The current mortality rate for premenopausal breast cancer is approximately fourfold higher in the Western World than in Far East Asian nations 1. Migrants from Asia to the USA typically acquire a breast cancer risk associated with their host nation by the second generation, suggesting a direct influence of environmental rather than genetic factors 23. The reduced prevalence of breast cancer in Far East Asian nations directly correlates with the consumption of a staple diet that is abundant in soy 4. Asian populations consume an approximate mean daily soy intake of 10–50 g, declining to just 1–3 g in the USA. The recent adoption of a more westernized diet correlates with an increased breast cancer incidence in urban areas of Japan, Singapore and China 5. Soy-containing foods are an abundant source of phyto-oestrogens, and research suggests that these compounds may exhibit chemoprotectant activity against a number of human cancers, including colon carcinoma and hormone-dependent cancers of the breast and prostate 6. This report focuses on the putative chemopreventive role of phyto-oestrogens in breast cancer, providing a comprehensive review of the published literature to date.

Phyto-oestrogens may be classified into a number of principal groups 2789: the isoflavones (genistein, daidzein, biochanin A), the lignans (enterolactone, enterodiol), the coumestans (coumestrol) and the stilbenes (resveratrol). As illustrated in Fig. 1, all are polyphenols sharing structural similarity with the principal mammalian estrogen 17β-oestradiol. Shared features include the presence of a pair of hydroxyl groups and a phenolic ring, which is required for binding to the oestrogen receptor (ER) subtypes α and β. The position of the hydroxyl groups appears to be important in determining ER binding ability and transcriptional activation, with maximal potency achieved at positions four, six and seven 101112. The isoflavones are naturally found in soybeans and soy-based food products, including tofu, soy milk, textured soy protein and miso. Lignans are present in flaxseed and most fruit and vegetables, and the predominant dietary source of stilbenes is peanuts, grapes and red wine 713. The coumestans are much less frequently consumed within the human diet, but they are more potent activators of ER signalling pathways than are the isoflavones genistein and daidzein 1014. By contrast, the stilbene resveratrol is the least potent activator of ER signalling 11.

The isoflavones are present in soy as β-glucosides. Metabolism by the gastrointestinal microflora yields a number of metabolites including equol and O-desmethyl-angolensin. Parental compounds and their metabolites are absorbed into the bloodstream, becoming rapidly detectable in the plasma and urine 1516171819. Plasma isoflavone concentrations are considerably elevated in Asian populations as compared with in western ones. A recent comparison of Japanese and UK females revealed an almost 20-fold increase in plasma genistein levels in the Japanese cohort, and daidzein concentrations were similarly elevated by 18-fold 20. Plasma isoflavone concentrations may accumulate to approximately 100- to 1000-fold higher than endogenous oestradiol levels following the ingestion of soy-rich meal. However, research suggests a decreased ER binding affinity of isoflavone compounds as compared with the mammalian oestrogens 9102122. Competition binding assays revealed a 50-fold lower binding affinity of genistein for cytoplasmic ER sites as compared to 17β-oestradiol 23.

The complete metabolic activation of soy isoflavones is proposed to occur locally within target tissue. In support of this hypothesis, the analysis of tissue culture supernatants from genistein and biochanin A treated MCF-7 and T47-D cells revealed the presence of hydroxylated and methylated isoflavone metabolites 24. Current research suggests a role for the CYP family of cytochrome P450 enzymes in the intratumour metabolism of phyto-oestrogen compounds 2526. The CYP1B1 enzyme is expressed in a wide range of human tumour types, including breast 27; however, expression is absent within normal tissue. CYP1B1 is proposed to catalyze the hydroxylation of resveratrol to yield the related stilbene piceatannol 26. Piceatannol is a tyrosine kinase inhibitor with antileukaemic properties, which differs in structure from resveratrol by the presence of an additional hydroxyl group. A number of plant flavonoids are also putative substrates for the CYP family of enzymes 25. Maubach and coworkers 28 recently reported the use of high-performance liquid chromatography to quantify isoflavones in normal breast biopsy tissue following consumption of soy for 5 consecutive days. Equol concentrations were approximately fivefold higher in the breast tissue homogenates than in serum, providing further evidence for the metabolism of phyto-oestrogen compounds within mammary tissue.

