Medical Hypothesis: BifunctionalGenetic-Hormonal Pathwaysto Breast CancerDevra Lee Davis,1 Nitin T. Telang," 2 Michael P. Osborne,2and H. Leon Bradlow21World Resources Institute, Washington, DC; 2Strang Cancer ResearchLaboratory, The Rockefeller University, New York, New York

As inherited germ line mutations, such as loss of BRCA1 or AT, account for less than 5% of allbreast cancer, most cases involve acquired somatic perturbations. Cumulative lifetime exposure

to bioavailable estradiol links most known risk factors (except radiation) for breast cancer. Basedon a series of recent experimental and epidemiologic findings, we hypothesize that the multistepprocess of breast carcinogenesis results from exposure to endogenous or exogenous hormones,including phytoestrogens that directly or indirectly alter estrogen metabolism. Xenohormones are

defined as xenobiotic materials that modify hormonal production; they can work bifunctionally,through genetic or hormonal paths, depending on the periods and extent of exposure. As forgenetic paths, xenohormones can modify DNA structure or function. As for hormonal paths, twodistinct mechanisms can influence the potential for aberrant cell growth: compounds can directlybind with endogenous hormone or growth factor receptors affecting cell proliferation or

compounds can modify breast cell proliferation altering the formation of hormone metabolites thatinfluence epithelial-stromal interaction and growth regulation. Beneficial xenohormones, such as

indole-3-carbinol, genistein, and other bioflavonoids, may reduce aberrant breast cell proliferation,and influence the rate of DNA repair or apoptosis and thereby influence the genetic or hormonalmicroenvironments. Upon validation with appropriate in vitro and in vivo studies, biologic markersof the risk for breast cancer, such as hormone metabolites, total bioavailable estradiol, and freeradical generators can enhance cancer detection and prevention. Environ Health Perspect105(Suppl 3):571-576 (1997)

Key words: estradiol metabolism, genetic and hormonal mechanisms, breast cancer,

xenohormones, xenoestrogen, environment

Introduction

As the most common cancer amongwomen in modern societies, breast canceris a complex and important disease. Theaverage patient who dies with the diseaseloses about two decades of life, so thatnearly 2 million women-years of life arelost annually to breast cancer in the UnitedStates and Europe (1). Although about87% of all cases survive for 5 years, nearly

half of all women die from to breast cancerby one decade after diagnosis (1-3).

About one-third of all cases of breastcancer can be attributed to recognized riskfactors. Neither changes in established riskfactors nor screening practices completelyaccount for the persisting 1% annualincrease in the incidence of breast cancer.Similarly, changes in risk factors or in

This paper was presented in part at the Workshop on Hormones, Hormone Metabolism, Environment, andBreast Cancer held 28-29 September 1995 in New Orleans, Louisiana. Manuscript received at EHP 6 June1996; manuscript accepted 2 August 1996.

The laboratory research programs have been funded in part by National Institutes of Health grants CA44741and CA 29502, Department of Defense grant DAMD 1 7-94-J-4208, and philanthropic support to the StrangCancer Prevention Center and to the World Resources Institute from Wallace-Global and the Jennifer AltmanFund.. The authors acknowledge the excellent editorial assistance of K.J. Brady.

Address correspondence to Dr. D.L. Davis, World Resources Institute, 1709 New York Avenue,Washington, DC 20006. Telephone: (202) 638-6300. Fax (202) 638-0036. E-mail: devra@wri.org

Abreviations used: ATM, ataxia telangiectasia, mutated; DMBA, 7,12-dimethylbenz[a]anthracene; E2, 17,Bestradiol; ER, estrogen receptor; Fapy-A, 4,6-diamino-5-formamidopyrimidine; -OH, hydroxyl; 2-OHEj, 2-hydrox-yestrone; 16a-OHE1, 16a-hydroxyestrone; PI-3-kinases, phosphatidylinositol-3-kinases; SHBG, sexhormone-binding globulin.

screening practices do not explain geo-graphic variations in prevalence of the dis-ease (1-4). Inherited germ cell mutationsoccur in about 5% of all cases and in about30% of cases under 40 years of age (3-6).

