Background: Vitamin D is hypothesized to lower the risk of breast cancer by inhibiting cell proliferation via the nuclear vitamin D receptor (VDR). Two common single nucleotide polymorphisms (SNP) in the VDR gene (VDR), rs1544410 (BsmI), and rs2228570 (FokI), have been inconsistently associated with breast cancer risk. Increased risk has been reported for the FokI ff genotype, which encodes a less transcriptionally active isoform of VDR, and reduced risk has been reported for the BsmI BB genotype, a SNP in strong linkage disequilibrium with a 3′-untranslated region, which may influence VDR mRNA stability.

Methods: We pooled data from 6 prospective studies in the National Cancer Institute Breast and Prostate Cancer Cohort Consortium to examine associations between these SNPs and breast cancer among >6,300 cases and 8,100 controls for each SNP using conditional logistic regression.

Results: The odds ratio (OR) for the rs2228570 (FokI) ff versus FF genotype in the overall population was statistically significantly elevated [OR, 1.16; 95% confidence interval (95% CI), 1.04-1.28] but was weaker once data from the cohort with previously published positive findings were removed (OR, 1.10; 95% CI, 0.98-1.24). No association was noted between rs1544410 (BsmI) BB and breast cancer risk overall (OR, 0.98; 95% CI, 0.89-1.09), but the BB genotype was associated with a significantly lower risk of advanced breast cancer (OR, 0.74; 95% CI, 0.60-0.92).

Conclusions: Although the evidence for independent contributions of these variants to breast cancer susceptibility remains equivocal, future large studies should integrate genetic variation in VDR with biomarkers of vitamin D status. (Cancer Epidemiol Biomarkers Prev 2009;18(1):297–305)

The geographic gradient in breast cancer incidence in North America suggests the possibility that sunlight and vitamin D may reduce breast cancer risk (1). Higher circulating 25-hydroxyvitamin D [25(OH)D] levels, which reflect vitamin D status from dietary intake, vitamin supplements, and sun exposure combined, have been associated with lower risk of breast cancer in a retrospective (2) and one (3) of two (3, 4) prospective studies. 25(OH)D can be converted to its active form, 1,25-hydroxyvitamin D in breast tissue, where it then binds to the vitamin D receptor (VDR), a nuclear transcription factor that regulates the expression of multiple genes, including some responsible for cell cycle regulation, differentiation, and apoptosis (5). The receptor is present in most cell types, including normal and neoplastic breast tissue (6). In the MMTV-neu transgenic mouse model of breast cancer, animals totally lacking the VDR gene exhibited abnormal mammary duct morphology and had reduced survival, whereas, in animals heterozygous for VDR, the incidence of mammary tumors was increased and the latency was shortened (7).

Two common genetic polymorphisms in the VDR, rs2228570 (FokI) and rs1544410 (BsmI), have been inconsistently associated with breast cancer risk in previous studies. In a large nested case-control study, the rs2228570 (FokI) ff genotype was associated with a 34% higher breast cancer risk (95% CI, 1.06-1.69; ref. 8), but other studies, mostly smaller in size (9-14), have not reported similar associations. The presence of the rs2228570 (FokI) f allele in the 5′-promoter region of the VDR results in production of a VDR protein that is three amino acids longer and less effective as a transcriptional activator (15).

Initial epidemiologic studies of the rs1544410 (BsmI) polymorphism suggested a potentially important role in breast cancer of variants in this single nucleotide polymorphism (SNP), especially for more aggressive forms of the disease (9, 20). The intronic rs1544410 (BsmI) SNP is located at the 3′-end of the VDR gene. This SNP is in strong linkage disequilibrium with a poly(A) microsatellite repeat in the 3′-untranslated region (12, 17, 18), which may influence VDR mRNA stability. Six (2, 9, 16, 19-21) of 11 (2, 8, 9, 11, 12, 16, 19-23) studies have reported higher breast cancer risk associated with the rs1544410 (BsmI) bb genotype.

There are several potential explanations for inconsistencies in findings for these common SNPs. Many of the individual studies have been based on <200 cases (9, 11, 19, 20, 22, 23). Allelic frequencies of the VDR polymorphisms, particularly BsmI, vary by ethnicity (16, 17, 20, 21), and few studies have been large enough to examine risk by ethnicity with precision. Associations that vary by tumor characteristics might also be missed in studies combining all cases. Effect modification by environmental factors, including calcium intake (12), which may influence vitamin D metabolism, has also been suggested. We pooled data from six cohorts collaborating in the National Cancer Institute Breast and Prostate Cancer Cohort Consortium (24) to determine in a very large series of cases and controls if these two widely studied SNPs in the VDR gene, rs2228570 (FokI) and rs1544410 (BsmI), contribute to susceptibility to breast cancer.

Study Population

The Breast and Prostate Cancer Cohort Consortium, particularly the breast cancer component (25), has been described in detail elsewhere (24). Briefly, the consortium includes large prospective cohorts (or consortia of smaller cohorts) assembled in the United States and Europe that have DNA for genotyping and extensive questionnaire data on all participants. This analysis included 6,473 cases and 8,397 controls for rs2228570 (FokI) and 6,355 breast cancer cases and 8,149 controls for rs1544410 (BsmI) from six cohorts that had genotyped these SNPs on the VDR gene. Cohorts included the American Cancer Society Cancer Prevention Study II (CPS-II) Nutrition Cohort, the European Prospective Investigation into Cancer and Nutrition (EPIC); the Harvard Nurses' Health Study (NHS); the Hawaii-Los Angeles Multiethnic Cohort (MEC); the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial (PLCO) cohort, and the Women's Health Study (WHS). With the exception of MEC, most women in these cohorts are Caucasian. The MEC includes U.S. Caucasians, African Americans, Latinos, Japanese Americans, and Native Hawaiians. Each study has been approved by its respective institutional review board.

Breast cancer cases were identified in each cohort primarily by self-report and subsequently verified by medical records or linkage with population-based tumor registries. Controls were individually or frequency matched to cases by age at entry and, depending on the cohort, additional criteria, as described below. Information on estrogen receptor (ER) and progesterone receptor (PR) status and on localized versus metastatic cancers was obtained for most cohorts. Information on breast cancer risk factors was obtained by questionnaire before cancer diagnosis in all cohorts. Diet was assessed using validated food frequency questionnaires; dietary and total calcium intake estimates, adjusted for calories by the residual method (26), were calculated using nutrient databases and analytic programs specific to each cohort's food frequency questionnaire. Information on vitamin D status or intake was not available for all cohorts. Blood samples were collected before diagnosis in all cohorts, except for MEC and CPS-II, in which most were collected after diagnosis.