Serum concentrations of 17β-estradiol are approximately 40% lower in Asian women than in their Caucasian counterparts 29. A low lifetime exposure to oestrogen is associated with a reduced risk for breast cancer. Human dietary intervention studies revealed a direct association between the modest consumption of soy products and a reduction in circulating steroid hormone levels. Daily consumption of 154 mg isoflavones for the duration of a single menstrual cycle correlated with substantially decreased plasma concentrations of 17β-oestradiol and progesterone in a cohort of premenopausal women 18. A longer-term study conducted by Kumar and coworkers 30 similarly reported a moderate decrease in serum oestradiol and oestrone levels following daily ingestion of 40 mg isoflavones for 3 months. Menstrual cycle length was increased by 3.52 days, and the follicular phase of the cycle was extended by 1.46 days. Increased menstrual cycle length may serve to reduce the total number of cycles per lifetime, therefore decreasing the total exposure of breast epithelia to endogenous oestrogens. Conversely, a year-long dietary intervention trial involving 34 premenopausal women failed to reveal a significant effect of 100 mg/day isoflavone consumption either on menstrual cycle length or on serum levels of various steroid hormones, including oestrone, oestradiol and progesterone 19.

As a possible explanation for these contradictory data, a study conducted by Duncan and coworkers 31 revealed differential hormonal effects of soy isoflavones depending on the ability to excrete the daidzein metabolite equol. Daily ingestion of 10 mg soy protein by premenopausal equol excretors resulted in a hormonal profile associated with reduced breast cancer risk, characterized by lowered plasma levels of oestrone, oestrone–sulphate and testosterone. Hormone levels however remained unchanged in the equol nonexcretors after soy ingestion.

The reduction in steroid hormone levels by phyto-oestrogens is proposed to occur via the direct regulation of 17β-oestradiol biosynthesis and metabolism. Phytochemicals isolated from vegetable extracts effectively suppress the activity of the aromatase enzymes, which are responsible for conversion of androgens to oestrogens 32. Isoflavone concentrations of 1–10 μmol/l similarly reduced by 50% the activity of the estradiol biosynthetic enzymes 3β-hydroxysteroid dehydrogenase and 17β-hydroxysteroid dehydrogenase 10. The daily ingestion of 113–202 mg isoflavones by premenopausal women correlated with a 40% increase in the urinary excretion of 2-hydroxyestrone, a putative anticancer metabolite of 17β-oestradiol 33. The above studies thus suggest a dual chemoprotectant mechanism of soy, in which the isoflavones suppress steroid hormone biosynthesis while promoting the metabolism of oestradiol to the protective 2-hydroxylated metabolites.

Despite their apparent effect on endogenous hormone levels, the role of phyto-oestrogens in breast cancer initiation and development is unclear. Few studies to date have addressed the effects of long-term phyto-oestrogen exposure in humans. Daily dietary supplementation with 45 mg soy isoflavone for 14 days correlated with increased proliferation of normal breast epithelia in a group of 48 premenopausal women 16. Expression of the ER target protein progesterone receptor was upregulated, suggesting an oestrogenic effect. An identical trial using a larger cohort of 84 premenopausal women conversely found no significant effect of soy consumption on the proliferation of normal breast tissue 17. A number of recent epidemiological studies similarly failed to correlate soyfood consumption with reduced breast cancer risk. A Japanese prospective study conducted in a cohort of approximately 35,000 women 34 revealed no significant association between soy consumption during adulthood and breast cancer incidence. The retrospective analysis of soy food intake in a multiethnic cohort of non-Asian breast cancer patients and control individuals residing in the USA similarly failed to correlate soy intake with breast cancer risk 35.