The common tie linking most of theestablished risk factors, aside from thesemutations, is greater cumulative exposureto bioavailable 17,-estradiol (E2) (4, 7-11).Bioavailable E2 is defined as a free hor-mone not bound to steroid hormone-binding globulin (SHBG) or weakly boundto albumin (9-12). Women with elevatedlevels of bioavailable E2 have a 2- to 4-foldexcess risk of breast cancer (10). Bio-available E2 can diffuse into cells and sub-sequently be taken into the nucleus whereit can bind to the estrogen receptor (ER).The hormone also can be converted in thecytoplasm into other biologically activemetabolites and free radicals (4,7). Inaddition, other hormones, such as andro-gens and progestagens, can influence theproduction and metabolism of E2(4,9-12). The hormone-SHBG complexbound to the cell surface receptor inducescAMP-mediated phosphorylation (12).Medical HypothesisWe have previously suggested thatcompounds functioning as xenoestrogensaffect the rate and type of estradiol metabo-lites formed. Xenoestrogens may also binddirectly with the ER to modulate breast cellproliferation and thereby influence thedevelopment of breast cancer and otherhormonally mediated diseases (9,11,12). Inthis report we expand the hypothesis toinclude a more detailed consideration ofpossible genetic-hormonal-environmentalinteractions (3,4), including the complexrelationship among estrogens, androgens,their antagonists, and other hormones inbreast cancer development.

Based on recent experimental andepidemiologic findings in this laboratoryand elsewhere (4,9,11,13), we hypothesizethat prenatal, adolescent, or midlife expo-sure to endogenous endocrine agents, xeno-hormones, or their metabolites can havebifunctional effects on the risk of developingbreast cancer.We also hypothesize that some xeno-

hormonal exposures can, through redoxcycling between estrogens and their corre-sponding quinones, yield reactive oxygenspecies that can cause structural oxidativedamage to DNA and increase rates of oxida-tive DNA base modifications (7,8,14-16).

Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997 571

DAVIS ET AL.

Lipid oxidation products may also functionas endogenous DNA damaging-agents(17). In addition, other types of reactivefunctions, such as methylation or phospho-rylation, can affect key functional regionsof DNA, induding cell cycle genes criticalfor cell proliferation, development, andgrowth. Exposure to xenohormonesthrough diet, pharmaceuticals, and envi-ronmental chemicals can alter theparenchymal environment, either by pro-moting already initiated breast cells intorelatively rapid proliferation or by imped-ing such growth, if the xenohormones areantioxidants, hormone antagonists, orantiangiogenics (18).

Figures 1 to 3 show the impact ofsteroid hormones on interacting geneticand hormonal pathways critical for breast

Xenohormones,endogenous hormones

Cyp450-dependentaromatase 17p-oxidoreductase

carcinogenesis and its prevention. INoccurring or synthetic xenohormoaffect the process of tumorigenic tmation at both genotoxic (initiatiepigenetic (postinitiational) levels.to the bifunctional effect of bioavaiis its enzymatic conversion to produdistinctive biological activity. The lmal enzymes, aromatase, 17p-oxictase and steroid hydroxylases Cyp41A2, iBi, and 2B2 are essentialfor steroid hormone metabolism11,13). The metabolites themselvesoxidative products may directlygenotoxic DNA damage and moduoncogene, tumor suppressor gene,cycle control gene expression. Addihormones, their metabolites, anestrogens may exert epigenetic, p

Xenohormones,endogenous hormones

Cyp450-dependentaromatase 17p-oxidore

4aturally effects on preinitiated cells via hormoneines may and growth factor receptors or via intercel-transfor- lular gap junctional communication. Theseional) or indirect effects predominantly operate dur-Central ing the postinitiational (promotional)

ilable E2 events of carcinogenic transformation.icts with Beneficial xenoestrogens, such asmicroso- genistein and other bioflavonoids thatioreduc- occur in vegetables, fruits, and grain prod-50 lAl, ucts, may reduce aberrant breast cell prolif-enzymes eration and inhibit angiogenesis, and may(4,7-9, increase DNA repair processes and enhanceor their cytodifferentiation and apoptosis. Someinduce phytoestrogens such as genistein alsolation of influence aberrant cell proliferation. At theor cell- pharmacological levels these compounds

itionally, operating via receptor-independent mecha-id xeno- nisms may inhibit kinases and DNAaracrine topopisomerases and thereby affect other

intracellular biochemical targets (14-16).About 15% of women who are carriers

of mutated BRCAI apparently do notdevelop breast cancer (5,6). In these cases,beneficial xenohormones or other exogenousfactors may play a positive role by promot-

iductase ing enzymatic detoxification of potentialcarcinogens, DNA repair processes, or

Total bioavailable estradiolE2 DEl

Cyp450-dependent hydroxylases

16a-OHE1, 4-OHE1,4-OHE2

Genotoxic DNA damage

Total bioavailable estradiolE2 I D E ]

Cyp450-dependent hydroxylases

2-OHE1, 2-OHE2,2-MeOHE,

(E2 antagonists)