Two VDR SNPs, rs2228570 (FokI) F/f and rs1544410 (BsmI) b/B, were genotyped in the breast cancer cases and controls. For CPS-II and NHS, genotyping was conducted as described previously (8, 12). For EPIC, MEC, PLCO, and WHS, genotyping was done in four laboratories (IARC; University of Southern California; Core Genotyping Facility, National Cancer Institute; and Harvard School of Public Health, respectively). A fluorescent 5-endonuclease assay and the ABI-PRISM 7900 for sequence detection (TaqMan) were used, with an assay success rate of >97% in each laboratory and a replication rate of >99% for the blinded duplicates inserted within each study's samples (5-10% depending on study). Assay characteristics for the two VDR SNPs are available on a public Web site.23

No interlaboratory variation in genotyping among IARC/University of Southern California/Core Genotyping Facility/Harvard School of Public Health, assessed by genotyping a designated set of 94 samples from the Coriell Biorepository (27) in each laboratory, was noted. We used a χ2 test to assess whether the rs2228570 (FokI) and rs1544410 (BsmI) genotype distributions were in Hardy-Weinberg equilibrium (HWE) within populations defined by cohort and ethnicity.

Statistical Analyses

We used conditional logistic regression to estimate odds ratios (OR) and 95% confidence intervals (95% CI) for disease by SNP genotype using PROC PHREG in SAS version 9.1. The heterozygous and homozygous carriers of the minor allele were each compared with the homozygous carriers of the more prevalent allele, which leads to a 2 df test for association between SNP genotypes and risk of disease. Ptrend values were calculated assuming a log-additive genetic model with 1 df. As noted, controls were individually or frequency matched to cases on age at entry and, depending on the cohort, additional characteristics, which could include study center, race/ethnicity, menopausal status, exogenous hormone use, phase of menstrual cycle, date of blood collection, time of day at blood collection, and fasting status at blood collection. Because PLCO did not match on race/ethnicity, this variable was included as a covariate in all models.

We considered conditional logistic regression models both without adjustment and with adjustment for known breast cancer risk factors, including age at menarche (≤12, 13-14, ≥15 years), parity (ever/never full-term pregnancy), menopausal status at blood draw (pre/post/unknown or missing), use of postmenopausal hormone therapy at blood draw (ever/never/unknown or missing), and body mass index (kg/m2 as a continuous variable). Data on other breast cancer risk factors, such as family history, history of benign breast disease, and age at menopause, were not available from all cohorts. Because the results were essentially unchanged with adjustment, we present results from the conditional model controlling only for race/ethnicity. The analyses presented include invasive and in situ breast cancer cases; exclusion of the in situ breast cancers produced similar results.

We examined the heterogeneity of associations across the cohorts and across racial/ethnic groups using the Q statistic (28). Logistic regression models to examine breast cancer associations with genotype by specific hormone receptor status included only cases classified as ER+/PR+, ER−/PR−, and their matched controls. We similarly examined risk of breast cancer by genotype among women with localized and advanced disease. For all cohorts, advanced disease was defined as metastases to distant organs (“distant” by Surveillance, Epidemiology and End Results Program staging) or regional metastases to lymph nodes or other adjacent tissues (“regional” by Surveillance, Epidemiology and End Results Program staging). The NHS and PLCO cohorts also included breast tumors >2 cm in diameter without lymph node involvement or other regional spread (American Joint Committee on Cancer stage II) among advanced cases according to American Joint Committee on Cancer staging guidelines. In the various EPIC recruitment centers, advanced tumors were defined as distant metastases only or regional plus distant metastases combined due to different coding practices at the cancer registries. Consequently, we conducted a sensitivity analysis restricting analysis of advanced cases from EPIC centers with >10% of cases in this category. We also examined results stratified by menopausal status at blood draw. To test for heterogeneity by outcome, we conducted case-only analyses using unconditional logistic regression with the tumor characteristic or menopausal status as the dependent variable.

Dietary and total calcium intakes were combined across cohorts using cohort-specific quintiles. We tested for heterogeneity in genetic effects across extremes of dietary and total calcium intake by comparing a model containing indicator variables for heterozygous and homozygous minor allele genotypes, two categories of increased calcium intake (the three middle quintiles combined and the top quintile), and their product interaction terms to a model with only the genotype and calcium intake variables (a 4 df test).

Descriptive characteristics of study participants are provided for each cohort in Table 1. The majority of women were postmenopausal and White, except for the MEC, in which there were roughly equal numbers of White, Hispanic, African American, and Japanese American women. The genotype distribution in controls of the rs2228570 (FokI) polymorphism obeyed HWE in all cohorts combined (P = 0.58) and in each of the five racial/ethnic subgroups within the MEC. In controls from all the cohorts, the distribution of genotypes for the rs1544410 (BsmI) polymorphism deviated from that expected by HWE (P = 0.0004). However, the minor allele frequency for this SNP varied noticeably among the White, Hispanic, African American, Japanese American, and Hawaiian controls in the MEC (Table 2). Within each of these racial/ethnic subgroups, genotype distributions respected HWE (P > 0.05 for all).

Table 1.

Descriptive characteristics of breast cancer cases and controls genotyped for the VDR FokI or BsmI SNP, by cohort

