Original Article

You have full text access to this OnlineOpen article

Age-specific association of steroid hormone pathway gene polymorphisms with breast cancer risk†

  1. David A. Ralph PhD1,
  2. Lue Ping Zhao PhD2,
  3. Christopher E. Aston PhD3,
  4. Sharmila Manjeshwar PhD1,
  5. Thomas W. Pugh MS1,
  6. Daniele C. DeFreese MS1,
  7. Bobby A. Gramling MS1,
  8. Craig D. Shimasaki PhD1,
  9. Eldon R. Jupe PhD1,3,‡,*

Article first published online: 13 APR 2007

DOI: 10.1002/cncr.22634

Issue

Cancer

Cancer

Volume 109, Issue 10, pages 1940–1948, 15 May 2007

Additional Information

How to Cite

Ralph, D. A., Zhao, L. P., Aston, C. E., Manjeshwar, S., Pugh, T. W., DeFreese, D. C., Gramling, B. A., Shimasaki, C. D. and Jupe, E. R. (2007), Age-specific association of steroid hormone pathway gene polymorphisms with breast cancer risk. Cancer, 109: 1940–1948. doi: 10.1002/cncr.22634

Author Information

  1. 1

    InterGenetics Incorporated, Oklahoma City, Oklahoma

  2. 2

    Division of Public Health Sciences, Fred Hutchinson Cancer Research Center, Seattle, Washington

  3. 3

    General Clinical Research Center and Department of Pediatrics, University of Oklahoma Health Sciences Center, Oklahoma City, Oklahoma

  1. ‡

    Fax: (405) 271-1725

*InterGenetics Incorporated, 655 Research Parkway, Suite 300, Oklahoma City, OK 73104

  1. †

    All authors are either employees or consultants of InterGenetics Incorporated

Publication History

  1. Issue published online: 25 APR 2007
  2. Article first published online: 13 APR 2007
  3. Manuscript Accepted: 12 FEB 2007
  4. Manuscript Revised: 28 JAN 2007
  5. Manuscript Received: 27 OCT 2006

Funded by

  • U.S. Army Breast Cancer Research Program. Grant Number: DAMD17-01-1-0358
  • American Cancer Society. Grant Number: RPG-97-167-01-MGO
  • Oklahoma Center for the Advancement of Science and Technology. Grant Numbers: AR05.1025, AR02.2032, AR99.2-007, AR01.1-050
  • National Institutes of Health
  • National Center for Research Resources
  • General Clinical Research Center. Grant Number: M01RR-14467
  • Presbyterian Health Foundation
  • Oklahoma Life Sciences Fund
  • Swisher Family Trust

SEARCH

SEARCH BY CITATION

ARTICLE TOOLS

Get PDF (173K)

Keywords:

  • age-specific genetic associations;
  • steroid hormone pathway;
  • single nucleotide polymorphisms;
  • breast cancer

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

BACKGROUND

Breast cancer (BC) is a complex disease, and the incidence rates for BC increase with age. Both environmental factors and genetics have an impact on the risk of BC. Although the effects of environmental factors may vary with age, it has been assumed generally that the penetrance of single nucleotide polymorphisms (SNPs) is constant throughout life. In the current study, the results demonstrated that certain SNPs exhibit BC risk associations that vary considerably with age.

METHODS

SNPs in 12 steroid hormone pathway genes were investigated for associations with BC risk in white women who were enrolled in an age-matched, case-control (1:2 for cases and controls, respectively) study that consisted of a discovery set (n = 5000 women) and an independent validation set (n = 1583 women).

RESULTS

Significant age-related trends were identified and confirmed for SNPs in 4 genes associated with BC risk. The cytosine/cytosine (C/C) genotype of cytochrome P450 XIB2 (CYP11B2) was associated with decreased risk at younger ages (ages 30–44 years) but an increased risk at older ages (ages 55–69 years). The homozygous cytosine-guanine (CG/CG) genotype of uridine phosphorylase glycosyltransferase 1A7 (UGT1A7) was associated with increased risk at younger ages but decreased risk at older ages. Associations in cytochrome P450 19 (CYP19) and progesterone receptor (PGR) were confined to middle age (ages 45–54 years).

CONCLUSIONS

The identification of age-specific genetic associations may have profound implications for future etiologic studies of BC and for the use of SNP genotyping to accurately predict the risk of BC in women. Cancer 2007. © 2007 American Cancer Society.

Worldwide, breast cancer (BC) is the most commonly diagnosed cancer, affecting 1 in 8 women in the U.S.1 For decades, studies of hormonally related risk factors contributed substantially toward understanding the origins of BC.2, 3 The modern era of BC risk assessment began with the identification of highly penetrant mutations in the BRCA1 and BRCA2 genes that explain the etiology of BC in some families with strong histories.4, 5 However, the majority of BC cases occur sporadically in individuals with little or no family history, and to our knowledge no clear role for the BRCA genes in sporadically occurring BC has emerged. It is likely that a predisposition to sporadic BC is associated with relatively common but weakly penetrant genetic variants in multiple genes.6, 7 Advances in genotyping technologies have enabled the identification of such weakly penetrant polymorphisms.8, 9 The relation of steroid hormone (SH) levels to BC risk led to the investigation of single nucleotide polymorphisms (SNPs) in genes that influence endogenous estrogen levels or the bioavailability and detoxification of reactive estrogen metabolites, and many have been associated with BC risk.10, 11

The influence of sex SH levels, especially estrogens, on BC risk is known to vary depending on a woman's age or menopausal status.12–17 This supports the hypothesis that SNPs in functional regions of genes involved in SH synthesis, signaling, and metabolism may differentially impact BC risk, depending on age or menopausal status. Indeed, some studies of SNPs in sporadic BC suggest that their influence on risk is related to menopausal status18–20 or age of onset.21, 22 However, in most studies, the risks associated with SNPs have been evaluated without respect to age, partly because statistical methods (eg, linkage analysis) assume age-invariant odds ratios (ORs) or because study sizes are too small to identify age-stratified associations.