Increasing epidemiological evidence suggests that the chemoprotectant effects of phyto-oestrogens are dependent on a lifelong exposure from childhood. A retrospective study revealed decreased soyfood intake during adolescence in a cohort of 1459 Chinese breast cancer patients, as compared with age-matched control individuals 36. Daily soy consumption between the ages of 13 and 15 was estimated at 6.45 g in the patient group, increasing to 7.23 g in the control cohort. A potential flaw in the study, however, concerns the ability of women up to the age of 64 years to recall accurately the precise soyfood quantities consumed many years earlier during adolescence. These observations may nonetheless explain the apparent lack of a growth inhibitory effect of soy isoflavones in the adult dietary intervention studies discussed above. Isoflavones are detectable in breast milk following soy consumption 15, implying that the lower breast cancer incidence in Asian countries may be attributable to phyto-oestrogen exposure from birth via breast-feeding. Rodent studies have accordingly revealed the effective transfer of genistein from maternal milk to offspring 37.

Most available information regarding the effects of phyto-oestrogens on tumour initiation and the growth of pre-existing tumours is derived from rodent studies. A number of similarities do exist between mammary gland development in rodents and humans. In both species the differentiation of breast tissue to form lobules and terminal end-bud structures occurs prepubertally. Further maturation does take place throughout adulthood, giving rise to alveolar buds, which become alveoli during pregnancy and lactation 38.

Rodent dietary intervention studies using phyto-oestrogens have reported chemopreventive activities when feeding is initiated before puberty, at a time when the mammary gland is undergoing development 3383940. The consumption of a resveratrol-supplemented diet by adolescent rats served to decrease sensitivity to the chemical carcinogens 7,12-dimethylbenz(a)anthracene (DMBA) and N-methyl-N-nitrosaurea 3940. DMBA treatment induced mammary tumours in 45% of the resveratrol-treated rodents, increasing to 75% in the group receiving a control diet. An extended tumour latency period in excess of 3 weeks was observed in the resveratrol treatment groups, with the resultant tumours retaining a more differentiated morphology as compared with control animals 1839. Resveratrol consumption was also associated with the reduced mammary expression of a number of proteins that are putatively involved in malignant progression, including cyclo-oxygenase-2, matrix metalloprotease-9 and nuclear factor-κB. Similar findings have been noted in prepubertal rats fed an isoflavone-containing diet before tumour initiation using DMBA. Despite having no effect on mammary tumour incidence, soy isoflavone consumption was associated with an increased tumour latency period. The resultant tumours excised from the soy-fed animals were smaller in size and exhibited a more differentiated phenotype compared with control animals 42.

It has been proposed that phyto-oestrogens protect against cancer development in adolescent rodents by promoting maturation of the mammary gland. Analysis of breast tissue from prepubertal rats injected with genistein revealed a decrease in the number of immature terminal end-buds, together with an increase in the more differentiated lobules type II 38. Genistein treatment of human breast cancer cell lines has similarly been found to induce the expression of a number of maturation markers, including casein, lipid droplets and intercellular adhesion molecule-1 43.

The effect of soy isoflavones on spontaneous tumour development was recently investigated using neu-ErbB2 over-expressing transgenic mice, which characteristically develop multiple mammary tumours during adulthood 40. Tumour initiation was temporarily delayed following the consumption of an isoflavone mix, but no chemoprotective effects were observed in mice consuming either genistein or daidzein in isolation. An equal rate of tumour growth was noted in the control and treatment groups, although the isoflavone group exhibited a lower incidence of lung metastases 40. Antimetastatic activities of the isoflavones were similarly revealed in a study in which mice were fed an isoflavone-supplemented diet before injection with the metastatic 4526 mammary carcinoma cell line 44. The isoflavone diet was continued following surgical excision of the resultant mammary tumour at a size of 1.0 cm diameter. Both the incidence and size of macroscopically detectable lung metastases were significantly reduced in the soy-fed mice, suggesting a potential clinical application of soy isoflavones in the prevention of metastasis. Putative anticancer mechanisms of soy and isolated phyto-oestrogen compounds using in vitro and in vivo breast cancer models is summarized in Table 1.