CPS-II
EPIC
MEC
NHS
PLCO
WHS
CasesControlsCasesControlsCasesControlsCasesControlsCasesControlsCasesControls
n 499 504 1,677 2,795 1,598 1,952 1,257 1,748 1,073 1,100 685 683 
Race/ethnicity, n (%)             
    White 488 (98) 497 (99) 1,677 (100) 2,795 (100) 399 (25) 437 (22) 1,179 (94) 1641 (94) 975 (91) 991 (90) 654 (95) 653 (96) 
    Hispanic 4 (1) 1 (0)   332 (21) 383 (20) 3 (0) 6 (0) 10 (1) 13 (1) 4 (1) 4 (1) 
    African American 4 (1) 4 (1)   338 (21) 426 (22) 10 (1) 11(1) 45 (4) 43 (4) 5 (1) 5 (1) 
    Asian   422 (26) 419 (21) 3 (0) 9 (1) 36 (3) 44 (4) 7 (1) 6 (1) 
    Hawaiian   107 (7) 287 (15) 6 (1) 3 (0) 
    Other/missing 1 (0)   62 (5) 81 (5) 1 (0) 6 (1) 15 (2) 15 (2) 
Age at diagnosis, mean (SD) 70 (6)  58 (8)  65 (9)  63 (7)  66 (6)  60 (8)  
Body mass index, kg/m2 (mean, SD) 25 (5) 26 (5) 26 (4) 26 (5) 27 (6) 27 (6) 25 (5) 26 (5) 27 (5) 27 (5) 25 (4) 26 (5) 
Menopausal status, n (%)*             
    Premenopausal   411 (25) 779 (28) 174 (11) 319 (16) 259 (21) 326 (19)   148 (22) 145 (21) 
    Postmenopausal 499 (100) 504 (100) 1,123 (67) 1,771 (63) 1,385 (87) 1,600 (82) 869 (69) 1,271 (73) 1,063 (99) 1,091 (99) 435 (64) 408 (60) 
    Unknown/missing   143 (9) 245 (9) 39 (2) 33 (2) 129 (10) 151 (9) 10 (1) 9 (1) 102 (15) 130 (19) 
Age at menarche, n (%)             
    ≤12 232 (46) 226 (45) 568 (34) 981 (35) 845 (53) 961 (49) 638 (51) 844 (48) 213 (20) 213 (20) 388 (57) 354 (52) 
    13-14 215 (43) 234 (46) 784 (47) 1,234 (44) 558 (35) 747 (38) 514 (41) 748 (43) 592 (55) 605 (55) 253 (37) 290 (42) 
    ≥15 46 (9) 40 (8) 257 (15) 496 (18) 166 (10) 226 (12) 97 (8) 145 (8) 268 (25) 279 (25) 44 (6) 39 (6) 
    Unknown/missing 6 (1) 4 (1) 68 (4) 84 (3) 29 (2) 18 (1) 8 (1) 11 (1) 3 (0) 
Age at menopause, n (%)             
    ≤44 100 (20) 110 (22) 127 (11) 235 (13) 419 (20) 554 (35) 184 (21) 283 (22) 273 (26) 288 (26) 70 (16) 88 (22) 
    45-49 100 (20) 139 (28) 253 (23) 481 (127) 354 (26) 432 (27) 243 (28) 363 (29) 224 (21) 263 (24) 129 (30) 123 (30) 
    50-54 226 (45) 195 (39) 419 (37) 663 (37) 441 (32) 451 (28) 395 (45) 551 (43) 413 (39) 395 (36) 180 (41) 142 (35) 
    ≥55 61 (12) 49 (10) 88 (8) 133 (8) 124 (9) 114 (7) 47 (5) 74 (6) 153 (14) 145 (13) 31 (7) 32 (8) 
    Unknown/missing 12 (2) 11 (2) 236 (21) 259 (15) 47 (3) 49 (3)     25 (6) 23 (6) 
Parity, n (%)             
    Nulliparous 45 (9) 43 (9) 214 (13) 355 (13) 230 (14) 218 (11) 95 (8) 119 (7) 113 (11) 93 (8) 106 (15) 94 (14) 
    ≤2 children 170 (34) 138 (27) 927 (55) 1,449 (52) 579 (36) 676 (35) 412 (33) 543 (31) 368 (34) 329 (30) 270 (39) 238 (35) 
    ≥3 children 270 (54) 311 (62) 432 (26) 842 (30) 767 (48) 1,038 (53) 737 (59) 1,077 (62) 592 (55) 676 (61) 309 (45) 351 (51) 
    Unknown/missing 14 (3) 12 (2) 104 (6) 149 (5) 22 (1) 20 (1) 13 (1) 9 (1) 2 (0) 
First-degree family history, n (%)             
    Yes 101 (20) 75 (15)   273 (17) 217 (11) 244 (19) 243 (14) 209 (19) 175 (16) 137 (20) 110 (16) 
    No 388 (78) 409 (81)   1,318 (82) 1,732 (89) 1,013 (81) 1,505 (86) 856 (80) 920 (84) 539 (79) 566 (83) 
    Unknown 10 (2) 20 (4) 1,677 (100) 2,795 (100) 7 (0) 3 (0) 8 (1) 5 (0) 9 (1) 7 (1) 
Postmenopausal hormone therapy, n (%)             
    Never 164 (33) 208 (41) 624 (56) 1,135 (64) 493 (36) 647 (40) 204 (23) 360 (28) 281 (26) 326 (30) 133 (31) 152 (37) 
    Ever 332 (67) 294 (58) 451 (40) 573 (32) 880 (63) 930 (58) 665 (77) 911 (72) 778 (73) 757 (69) 283 (65) 234 (57) 
    Unknown/missing 3 (1) 2 (0) 48 (4) 63 (4) 12 (1) 23 (2) 4 (0) 8 (1) 19 (4) 22 (5) 
ER/PR status, n (%)             
    ER+/PR+ 219 (44)    763 (48)  597 (47)  434 (40)  475 (69)  
    ER−/PR− 27 (5)    216 (14)  160 (13)  69 (6)  83 (12)  
    ER+/PR− 33 (7)    135 (8)  124 (10)  54 (5)  60 (9)  
    ER−/PR+ 3 (1)    37 (2)  33 (3)  4 (0)  17 (2)  
    ER/PR borderline 7 (1)    21 (1)  27 (2)  35 (3)  8 (1)  
    Not available 210 (42)  1,677 (100)  426 (27)  315 (25)  477 (44)  42 (6)  
Stage of breast cancer, n (%)             
    In situ 108 (22)  109 (6)  15 (1)  209 (17)  172 (16)   
    Localized invasive 302 (61)  803 (48)  1,160 (73)  658 (52)  394 (37)  476 (69)  
    Advanced 69 (14)  274 (16)  417 (26)  368 (29)  223 (21)  170 (25)  
    Unknown 20 (4)  491 (29)  6 (0)  22 (2)  284 (26)  39 (6)  
CPS-II
EPIC
MEC
NHS
PLCO
WHS
CasesControlsCasesControlsCasesControlsCasesControlsCasesControlsCasesControls
n 499 504 1,677 2,795 1,598 1,952 1,257 1,748 1,073 1,100 685 683 
Race/ethnicity, n (%)             
    White 488 (98) 497 (99) 1,677 (100) 2,795 (100) 399 (25) 437 (22) 1,179 (94) 1641 (94) 975 (91) 991 (90) 654 (95) 653 (96) 
    Hispanic 4 (1) 1 (0)   332 (21) 383 (20) 3 (0) 6 (0) 10 (1) 13 (1) 4 (1) 4 (1) 
    African American 4 (1) 4 (1)   338 (21) 426 (22) 10 (1) 11(1) 45 (4) 43 (4) 5 (1) 5 (1) 
    Asian   422 (26) 419 (21) 3 (0) 9 (1) 36 (3) 44 (4) 7 (1) 6 (1) 
    Hawaiian   107 (7) 287 (15) 6 (1) 3 (0) 
    Other/missing 1 (0)   62 (5) 81 (5) 1 (0) 6 (1) 15 (2) 15 (2) 
Age at diagnosis, mean (SD) 70 (6)  58 (8)  65 (9)  63 (7)  66 (6)  60 (8)  
Body mass index, kg/m2 (mean, SD) 25 (5) 26 (5) 26 (4) 26 (5) 27 (6) 27 (6) 25 (5) 26 (5) 27 (5) 27 (5) 25 (4) 26 (5) 
Menopausal status, n (%)*             
    Premenopausal   411 (25) 779 (28) 174 (11) 319 (16) 259 (21) 326 (19)   148 (22) 145 (21) 
    Postmenopausal 499 (100) 504 (100) 1,123 (67) 1,771 (63) 1,385 (87) 1,600 (82) 869 (69) 1,271 (73) 1,063 (99) 1,091 (99) 435 (64) 408 (60) 
    Unknown/missing   143 (9) 245 (9) 39 (2) 33 (2) 129 (10) 151 (9) 10 (1) 9 (1) 102 (15) 130 (19) 
Age at menarche, n (%)             
    ≤12 232 (46) 226 (45) 568 (34) 981 (35) 845 (53) 961 (49) 638 (51) 844 (48) 213 (20) 213 (20) 388 (57) 354 (52) 
    13-14 215 (43) 234 (46) 784 (47) 1,234 (44) 558 (35) 747 (38) 514 (41) 748 (43) 592 (55) 605 (55) 253 (37) 290 (42) 
    ≥15 46 (9) 40 (8) 257 (15) 496 (18) 166 (10) 226 (12) 97 (8) 145 (8) 268 (25) 279 (25) 44 (6) 39 (6) 
    Unknown/missing 6 (1) 4 (1) 68 (4) 84 (3) 29 (2) 18 (1) 8 (1) 11 (1) 3 (0) 
Age at menopause, n (%)             
    ≤44 100 (20) 110 (22) 127 (11) 235 (13) 419 (20) 554 (35) 184 (21) 283 (22) 273 (26) 288 (26) 70 (16) 88 (22) 
    45-49 100 (20) 139 (28) 253 (23) 481 (127) 354 (26) 432 (27) 243 (28) 363 (29) 224 (21) 263 (24) 129 (30) 123 (30) 
    50-54 226 (45) 195 (39) 419 (37) 663 (37) 441 (32) 451 (28) 395 (45) 551 (43) 413 (39) 395 (36) 180 (41) 142 (35) 
    ≥55 61 (12) 49 (10) 88 (8) 133 (8) 124 (9) 114 (7) 47 (5) 74 (6) 153 (14) 145 (13) 31 (7) 32 (8) 
    Unknown/missing 12 (2) 11 (2) 236 (21) 259 (15) 47 (3) 49 (3)     25 (6) 23 (6) 
Parity, n (%)             
    Nulliparous 45 (9) 43 (9) 214 (13) 355 (13) 230 (14) 218 (11) 95 (8) 119 (7) 113 (11) 93 (8) 106 (15) 94 (14) 
    ≤2 children 170 (34) 138 (27) 927 (55) 1,449 (52) 579 (36) 676 (35) 412 (33) 543 (31) 368 (34) 329 (30) 270 (39) 238 (35) 
    ≥3 children 270 (54) 311 (62) 432 (26) 842 (30) 767 (48) 1,038 (53) 737 (59) 1,077 (62) 592 (55) 676 (61) 309 (45) 351 (51) 
    Unknown/missing 14 (3) 12 (2) 104 (6) 149 (5) 22 (1) 20 (1) 13 (1) 9 (1) 2 (0) 
First-degree family history, n (%)             
    Yes 101 (20) 75 (15)   273 (17) 217 (11) 244 (19) 243 (14) 209 (19) 175 (16) 137 (20) 110 (16) 
    No 388 (78) 409 (81)   1,318 (82) 1,732 (89) 1,013 (81) 1,505 (86) 856 (80) 920 (84) 539 (79) 566 (83) 
    Unknown 10 (2) 20 (4) 1,677 (100) 2,795 (100) 7 (0) 3 (0) 8 (1) 5 (0) 9 (1) 7 (1) 
Postmenopausal hormone therapy, n (%)             
    Never 164 (33) 208 (41) 624 (56) 1,135 (64) 493 (36) 647 (40) 204 (23) 360 (28) 281 (26) 326 (30) 133 (31) 152 (37) 
    Ever 332 (67) 294 (58) 451 (40) 573 (32) 880 (63) 930 (58) 665 (77) 911 (72) 778 (73) 757 (69) 283 (65) 234 (57) 
    Unknown/missing 3 (1) 2 (0) 48 (4) 63 (4) 12 (1) 23 (2) 4 (0) 8 (1) 19 (4) 22 (5) 
ER/PR status, n (%)             
    ER+/PR+ 219 (44)    763 (48)  597 (47)  434 (40)  475 (69)  
    ER−/PR− 27 (5)    216 (14)  160 (13)  69 (6)  83 (12)  
    ER+/PR− 33 (7)    135 (8)  124 (10)  54 (5)  60 (9)  
    ER−/PR+ 3 (1)    37 (2)  33 (3)  4 (0)  17 (2)  
    ER/PR borderline 7 (1)    21 (1)  27 (2)  35 (3)  8 (1)  
    Not available 210 (42)  1,677 (100)  426 (27)  315 (25)  477 (44)  42 (6)  
Stage of breast cancer, n (%)             
    In situ 108 (22)  109 (6)  15 (1)  209 (17)  172 (16)   
    Localized invasive 302 (61)  803 (48)  1,160 (73)  658 (52)  394 (37)  476 (69)  
    Advanced 69 (14)  274 (16)  417 (26)  368 (29)  223 (21)  170 (25)  
    Unknown 20 (4)  491 (29)  6 (0)  22 (2)  284 (26)  39 (6)  
*