Thus, although the contribution of SH-related factors to BC risk vary with age, age-specific genetic contributions to risk have not been investigated extensively. In this study, the potential association with BC risk was investigated in 18 candidate SNPs in 12 genes from pathways related to SH synthesis, signaling, or metabolism. These SNPs have been studied previously for association with the risk of breast or other cancers. To detect potential age-specific genetic associations (ASGAs) in white women, we analyzed a large discovery set of 1667 cases and 3333 age-matched, cancer-free controls and stratified them into 3 age groups (ages 30–44 years, ages 45–54 years, and ages 55–69 years). Our findings were validated using an independent set of 526 cases and 1057 controls. Four gene polymorphisms that exhibited ASGAs with BC were identified and validated.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Study Description

Women were enrolled in 6 geographically distinct regions of the U.S. Approximately half were enrolled in the greater Oklahoma City area (1996–2006), and the remaining women were recruited from Seattle, Southern California, Kansas City, Florida, and South Carolina (2003–2006). Patients were approached consecutively, without prior knowledge of their disease status, as they presented for appointments at mammography centers. For cases enrolled in oncology clinics, the controls were obtained in general practice clinics in the same medical complex. Some cases and controls were enrolled at Komen Races or other community-based events. At all collection sites, the majority of individuals who were approached enrolled in the study. The characteristics of cases and controls were similar for epidemiologic factors related to lifetime SH exposure, but variations in hormone-replacement therapy (HRT) use were observed (see Table A in the web supplement at URL: www.intergenetics.com).

Cases were defined as women with a self-reported diagnosis of BC, whereas controls had never been diagnosed with any cancer. No exclusions were in effect for enrollment in the study. All enrolled participants provided informed consent, completed a questionnaire on personal medical history and family history of cancer, and provided a buccal cell sample that was collected in commercial mouthwash. All patients were enrolled under Institutional Review Board (IRB)-approved informed consent, and all study protocols were IRB approved, monitored, and performed as described previously.23

This report focuses only on the genetic analysis of white women. The age range for women in the study was from 30 years to 69 years: The age at diagnosis was used for cases, and the age at enrollment was used for controls. The primary discovery set consisted of 5000 women (1667 BC cases and 3333 cancer-free controls) who were age-matched to the cases within 1 year. Age matching was done to adjust for potential confounding effects caused by age-related risk factors when assessing ORs across different ages. An independent validation set consisting of 526 cases and 1057 controls was used to confirm the discovered associations.

SNPs

Samples were genotyped for the following 18 SNPs in 12 SH pathway candidate genes: catechol-O-methyltransferase (COMT) (rs4680); Cytochrome P450, family 1A, polypeptide 1 CYP1A1 (rs4646903, rs1048943); cytochrome P450, family XIB, polypeptide 2 (CYP11B2) (rs1799998); Cytochrome P450, family 1A, polypeptide 1 CYP1B1 (rs10012, rs1056836); CYP17 (rs743572); cytochrome P450, family 19, subfamily A, polypeptide 1 (CYP19) (rs10046, rs700519); microsomal epoxide hydrolase (EPHX1) (rs1051740), estrogen receptor α (ERA) (rs2077647); progesterone receptor α (PGR) (rs1042838, rs10895068); SH-binding globulin (SHBG) (rs6529, rs1799941); manganese superoxide dismutase (SOD2) (rs1799725); and uridine diphosphate glycosyltransferase 1 family, polypeptide A7 (UGT1A7) (rs17868324, rs11692021). Information concerning the 6 SNPs that are most relevant to this report is shown in Table 1.18, 19, 24–36 This information is available in the web supplement for all 18 SNPs (available at URL: www.intergenetics.com, Table B).

Table 1. Single Nucleotide Polymorphisms in Steroid Hormone Pathway Genes
GeneNameFunctiondbSNP IDPolymorphismFunctional effectReferences*
FunctionalEpidemiologic

  1. dbSNP ID indicates single nucleotide polymorphism data base identification number; ↓, decreased; ↑, increased; UTR, untranslated region; PR mRNA, progesterone receptor messenger RNA; UDP, uridine diphosphate.

COMTCatechol-O-methyltransferaseInactivates catechol estrogens by methylationrs4680G→A, Val158Met↓ Methylation activity2418,19,25
CYP11B2Cytochrome P450 family XIB, polypeptide 2Synthesis of aldosterone in renin-angiotensin systemrs1799998C→T, promoter −344↑ Aldosterone secretion26,2728
CYP19Cytochrome P450, family 19, subfamily A, polypeptide 1Terminal enzyme in estrogen synthesis that catalyzes formation of C18 estrogens from C19 androgensrs10046C→T, 3′UTR↑ Activity phenotype2929
PGRProgesterone receptorMediates the effects of progesterone during breast development; 2 isoforms: PGR-A (opposes the effects of PGR-B) and PGR-B (promotes breast cell proliferation)rs1042838G→T, Val660Leu↑ Half-life of PR mRNA3031
SOD2Manganese superoxide dismutaseIntramitochondrial, manganese-dependent, free radical scavenger that metabolizes reactive oxygen species to hydrogen peroxiders1799725C→T, Val16AlaMay affect protein transport3233
UGT1A7UDP glycosyltransferase 1 family, polypeptide A7Detoxification of lipophilic xenobiotics, hormones, and drugs by glucuronidationrs17868324AA→CG, Lys131Arg→ Enzyme activity3435,36

Genotyping

Genomic DNA was isolated, and the majority of the samples were genotyped by microbead-based, allele-specific primer extension, as previously described.23 Primer sequences and genotyping conditions are available from the authors upon request. Allele-specific primary extension (ASPE) assays had reproducibility rates >99.4%. Restriction fragment-length polymorphism assays were done on approximately 3% of samples and had reproducibility rates >98%. Genotyping was done blinded to the case-control status, and internal reproducibility was confirmed by examining ≥5% of the specimens in duplicate.