The selective ER modulator tamoxifen is used clinically in the adjuvant treatment of oestrogen-dependent breast cancer. The drug is also administered as a prophylactic to individuals who are at high risk for developing the disease 4546. Side effects associated with tamoxifen therapy include menopause-like symptoms such as hot flushes, joint pain, sleep disorders and depression, which may be reduced by the use of hormone replacement therapy (HRT) 4748. Long-term HRT is associated with an increased risk for mammary carcinogenesis, and its use by breast cancer patients is therefore discouraged. As a natural alternative, patients may self-medicate with soy isoflavone supplements to alleviate the tamoxifen-induced menopausal symptoms 49. Published literature regarding the ingestion of dietary phyto-oestrogens by breast cancer patients and survivors is, however, controversial 5051.

The consumption of genistein by athymic mice antagonized the ability of tamoxifen to inhibit the proliferation of oestrogen-dependent mammary tumours 52. Tumour suppression by tamoxifen correlated with decreased expression of the ER-inducible genes presenelin-2 (pS2) and cyclin D1. Tumour growth was significantly enhanced in mice simultaneously exposed to tamoxifen and genistein, whereas levels of pS2 and cyclin D1 expression were increased. Physiological concentrations of genistein were similarly found to reverse the antagonistic effects of 4-hydroxytamoxifen on ER signalling pathways 53, promoting the binding of ER-α to the positively acting steroid receptor coactivator (SRC)-1. A recent tissue culture study conversely reported a synergistic antiproliferative effect of tamoxifen and genistein 54. The proliferation of a panel of dysplastic and cancerous breast cell lines was inhibited by tamoxifen in a dose-dependent manner, and growth was more potently suppressed by combined treatment with tamoxifen and genistein.

Oestrogen signalling typically involves the diffusion of ligand through the cell cytoplasm and subsequent binding to the nuclear receptor subtypes ER-α and ER-β. Ligand-bound receptors dimerise and associate with oestrogen response element (ERE) and activator protein-1 element located in the promoter region of target genes, thereby activating transcription. The association between receptor dimers and DNA response elements is enhanced by the binding of cofactor proteins, such as amplified in breast cancer-1, thyroid hormone receptor-associated protein, SRC-1, glutamate receptor interacting protein-1 and translation initiation factor 2 55. Examples of ERE-induced genes include PR, c-fos, bcl-2 and cathepsin D, whereas pS2 and cyclin D1 are transcribed via the activator protein-1 response element.

In breast carcinoma cell lines containing functional ER subtypes the isoflavones exert a biphasic growth effect, stimulating cellular proliferation at concentrations below 5 μmol/l and inhibiting growth in a dose-dependent manner at elevated doses 23435657. Growth inhibition correlates with decreased DNA synthesis and cell cycle arrest at the G₂/M checkpoint 235456585960. Current research suggests a principal signalling role of ER-β in response to isoflavone exposure 61. Whereas 17β-oestradiol binds to ER-α and ER-β with equal affinity, the soy isoflavones selectively associate with ER-β 62. Receptor binding assays revealed an eightfold to 16-fold increase in the affinity of genistein, daidzein and biochanin A for ER-β as compared with ER-α 63. In ER-negative breast cancer cells transfected to express ER-α alone, genistein was only weakly able to stimulate gene transcription through the ERE. By contrast, genistein effectively bound to ER-β and promoted the association of cofactor proteins, thereby regulating downstream ER-β-mediated gene transcription 62. The preferential binding affinity of genistein for ER-β similarly resulted in a respective 12,000-fold and 33-fold increase in the recruitment of translation initiation factor-2 and SRC-1a to ER-β as compared with ER-α 22. An enhanced transcriptional activity in response to genistein was, however, noted in cells transfected to express both receptor subtypes as compared with cells solely expressing ER-β 64. Although ER-α is itself unable to mediate isoflavone signal transduction, it was postulated that the presence of the receptor subtype may enhance ER-β signalling via the formation of ER-α/β heterodimers. These observations imply that the precise tissue-specific effects of the soy isoflavones are dependent on the expression levels and ratios of ER-α and ERβ. The various cofactors are similarly expressed in a tissue-specific manner, therefore further influencing the cellular response to dietary phyto-oestrogens.