16a-OHE1, 4-OHE1,4-OHE2

(E2 agonists)

Altered expression of cell cyclerelated genes, oncogenes,tumor suppressor genes

Aberrant proliferation

Gap junction inhibition

Breast cancer

Figure 1. Impact of steroid hormones on genotoxicpathways of breast carcinogenesis. In the genotoxicpathway, bioavailable 17p-estradiol is converted viaCyp450-dependent hydroxylases to 16a-OHE1, 4-OHE1,or 4-OHE2. These metabolites, by virtue of their directeffect on DNA, cause genotoxic DNA damage. Thisdamage alters the expression of cell cycle related genes,oncogenes, and tumor suppressor genes leading toaberrant proliferation and breast cancer development.

ER, PgR, AR, EGFR, TGFa, TGF-PRIGFR-mediated responses

Aberrant proliferation

Gap junction inhibition

Breast cancer

Figure 2. Impact of steroid hormones on hormonalpathways of breast carcinogenesis. In the hormonalpathway, the E2 metabolites having estrogenic proper-ties (16a-OHE1, 4-OHE1, or 4-OHE2) or antiestrogenicproperties (2-OHE1, 2-OHE2, and 2-MeOHEj) exert theirgrowth modulatory effects indirectly via receptor-medi-ated mechanisms. These alterations lead to modula-tion of aberrant proliferation and of breast cancerdevelopment.

rE2 * oEl_

l \ I~~~~~~~~~~

DNA

Structure Function

Aberrant proliferation

Gap junction inhibition

Breast cancer

Figure 3. Bifunctional pathways to breast cancer. Inthe bifunctional pathway, the E2 metabolites affect cellproliferation and breast cancer development eitherdirectly via receptor-independent mechanisms involvingstructural/functional alterations in DNA, or indirectlyvia receptor-dependent mechanisms involving pheno-typic growth regulation. Both mechanisms eventuallyupregulate aberrant proliferation and development ofbreast cancer.

Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997572

BIFUNCTIONAL PATHWAYS TO BREAST CANCER

antioxidant formation, or by otherwiseenhancing the ability of the cell to overridesignals that would produce uncontrolledgrowth. Recently several researchersreported that the average age of onset of thedisease has fallen in carriers of BRC41 aswell (5,6). Thus exogenous factors appearto affect the timing and expression of breastcancer in those with predisposing germ cellmutations. Under this hypothesis, xenohor-monal-genetic interactions could accountfor variations in gene expression.

Hormone Metabolismand Breast CancerMany steroid hormones are involved in thedevelopment of the human breast. Themammotropic steroids estrogen and prog-esterone are as crucial for breast-cell prolif-eration as the lactogenic hormonesprolactin and glucocorticoids are for cytod-ifferentiation (3,4,11,13,19). Thus hor-mone-mediated alterations in proliferativeand cytodifferentiative status may regulatecarcinogenesis. Receptors for most of thesteroid hormones are reported to be upreg-ulated in clinical breast cancer (20-22).The complex interacting influences ofsteroid hormones on breast carcinogenesis,however, remain to be elucidated.

Although a role of androgens in breastcancer remains to be unequivocally estab-lished, researchers have long observed thatthe higher the cumulative levels ofbioavailable E2, the greater the endogenousrate of breast-cell division and the cumulativerisk of breast cancer (4,11). Endogenous E2and most natural plant estrogens (phyto-estrogens) are metabolized and excretedrelatively rapidly and readily bind to SHBG,whereas most xenoestrogens do not appear tohave this binding capacity (11,12,23-25).Moreover, the half-life of some lipophilicxenoestrogens, such as the organochlorinepesticides, can extend over several decades,in contrast to most natural estrogens, whichare metabolized completely within severalminutes or hours.We and others (9-13,23-24) have

suggested that two competing, mutuallyexclusive enzymatic pathways can alter theproduction of bioavailable estradiol.Pathway 1 inserts a hydroxyl (-OH) func-tion at the C2- position and yields thecatechol estrogen 2-hydroxyestrone (2-OHE1), a weakly estrogenic metabolite.Pathway 2 adds an OH at the C16a posi-tion and yields 16a-hydroxyestrone (16a-OHE1); this creates a fully potent estrogenicmetabolite that can covalently bind to theER (Figure 4).