Statistical Methods

SNP associations and their age interactions were evaluated by using both descriptive and analytic statistics. Genotype frequencies were summarized, and a chi-square test was used to evaluate Hardy-Weinberg equilibrium (HWE) for individual SNP genotypes in the controls.37 SNP associations with case and control status were assessed using chi-square test statistics with 2 degrees of freedom in 3 × 2 contingency table analyses, which are equivalent to unconditional logistic regression.38 This analysis also was used to compute ORs and related statistics using the most common homozygote as the reference genotype.39 Analyses were performed first without age stratification; then, stratification of the entire sample set into 3 age groups (ages 30–44 years, ages 45–54 years, and ages 55–69 years) was done to identify ASGAs. The mean age of menopause for women in the U.S. is approximately 50.5 years,40 but the endocrine hormonal transition exhibits approximately a window of ±5 years.41 Thus, we chose age categories that were likely to be representative of premenopausal, perimenopausal, and postmenopausal life stages to minimize the effect of differing ages of menopause in individuals. Raw P values without correcting for multiple comparisons are reported. For SNPs that demonstrated ASGAs, adjusted analyses were performed to control for HRT use. Analysis of the discovery set was followed by validation in the independent set of cases and controls. Finally, to fully quantify ASGAs as a function of SNP penetrance, a sliding 10-year window strategy was used to estimate genotypic ORs and related statistics for 1-year incremental age groups (ages 30–39 years, ages 31–40 years,…, ages 59–68 years, and ages 60–69 years). Because precise age associations may vary from individual to individual, this 10-year sliding window strategy was used to increase power and to reduce the noise of random variation between single year differences in each individual.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Overall Associations With BC Risk

Overall, age-independent analyses of associations were performed on the discovery set. All SNPs conformed to HWE (P > .05) in the control population, as would be expected in a general population at steady state, and were used in subsequent association analyses. Table 2 shows the results for the SNPs that are relevant to this report, and the results for all SNPs are presented in the web supplement (available at URL: www.intergenetics.com, Table C). The only significant association (P < .05) with BC risk was for the cytosine/cytosine (C/C) genotype of the SOD2 gene (OR, 1.2; P = .02). SNP genotypes for COMT, CYP11B (Table 2), and the 3′-untranslated region (3′UTR) of CYP1A1 (available at URL: www.intergenetics.com, Table C) exhibited suggestive associations (.05 < P < .01). In the validation set, the SOD2 association was not replicated (OR, 1.0), suggesting a possible false discovery.

Table 2. Overall Associations With Breast Cancer Risk in the Discovery Set
SNP/GenotypeNo. (%)OR (95% CI)P (HWE)
CasesControls

  • SNP indicates single nucleotide polymorphism; OR, odds ratio; 95% CI, 95% confidence interval; HWE, Hardy-Weinberg equilibrium;

  • A, adenine; G, guanine; T, thymine; C, cytosine; (Ref), reference genotype.

  • *

    .05 < P < .1.

  • †

    P ≤ .05.

COMT
 A/A405 (25)900 (27)1 (Ref).8
 G/A825 (51)1631 (50)1.1 (0.9–1.3)
 G/G396 (24)755 (23)1.2 (0.9–1.4)*
CYP11B2
 T/T486 (29)1044 (32)1.5
 C/T842 (51)1613 (48)1.1 (0.9–1.3)*
 C/C323 (20)651 (20)1.1 (0.9–1.3)
CYP19
 T/T461 (28)883 (27)1.6
 C/T830 (51)1650 (50)1 (0.8–1.1)
 C/C349 (21)758 (23)0.9 (0.7–1)
PRG
 G/G1140 (69)2344 (71)1.3
 T/G454 (28)879 (27)1.1 (0.9–1.2)
 T/T47 (3)71 (2)1.4 (0.9–2)
SOD2
 T/T392 (24)861 (26)1.7
 C/T816 (49)1667 (50)1.1 (0.9–1.2)
 C/C440 (27)786 (24)1.2 (1–1.5)†
UGT1A7
 AA/AA645 (41)1335 (42)1.2
 AA/CG727 (46)1446 (45)1 (0.9–1.2)
 CG/CG211 (13)430 (13)1 (0.8–1.2)

Age-stratified Associations with BC Risk

Because of our focus on potential ASGAs for these SNPs, we matched cases and controls by age within 1 year. To investigate potential age-specific SNP associations, we computed ORs for each SNP within 3 age groups: young (ages 30–44 years), middle (ages 45–54 years), and old (ages 55–69 years). Table 3 shows the ORs with 95% confidence intervals (95% CIs) and genotype frequencies that were determined in the discovery set for 5 genes (COMT, CYP11B2, CYP19, PGR, and UGT1A7) with SNP genotypes that exhibited significant associations with BC risk in ≥1 age group(s). Risk associations for SNPs in 3 genes (COMT, CYP19, and PGR) were limited to only 1 age group. For COMT, both homozygous guanine/guanine (G/G) (OR, 1.5; P = .02) and heterozygous guanine/adenine (G/A) (OR,1.4; P = .01) were associated with elevated BC risk in the young group. For CYP19, the homozygous C/C genotype was associated with reduced risk (OR, 0.7; P = .02) only in the middle group. The homozygous thymine/thymine (T/T) genotype in PGR was associated with significantly increased BC risk confined to the middle group (OR, 2.0; P = .04).