The stilbene resveratrol is structurally similar to the synthetic oestrogen diethylstilbestrol. Treatment of breast cancer cell lines with resveratrol represses proliferation in a dose-dependent manner, inducing G₂/M phase cell cycle arrest 39. Resveratrol exhibits a relatively weak ER-binding affinity as compared with oestradiol 65; however, unlike the soy isoflavones, it is able to bind to both ER-α and ER-β with equal affinity. In cells transfected to express either ER-α or ER-β resveratrol was found to act as an agonist for both receptor subtypes, stimulating ERE transcriptional activity through either ER-α or ER-β alone 21. Similar agonist activity was observed in MCF-7 breast cancer cells, which predominantly express the ER-α isoform. Resveratrol induced the dose-dependent activation of ERE-mediated transcription, also upregulating the expression of the ER target genes pS2 and PR 65. Recent studies proposed that the cell cycle inhibitor protein p21^WAF1 is a potential downstream target of resveratrol-induced ER signalling pathways 66. The treatment of ER-α-expressing breast cancer cells with resveratrol resulted in a 23-fold increase in p21^WAF1 gene expression, as determined by cDNA microarray analysis. The resveratrol-mediated induction of p21^WAF1 was blocked by treatment with the pure anti-oestrogen ICI 182,780, confirming p21^WAF1 gene regulation as an ER-mediated event.

At concentrations in excess of 25 μmol/l, the soy isoflavones are capable of inducing apoptosis in human breast cancer cells 23676869. ER-negative cell lines retain sensitivity to the apoptotic effects of soy isoflavones, thereby confirming that apoptosis occurs in a hormone-independent manner. Apoptosis was effectively induced in the ER-α-negative MDA-MB-231 breast cancer cell line by genistein and daidzein concentrations of 50–100 μmol/l 5960. In MCF-7 cell cultures the induction of cell death by treatment with genistein coincided with the increased expression of the proapoptotic proteins Bax and p53 67. Breast cancer cell lines expressing mutant p53 also undergo apoptosis in response to phyto-oestrogen treatment, thereby implying apoptosis induction by both p53-dependent and p53-independent mechanisms 6770. The polyphenol epigallocatechin (EGC) is principally found in green tea and is proposed to have anticancer properties. Treatment of p53-mutant breast cancer cells with 100 μmol/l EGC induced a 40% increase in apoptosis, correlating with increased Bax expression and reduced levels of the antiapoptotic protein Bcl-2 70. EGC-induced apoptosis was abolished following treatment with anti-Fas neutralizing antibodies or caspase inhibitors, suggesting the involvement of Fas signalling pathways. Although phyto-oestrogen compounds are effective inducers of apoptosis in cell culture models, it is unlikely that plasma isoflavone concentrations would accumulate to the required levels for the activation of apoptotic pathways in vivo. It is estimated that plasma phyto-oestrogen concentrations may reach a maximum of 2–4 μmol/l following the moderate consumption of soy products 1552, although it is possible that higher levels may be present in target tissues. In a recently reported study, equol concentrations within breast tissue were found to exceed serum levels; however, the reverse was true for genistein and daidzein 28. A recent in vitro study 68 revealed the flavone baicalein to be a more potent inducer of apoptosis than genistein. Baicalein is isolated from the plant Scutellariae radix and is a common ingredient in herbal tea preparations. A concentration of 10 μmol/l baicalein induced significant cell death in MCF-7 cell cultures, suggesting baicalein as a potentially useful pharmacological agent in breast cancer therapy.

Dietary phyto-oestrogens are capable of inhibiting the proliferation of hormone-independent breast cell lines 43545869. It has been proposed that growth inhibition in the absence of functional ER occurs via the inhibition of tyrosine kinase activity. The protein tyrosine kinases are involved in a number of growth factor signalling pathways, including transforming growth factor (TGF)-α, insulin-like growth factor (IGF)-I, IGF-II and epidermal growth factor (EGF). In ER-negative breast cancer cultures 5 μmol/l genistein negated the stimulatory effects of TGF-α, IGF-I and IGF-II, implying the inhibition of tyrosine kinase activity 23. The human EGF receptor-2 oncogene (Her-2) is constitutively overexpressed in approximately 30% of breast cancers and is associated with a poor patient prognosis 71. Research using breast cancer cell lines suggests that dietary phyto-oestrogens are capable of repressing EGF receptor activity. The inhibition of tyrosine kinase activity by 5 μmol/l genistein in MCF-7 cells correlated with the repression of EGF receptor tyrosine phosphorylation in response to EGF stimulation 23. Similar findings were reported in a recent study investigating the chemoprotective effects of the green tea polyphenol epigallocatechin-3 gallate (EGCG). The treatment of Her-2/neu over-expressing mouse mammary cells with 20–80 μg/ml EGCG inhibited proliferation in a dose-dependent manner, correlating with a reduction in Her-2/neu signalling activity 72. The basal tyrosine phosphorylation of Her-2/neu was decreased by approximately 96% following treatment with 80 μg/ml EGCG. Downstream activities of the signalling proteins phosphoinositide 3-kinase, Akt and nuclear factor-κB were similarly repressed, suggesting a potential clinical application of EGCG in breast cancer therapy.

