Open Access

Future possibilities in the prevention of breast cancer: Role of genetic variation in breast cancer prevention

  • Heather Spencer Feigelson1 and
  • Brian E Henderson2
Breast Cancer Research20002:277

https://doi.org/10.1186/bcr69

Received: 23 November 1999

Accepted: 30 March 2000

Published: 1 August 2000

Abstract

Risk factors for breast cancer are related to endogenous hormonesand reproductive events. As such, traditional cancer prevention strategies arenot easily applicable. Tamoxifen and other selective estrogen receptormodulators (SERMs) offer a new preventive strategy for some high-risk women,but have not yet been shown to be efficacious for all women. New tools toidentify high-risk women are needed. One such tool is the development of amultigenic model of breast cancer susceptibility that can be used to screenwomen in order to identify those who carry a combination of alleles that putsthem at significantly increased risk.

Keywords

breast cancer epidemiology polymorphism prevention risk factors

Introduction

The traditional view of public health prevention strategies considers three broad categories. Primary prevention involves activities that are aimed at reducing and removing agents that increase risk. Secondary prevention includes early detection activities that are designed to reduce mortality. Finally, tertiary prevention aims to minimize or reduce long-term disability and suffering. Cancer prevention usually implies activities involved in primary or secondary prevention. For many cancer sites, we have made significant contributions in prevention. For example, the Pap smear has had a great impact on cervical cancer, and refrigeration and other safe food preservation practices have markedly reduced the incidence of stomach cancer. Largely because of the underlying hormonal etiology, however, primary prevention strategies for breast cancer have been limited.

Breast cancer risk factors

The biggest determinants of breast cancer risk are related to endogenous hormone levels and major reproductive events, and thus do not lend themselves to traditional prevention strategies. Table 1 lists the established breast cancer risk factors. Those with the greatest impact on risk are listed first, and unfortunately are the most difficult to modify by traditional public health measures. For example, surgical removal of the ovaries can hasten the onset of menopause, but is not considered a reasonable approach to reduce the risk of breast cancer. Age at menarche and the establishment of regular menstrual cycles may be delayed by vigorous physical activity and possibly diet, but rarely is breast cancer prevention a concern before menarche occurs.
Table 1

Breast cancer risk factors

 

Modifiable using traditional intervention strategies

 

Yes

No

Factors that confer highest risk

  

   Age

 

   Inherited susceptibility

 

   Atypical hyperplasia

 

   Endogenous hormone levels

 

   Early age at menarche

 

   Late age at menopause

 

   Late age at first pregnancy

 

Factors with small effects on risk

  

   Physical activity

 

   Lactation

 

   Weight/body mass

 

   HRT

 

   Alcohol consumption

 

Possible risk factors that confer small risk

  

   Oral contraceptive use

 

   High dietary fat

 

   Low dietary fiber

 

Limiting the use of alcohol, hormone replacement therapy (HRT), and oral contraceptives may reduce the risk of breast cancer, but the impact of these factors on breast cancer risk are modest. Furthermore, these agents have significant beneficial effects on the risks of other chronic diseases, and these risk-benefit ratios must be carefully weighed. Prolonged lactation may offer a small reduction in breast cancer risk, but is not likely to be readily accepted as a risk reduction strategy. The benefit of physical activity after menarche for the purpose of reducing breast cancer risk is unclear. Because physical activity can curb weight gain, it can be seen as an important preventive measure because postmenopausal obesity increases breast cancer risk. Although the literature is mixed on whether dietary fat and fiber are important in breast cancer etiology, these factors are readily modifiable. However, the effect of moderate dietary changes on breast cancer risk are likely to be small, at best.

Selective estrogen receptor modulators for breast cancerprevention

The newest hope in the area of breast cancer prevention came from the findings of a highly publicized tamoxifen trial reported in 1998 [1]. These data provided the first information from a randomized clinical trial to support the hypothesis that breast cancer can be prevented among high-risk women. Tamoxifen reduced the risk of invasive breast cancer by 49% (P <0.00001) and noninvasive breast cancer by 50% (P <0.002). The effect was limited to estrogen receptor-positive tumors. Two smaller trials [2,3] failed to find a benefit with tamoxifen; however, these seemingly different results may be due largely to differences in the age and family history of the populations studied, and (at least in one study) poor compliance with the treatment regimen [4].