Table 3. Age-specific Genetic Associations With Breast Cancer Risk in the Discovery Set
SNP/GenotypeYoung (Ages 30–44 Years)Middle (Ages 45–54 Years)Old (Ages 55–69 Years)
No. (%)OR (95% CI)No. (%)OR (95% CI)No. (%)OR (95% CI)
CasesControlsCasesControlsCasesControls

  • OR indicates odds ratio; 95% CI, 95% confidence interval; SNP, single nucleotide polymorphism; (Ref), reference genotype; UTR, untranslated region.

  • *

    .05 < P < .1.

  • †

    P ≤ .05.

COMT
 A/A106 (21)282 (27)1 (Ref)160 (27)339 (28)1130 (28)253 (27)1
 G/A272 (53)512 (50)1.4 (1.1– 1.8)†289 (49)593 (50)1 (0.8– 1.3)236 (50)473 (50)0.9 (0.7– 1.3)
 G/G131 (26)238 (23)1.5 (1.1– 2.0)†145 (24)261 (22)1.2 (0.9– 1.6)104 (22)226 (23)0.9 (0.6– 1.2)
CYP11B2
 T/T163 (32)306 (30)1183 (30)372 (31)1124 (26)332 (35)1
 C/T272 (53)499 (48)1 (0.8– 1.3)293 (49)599 (50)1 (0.8– 1.2)251 (52)460 (48)1.5 (1.1– 1.9)†
 C/C79 (15)228 (22)0.6 (0.5– 0.9)†125 (21)232 (19)1.1 (0.8– 1.5)106 (22)170 (17)1.7 (1.2– 2.3)†
CYP19 (3′UTR)
 T/T142 (28)267 (26)1181 (30)303 (25)1127 (26)286 (30)1
 C/T254 (50)538 (53)0.9 (0.7– 1.1)297 (50)614 (52)0.8 (0.6– 1)*251 (53)444 (46)1.3 (1– 1.6)*
 C/C114 (22)220 (21)0.9 (0.7– 1.3)119 (20)279 (23)0.7 (0.5– 0.9)†101 (21)230 (24)1 (0.7– 1.4)
PGR (V660L)
 G/G368 (71)724 (71)1419 (70)870 (73)1319 (67)669 (70)1
 T/G134 (26)275 (27)1 (0.7– 1.2)159 (27)309 (26)1.1 (0.8– 1.3)144 (30)267 (28)1.1 (0.9– 1.4)
 T/T13 (3)26 (2)1 (0.5– 1.9)19 (3)20 (1)2 (1– 3.7)†13 (3)25 (2)1.1 (0.5– 2.2)
UGT1A7 (K131R)
 AA/AA188 (38)439 (44)1226 (39)488 (42)1207 (46)366 (39)1
 AA/CG229 (46)447 (44)1.2 (0.9– 1.5)289 (49)524 (45)1.2 (0.9– 1.5)188 (42)425 (46)0.8 (0.6– 1)†
 CG/CG77 (16)122 (12)1.5 (1– 2)†71 (12)155 (13)1 (0.7– 1.4)56 (12)139 (15)0.7 (0.5– 1)*

The SNPs in the other 2 genes (CYP11B2 and UGT1A7) exhibited an unexpected pattern of ASGAs. For both genes, risk associations reversed between the young and old groups, such that genotypes associated with increased risk in the young group became protective in the old group, or vice versa. In CYP11B2, the homozygous C/C genotype was associated with significantly reduced risk (OR, 0.6; P = .008) in the young group. In contrast, the homozygous C/C genotype (OR, 1.7; P = .002) and the heterozygous C/T genotype (OR, 1.5; P = .004) both were associated with increased risk in the old group. Similarly, the homozygous CG/CG genotype at UGT1A7 was associated with a gradual decline in risk from young, to middle age, to old age (OR, 1.5, 1.0, and 0.7, respectively). In addition, the heterozygous AA/CG genotype was associated with reduced risk confined to the old group. Four additional genes with ASGAs that approached significance (.05<P<.1) are shown in the web supplement (available at URL: www.intergenetics.com, Table D). To determine whether the use of HRT may be a confounding factor affecting the observed ASGAs, analyses were repeated adjusting for HRT use. The results were similar to the unadjusted analyses for the majority of the ASGAs (data not shown). The only exception was for the T/T genotype of PGR, which was suggestive of an increased risk in the young group; however, this result was far from statistically significant (P = .22). Thus, we conclude that HRT use had little or no impact these ASGAs.

Validation of Discovered Associations

To validate these 5 potential ASGAs, their ORs, 95% CIs, and pertinent genotype frequencies were computed for the independent validation set (Table 4). Although none of the ORs reach statistical significance, for CYP11B2, the age-specific patterns of the C/C and C/T genotypes were consistent with those observed in the discovery set. For the C/C genotype of CYP19, a decreased OR (0.8) in the middle age group was similar to the OR of 0.7 in the discovery set. The result for the homozygous T/T genotype for PGR also replicated with an estimated OR of 1.5 in the middle age group. The estimated ORs for the homozygous CG/CG genotype of UGT1A7 followed the same pattern in the validation set that was observed in the original discovery data set, gradually decreasing from 1.1, to 0.7, and to 0.5 (P = .05) over the young, middle, and old age groups, respectively. Although the absolute magnitudes of the OR values differed, this age-specific pattern was consistent with that observed in the discovery set (1.5, to 1.0, and to 0.7). Finally, the results for COMT failed to replicate in the validation set.