The soy isoflavones have additionally been proposed to regulate the proliferation of breast epithelia via an alternative mechanism involving the modulation of TGF-β synthesis 73. In normal mammary tissue TGF-β maintains proliferative homeostasis by inhibiting the growth of epithelial cells 7475. The incubation of human mammary epithelial cells with 5 μmol/l genistein induced a fivefold increase in the level of TGF-β secretion 76. The further analysis of media conditioned with human mammary epithelial cells revealed the presence of the active as opposed to latent form of TGF-β, thus implying a direct link between soy isoflavones and the TGF-β signalling pathway.

This review highlights current ambiguities regarding the precise role of phyto-oestrogens in breast cancer chemoprevention. Phyto-oestrogens were originally proposed as cancer protective agents following epidemiological observations revealing a low breast cancer incidence in soy-consuming populations. However, later research failed to ascertain a preventive role of phyto-oestrogens in breast cancer, possibly because of lack of long-term human dietary intervention trials. Although existing short-term studies do suggest an association between phyto-oestrogen consumption and a favourable steroid hormone profile, the precise effects on mammary cell proliferation remain unclear. The apparent lack of proliferative effects in adult dietary trials may be explained by both rodent studies and retrospective dietary recall studies, which suggest that breast cancer chemoprevention during adulthood is critically dependent on the prior consumption of phyto-oestrogens during adolescence. The physiological effects of phyto-oestrogens may be further influenced by additional factors such as ER expression status within the mammary tissue, precise quantities of phyto-oestrogens ingested and the ability to produce active metabolites.

Although the role of dietary phyto-oestrogens in breast cancer chemoprevention remains inconclusive, rodent studies encouragingly suggest that soy products may possess antitumour activities. Mammary tumours in phyto-oestrogen-fed rodents typically develop with an increased latency period, are smaller in size, are better differentiated and have a reduced capacity for metastasis. Perhaps the most clinically promising data concern the ability of phyto-oestrogen compounds to inhibit the proliferation of ER-negative breast cancer cultures and tumour xenografts in rodents. Although ER-positive cells are often growth stimulated by low phyto-oestrogen concentrations, stimulatory effects have not been reported in cells devoid of functional ER. Phytoestrogen compounds may thus potentially prove to be clinically useful for the treatment of ER-negative breast tumours that are typically unresponsive to traditional tamoxifen hormonal therapy. Although current data regarding self-medication with phyto-oestrogens by breast cancer patients remain controversial, the published literature does imply that these compounds may prove to be promising agents for future use in cancer therapy. Clinical trials are warranted to compare the efficacy of phyto-oestrogens with that of conventional endocrine therapies, and to assess their efficacy as alternatives to HRT to alleviate menopausal symptoms. The latter is especially pertinent given that the premature termination of the HABITS (Hormonal Replacement Therapy After Breast Cancer – Is It Safe?) trial 77 may lead to increased phyto-oestrogen use.

None declared.

DMBA = 7,12-dimethylbenz(a)anthracene; EGC = epigallocatechin; EGF = epidermal growth factor; EGCG = epigallocatechin-3 gallate; ER = oestrogen receptor; ERE = oestrogen response element; HRT = hormone replacement therapy; IGF = insulin-like growth factor; SRC = steroid receptor coactivator; TGF = transforming growth factor.