Our studies on murine mammaryepithelial cell cultures have shown that ini-tiators of rodent mammary carcinogenesis(chemical carcinogens, oncogenes, andtransforming retroviruses) upregulate path-way 2 at the expense of pathway 1(3,26-29). Exposure of immortalized butnontumorigenic murine mammary epithe-lial cells to 16a-OHE1 results in genotoxicDNA damage and increased cellular prolif-eration in anchorage dependent andanchorage independent growth conditionsin a manner similar to that induced bytreatment with the complete carcinogen7, 12-dimethylbenz[a]anthracene (DMBA)(25). In contrast, 2-OHE1 lacks theseactivities and downregulates the effects ofDMBA (28-29). In carcinogen initiatedand in tumor-derived cells, 16a-OHE1enhances, while 2-OHE1 inhibits, theexpression of transformed phenotype(11,13,24,26). Similar changes have beenreported in the human mammary explantand cell culture models (13,28). It is there-fore conceivable that the cellular metabo-lism of E2 is altered during rodent andhuman mammary carcinogenesis and thatindividual metabolites may exert distinctbiological effects during initiation or pro-motion of carcinogenesis. Furthermore,experiments on laboratory models of mam-mary carcinogenesis have shown that thenaturally occurring phytochemical indole-3-carbinol and omega-3 polyunsaturatedfatty acid prevent carcinogenic transforma-tion, largely due to enhancement of theC2-hydroxylation of E2 that leads toincreased formation of antiproliferative2-OHE1 (11,24,31).

The ratio of 16a-hydroxyestrone to2-hydroxyestrone has been found to be ele-vated in women and experimental animalswith high rates of mammary tumors(13,30,32). In human mammary carci-noma cell cultures, some organochlorinepesticides activate the type of Cyp450 thatis responsible for 16a-hydroxyestrone for-mation and produce elevated metaboliteratios, comparable to that induced by theknown rodent carcinogen DMBA (24).A number of studies indicate that

16a-hydroxylation of E2 plays a bifunc-tional role in the development of breastcancer. The estrogen receptor as a recog-nized transcription factor may be central tothe process by which estrogen metabolitesinduce genomic changes. This processindudes the ability ofER to bind to appro-priate DNA response elements, enhancetranscriptional activation, and initiate acascade of events involving expression of

several estrogen-responsive genes such aspS2, c-fos, c-jun, and c-myc, which code forpositive growth regulatory nuclear proteins(33). Other metabolites of estradiol exhibitreversible binding to ER. Work from thislaboratory indicates that 16a-OHE1 hasthe unique capacity to bind covalently andirreversibly with the ER (13,34).

Natural variations in the levels ofendogenous estrogen metabolism leading tothe formation of metabolites with distinc-tive biological activity may explain somereported ethnic and geographic variations inbreast cancer. Asian women, who havemuch lower rates of breast cancer and breastsecretions, generally have higher levels of2-OHE1 and low levels of 16a-OHE1(9,11,23,35). Diets rich in vegetables, fruits,and grain products could represent an addi-tional source of phytoestrogens capable ofmodifying breast carcinogenesis.

The relative importance of variousmetabolites of E2 in the process of carcino-genesis has been addressed in severalmodels. The catechol estrogens 2-OHE1, 4-OHE1, 2-OHE2, and 4-OHE2 have docu-mented pleiotropic effects on organ-sitecarcinogenesis. For example, the catecholestrogens 4-OHE2 and 4-OHE1 arereported to function as genotoxic agents inthe hamster kidney model, in part due totheir ability to induce oxidative DNA dam-age via free radical generation (7,8,36-38).Microsomes from human fibroadenomaand adenocarcinoma exhibit higher levels of4-OHE2 than of 2-OHE2 compared tonormal breast tissue (7,8). In vivo andin vitro studies from our laboratory haveshown that during mammary carcinogenesis,formation of2-OHE1 is decreased while thatof 16a-OHE1 is increased, and agents thatincrease 2-OHE1 inhibit carcinogenesis(26,28,29,39). Thus positive regulation ofgrowth by 16a-OHE1, 4-OHE1, and 4-OHE2 and negative regulation by 2-OHE1 and 2-OHE2 may be consistentwith the estrogenic or antiestrogenic prop-erties of specific metabolites of E2. Despitethe pleiotropic effects of E2 metabolites,the general consensus is that bioavailableE2 has an important role in the risk forbreast cancer (4,9,11,23,35). The role ofother hormones in breast carcinogenesis,however, remains to be elucidated.

Structural And FunctionalDamage To DNADistinct endogenous processes can alterthe structure or function of DNA, includ-ing oxidation, methylation, deamination,phosphorylation, and depurination.

Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997 573

DAVIS ET AL.