Table 4. Age-specific Genetic Associations With Breast Cancer Risk in the Validation Set
SNP/GenotypeYoung (Ages 30–44 Years)Middle (Ages 45–54 Years)Old (Ages 55–69 Years)
No. (%)OR (95% CI)No. (%)OR (95% CI)No. (%)OR (95% CI)
CasesControlsCasesControlsCasesControls

  • OR indicates odds ratio; 95% CI, 95% confidence interval; SNP, single nucleotide polymorphism; (Ref), reference genotype; UTR, untranslated region.

  • * .05 < P < .1.

  • †

    P ≤ .05.

COMT
 A/A48 (27)84 (24)1 (Ref)57 (31)94 (25)135 (26)76 (27)1
 G/A89 (50)173 (49)0.8 (0.5–1.3)95 (51)180 (48)0.9 (0.6–1.3)68 (49)140 (50)1.1 (0.6–1.7)
 G/G41 (23)95 (27)0.8 (0.4–1.3)33 (18)100 (27)0.5 (0.3–0.9)†34 (25)63 (23)1.2 (0.7–2.1)
CYP11B2
 T/T59 (33)99 (28)160 (32)125 (33)140 (29)93 (33)1
 C/T85 (47)183 (51)0.8 (0.5–1.2)91 (49)178 (47)1.1 (0.7–1.6)72 (53)141 (50)1.2 (0.7–1.9)
 C/C35 (20)75 (21)0.8 (0.5–1.3)36 (19)74 (20)1 (0.6–1.7)25 (18)47 (17)1.2 (0.7–2.3)
CYP19 (3′UTR)
 T/T49 (28)99 (28)157 (31)109 (29)136 (26)66 (24)1
 C/T81 (45)176 (50)0.9 (0.6–1.4)90 (48)176 (47)0.9 (0.6–1.5)60 (43)151 (54)0.7 (0.4–1.2)
 C/C47 (27)78 (22)1.2 (0.7–2)39 (21)87 (23)0.9 (0.5–1.4)43 (31)62 (22)1.3 (0.7–2.2)
PGR (V660L)
 G/G116 (65)248 (70)1127 (69)266 (72)198 (70)202 (72)1
 T/G58 (33)97 (27)1.3 (0.9–1.9)51 (27)94 (25)1.1 (0.8–1.7)38 (27)74 (26)1.1 (0.7–1.7)
 T/T4 (2)10 (3)0.8 (0.3–2.8)7 (4)10 (3)1.5 (0.6–3.9)4 (3)5 (2)1.7 (0.4–6.3)
UGT1A7 (K131R)
 AA/AA63 (36)138 (40)186 (48)138 (38)162 (47)117 (43)1
 AA/CG84 (48)152 (44)1.2 (0.8–1.8)69 (38)169 (46)0.7 (0.4–0.9)†61 (46)121 (44)1 (0.6–1.5)
 CG/CG29 (16)57 (16)1.1 (0.6–1.9)25 (14)59 (16)0.7 (0.4–1.2)9 (7)37 (13)0.5 (0.2–1)†

Nonparametric Evaluation of ASGAs

The ASGAs were delineated further using a sliding window strategy that employed decade increments to analyze ORs nonparametrically. Figure 1 shows the relation between OR and age for SNP genotypes from the 4 validated associations (CYP11B2, UGT1A7, CYP19, and PGR). The homozygous C/C genotype of CYP11B2 is associated with a gradual increase in the OR from 0.5 at approximately age 35 years to 1.7 at approximately age 65 years (Fig. 1A). The trend appears linear with a correlation coefficient (R2) = 0.95. In contrast, the OR for the heterozygous C/T genotype does not appear to vary with age until age 50 years, and then exhibits a local increase from 1.0 to 1.7 from ages 50 years to 69 years (local R2 = 0.95). The ORs associated with the CG/CG and AA/CG genotypes of UGT1A7 exhibit a similar gradual decline over age from age 35 years to age 65 years (R2 = 0.88) (Fig. 1B). For CYP19, the ORs of both the T/T and C/T genotypes increase linearly beginning at age 50 years and continuing until age 69 years (R2 = 0.95 for C/T and R2 = 0.69 for C/C) (Fig. 1C). Individuals with the T/T genotype at the PGR locus exhibit an elevated BC risk only in middle age (Fig. 1D).

Figure 1. These charts illustrate age-specific genetic association trends for single nucleotide polymorphism genotypes in steroid hormone pathway genes. Odd ratios (ORs) and P values for 10-year sliding windows were calculated individually for the least common homozygous and heterozygous genotypes by using the most common homozygous genotype as a reference. The relationship between OR (y-axis) and age for each increment (x-axis) is shown for the C/C and C/T genotypes of cytochrome P450 XIB2 (CYP11B2) (A), the CG/CG and AA/CG genotypes of uridine phosphorylase glycosyltransferase (1A7 UGT1A7) (B), the C/C and C/T genotypes of cytochrome P450 19 (CYP19) (C), and progesterone receptor (PGR) (D). The genotypes and number of individuals analyzed are shown at the top of each chart and are shaded to match their respective lines on the plots. Estimated ORs were plotted against the middle age point within each age interval, and points are designated with solid triangles if the corresponding, unadjusted P value was ≤ .05 and with open triangles if .05 < P < .1. The temporal pattern was examined for the trend, and the correlation coefficient (R2) and age interval are shown for each trend line. The solid horizontal line on each plot shows reference (Ref) to an OR = 1. 3′UTR indicates 3′-untranslated region.