Menopausal status at time of blood donation.

Age at menarche in PLCO; categories for ages 12 to 13 were modeled as ages 13 to 14, and ages 14 to 15 and ≥16 were combined with ≥15.

Among postmenopausal women only.

Table 2.

Association of the rs2228570 (FokI) SNP with breast cancer risk by cohort, overall, and by ethnicity

CohortGenotypeCasesControlsHWE controlsMinor allele frequency controlsOR (95% CI)PtrendPheterogeneity
CPS-II FF 185 178     
 Ff 200 214 0.9 0.38 0.90 (0.68-1.18)   
 ff 73 66   1.05 (0.71-1.56) 0.93  
EPIC FF 643 1,070     
 Ff 754 1,245 0.5 0.37 1.00 (0.88-1.14)   
 ff 224 383   0.97 (0.80-1.18) 0.83  
MEC FF 657 844     
 Ff 668 834 0.44 0.34 1.04 (0.89-1.21)   
 ff 201 223   1.17 (0.93-1.46) 0.22  
NHS FF 422 655     
 Ff 578 765 0.85 0.37 1.16 (0.98-1.36)   
 ff 205 228   1.40 (1.11-1.75) 0.003  
PLCO FF 380 441     
 Ff 505 512 0.69 0.36 1.16 (0.96-1.40)   
 ff 180 141   1.49 (1.15-1.93) 0.003  
WHS FF 225 219     
 Ff 292 288 0.82 0.39 0.99 (0.77-1.26)   
 ff 81 91   0.87 (0.62-1.23) 0.5  
All cohorts, excluding NHS FF 2,090 2,752     
 Ff 2,419 3,093 0.45 0.36 1.03 (0.95-1.11)   
 ff 759 904   1.10 (0.98-1.24) 0.12  
All cohorts FF 2,512 3,407     
 Ff 2,997 3,858 0.45 0.36 1.05 (0.98-1.13)   
 ff 964 1,132   1.16 (1.04-1.28) 0.006 0.03 
         