Oncogenes have been considered moleculartargets whose gain of function is associatedwith cancer. An array of tumor suppressorgenes such as BRCA1, Rb, DCC, p53, andataxia telangiectasia, mutated (ATM) allowcancer to proceed undisturbed when theirediting and review functions are lostthrough phosphorylation or other changes(5,6,26,33). In addition, growth factors andhormone receptors can also independentlystimulate cell proliferation or activate hor-mone-responsive genes without causingstructural damage to DNA (3,4,17,26,33).

Metabolism of oxygen ipvolving achain of one-electron reductions and theformation of DNA-reactive free radicalsplays a pivotal role in the cause of extensivestructural or functional damage to DNA.One-electron oxidation of the catecholestrogen 4-OHE2 leads to the formation ofa semiquinone that is a reactive species.Semiquinones can be further oxidized toquinones. Alternatively, semiquinones canreact to give superoxide ions, which in turnyield H202. The hydrogen peroxide thusformed then yields the reactive -OHmoiety. Free radicals produced fromsemiquinones can directly damage thephosphodiester bonds or alter the DNAnucleotide base sequence, thereby impair-ing normal transcriptional processes(7,8,14-16,36-38). Free-radical induceddamage may also cause the loss of tumorsuppressor gene function, reduce naturalkiller cells, impair enzymatic detoxificationprocesses, or alter growth factor synthesis.

Several investigations demonstrate thathydroxyl (-OH) radicals and oxidativemetabolites of aromatic hydrocarbons canirreparably and specifically modify thestructure of DNA, leading to mutagenesisand carcinogenesis (7,8,14-16,40).Distinctive types of -OH-induced modifi-cations in DNA bases have been detectedin women with breast cancer and in thoseat risk for the disease, compared to con-trols, who have undergone reductionmammoplasty. DNA from mammoplastypatients had relatively higher proportionsof ring opening product of adenine,4,6-diamino-5-formamidopyrimidine(Fapy-A), whereas cancer patients hadmarkedly lower levels of Fapy-A and higherlevels of 8-hydroxyquanine (16). It has beensuggested that the relevance of oxidativeDNA damage in breast cancer reduction ispredominantly due to H202 generation.H202 can cross the nuclear membrane,where it is converted by the iron- or copper-catalyzed Fenton reaction to the free radical*OH. Some xenoestrogens and estrogen

metabolites may promote the productionof free radicals by redox cycling of H202mediated by cytochrome P450 oxidaseand reductase, leading to production ofDNA-damaging reactive oxygen species.

Another critical pathway to breastcancer can arise from agents that affectgenes for regulatory proteins, such as phos-phatidylinositol-3-kinases (PI-3-kinases).These proteins are essential for detectingDNA damage. PI-3-kinase may functionby halting improper cell growth and divi-sion until damage is repaired. Persons whohave inherited ATM and are lacking thiskey regulatory gene are unable to recoverfrom radiation and possibly other DNA-damaging exposures. Consequently, theyaccumulate harmful mutations much morereadily than healthy persons. Some experi-mental and epidemiologic evidence sug-gests that women who inherit a singledefective copy ofATM are at increased riskof breast cancer compared with those withnormal AT. Experimental cell culturestudies reveal that cells containing a singlefaulty copy ofATM incur a higher deathrate following radiation exposure.

DiscussionDevelopment of cancer is the consequenceof complex and subtle interactions betweenthe environment and the genome. As 95%of human breast cancer is not due to inher-ited mutations, the question becomes oneof what causes people who have inherited ahealthy array of genes to acquire the dis-ease. It appears likely that a wide variety ofxenobiotic compounds may alter bioactiveestrogen, androgen, or progesterone inbreast cancer cases, in part via Cyp450-mediated enzymatic reactions. Persons whohave inherited genetic susceptibility tobreast cancer might be especially sensitiveto the proliferative effects of some xenohor-mones from the environment. Persons withthese defects who do not develop the dis-ease may have had greater exposure to bene-ficial xenohormones or to other protectivefactors. Both genetic and hormonal path-ways that act in a bifunctional manner mayinduce aberrant proliferation leading to thedevelopment of breast cancer. Some xeno-estrogens affect the production andmetabolism of estradiol and thereby regu-late cumulative levels of the bioavailablehormone. Some xenohormones directlymodify DNA structure or function.Structural damage can result from geno-toxic E2 metabolites and from redox cyclingof harmful xenoestrogens or other agentsthat may produce reactive oxygen species.