Download figure to PowerPoint

thumbnail image

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

It long has been recognized that SH exposure, especially to estrogens, contributes significantly to BC risk in a manner that is dependent on menopausal status. The levels of specific hormones associated with premenopausal risk differ from those associated with postmenopausal risk.17, 42–45 Epidemiologic factors that are surrogate markers for lifetime exposure to estrogen also have a differential impact on BC risk in relation to age and menopausal status.12–15, 46 For some SH-related factors, the risk associated with BC reverses, depending on menopausal status. Both nulliparity and obesity have been associated with lower BC risk in premenopausal women and with increased BC risk later in life.14, 47–50 Because SH-related factors affect BC risk differentially with age or menopausal status, perhaps it is not surprising that the association of certain SNPs in SH pathway genes with BC risk were similarly influenced. Our most striking finding was that, for SNP genotypes in CYP11B2 and UGT1A7, not only does the magnitude of BC risk associations for some SH pathway gene SNPs vary with age, but the direction of risk changed with age. In addition, associations confined to the middle-aged group were observed for SNP genotypes in CYP19 and PGR. The validity of these unanticipated ASGAs required careful scrutiny, because the discovery set findings were based on statistical assessment of P values without correcting for multiple comparisons. To minimize the possibility of false-positive discoveries, an independent data set was analyzed, and the ASGAs that were discovered for CYP11B, UGT1A7, CYP19, and PGR were confirmed.

The observed patterns of ASGAs become evident only by analysis of a study cohort large enough to permit age stratification. Clearly, these significant ASGAs were not apparent in overall analyses, either because they were balanced by opposite risk associations that occurred in the young group versus the older group (CYP11B2, UGT1A7) or because they were diluted by lack of association in the entire sample set (CYP19, PGR). The size of our study enabled us to characterize ASGAs further by using nonparametric analyses to evaluate ORs continuously over all ages ranging from 35 years to 65 years. These results clearly supported the conclusion that ORs associated with these SNPs vary significantly with age.

Our findings concerning ASGAs with SNPs in CYP11B2, UGT1A7, CYP19, and PGR may indicate an informative new way to evaluate genetic associations with BC risk. Prior studies of polymorphisms in SH pathway genes suggested that some were associated differentially with BC risk, depending on menopausal status.18–20 Although menopausal status certainly is correlated with age, examining the type of age-specific penetrance that we describe may be more informative. Considering the variation in a woman's hormone status with age, perhaps our findings are not surprising; however, there are no clear indications in the current literature why the phenomenon occurs in these particular genes. CYP11B2 is a key enzyme that ultimately converts 11-deoxycorticosterone to aldosterone. Earlier studies had reported an association of the C/C genotype with increased risk of type II diabetes, which, in turn, has been associated with increased risk of BC in postmenopausal women.51–53 Our finding of an increased risk of BC associated with the C/C genotype in the older age group is consistent with these earlier studies. Although the function of UGT1A7 in conjugating a wide variety of substrates, including steroids, environmental mutagens, and pharmaceuticals, suggests its potential to influence BC risk,54 this gene has not been investigated previously in BC, but the low-activity allele has been associated with increased colon and orolaryngeal cancer risk.34–36 The CYP19 gene encodes the terminal enzyme in the estrogen biosynthetic pathway, and 1 study has reported an overall protective effect of the C/C polymorphism in BC.29 Finally, the missense polymorphism in PGR is in complete linkage disequilibrium with several other polymorphisms in the gene, including the progesterone receptor polymorphism PROGINS,30, 31 and has been associated with decreased BC risk, especially in young premenopausal women in some studies,31 but not in others.55