At present, we do not know whether the Breast Cancer Prevention Trial (BCPT, NSABP-P1) [1] tamoxifen findings are generalizable to all groups of women, such as those with BRCA mutations or women from ethnic minority groups. We do not know whether the decreased incidence observed among tamoxifen users was due to a delay in the development of occult tumors that would ultimately be diagnosed after cessation of tamoxifen use. If so, how long can tamoxifen be safely administered. Finally, none of the studies to date have provided reliable data on mortality. The risks of tamoxifen chemoprevention must be weighed carefully against its benefits. Although tamoxifen reduces the risk of certain breast cancers and bone fractures, it increases the risk of endometrial cancer. Whether tamoxifen can reduce the incidence of heart disease is unknown, and data from the BCPT [1] suggest that the drug may increase the risk of stroke and cataracts. (See Gail et al [5] for a complete discussion of the risks and benefits of tamoxifen chemoprevention.)

Raloxifene is another SERM that has recently been suggested to reduce the risk of estrogen receptor-positive breast cancer, but without the subsequent increased risk of endometrial cancer [6]. However, that study also needs further confirmation and careful follow up.

Multigenic model of breast cancer susceptibility

What more can be done to reduce the risk of breast cancer? A long-standing strategy of risk reduction involves targeting intervention efforts toward high-risk groups, but who is at high risk for breast cancer? One in eight women in the USA will develop breast cancer, and we must find more effective ways to identify these women early. One possibility is to develop a multigenic model of breast cancer susceptibility, and then screen women to determine who carries a combination of alleles that puts them at significantly increased risk. Given the large and compelling body of epidemiologic and experimental evidence that implicates estrogens in the etiology of human breast cancer, we proposed a multigenic model of breast cancer predisposition that includes genes that are involved in estrogen biosynthesis and intracellular binding [7*]. We hypothesized that functionally relevant sequence variants in such genes would act together, and also interact with well-known hormonally related risk factors, to define a high-risk profile for breast cancer. The term 'functionally relevant sequence variants' refers to those mutations or polymorphisms that can be shown through laboratory experiments to alter encoded protein structure or function, interaction with other proteins, or half-life and stability within the cell.

We assumed that variation in genes that encode critically important enzymes in estradiol biosynthesis would, individually, result in only modest differences in the rate of biosynthesis. Presumably, there would be limited evolutionary tolerance for major variation in hormone synthesis, which could disrupt reproductive ability. However, a combination of genes, each with minor variation in expressed activity, could provide a degree of separation of risk that would be clinically useful. These small variations could result in a large cumulative effect after several decades. For example, the model of breast tissue age by Pike et al [8**] demonstrates that a 20% difference in levels of circulating estrogen can result in a more than twofold increase in lifetime breast cancer risk.

As shown in Figure 1, our original presentation of this model included three genes of interest: the 17β -hydroxysteroid dehydrogenase 1 gene (HSD17B1), the cytochrome p459c17α gene (CYP17), and the estrogen receptor alpha gene (ESR1). We later added the aromatase gene (CYP19) to the model, although a functional polymorphism was not known. Huang et al [9*] recently reported their findings of a similar model with the estrogen metabolizing genes CYP17; CYP1A1, which participates in estrogen hydroxylation; and the catechol-O-methyltransferase gene (COMT), which encodes the enzyme that is responsible for O-methylation leading to inactivation of catechol estrogen. The numerous studies of other polymorphisms are reviewed elsewhere [10].
Figure 1

Estrogen metabolism in the ovaries and breast epithelium and fourcandidate genes that may play a role in breast cancer etiology. The genes ofinterest are the cytochrome P450c17α gene (CYP17), the aromatasecytochrome P450 gene (CYP19), the 17β -hydroxysteroiddehydrogenase 1 gene (HSD17B1), and the estrogen receptor alpha gene(ESR1).

At present, data to support a role for CYP17 are most compelling. As summarized below, this gene has been shown to be associated with the risk of breast cancer, serum hormone levels, age at menarche, and use of HRT.

CYP17 encodes the cytochrome p450c17α enzyme, which functions at key branch points in human steroidogenesis [11]. The 5' -untranslated region of CYP17 contains a single base-pair polymorphism that is 34 bp upstream from the initiation of translation, and 27 bp downstream from the transcription start site [12]. This base-pair change creates a recognition site for the MspAI restriction enzyme, and has been used to designate two alleles: A1 (the published sequence) and A2.