Functional damage can include alteredtranscriptional activities related to onco-genes and tumor-suppressor genes, whichare responsible for positive and negativeregulation of growth. Phosphorylation ofp53 gene product or of PI-3-kinase is alsoindicative of functional DNA damage thatimpedes cell repair by hindering recogni-tion of damaged cells and allowing theaccumulation of harmful mutations. Thegrowth advantage of aberrant cells, possiblyby activation of genes involved in cell-cycleprogression, leads to the development ofbreast cancer (3,4,13,19,33).

The bifunctional genetic-hormonalhypothesis for breast cancer also provides ameans by which to link the interactiveinfluence of dietary factors on disease pro-gression. The effect of dietary fat on breastcancer still remains equivocal. The type ofdietary fat and the levels of xenoestrogeniccontaminants in the diet may provideimportant leads to resolve the observedinconsistencies. Furthermore, such ananalysis should also provide mechanism-based interpretation for experimental andclinical evidence about the role of dietaryfat in breast cancer development. Poly-unsaturated fatty acids are considered tomodulate breast carcinogenesis in part byinterfering with membrane-mediated gap-junctional intercellular communication(41). This process is critical for cellularhomeostasis whose impairment may pro-mote growth advantage of the cancer cell.Lipophilic organochlorine pesticides operat-ing via synergistic interactions inhibit gap-junctional communication in humanmammary epithelial cells derived fromreduction mammoplasty (42). Xenobioticssuch as the food colorant Red Dye No. 3have been reported to stimulate cell-cycleprogression in human mammary carcinomaT47D cells acting via the cyclin-dependentkinase 2 and ER binding (14,15). Xeno-biotics that have a carboxylic acid functioncan be incorporated structurally into com-plex lipids such as triglyceride (fat) andphospholipids (membranes). When mobi-lized some of these metabolites may serve assignaling molecules similar to phorbolesters, which are well-established tumorpromoters (41,43). We need to determinewhether bioaccumulated lipophilic xeno-hormones, xenobiotics such as aromatichydrocarbons, and organochlorines areelevated in women at risk for breast cancer.

This genetic-hormonal hypothesis alsomay account for one of the discrepanciesin cancer patterns that was first noted bythe distinguished Danish researcher

Environmental Health Perspectives - Vol 105, Supplement 3 * April 1997574

BIFUNCTIONAL PATHWAYS TO BREAST CANCER

J. Clemmensen (44). Commonly referredto as Clemmensen's hook, this discrepancyoccurs in the relationship between age andbreast cancer incidence noted for severalcountries. Incidence of breast cancer risesprogressively with age up to about 45 yearsof age, after which the rate of increaseforms a hook and levels off or declines forabout 10 years and then resumes anincreasing, but more modest, slope. Theages of this plateau correspond to theperiod of perimenopause, when the ovariesbegin to produce less estrogen and prog-estin (44). It is tempting to speculate thatthe renewed surge in breast cancer aftermenopause, especially in obese women,

might be linked with xenoestrogens and withthe production of endogenous estrogens thatwould be greatest in those with proportion-ally more body fat. Also, obese womenwould have higher rates of membrane lipiddamage (9-11,23,35,42,45).

With respect to breast cancer, most ofthe confirmed risk factors, which relate toreproductive behavior and dietary factors,are not easily changed by social policy.Many of the proposed interventions toreduce breast cancer involve the lifelonguse of pharmaceutical agents to changehormonal metabolism or the advocacy ofradical changes in diet, lifestyle, or evenreproductive behavior. As for the latter

point, a generation of women that hasstruggled long for reproductive freedom isunlikely to accept constraints on theirreproductive choices.

This hypothesis has major implicationsfor breast cancer screening, prevention,treatment, and management. Biologicmarkers of structural and functional dam-age to DNA and of estradiol metabolismcould prove useful for identifying personsat risk of developing breast cancer, assist inprognostic predictions, and provide base-lines to assess the efficacy of potential ther-apeutic and nutritional interventions forprevention, treatment, and management ofthe disease.

REFERENCES

1. Kohlmeier L, Rehm J, Hoffmeister H. Lifestyle and trends inworldwide breast cancer rates. Ann NY Acad Sci 609:259-269(1990).

2. Kosary CL, Gloeckier Ries LA, Miller BA, Hankey BF, HarrasA, Edwards BK, eds. SEER Cancer Statistics Review1973-1991. NIH Publ 94-2789. Bethesda, MD:NationalCancer Institute, 1994.

3. Harris JR, Lippman ME, Veronesi U, Willett W. Breast cancer(first of three parts). N Engl J Med 327:319-328 (1992).

4. Pike MC, Spicer DV, Dahmoush L, Press MF. Estrogens,progestogins, normal breast cells, and breast cancer risk.Epidemiol Rev 15:17-35 (1993).