Hopefully, these intriguing results will provide an impetus for other investigators to search for ASGAs in breast cancer and in other cancers. This study focused on white women, and further studies will be necessary to determine the relevance of ASGAs in other ethnicities. If ASGAs are validated in independent studies, then they could have significant impact on the application of SNP technologies for use in BC risk prediction and in cancer screening and prevention. ASGAs may explain the lack of reproducibility of genetic associations across different studies,56 either because of variability in the mean age of the sample sets that are studied or because overall analyses in a study average the risk across all ages. Thus, consideration of the age distribution of participants may have significant implications in the design of future SNP association studies. Certainly, when designing studies to develop risk-predictive models, there is a clear need to recruit clinical populations that are large enough to permit examination of potential ASGAs and to replicate the studies in a target population with a similar age distribution. Finally, when developing a genetic test for BC risk, there must be a suitable age distribution in the target population to be sure that ASGAs are not missed in the analysis. Indeed, we hope that our discovery brings us a step closer to the implementation of personalized medicine and accurately assessing BC risk in all women.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We thank Drs. Linda Thompson and John Mulvihill for critical comments; Laura Blaylock, Jean Kay, John Seagraves, and Billie Palacios for technical assistance; and the many clinicians and their patients who participated in the study.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES
  • 1
    Althuis MD, Dozier JM, Anderson WF, Devesa SS, Brinton LA. Global trends in breast cancer incidence and mortality 1973–1997. Int J Epidemiol. 2005; 34: 405–412.
  • 2
    Kelsey JL, Bernstein L. Epidemiology and prevention of breast cancer. Annu Rev Public Health. 1996; 17: 47–67.
  • 3
    Bernstein L. Epidemiology of endocrine-related risk factors for breast cancer. J Mammary Gland Biol Neoplasia. 2002; 7: 3–15.
  • 4
    Newman B, Mu H, Butler LM, Millikan RC, Moorman PG, King MC. Frequency of breast cancer attributable to BRCA1 in a population-based series of American women. JAMA. 1998; 279: 915–921.
  • 5
    Malone KE, Dailing JR, Neal C, et al. BRCA1 mutations and breast cancer in the general population: analyses in women before age 35 years and in women before age 45 years with first-degree family history. JAMA. 1998; 279: 922–929.
  • 6
    Pharoah PD, Antoniou A, Bobrow M, Zimmern RL, Easton DF, Ponder BA. Polygenic susceptibility to breast cancer and implications for prevention. Nat Genet. 2002; 31: 33–36.
  • 7
    Ponder BA, Antoniou A, Dunning A, Easton DF, Pharoah PD. Polygenic inherited predisposition to breast cancer. Cold Spring Harb Symp Quant Biol. 2005; 70: 35–41.
  • 8
    Dunning AM, Healey CS, Pharoah PD, et al. A systematic review of genetic polymorphisms and breast cancer risk. Cancer Epidemiol Biomarkers Prev. 1999; 8: 843–854.
  • 9
    de Jong MM, Nolte IM, te Meerman GJ, et al. Genes other than BRCA1 and BRCA2 involved in breast cancer susceptibility. J Med Genet. 2002; 39: 225–242.
  • 10
    Kristensen VN, Borresen-Dale AL. Molecular epidemiology of breast cancer: genetic variation in steroid hormone metabolism. Mutat Res. 2000; 462: 323–333.
  • 11
    Thompson PA, Ambrosone C. Molecular epidemiology of genetic polymorphisms in estrogen metabolizing enzymes in human breast cancer. J Natl Cancer Inst Monogr. 2000; 27: 125–134.
  • 12
    Stravraky K, Emmons S. Breast cancer in premenopausal and postmenopausal women. J Natl Cancer Inst. 1974; 53: 644–654.
  • 13
    Paffenbarger RSJr, Kampert JB, Chang H-G. Characteristics that predict risk of breast cancer before and after the menopause. Am J Epidemiol. 1980; 112: 258–268.
  • 14
    Kampert JB, Whittemore AS, Paffenbarger RSJr. Combined effect of childbearing, menstrual events, and body size on age-specific breast cancer risk. Am J Epidemiol. 1988; 128: 962–979.
  • 15
    Clavel-Chapelon F, E3N-EPIC Group. Differential effects of reproductive factors on the risk of pre- and postmenopausal breast cancer. Results from a large cohort of French women. Br J Cancer. 2002; 86: 723–727.
  • 16
    Hankinson SE, Willett WC, Manson JE, et al. Plasma sex steroid hormone levels and risk of breast cancer in postmenopausal women. J Natl Cancer Inst. 1998; 90: 1292–1299.
  • 17
    Kaaks R, Berrino F, Key T, et al. Serum sex steroids in premenopausal women and breast cancer risk within the European Prospective Investigation into Cancer and Nutrition (EPIC). J Natl Cancer Inst. 2005; 97: 755–765.
  • 18
    Thompson PA, Shields PG, Freudenheim JL, et al. Genetic polymorphisms in catechol-O-methyltransferase, menopausal status, and breast cancer risk. Cancer Res. 1998; 58: 2107–2110.
  • 19
    Wedren S, Rudqvist TR, Granath F, et al. Catechol-O-methyltransferase gene polymorphism and post-menopausal breast cancer risk. Carcinogenesis. 2003; 24: 681–687.
  • 20
    Zhu Y, Brown HN, Zhang Y, Holford TR, Zheng T. Genotypes and halotypes of the methyl-CpG-binding domain 2 modify breast cancer risk dependent upon menopausal status. Breast Cancer Res. 2005; 7: R745–R752.
  • 21
    Bergman-Jungestrom M, Gentile M, Lundin AC, Wingren S. Association between CYP17 gene polymorphism and risk of breast cancer in young women. Int J Cancer. 1999; 84: 350–353.
    Direct Link:
  • 22
    Spurdle AB, Hopper JL, Dite GS, et al. CYP17 promoter polymorphism and breast cancer in Australian women under age forty years. J Natl Cancer Inst. 2000; 92: 1674–1681.
  • 23
    Aston CE, Ralph DA, Lalo DP, et al. Oligogenic combinations associated with breast cancer risk. Hum Genet. 2005; 116: 208–221.
  • 24
    Lachman HM, Papolos DF, Saito T, Yu YM, Szumlanski CL, Weinshilboum RM. Human catechol-O-methyltransferase pharmacogenetics: description of a functional polymorphism and its potential application to neuropsychiatric disorders. Pharmacogenetics. 1996; 6: 243–250.