Ethnic-specific results from multiethnic cohort only         
    Hispanic FF 106 134     
 Ff 146 180 0.8 0.41 1.05 (0.75-1.48)   
 ff 66 64   1.34 (0.87-2.07) 0.22  
    African American FF 197 257     
 Ff 115 135 0.22 0.23 1.13 (0.83-1.56)   
 ff 13 27   0.65 (0.32-1.30) 0.81  
    Japanese American FF 163 173     
 Ff 164 179 0.67 0.35 1.06 (0.78-1.45)   
 ff 74 53   1.63 (1.07-2.49) 0.05  
    Hawaiian FF 39 121     
 Ff 52 126 0.77 0.34 1.02 (0.61-1.71)   
 ff 13 31   1.02 (0.46-2.27) 0.94  
    Caucasian FF 152 159     
 Ff 191 214 0.06 0.37 0.94 (0.69-1.28)   
 ff 35 48   0.80 (0.48-1.33) 0.42 0.31 
         
All cohorts combined         
    Caucasian FF 1,900 2,576     
 Ff 2,406 3,085 0.89 0.38 1.05 (0.97-1.14)   
 ff 774 917   1.15 (1.02-1.28) 0.02 0.008 
CohortGenotypeCasesControlsHWE controlsMinor allele frequency controlsOR (95% CI)PtrendPheterogeneity
CPS-II FF 185 178     
 Ff 200 214 0.9 0.38 0.90 (0.68-1.18)   
 ff 73 66   1.05 (0.71-1.56) 0.93  
EPIC FF 643 1,070     
 Ff 754 1,245 0.5 0.37 1.00 (0.88-1.14)   
 ff 224 383   0.97 (0.80-1.18) 0.83  
MEC FF 657 844     
 Ff 668 834 0.44 0.34 1.04 (0.89-1.21)   
 ff 201 223   1.17 (0.93-1.46) 0.22  
NHS FF 422 655     
 Ff 578 765 0.85 0.37 1.16 (0.98-1.36)   
 ff 205 228   1.40 (1.11-1.75) 0.003  
PLCO FF 380 441     
 Ff 505 512 0.69 0.36 1.16 (0.96-1.40)   
 ff 180 141   1.49 (1.15-1.93) 0.003  
WHS FF 225 219     
 Ff 292 288 0.82 0.39 0.99 (0.77-1.26)   
 ff 81 91   0.87 (0.62-1.23) 0.5  
All cohorts, excluding NHS FF 2,090 2,752     
 Ff 2,419 3,093 0.45 0.36 1.03 (0.95-1.11)   
 ff 759 904   1.10 (0.98-1.24) 0.12  
All cohorts FF 2,512 3,407     
 Ff 2,997 3,858 0.45 0.36 1.05 (0.98-1.13)   
 ff 964 1,132   1.16 (1.04-1.28) 0.006 0.03 
         
Ethnic-specific results from multiethnic cohort only         
    Hispanic FF 106 134     
 Ff 146 180 0.8 0.41 1.05 (0.75-1.48)   
 ff 66 64   1.34 (0.87-2.07) 0.22  
    African American FF 197 257     
 Ff 115 135 0.22 0.23 1.13 (0.83-1.56)   
 ff 13 27   0.65 (0.32-1.30) 0.81  
    Japanese American FF 163 173     
 Ff 164 179 0.67 0.35 1.06 (0.78-1.45)   
 ff 74 53   1.63 (1.07-2.49) 0.05  
    Hawaiian FF 39 121     
 Ff 52 126 0.77 0.34 1.02 (0.61-1.71)   
 ff 13 31   1.02 (0.46-2.27) 0.94  
    Caucasian FF 152 159     
 Ff 191 214 0.06 0.37 0.94 (0.69-1.28)   
 ff 35 48   0.80 (0.48-1.33) 0.42 0.31 
         
All cohorts combined         
    Caucasian FF 1,900 2,576     
 Ff 2,406 3,085 0.89 0.38 1.05 (0.97-1.14)   
 ff 774 917   1.15 (1.02-1.28) 0.02 0.008 

Associations for individual SNPs are displayed in Table 2 by cohort, by race/ethnicity, and overall; ORs are shown by tumor characteristics and menopausal status in Fig. 1. We observed a modest, positive, statistically significant association between the rs2228570 (FokI) ff genotype and relative risk of breast cancer (OR, 1.16; 95% CI, 1.04-1.28; Ptrend = 0.006; Pheterogeneity = 0.03). However, this association was no longer statistically significant (OR, 1.10; 95% CI, 0.98-1.24; Ptrend = 0.12) after exclusion of the NHS, in which a positive association with the rs2228570 (FokI) f polymorphism was reported previously (7). Exclusion of the 616 and 603 in situ breast cancers included in the FokI and BsmI analyses, respectively, produced similar results (OR, 1.15; 95% CI, 1.03-1.28 and OR, 0.98; 95% CI, 0.88-1.09, respectively).

Figure 1.

Risk of breast cancer by tumor characteristics and menopausal status for rs2228570 (FokI) and rs1544410 (BsmI). OR (95% CI) were calculated by conditional logistic regression and compared subjects homozygous for the less common variant (ff and BB, respectively) with subjects homozygous for the more common variant (FF and bb, respectively). Diamond and dashed line, overall OR.

Figure 1.

Risk of breast cancer by tumor characteristics and menopausal status for rs2228570 (FokI) and rs1544410 (BsmI). OR (95% CI) were calculated by conditional logistic regression and compared subjects homozygous for the less common variant (ff and BB, respectively) with subjects homozygous for the more common variant (FF and bb, respectively). Diamond and dashed line, overall OR.

Close modal

In analyses within the racial/ethnic subgroups in the MEC, the association for the FokI ff genotype was highest among Japanese American women (OR, 1.63; 95% CI, 1.07-2.49); however, the Pheterogeneity across racial/ethnic subgroups was not statistically significant (P = 0.31; Table 2). In our complete data set, we observed for the FokI ff genotype a marginally stronger association for localized invasive tumors than advanced tumors (OR, 1.23; 95% CI, 1.08-1.41 and OR, 1.06; 95% CI, 0.86-1.29, respectively) and for ER−/PR− tumors than for ER+/PR+ tumors (OR, 1.35; 95% CI, 0.99-1.86 and OR, 1.20; 95% CI, 1.02-1.41, respectively (Fig. 1). However, the Pheterogeneity values by tumor characteristic were all >0.05).

No association was seen between the rs1544410 (BsmI) SNP and breast cancer in all races combined or analyses confined to Caucasians (Table 3). Japanese American women with the B allele were at lower breast cancer risk, and the Pheterogeneity across racial/ethnic groups was borderline (P = 0.08). In a subanalysis, women of all races with the rs1544410 (BsmI) BB genotype had a statistically significantly lower risk of advanced breast cancer (OR, 0.74; 95% CI, 0.60-0.92; Ptrend = 0.015), which persisted when only Caucasian women were considered (OR, 0.77; 95% CI, 0.61-0.97; Ptrend = 0.045; Fig. 1). These results were identical in a sensitivity analysis that included only EPIC centers with (>10%) advanced cases (regional and distant metastases, combined).

Table 3.