5. Mild Y, Swensen J, Shattuck-Eidens D, Futreal PA, HarshmanK, Tavtigian S, Liu QY, Cochran C, Bennett LM, Ding W, etal. A strong candidate gene for the breast and ovarian cancersusceptibility gene BRC41. Science 266:66-71 (1994 ).

6. Hall JM, Lee MK, Newman B, Morrow JE, Anderson LA,Huey B, King MC. Linkage of early-onset breast cancer tochromosome 17q21. Science 250:1684-1689 (1990).

7. Liehr JG, Ricci MJ, Jefcoate CR, Hannigan EV, Hokanson JA,Zhu BT. 4-hydroxylation of estradiol by human uterinemyometrium and myoma microsomes: implications for themechanism of uterine tumorigenesis. Proc Natl Acad Sci USA92:9220-9224 (1995).

8. Liehr JG, Ricci MJ. 4-hydroxylation of estrogens as marker forhuman mammary tumors. Proc Natl Acad Sci USA93:3294-3307 (1996).

9. Aldercreutz H, Gorbach SL, Goldin BR, Woods MN, DwyerJT, Hockerstedt K, Wahala K, Hase T, Hamalainen E, FoslisT. Estrogen and metabolite levels in women at risk for breastcancer. Proc Am Assoc Cancer Res 35:703 (1994).

10. Toniolo PG, Levitz M, Zeleniuch-Jacquotte A, Banerjec S,Koenig KL, Shore RE, Strax P, Paternack BS. A prospectivestudy of endogenous estrogens and breast cancer in post-menopausal women. J Natl Cancer Inst 87:190-197 (1995)

11. Davis DL, Bradlow, HL Wolff M, Woodruff T, Hoel DG,Anton-Culvier H. Medical hypothesis: xenoestrogens as pre-ventable causes of breast cancer. Environ Health Perspect101:372-377 (1993).

12. Rosner W, Hryb DJ, Khan MS, Nakhla AM, Romas NA. Sexhormone binding globulin: binding to cell membrane and gen-eration of a second messenger. J Androl 3:101-106 (1992).

13. Fishman J, Osborne MP, Telang NT. The role of estrogen inmammary carcinogenesis. Ann NY Acad Sci 768:91-100(1995).

14. Effects of estradiol, xenoestrogens, and EMF on breast cancercells. Radiation Res USA 146:444-452 (1996).

15. Dees C, Askari M, Garrett S, Gehrs K, Henley D, Ardies CM,Travis C. Estrogenic and DNA-damaging activity of red no. 3in human breast cancer cells. Environ Health Perspect, in press.

16. Malins D. Identification of hydroxyl radical-induced lesions inDNA base structure: biomarkers with a putative link to cancerdevelopment. J Toxicol Environ Health 40:247-261 (1993).

17. Ames B, Gold LS, Willett WC. The cause and prevention ofcancer. Proc Natl Acad Sci USA 92:5258-5265 (1995).

18. Davis DL, Axelrod D, Osborne MP, Telang NT. Environmentalinfluences on breast cancer risk. Sci Med, in press.

19. Nandi S, Guzman RC, Yang J. Hormones and mammary car-cinogenesis in mice, rats, and humans: a unifying hypothesis.Proc Natl Acad Sci USA 92:3650-3657 (1995).

20. Lea OA, Kvinnsland S, Thorsen T. Improved measurement ofandrogen receptors in human breast cancer. Cancer Res49:7162-7167 (1989).

21. Maclndoe JH, Etre LA. An antiestrogenic action of androgensin human breast cancer cells. J Clin Endocrinol Metab53:836-842 (1981).

22. Thomas BS, Bullbrook RD, Harward JL. Urinary androgenmetabolites and recurrence rates in early breast cancer. Eur JClin Oncol 18:447-451 (1982).

23. Aldercreutz H, Mousavi Y, Hockerstedt K. Diet and breastcancer. Acta Oncol 31:175-181 (1992).

24. Bradlow HL, Davis DL, Lin G, Sepkovic D, Tiwari RK.Effects of pesticides on the ratio of 16a/2-hydroxyestrone: abiologic marker of breast cancer risk. Environ Health Perspect103:147-150 (1995).

25. Nagel SC, vom Saal FS, Thayer, KA, Dhar MG, Boechler M,Welshons WV. Relative binding affinity-serum modified access(RBA-SMA) assay predicts the relative in vivo bioactivity of thexenoestrogens bisphenol A and octylphenol. Environ HealthPerspect 105:70-76 (1997).

