
Z
et al
.
2007
Aberrant allele frequencies of the SNPs located in microRNA target sites are potentially associated with human cancers
Nucleic Acids Res.
  ,
35
,
4535
4541
.
18.
Xiong
F
et al
.
2011
Genetic variation in an miRNA-1827 binding site in MYCL1 alters susceptibility to small-cell lung cancer
Cancer Res.
  ,
71
,
5175
5181
.
19.
Song
F
et al
.
2009
An miR-502-binding site single-nucleotide polymorphism in the 3’-untranslated region of the SET8 gene is associated with early age of breast cancer onset
Clin. Cancer Res.
  ,
15
,
6292
6300
.
20.
Zhang
W
et al
.
2011
A let-7 microRNA-binding site polymorphism in 3’-untranslated region of KRAS gene predicts response in wild-type KRAS patients with metastatic colorectal cancer treated with cetuximab monotherapy
Ann. Oncol.
  ,
22
,
104
109
.
21.
Jiang
Y
et al
.
2011
Genetic variants at 1p11.2 and breast cancer risk: a two-stage study in Chinese women
PLoS ONE
  ,
6
,
e21563
.
22.
Dai
J
et al
.
2012
Breast cancer risk assessment with five independent genetic variants and two risk factors in Chinese women
Breast Cancer Res.
  ,
14
,
R17
.
23.
Yu
F
et al
.
2007
let-7 regulates self renewal and tumorigenicity of breast cancer cells
Cell
  ,
131
,
1109
1123
.
24.
Ma
L
et al
.
2010
miR-9, a MYC/MYCN-activated microRNA, regulates E-cadherin and cancer metastasis
Nat. Cell Biol.
  ,
12
,
247
256
.
25.
Iorio
M.V
et al
.
2005
MicroRNA gene expression deregulation in human breast cancer
Cancer Res.
  ,
65
,
7065
7070
.
26.
Yu
Z
et al
.
2008
A cyclin D1/microRNA 17/20 regulatory feedback loop in control of breast cancer cell proliferation
J. Cell Biol.
  ,
182
,
509
517
.
27.
Yan
L.X
et al
.
2008
MicroRNA miR-21 overexpression in human breast cancer is associated with advanced clinical stage, lymph node metastasis and patient poor prognosis
RNA
  ,
14
,
2348
2360
.
28.
Valastyan
S
et al
.
2009
A pleiotropically acting microRNA, miR-31, inhibits breast cancer metastasis
Cell
  ,
137
,
1032
1046
.
29.
Kato
M
et al
.
2009
The mir-34 microRNA is required for the DNA damage response in vivo in C. elegans and in vitro in human breast cancer cells
Oncogene
  ,
28
,
2419
2424
.
30.
Varambally
S
et al
.
2008
Genomic loss of microRNA-101 leads to overexpression of histone methyltransferase EZH2 in cancer
Science
  ,
322
,
1695
1699
.
31.
Tavazoie
S.F
et al
.
2008
Endogenous human microRNAs that suppress breast cancer metastasis
Nature
  ,
451
,
147
152
.
32.
Kondo
N
et al
.
2008
miR-206 Expression is down-regulated in estrogen receptor alpha-positive human breast cancer
Cancer Res.
  ,
68
,
5004
5008
.
33.
Di Leva
G
et al
.
2010
MicroRNA cluster 221-222 and estrogen receptor alpha interactions in breast cancer
J. Natl. Cancer Inst.
  ,
102
,
706
721
.
34.
Huang
Q
et al
.
2008
The microRNAs miR-373 and miR-520c promote tumour invasion and metastasis
Nat. Cell Biol.
  ,
10
,
202
210
.
35.
Hiard
S
et al
.
2010
Patrocles: a database of polymorphic miRNA-mediated gene regulation in vertebrates
Nucleic Acids Res.
  ,
38
,
D640
651
.
36.
Kruger
J
et al
.
2006
RNAhybrid: microRNA target prediction easy, fast and flexible
Nucleic Acids Res.
  ,
34
,(Web Server issue)
W451
454
.
37.
Tchatchou
S
et al
.
2009
A variant affecting a putative miRNA target site in estrogen receptor (ESR) 1 is associated with breast cancer risk in premenopausal women
Carcinogenesis
  ,
30
,
59
64
.
38.
Saetrom
P
et al
.
2009
A risk variant in an miR-125b binding site in BMPR1B is associated with breast cancer pathogenesis
Cancer Res.
  ,
69
,
7459
7465
.
39.
Lao
J.P
et al
.
2008
Rad52 promotes postinvasion steps of meiotic double-strand-break repair
Mol. Cell
  ,
29
,
517
524
.
40.
Danoy
P
et al
.
2008
Variants in DNA double-strand break repair and DNA damage-response genes and susceptibility to lung and head and neck cancers
Int. J. Cancer
  ,
123
,
457
463
.
41.
Lee
K.M
et al
.
2005
Genetic polymorphisms of selected DNA repair genes, estrogen and progesterone receptor status, and breast cancer risk
Clin. Cancer Res.
  ,
11
,
4620
4626
.
42.
Park
M.S
et al
.
1996
Physical interaction between human RAD52 and RPA is required for homologous recombination in mammalian cells
J. Biol. Chem.
  ,
271
,
18996
19000
.
43.
Johnson
C.D
et al
.
2007
The let-7 microRNA represses cell proliferation pathways in human cells
Cancer Res.
  ,
67
,
7713
7722
.
44.
Sampson
V.B
et al
.
2007
MicroRNA let-7a down-regulates MYC and reverts MYC-induced growth in Burkitt lymphoma cells
Cancer Res.
  ,
67
,
9762
9770
.
45.
Lee
Y.S
et al
.
2007
The tumor suppressor microRNA let-7 represses the HMGA2 oncogene
Genes Dev.
  ,
21
,
1025
1030
.
46.
Warren
A.S
et al
.
Reproductive factors and histologic subtype in relation to mortality after a breast cancer diagnosis
Breast Cancer Res. Treat
  ,
130
,
975
980
.
47.
Parsa
P
et al
.
2009
Effects of reproductive factors on risk of breast cancer: a literature review
Asian Pac. J. Cancer Prev.
  ,
10
,
545
550
.
48.
Rajkumar
L
et al
.
2001
Short-term exposure to pregnancy levels of estrogen prevents mammary carcinogenesis
Proc. Natl. Acad. Sci. U.S.A.
  ,
98
,
11755
11759
.
49.
MacMahon
B
et al
.
1970
Age at first birth and breast cancer risk
Bull. World Health Organ.
  ,
43
,
209
221
.

Abbreviations:

    Abbreviations:
  • CI

    confidence interval

  • LD

    linkage disequilibrium

  • OR

    odds ratio

  • SNP

    single nucleotide polymorphisms

  • UTR

    untranslated regions