Association of rs1544410 (BsmI) SNP with breast cancer risk by cohort, overall, and by ethnicity

CohortGenotypeCasesControlsHWE controlsMinor allele frequency controlsOR (95% CI)PtrendPheterogeneity
CPS-II bb 142 162     
 Bb 212 200 0.53 0.39 1.22 (0.90-1.66)   
 BB 78 70   1.28 (0.86-1.92) 0.16  
EPIC bb 573 951     
 Bb 767 1,219 0.08 0.4 1.02 (0.89-1.18)   
 BB 256 450   0.93 (0.77-1.13) 0.59  
MEC bb 903 1,051     
 Bb 518 672 0.0009 0.26 0.93 (0.79-1.08)   
 BB 115 158   0.89 (0.68-1.17) 0.26  
NHS bb 407 550     
 Bb 555 723 0.86 0.4 1.01 (0.85-1.20)   
 BB 160 242   0.91 (0.72-1.15) 0.55  
PLCO bb 405 407     
 Bb 468 533 0.41 0.39 0.87 (0.72-1.05)   
 BB 192 157   1.21 (0.94-1.56) 0.43  
WHS bb 201 200     
 Bb 303 298 0.78 0.42 1.01 (0.78-1.31)   
 BB 100 106   0.94 (0.68-1.32) 0.77  
All cohorts bb 2,631 3,321     
 Bb 2,823 3,645 0.0004 0.37 0.98 (0.91-1.06)   
 BB 901 1,183   0.98 (0.89-1.09) 0.66 0.5 
         
Ethnic-specific results from multiethnic cohort only         
    Hispanic bb 184 207     
 Bb 115 141 0.26 0.25 0.91 (0.66-1.26)   
 BB 24 21   1.32 (0.70-2.48) 0.85  
    African American bb 163 217     
 Bb 126 155 0.34 0.29 1.10 (0.80-1.51)   
 BB 27 40   0.91 (0.53-1.57) 0.94  
    Japanese American bb 341 299     
 Bb 71 106 0.86 0.14 0.60 (0.42-0.85)   
 BB   0.52 (0.12-2.26) 0.003  
    Hawaiian bb 65 175     
 Bb 35 86 0.7 0.2 1.23 (0.74-2.04)   
 BB 13   1.05 (0.31-3.53) 0.52  
    Caucasian bb 150 153     
 Bb 171 184 0.08 0.41 1.02 (0.74-1.41)   
 BB 57 79   0.85 (0.56-1.30) 0.55 0.08 
         
All cohorts combined         
    Caucasian bb 1,751 2,271     
 Bb 2,381 3,010 0.14 0.41 1.01 (0.93-1.10)   
 BB 821 1,077   0.99 (0.89-1.11) 0.97 0.62 
CohortGenotypeCasesControlsHWE controlsMinor allele frequency controlsOR (95% CI)PtrendPheterogeneity
CPS-II bb 142 162     
 Bb 212 200 0.53 0.39 1.22 (0.90-1.66)   
 BB 78 70   1.28 (0.86-1.92) 0.16  
EPIC bb 573 951     
 Bb 767 1,219 0.08 0.4 1.02 (0.89-1.18)   
 BB 256 450   0.93 (0.77-1.13) 0.59  
MEC bb 903 1,051     
 Bb 518 672 0.0009 0.26 0.93 (0.79-1.08)   
 BB 115 158   0.89 (0.68-1.17) 0.26  
NHS bb 407 550     
 Bb 555 723 0.86 0.4 1.01 (0.85-1.20)   
 BB 160 242   0.91 (0.72-1.15) 0.55  
PLCO bb 405 407     
 Bb 468 533 0.41 0.39 0.87 (0.72-1.05)   
 BB 192 157   1.21 (0.94-1.56) 0.43  
WHS bb 201 200     
 Bb 303 298 0.78 0.42 1.01 (0.78-1.31)   
 BB 100 106   0.94 (0.68-1.32) 0.77  
All cohorts bb 2,631 3,321     
 Bb 2,823 3,645 0.0004 0.37 0.98 (0.91-1.06)   
 BB 901 1,183   0.98 (0.89-1.09) 0.66 0.5 
         
Ethnic-specific results from multiethnic cohort only         
    Hispanic bb 184 207     
 Bb 115 141 0.26 0.25 0.91 (0.66-1.26)   
 BB 24 21   1.32 (0.70-2.48) 0.85  
    African American bb 163 217     
 Bb 126 155 0.34 0.29 1.10 (0.80-1.51)   
 BB 27 40   0.91 (0.53-1.57) 0.94  
    Japanese American bb 341 299     
 Bb 71 106 0.86 0.14 0.60 (0.42-0.85)   
 BB   0.52 (0.12-2.26) 0.003  
    Hawaiian bb 65 175     
 Bb 35 86 0.7 0.2 1.23 (0.74-2.04)   
 BB 13   1.05 (0.31-3.53) 0.52  
    Caucasian bb 150 153     
 Bb 171 184 0.08 0.41 1.02 (0.74-1.41)   
 BB 57 79   0.85 (0.56-1.30) 0.55 0.08 
         
All cohorts combined         
    Caucasian bb 1,751 2,271     
 Bb 2,381 3,010 0.14 0.41 1.01 (0.93-1.10)   
 BB 821 1,077   0.99 (0.89-1.11) 0.97 0.62 

We evaluated whether the associations between the two VDR polymorphisms and breast cancer risk were modified by total calcium intake. Although risk was highest among women with the rs2228570 (FokI) ff genotype in the top study-specific quintile of total calcium intake (OR, 1.37; 95% CI, 1.04-1.81) versus FF/bottom quintile of total calcium, the test for interaction was of borderline statistical significance (P = 0.08; Table 4A). Analyses stratified by extreme quintiles of total calcium intake showed no effect modification of the association between the rs1544410 (BsmI) genotype and breast cancer risk (Table 4B).

Table 4.

(A) Association among VDR FokI genotype, total calcium intake, and breast cancer risk
FokI genotypeQuintiles of total calcium intake
12-45
FF 368/468 1,066/1,277 306/425 
 1.00 1.04 (0.88-1.22) 0.83 (0.67-1.02) 
Ff 405/469 1,236/1,455 449/494 
 1.09 (0.90-1.33) 1.05 (0.89-1.23) 1.07 (0.89-1.30) 
ff 125/127 404/438 157/135 
 1.25 (0.94-1.67) 1.13 (0.93-1.38) 1.37 (1.04-1.81) 
    
(B) Association among VDR BsmI genotype, total calcium intake, and breast cancer risk
 
   
BsmI genotype Quintiles of total calcium intake
 
  

 
1
 
2-4
 
5
 
bb 417/457 1,139/1,325 370/428 
 1.00 0.91 (0.78-1.07) 0.85 (0.70-1.04) 
Bb 366/471 1,147/1,336 390/435 
 0.82 (0.68-1.00) 0.89 (0.76-1.04) 0.91 (0.74-1.10) 
BB 100/115 357/410 141/149 
 0.94 (0.69-1.28) 0.90 (0.74-1.11) 0.96 (0.73-1.26) 
(A) Association among VDR FokI genotype, total calcium intake, and breast cancer risk
FokI genotypeQuintiles of total calcium intake
12-45
FF 368/468 1,066/1,277 306/425 
 1.00 1.04 (0.88-1.22) 0.83 (0.67-1.02) 
Ff 405/469 1,236/1,455 449/494 
 1.09 (0.90-1.33) 1.05 (0.89-1.23) 1.07 (0.89-1.30) 
ff 125/127 404/438 157/135 
 1.25 (0.94-1.67) 1.13 (0.93-1.38) 1.37 (1.04-1.81) 
    