26. Telang NT. Oncogenes, estradiol biotransformation and mam-mary carcinogenesis. Ann NYAcad Sci 784:277-287 (1996).

27. Telang NT, Suto A, Wong GY, Osborne MP, Bradlow HL.Induction by estrogen metabolite 16a-hydroxyestrone of geno-toxic damage and aberrant proliferation in mouse mammaryepithelial cells. J Natl Cancer Inst 84:634-638 (1992).

28. Suto A, Bradlow HL, Wong GYC, Osborne MP, Telang NT.Experimental down-regulation of intermediate biomarkers ofcarcinogenesis in mouse mammary epithelial cells. BreastCancer Res Treat 27:193-202 (1993).

29. Telang NT, Suto A, Wong GY, Bradlow HL, Osborne MP.Genotoxic damage and aberrant proliferation by estrogenmetabolites in mammary epithelial cells [Abstract]. Proc AmAssoc Cancer Res 33:278 (1992).

Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997 575

DAVIS ET AL.

30. Osborne MP, Bradlow HL, Wonf GYC, Telang NT. Up-regu-lation of estradiol CI6a-hydroxy ation in human breast tissue:a potential biomarker for breast cancer risk. J Natl Cancer Inst85:1917-1920 (1993).

31. Osborne MP, Karmali RA, Bradlow HL, Hershcopf RJ,Kourides IA, Williams WR, Rosen PP, Fishman J. Omega-3fatty acids: modulation of estrogen metabolism and potentialfor breast cancer prevention. Cancer Invest 6:629-631 (1988).

32. Bradlow HL, Hershcopf RJ, Martucci CP, Fishman J. Estradiol16a-hydroxylation correlates with mammary tumor incidenceand presence of murine mammary tumor virus: a possiblemodel for hormonal etiology of breast cancer in human. ProcNatl Acad Sci USA 82:6295-6299 (1985).

33. Weisz A, Bresciani F. Estrogen regulation of protooncogenescoding for nuclear proteins. Crit Rev Oncog 4:361-388 (1993).

34. Swaneck GE, Fishman J. Covalent binding of the endogenousestrogen 166a-hydroxyestrone to estradiol receptor in humanbreast cancer cells: characterization and intranuclear localiza-tion. Proc Natl Acad Sci USA 85:7831-7835 (1988).

35. Goldin BR, Gorbach SL. Effect of diet on the plasma levels,metabolism and excretion of estrogens. Am J Nutrition48:787-790 (1988;)

36. Weisz J, Bui QD, Roy D, Liehr JG. Elevated 4-hyroxylation ofestradiol by hamster kidney microsomes: a potential pathway ofmetabolic activation of estrogens. Endocrinology 131:655-661(1992).

37. Han X, Liehr JG. Microsome-mediated 8-hydroxylation ofguanine bases of DNA by steroid estrogens: correlation of

DNA damage by free radicals with metabolic activation toquinones. Carcinogenesis 16:2571-2574 (1995).

38. Li JJ, Li SA. Estrogen carcinogenesis in hamster tissues: acritical review. Endocrine Rev 11:524-531 (1990).

39. Tiwari RK, Guo L, Bradlow HL, Telang NT, Osborne MP.Selective responsiveness of breast cancer cells to indole-3-carbinol, a chemopreventive agent. J Natl Cancer Inst86:126-131 (1994).

40. Li D, Wang M, Dhingra K, Hittleman WN. Aromatic DNAadducts in adjacent tissues of breast cancer patients: clues tocancer etiology. Cancer Res 56:287-293 (1996 ).

41. Aylsworth CF. Effects of lipids on gapjunctionally mediatedintercellular communication: possible role in the promotion oftumorigenesis by dietary fat. In: Dietary Fats and Cancer (Ip C,Birt DF, Rogers AE, Metlin C, eds). New York:Alan R. Liss,1986;607-622.

42. Kang KS, Wilson MR, Hayashi T, Chang CC, Trosko JE.Inhibition of gapjunctional communication in normal humanbreast epithelial cells after treatment with pesticides, PCB, PBBalone or in mixture. Environ Health Perspect 104:192-200(1996).

43. Coleman, RA. Identifying xenobiotic acids that can form PKC-activating diacylglycerols. J Lipid Res USA 36:2493-2503(1995).

44. Clemmesen J. Carcinoma of the breast: results from statisticalresearch. BrJ Radiol 21:583-590 (1948).

45. Colditz GA. Fat, estrogens and the time frame for preventionof breast cancer (editorial). Epidemiology 6:207-208 (1995).

576 Environmental Health Perspectives * Vol 105, Supplement 3 * April 1997