An association between risk for breast cancer and this CYP17 polymorphism was first shown in 1997 [13]. In a case-control study of incident breast cancer among Asian, African-American and Latina women [13], we found a 2.5-fold increased risk of advanced breast cancer associated with the CYP17 A2 allele. That study also presented preliminary evidence suggesting that the CYP17 gene may be associated with age at menarche. The reduced risk of breast cancer associated with a later age at menarche was largely limited to A1/A1 women (odds ratio 0.47, 95% confidence 0.22-0.98 for breast cancer and later age at menarche) compared with women who carried the A2 allele (odds ratio 0.80, 95% confidence interval 0.51-1.27).

These results suggested that serum hormone levels may differ by CYP17 genotype. We pursued this finding in a separate study and found that the CYP17 genotype was associated with serum estradiol and progesterone levels among young nulliparous women [14]. As shown in Figure 2, serum estradiol measured around day 11 of the menstrual cycle was 11 and 57% higher (P =0.04), respectively, among women heterozygous and homozygous for the CYP17 A2 allele compared with A1/A1 women. Similarly, around cycle day 22, serum estradiol was 7 and 28% higher (P = 0.06) and progesterone (Fig. 3) was 24 and 30% higher (P =0.04). These data provide direct evidence of genetic control of serum hormone levels, but the sample size was small and the results need confirmation.
Figure 2

Geometric mean serum estradiol (E2) concentrations among youngnulliparous women on days 11 and 22 of the menstrual cycle by CYP17genotype. Day 11 E2, P for trend 0.04; day 22 E2, P for trend0.06.

Figure 3

Geometric mean serum progesterone concentrations among youngnulliparous women on day 22 of the menstrual cycle by CYP17 genotype(P for trend 0.04).

Finally, we recently reported that CYP17 is associated with use of HRT [15*]. Among 749 postmenopausal women aged 44-75 years at baseline randomly selected from a larger multiethnic cohort, women who carry the CYP17 A2/A2 genotype were about half as likely as women with the A1/A1 genotype to be current HRT users (odds ratio 0.52, 95% confidence interval 0.31-0.86). This association was present in all four of the racial/ethnic groups that were included in the study (white, African-American, Latino, and Japanese), and for women above and below the median weight of 150 pounds. Presumably, women with the A2/A2 genotype have fewer indications for HRT use due to their higher level of endogenous hormones.

Since this original study of CYP17 was published [13], at least six other studies have reported on CYP17 and breast cancer (Table 2) [9*,16,17,18,19*,20]. The results of these subsequent studies were largely negative and suggest heterogeneity by ethnicity. However, there may be several reasons for the discrepant results.
Table 2

Summary of published studies of CYP17 and breast cancer

       

CYP17 (A2/A2 versus A1/A1) and breast cancer

 
       

All cases

Advanced cases

 

Ref

Study design

Race/ethnicity

Age of patients (years)

Age of control individuals (years)

Cases (n)

Advanced cases (n)

OR

95% CI

OR

95% CI

HRT use considered?

[7*]

Nested case-control

Afro-American, Asian, Latina

63.0 ± 8.4

61.4 ± 8.3

174*

40

1.32

0.87-2.0

2.5

1.07-5.94

No

[16]

Case-control

White

<55

45-74

835

24

1.17

0.92-1.49

0.88

0.38-2.01

No

[17]

Nested case-control

White

60.4 ± 11.7

60.2 ± 11.5

115

30

0.89

0.41-1.95

1.39

0.26-7.28

Yes

[18]

Case-control

Afro-American

Not given

Not given

76

21

1.40

0.44-4.38

0.6

0.1-4.0

No

  

Latina

  

20

7

1.93

0.75-5.01

0.2

0.0-1.3

 
  

White

  

27

10

0.80

0.45-1.43

1.7

0.6-5.1

 

[19*]

Nested case-control

White

58.3 ± 7.1

 

464

107

0.91

0.61-1.34

0.84

0.54-1.32

Yes

[9*]

Case-control

Taiwanese

Not given

Not given

150

Not given

1.28

0.73-2.27

  

Yes

[20]

Case-control

White

59 (27-91)

20-44

510

93§

0.94

0.54-1.65

1.38§

0.62-3.06

No

Ages are expressed as mean ± standard error, mean (range), orrange. *Total cases, not specified by ethnicity;odds ratio (OR) reflects A1/A2 + A2/A2 combined;age not specified by case-control status;§advanced cases over age 55 years.

The study by Dunning et al [16] is the largest to date but had few advanced cases, and it did not examine the possible confounding effects of HRT. Four other smaller studies [9*,17,18,20] found a modest elevation in breast cancer risk with the CYP17 A2 allele in some subgroups, but those findings did not reach statistical significance. Kristensen et al [20] suggested that the effect of CYP17 might be limited to older patients (ie those aged over 55 years at diagnosis), which also may explain some of the inconsistent results.