(B) Association among VDR BsmI genotype, total calcium intake, and breast cancer risk
 
   
BsmI genotype Quintiles of total calcium intake
 
  

 
1
 
2-4
 
5
 
bb 417/457 1,139/1,325 370/428 
 1.00 0.91 (0.78-1.07) 0.85 (0.70-1.04) 
Bb 366/471 1,147/1,336 390/435 
 0.82 (0.68-1.00) 0.89 (0.76-1.04) 0.91 (0.74-1.10) 
BB 100/115 357/410 141/149 
 0.94 (0.69-1.28) 0.90 (0.74-1.11) 0.96 (0.73-1.26) 

NOTE: Pinteraction = 0.08. Pinteraction = 0.43.

In this large pooled analysis of data from 6 prospective studies, we observed a small, statistically significant increase in breast cancer risk associated with the ff genotype of the VDR rs2228570 (FokI) SNP, with a 16% increase in risk in homozygotes for the minor allele, relative to homozygotes for the more common allele (FF). However, there was evidence of heterogeneity of findings across cohorts (P = 0.03), which was not explained by differences in race/ethnicity. In addition, the association was weakened and lost statistical significance when excluding the NHS, in which a positive association had been reported previously.

The presence of the rs2228570 (FokI) f allele in the 5′-promoter region of the VDR results in production of a VDR protein that is less effective as a transcriptional activator (15). The cellular consequences of the less active ff genotype would be expected to mimic those of lower vitamin D status. Both the geographic gradient in breast cancer incidence and prospective and retrospective studies of circulating 25(OH)D and breast cancer risk suggest an inverse relationship between vitamin D status and breast cancer risk (2, 3), although findings are mixed (4). Our results, which found a modest increase in breast cancer risk associated with the ff genotype, although far from dramatic, are consistent with a role for vitamin D in breast cancer etiology.

The lack of an association between the rs1544410 (BsmI) SNP and breast cancer risk argues against a major role of this polymorphism in breast cancer susceptibility, a result consistent with the indeterminate associations observed between this SNP and breast cancer risk (2, 8, 9, 11, 12, 16, 19-23). Although 6 studies reported increased risk of breast cancer with the BsmI bb genotype (2, 9, 16, 19-21), mostly among Caucasian women, 5 other studies did not report a similar association (8, 11, 12, 22, 23). Two studies have suggested a higher risk of metastatic breast disease among homozygotes for the more common b allele (9, 20). In support of these findings, we did find that the BB genotype was statistically significantly and inversely associated with risk of advanced breast cancer tumors overall and when we restricted the analysis to Caucasian women. This result is consistent with preliminary reports suggesting a protective role for vitamin D in lung cancer survival and prognosis (29) and ecologic correlations of cancer survival with greater sun exposure or season of diagnosis (30, 31). The hypothesized mechanisms for better prognosis with more favorable vitamin D status are based on animal models and involve modulation of cell cycle progression, apoptosis, and cell signaling leading to reduced tumor invasiveness and angiogenesis (32). The ethnic differences in allele frequency for BsmI also raise the possibility of confounding by population stratification, particularly as ethnic differences in breast cancer survival have been suggested (33, 34).

The potential modification of genotype-breast cancer associations by environmental factors, such as diet, is worthy of consideration. Calcium and vitamin D metabolism are closely linked and both nutrients have favorable effects on cell proliferation and differentiation of several cancer cell lines in vitro (35). Dietary factors including calcium are known to affect the vitamin D endocrine system (36), and diet may also influence autocrine/paracrine vitamin D metabolism (37). In previous studies of breast cancer (11) and colorectal adenoma (38), the risk by VDR BsmI genotype varied by calcium intake, but no studies have reported an interaction with the VDR FokI SNP. We did not observe an interaction with the BsmI SNP, and observed only a weak interaction (P = 0.08) between total calcium intake and FokI genotype, using a conservative test for interaction. The association of the FokI ff genotype with increased risk of breast cancer was seen across all levels of total calcium intake, but the association between total calcium and breast cancer was limited to the FokI FF genotype. Although this finding may be due to chance, it may also indicate an interplay between calcium intake and VDR function. We were unable to examine VDR genotype interactions with circulating levels of 25(OH)D, the integrated marker of vitamin D status from diet, supplements, and UVB exposure. However, in a population-based case-control study (14) and a nested case-control study (8) no significant interaction among FokI genotype, 25(OH)D, and breast cancer risk was observed.

These results from a pooled analysis of data from six large cohorts suggest that the rs2228570 (FokI) and rs1544410 (BsmI) polymorphisms in VDR may have a small role in breast cancer susceptibility. Although both genetic findings support the hypothesis that vitamin D status plays a role in breast cancer etiology, the associations were modest and may be due to chance. In future studies, VDR genetic variation should be integrated with prediagnostic biomarkers of vitamin D status.

No potential conflicts of interest were disclosed.

Grant support: National Cancer Institute, NIH cooperative agreements UO1-CA98233, UO1-CA98710, UO1-CA98216, and UO1-CA98758 and contract N01-CO-12400.