  • 25
    Lavigne JA, Helzlsouer KJ, Huang HY, et al. An association between the allele coding for a low activity variant of catechol-O-methyltransferase and the risk for breast cancer. Cancer Res. 1997; 57: 5493–5497.
  • 26
    Connell JM, Fraser R, MacKenzie SM, et al. The impact of polymorphisms in the gene encoding aldosterone synthase (CYP11B2) on steroid synthesis and blood pressure regulation. Mol Cell Endocrinol. 2004; 217: 243–247.
  • 27
    Barbato A, Russo P, Siani A, et al. Aldosterone synthase gene (CYP11B2) C-344T polymorphism, plasma aldosterone, renin activity and blood pressure in a multi-ethnic population. J Hypertens. 2004; 22: 1895–1901.
  • 28
    Listgarten J, Damaraju S, Poulin B, et al. Predictive models for breast cancer susceptibility from multiple single nucleotide polymorphisms. Clin Cancer Res. 2004; 10: 2725–3727.
  • 29
    Kristensen VN, Harada N, Yoshimura N, et al. Genetic variants of CYP19 (aromatase) and breast cancer risk. Oncogene. 2000; 19: 1329–1333.
  • 30
    Agoulnik IU, Tong XW, Fischer DC, et al. A germline variation in the progesterone receptor gene increases transcriptional activity and may modify ovarian cancer risk. J Clin Endocrinol Metab. 2004; 89: 6430–6347.
  • 31
    De Vivo I, Hankinson SE, Colditz GA, Hunter DJ. The progesterone receptor Val660Leu polymorphism and breast cancer risk. Breast Cancer Res. 2004; 6: R636–R639.
  • 32
    Rosenblum JS, Gilula NB, Lerner RA. On signal sequence polymorphisms and diseases of distribution. Proc Nat Acad Sci USA. 1996; 93: 4471–4473.
  • 33
    Mitrunen K, Sillanpaa P, Kataja V, Eskelinen M, et al. Association between manganese superoxide dismutase (MnSOD) gene polymorphism and breast cancer risk. Carcinogenesis. 2001; 22: 827–829.
  • 34
    Strassburg CP, Vogel A, Kneip S, Tukey RH, Manns MP. Polymorphisms of the human UDP-glucuronosyltransferase (UGT)1A7 gene in colorectal cancer. Gut. 2002; 50: 851–856.
  • 35
    Zheng Z, Park JY, Guillemette C, Schantz SP, Lazarus P. Tobacco carcinogen-detoxifying enzyme UGT1A7 and its association with orolaryngeal cancer risk. J Natl Cancer Inst. 2001; 93: 1411–1418.
  • 36
    Van der Logt EM, Bergevoet SM, Roelofs HM, et al. Genetic polymorphisms in UDP-glucuronosyltransferases and glutathione S-transferases and colorectal cancer risk. Carcinogenesis. 2004; 25: 2407–2415.
  • 37
    Hartl DL, Clark AG. Principles of Population Genetics. Sunderland, Mass: Sinauer Associates, Inc.; 1997.
  • 38
    Everitt BS. The Analysis of Contingency Tables. ed. 2. London: Chapman and Hall; 1986.
  • 39
    Breslow N E, Day NE. Statistical Methods in Cancer Research. Lyon: International Agency for Research on Cancer 1980.
  • 40
    Nichols HB, Trentham-Dietz A, Hampton JM, et al. From menarche to menopause: trends among US women born from 1912 to 1969. Am J Epidemiol. 2006; 164: 1003–1011.
  • 41
    Overlie I, Moen MH, Morkrid L, Skjaeraasen JS, Holte A. The endocrine transition around menopause—a five years prospective study with profiles of gonadotropins, estrogens, androgens and SHBG among healthy women. Acta Obstet Gynecol Scand. 1999; 78: 642–647.
    Direct Link:
  • 42
    Tworoger S, Missmer S, Eliassen A, et al. The association of plasma DHEA and DHEA sulfate with breast cancer risk in predominantly premenopausal women. Cancer Epidemiol Biomarkers Prev. 2006; 15: 967–971.
  • 43
    Missmer S, Eliassen A, Barbieri R, Hankinson S. Endogenous estrogen, androgen, and progesterone concentrations and breast cancer risk among postmenopausal women. J Natl Cancer Inst. 2004; 96: 1856–1865.
  • 44
    Eliassen AH, Missmer SA, Tworoger SS, et al. Endogenous steroid hormone concentrations and risk of breast cancer among premenopausal women. J Natl Cancer Inst. 2006; 98: 1406–1415.
  • 45
    Key T, Appleby P, Barnes I, Reeves G. Endogenous sex hormones and breast cancer in postmenopausal women: reanalysis of nine prospective studies. J Natl Cancer Inst. 2002; 94: 606–616.
  • 46
    Gail M, Brinton L, Byar D, et al. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst. 1989; 81: 1879–1886.
  • 47
    Anderson W, Matsuno R, Sherman M, et al. Estimating age-specific breast cancer risks: a descriptive tool to identify age interactions. Cancer Causes Control. 2007 Jan 9; [Epub ahead of print].
  • 48
    Largent J, Ziogas A, Anton-Culver H. Effect of reproductive factors on stage, grade and hormone receptor status in early-onset breast cancer. Breast Cancer Res. 2005; 7: R541–R554.
  • 49
    Huang Z, Hankinson S, Colditz G, et al. Dual effects of weight and weight gain on breast cancer risk. JAMA. 1997; 278: 1407–1411.
  • 50
    Lahmann P, Hoffmann K, Allen N, et al. Body size and breast cancer risk: findings from the European Prospective Investigation into Cancer and Nutrition (EPIC). Int J Cancer. 2004; 111: 762–71.
    Direct Link:
  • 51
    Ranade K, Wu KD, Risch N, et al. Genetic variation in aldosterone synthase predicts plasma glucose levels. Proc Natl Acad Sci USA. 2001; 98: 13219–13224.
  • 52
    Michels KB, Solomon CG, Hu FB, et al. Type 2 diabetes and subsequent incidence of breast cancer in the Nurses' Health Study. Diabetes Care. 2003; 26: 1752–1758.
  • 53
    Lipscombe LL, Goodwin PJ, Zinman B, McLaughlin JR, Hux JE. Diabetes mellitus and breast cancer: a retrospective population-based cohort study. Breast Cancer Res Treat. 2006; 98: 349–356.
  • 54
    Ogura K, Ishikawa Y, Kaku T, et al. A. Quaternary ammonium-linked glucuronidation of trans-4-hydroxytamoxifen, an active metabolite of tamoxifen, by human liver microsomes and UDP-glucuronosyltransferase 1A4. Biochem Pharmacol. 2006; 71: 1358–1369.
  • 55
    Feigelson HS, Rodriguez C, Jacobs EJ, Diver WR, Thun MJ, Calle EE. No association between the progesterone receptor gene +331G/A polymorphism and breast cancer. Cancer Epidemiol Biomarkers Prev. 2004; 13: 1084–1085.
  • 56
    Ioannidis JP, Ntzani EE, Trikalinos TA, Contopoulos-Ioannidis DG. Replication validity of genetic association studies. Nat Genet. 2001; 29: 306–309.
Get PDF (173K)