Only one study, that by Haiman et al [19*], was both significantly larger than our original study and gave adequate consideration to potential confounding. Those investigators did not find an association between CYP17 and breast cancer. However, their results are compatible with the potential modification of breast cancer risk due to late age of menarche. As in our study (and one other [18]) the protective effect of later onset of menarche was limited to women with the A1/A1 genotype (odds ratio 0.57, 95% confidence interval 0.36-0.90 for breast cancer and later age at menarche). Furthermore, among 297 postmenopausal women, those with the A2/A2 genotype had statistically significantly elevated levels of estrone (+14.2%, P =0.01) and dehydroepiandrosterone (+14.4%, P =0.02), and modest nonsignificant elevations in estradiol (+18.8%, P = 0.08), testosterone (+8.6%, P =0.34), androstenedione (+17.1%, P =0.06), and dehydroepiandrosterone sulfate (+7.2%, P =0.26) compared with women with the A1/A1 genotype. That study did not show evidence of an association between CYP17 and HRT use. However, the patterns of HRT use among a cohort of predominately white nurses may be substantially different from patterns of use among women in our multi-ethnic cohort [15*]. Additional studies are needed, but they must be of sufficient size and quality to evaluate adequately the role of CYP17 in advanced breast cancer while examining the influence of HRT use and other potentially important confounders and effect modifiers.

An important piece of information about CYP17 is still missing. What is the functional relevance of this polymorphism, or what is it marking? It was recently shown [20] that this T27C polymorphism in CYP17 converting the sequence CACT into CACC does not influence Sp-1 binding, as had been suggested based on its similarity to other known Sp-1 binding sequences. Gene function studies are needed to determine whether the A2 allele confers specifically a higher expression level of CYP17 [20]. Such studies will need to be carefully designed and evaluated because we would not expect these polymorphisms to result in a large difference in circulating hormone levels. Standard assays may not detect the relatively small differences in activity we would predict from the epidemiologic data.

Conclusion

The primary risk factors for breast cancer are not easily modifiable because they stem from prolonged endogenous hormonal exposures. The genetic basis of hormone levels as an important risk factor for breast cancer has only recently been recognized. It is now increasingly obvious that genetic susceptibility, acting through germline polymorphisms in metabolic genes, plays a critical role. Certainly the 'complete' multigenic model of breast cancer susceptibility would include several important genes. Further study is necessary to determine which genes consistently predict known breast cancer risk factors, serum hormone levels, and breast cancer.

The prevention and control of breast cancer is multidimensional. First, we must identify the key genetic components and use our knowledge of underlying genetic susceptibility to identify those women who are at highest risk. Second, we must identify effective lifestyle modifications, if any, that demonstrate a measurable and sustainable reduction in risk. Third, we must develop earlier screening tools using genetic markers of risk. Finally, we must develop suitable chemopreventive agents, such as tamoxifen, that can prevent tumor development or progression.

Authors’ Affiliations

(1)
American Cancer Society
(2)
University of Southern California
(3)
American Cancer Society, National Home Office