Note: J.D. McKay and M.L. McCullough contributed equally to this work. J.D. McKay is a C.J. Martin fellow (National Health and Medical Research Council Australia).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Garland FC, Garland CF, Gorham ED, Young JF. Geographic variation in breast cancer mortality in the United States: a hypothesis involving exposure to solar radiation.
Prev Med
1990
;
19
:
614
–22.
2
Lowe LC, Guy M, Mansi JL, et al. Plasma 25-hydroxy vitamin D concentrations, vitamin D receptor genotype and breast cancer risk in a UK Caucasian population.
Eur J Cancer
2005
;
41
:
1164
–9.
3
Bertone-Johnson ER, Chen WY, Holick MF, et al. Plasma 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D and risk of breast cancer.
Cancer Epidemiol Biomarkers Prev
2005
;
14
:
1991
–7.
4
Freedman DM, Chang SC, Falk RT et al. Serum levels of vitamin D metabolites and breast cancer risk in the prostate, lung, colorectal, and ovarian cancer screening trial.
Cancer Epidemiol Biomarkers Prev
2008
;
17
:
889
–94.
5
Colston K, Welsh J. Vitamin D and breast cancer. In: Feldman D, editor. Vitamin D. Stanford (CA): Elsevier Academic Press; 2005. p. 1663–77.
6
Townsend K, Banwell CM, Guy M, et al. Autocrine metabolism of vitamin D in normal and malignant breast tissue.
Clin Cancer Res
2005
;
11
:
3579
–86.
7
Zinser GM, Welsh J. Vitamin D receptor status alters mammary gland morphology and tumorigenesis in MMTV-neu mice.
Carcinogenesis
2004
;
25
:
2361
–72.
8
Chen WY, Bertone-Johnson ER, Hunter DJ, Willett WC, Hankinson SE. Associations between polymorphisms in the vitamin D receptor and breast cancer risk.
Cancer Epidemiol Biomarkers Prev
2005
;
14
:
2335
–9.
9
Bretherton-Watt D, Given-Wilson R, Mansi JL, Thomas V, Carter N, Colston KW. Vitamin D receptor gene polymorphisms are associated with breast cancer risk in a UK Caucasian population.
Br J Cancer
2001
;
85
:
171
–5.
10
Curran JE, Vaughan T, Lea RA, Weinstein SR, Morrison NA, Griffiths LR. Association of a vitamin D receptor polymorphism with sporadic breast cancer development.
Int J Cancer
1999
;
83
:
723
–6.
11
Ingles SA, Garcia DG, Wang W, et al. Vitamin D receptor genotype and breast cancer in Latinas.
Cancer Causes Control
2000
;
11
:
25
–30.
12
McCullough ML, Stevens VL, Diver WR, et al. Vitamin D pathway gene polymorphisms, calcium intake, and risk of postmenopausal breast cancer. Breast Cancer Res 2007;9:doi:10.1186/bcr642.
13
John EM, Schwartz GG, Koo J, Wang W, Ingles SA. Sun exposure, vitamin D receptor gene polymorphisms, and breast cancer risk in a multi-ethnic population.
Am J Epidemiol
2007
;
166
:
1409
–19.
14
Abbas S, Nieters A, Linseisen J, et al. Vitamin D receptor gene polymorphisms and haplotypes and postmenopausal breast cancer risk.
Breast Cancer Res
2008
;
10
:
R31
. Epub ahead of print.
15
Uitterlinden AG, Fang Y, van Meurs JBJ, et al. Genetics and biology of vitamin D receptor polymorphisms.
Gene
2004
;
338
:
143
–56.
16
Guy M, Lowe LC, Bretherton-Watt D, et al. Vitamin D receptor gene polymorphisms and breast cancer risk.
Clin Cancer Res
2004
;
10
:
5472
–81.
17
Ingles SA, Haile RW, Henderson BE, et al. Strength of linkage disequilibrium between two vitamin D receptor markers in five ethnic groups: implications for association studies.
Cancer Epidemiol Biomarkers Prev
1997
;
6
:
93
–8.
18
Slattery ML, Yakumo K, Hoffman M, Neuhausen S. Variants of the VDR gene and risk of colon cancer.
Cancer Causes Control
2001
;
12
:
359
–64.
19
Yamagata Z, Zhang Y, Asaka A. Association of breast cancer with vitamin D receptor gene polymorphism [abstract].
Am J Hum Genet
1997
;
61
:
A388
.
20
Ruggiero M, Pacini S, Aterini S, Fallai C, Ruggiero C, Pacini P. Vitamin D receptor gene polymorphism is associated with metastatic breast cancer.
Oncol Res
1998
;
10
:
43
–6.
21
Trabert B, Malone KE, Daling JR, et al. Vitamin D receptor polymorphisms and breast cancer risk in a large population-based case-control study of Caucasian and African-American women. Breast Cancer Res 2007;9:R84. doi:10.1186/ber 1833.
22
Buyru N, Tezol A, Yosunkaya-Fenerci E, Dalay N. Vitamin D receptor gene polymorphisms in breast cancer.
Exp Mol Med
2003
;
35
:
550
–5.
23
Hou M, Tien Y, Lin G, et al. Association of vitamin D receptor gene polymorphism with sporadic breast cancer in Taiwanese patients.
Breast Cancer Res Treat
2002
;
74
:
1
–7.
24
Hunter DJ, Riboli E, Haiman CA, et al. A candidate gene approach to searching for low-penetrance breast and prostate cancer genes.
Nat Rev Cancer
2005
;
5
:
977
–85.
25
Feigelson HS, Cox DG, Cann HM, et al. Haplotype analysis of the HSD17B1 gene and risk of breast cancer: a comprehensive approach to multicenter analyses of prospective cohort studies.
Cancer Res
2006
;
66
:
2468
–75.
26
Willett W, Stampfer MJ. Total energy intake: implications for epidemiologic analyses.
Am J Epidemiol
1986
;
124
:
17
–27.
27
Packer BR, Yeager M, Staats B, et al. SNP500 Cancer: a public resource for sequence validation and assay development for genetic variation in candidate genes.
Nucleic Acids Res
2004
;
1
:
D528
–32.
28
Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis.
Stat Med
2002
;
21
:
1539
–58.
29
Zhou W, Heist RS, Liu G, et al. Circulating 25-hydroxyvitamin D levels predict survival in early-stage non-small-cell lung cancer patients.
J Clin Oncol
2007
;
25
:
479
–85.
30
Moan J, Porojnicu AC, Dahlback A, Setlow RB. Addressing the health benefits and risks, involving vitamin D or skin cancer, of increased sun exposure.
Proc Natl Acad Sci U S A
2008
;
105
:
668
–73.
31
Porojnicu AC, Lagunova Z, Robsahm TE, Berg JP, Dahlback A, Moan J. Changes in risk of death from breast cancer with season and latitude.
Breast Cancer Res Treat
2007
;
102
:
323
–8.
32
Colston KW, Hansen CM. Mechanisms implicated in the growth regulatory effects of vitamin D in breast cancer.
Endocr Relat Cancer
2002
;
9
:
45
–59.
33
Porter PL, Lund MJ, Lin MG, et al. Racial differences in the expression of cell cycle-regulatory proteins in breast cancer.
Cancer
2004
;
100
:
2533
–42.
34
Chlebowski RT, Chen Z, Anderson GL, et al. Ethnicity and breast cancer: factors influencing differences in incidence and outcome.
J Natl Cancer Inst
2005
;
97
:
439
–48.
35
Lipkin M, Newmark HL. Vitamin D, calcium, and prevention of breast cancer: a review.
J Am Coll Nutr
1999
;
18
:
392
–7S.
36
Holick M. Vitamin D. In: Shils ME, Olson JA, Shike M, Ross AC, editors. Modern nutrition in health and disease. 9th ed. Philadelphia (PA): Lippincott Williams & Wilkins; 1999.
37
Cross HS, Kallay E, Lechner D, Gerdenitsch W, Adlercreutz H, Armbrecht HJ. Phytoestrogens and vitamin D metabolism: a new concept for the prevention and therapy of colorectal, prostate, and mammary carcinomas.
J Nutr
2004
;
134
:
1207
–12S.
38
Kim HS, Newcomb PA, Ulrich CM, et al. Vitamin D receptor polymorphism and the risk of colorectal adenomas: evidence of interaction with dietary vitamin D and calcium.
Cancer Epidemiol Biomarkers Prev
2001
;
10
:
869
–74.