References

  1. Fisher B, Costantino JP, Wickerham DL, et al: Tamoxifen for prevention of breast cancer: report of the NationalSurgical Adjuvant Breast and Bowel Project P-1 Study. J Natl Cancer Inst. 1998, 90: 1371-1388. 10.1093/jnci/90.18.1371.View ArticlePubMedGoogle Scholar
  2. Powles T, Eeles R, Ashley S, et al: Interim analysis of the incidence of breast cancer in the RoyalMarsden Hospital tamoxifen randomised chemoprevention trial. Lancet. 1998, 352: 98-101.View ArticlePubMedGoogle Scholar
  3. Veronesi U, Maisonneuve P, Costa A, et al: Prevention of breast cancer with tamoxifen: preliminary findingsfrom the Italian randomised trial among hysterectomised women. Lancet. 1998, 352: 93-97.View ArticlePubMedGoogle Scholar
  4. Pritchard KI: Is tamoxifen effective in prevention of breast cancer?[commentary]. Lancet. 1998, 352: 80-81. 10.1016/S0140-6736(98)85002-2.View ArticlePubMedGoogle Scholar
  5. Gail MH, Costantino JP, Bryant J, et al: Weighing the risks and benefits of tamoxifen treatment forpreventing breast cancer. J Natl Cancer Inst. 1999, 91: 1829-1846. 10.1093/jnci/91.21.1829.View ArticlePubMedGoogle Scholar
  6. Cummings SR, Eckert S, Krueger KA, et al: The effect of raloxifene on risk of breast cancer in postmenopausalwomen. Results from the MORE randomized trial. JAMA. 1999, 281: 2189-2197. 10.1001/jama.281.23.2189.View ArticlePubMedGoogle Scholar
  7. Feigelson HS, Ross RK, Yu MC, Coetzee GA, Reichardt JK, Henderson BE: Genetic susceptibility to cancer from exogenous and endogenousexposures. J Cell Biochem. 1996, 25: 15-22. 10.1002/(SICI)1097-4644(1996)25+<15::AID-JCB2>3.0.CO;2-5.View ArticleGoogle Scholar
  8. Pike MC, Krailo MD, Henderson BE, Casagrande JT, Hoel DG: Hormonal risk factors, breast tissue age and age-incidence of breastcancer. Nature. 1983, 330: 767-770.View ArticleGoogle Scholar
  9. Huang C-S, Chern H-D, Chang K-J, Cheng C-W, Hsu S-M, Shen C-Y: Breast cancer risk associated with genotype polymorphisms of theestrogen-metabolizing genes CYP17, CYP1A1, and COMT: A multigenic study oncancer susceptibility. Cancer Res. 1999, 59: 4870-4875.PubMedGoogle Scholar
  10. Dunning AM, Healey CS, Pharoah PDP, Teare MD, Ponder BAJ, Easton DFA: Systematic review of genetic polymorphisms and breast cancer risk. Cancer Epidemiol Biomarkers Prev. 1999, 8: 843-854.PubMedGoogle Scholar
  11. Brentano ST, Picado-Leonard J, Mellon SH, Moore CC, Miller WL: Tissue-specific, cyclic adenosine3',5'-monophosphate-induced, and phorbol ester-repressedtranscription from the human P450c17 promoter in mouse cells. Mol Endocrinol. 1990, 4: 1972-1979.View ArticlePubMedGoogle Scholar
  12. Carey AH, Waterworth D, Patel K, et al: Polycystic ovaries and premature male pattern baldness areassociated with one allele of the steroid metabolism gene CYP17. Hum Mol Genet. 1994, 3: 1873-1876.View ArticlePubMedGoogle Scholar
  13. Feigelson HS, Coetzee GA, Kolonel LN, Ross RK, Henderson BEA: Polymorphism in the CYP17 gene increases the risk of breast cancer. Cancer Res. 1997, 57: 1063-1065.PubMedGoogle Scholar
  14. Feigelson HS, Shames LS, Pike MC, Coetzee GA, Stanczyk FZ, Henderson BE: Cytochrome P450c17alpha gene (CYP17) polymorphism is associated withserum estrogen and progesterone concentrations. Cancer Res . 1998, 58: 585-587.PubMedGoogle Scholar
  15. Feigelson HS, McKean-Cowdin R, Pike MC, et al: Cytochrome P450c17alpha gene (CYP17) polymorphism predicts use ofhormone replacement therapy. Cancer Res. 1999, 59: 3908-3910.PubMedGoogle Scholar
  16. Dunning AM, Healey CS, Pharoah PD, et al: No association between a polymorphism in the steroid metabolism geneCYP17 and risk of breast cancer. Br J Cancer. 1998, 77: 2045-2047.View ArticlePubMedPubMed CentralGoogle Scholar
  17. Helzlsouer KJ, Huang HY, Strickland PT, et al: Association between CYP17 polymorphisms and the development ofbreast cancer. Cancer Epidemiol Biomarkers Prev. 1998, 7: 945-949.PubMedGoogle Scholar
  18. Weston A, Pan CF, Bleiweiss IJ, et al: CYP17 genotype and breast cancer risk. Cancer Epidemiol Biomarkers Prev. 1998, 7: 941-944.PubMedGoogle Scholar
  19. Haiman CA, Hankinson SE, Spiegelman D, et al: The relationship between a polymorphism in CYP17 with plasma hormonelevels and breast cancer. Cancer Res. 1999, 59: 1015-1020.PubMedGoogle Scholar
  20. Kristensen VN, Haraldsen EK, Anderson KB, et al: CYP17 and breast cancer risk: the polymorphism in the 5'flanking area of the gene does not influence binding to Sp-1. Cancer Res. 1999, 59: 2825-2828.Google Scholar

Copyright

© Current Science Ltd 2000