![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Nutrition Matters, Inc., Port Townsend, WA 98368 and Medical Oncology, Mayo Clinic, Rochester, MN 55905 *
2To whom correspondence should be addressed. E-mail: markm{at}olympus.net.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRecent historical perspectiveBackground on isoflavonesEffect of soy and...Soy and breast cancer...Markers of breast cancer...Insights from HRTTamoxifen and soyDoes the form of...Is it a question...ConclusionsLITERATURE CITED |
|---|
KEY WORDS: • soy • isoflavones • genistein • breast cancer patients • survival • risk
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRecent historical perspectiveBackground on isoflavonesEffect of soy and...Soy and breast cancer...Markers of breast cancer...Insights from HRTTamoxifen and soyDoes the form of...Is it a question...ConclusionsLITERATURE CITED |
|---|
27% of U.S. consumers reported using soy products at least once per week, which is nearly double the 1998 figure (1
), and soy foods sales are projected to be $6 billion in 2005, a nearly threefold increase from 1999 (2
). Several large food companies are now marketing user friendly soy products, such as breakfast cereals and energy bars, and several others have recently purchased soy food companies. Therefore, consumers will have even greater access to a wide variety of soy foods, which will likely result in a further mainstreaming of soy products.
Interest in soy and soybean constituents is driven primarily by reported potential health benefits in a variety of areas including cancer (3
–5
), osteoporosis (6
,7
) and coronary heart disease (CHD)4
(8
). Although in 1999, the Food and Drug Administration approved a health claim for the cholesterol-lowering properties of soy protein (9
) and soybeans are known to contain several bioactive components (10
), unarguably, it is the soybean isoflavones that have attracted most attention. Isoflavones have a very limited distribution in nature, and soybeans are essentially the only natural dietary source of these compounds (11
–14
). Recently, a work group convened by the NIH reviewed the isoflavone literature, a difficult task given that 600 papers on isoflavones are published annually (15
).
Isoflavones, classically viewed as possible selective estrogen receptor modulators (16
–18
), possess other nonestrogen receptor–mediated properties that also may contribute to their biological activity (19
–21
). For these reasons, some experts object to referring to isoflavones as phytoestrogens. The possibility that soy might exert the same hypothesized advantages of estrogen, especially in regard to CHD, osteoporosis and relief of vasomotor symptoms, but without the disadvantages, has led many women to view soy as an alternative to conventional hormone replacement therapy (HRT). For a variety of reasons, including fear of breast cancer, many women are reluctant to use or continue to use HRT (22
–24
). Only 35–40% of all women ever begin estrogen therapy and only 15% continue it long term (25
,26
). Estrogen discontinuation is associated with a loss of protection against CHD (27
) and osteoporosis (28
). Recent findings from studies involving older women challenging the protective effects of HRT against CHD (29
,30
) and even osteoporosis (31
,32
) will likely result in yet more women seeking alternatives.
One group of women most in need of and who frequently use alternatives consists of breast cancer patients (33
,34
). Nearly two thirds of breast cancer patients report experiencing hot flashes and most are reluctant to use estrogen replacement or have physicians who discourage them from doing so (35
). As the number of women taking tamoxifen to prevent or treat breast cancer increases, more women will experience menopausal symptoms for which conventional therapy may not be used (36
). In addition, 25% of the 180,000 women diagnosed annually with breast cancer are in their reproductive years, and 70% of them may develop premature menopause from chemotherapy (37
). For these reasons and because there is an intriguing body of literature suggesting soy consumption may reduce breast cancer risk and possibly improve survival, many breast cancer patients have embraced soy products, isoflavone supplements and foods to which isoflavones have been added. Nevertheless, the estrogenic effects of isoflavones observed in some experimental systems have recently led to considerable controversy among health professionals over the use of soy by breast cancer patients and to confusion among survivors.
The objective of this paper is to highlight studies that pertain most directly to this controversy in hopes of both offering guidance to healthcare professionals and stimulating further debate and research so that this issue can be resolved as quickly as possible. In an excellent review article, Bouker and Hilakiva-Clarke (38
) discussed several aspects of this issue (especially how the stage of life during which soy/isoflavone exposure occurs may affect breast cancer risk). Several organizations, including the American Cancer Society (39
), American Dietetic Association (40
) and American College of Obstetricians and Gynecologists (41
), have commented on the consumption of soy by breast cancer patients. However, in none of these cases do these statements appear to be the result of a comprehensive evaluation of the relevant literature.
|
Recent historical perspective |
|---|
|
TOPABSTRACTINTRODUCTIONRecent historical perspectiveBackground on isoflavonesEffect of soy and...Soy and breast cancer...Markers of breast cancer...Insights from HRTTamoxifen and soyDoes the form of...Is it a question...ConclusionsLITERATURE CITED |
|---|
) prompted the National Cancer Institute to hold a workshop on this subject (10
). Recommendations of this gathering led directly to the National Cancer Institute shortly thereafter issuing a request for applications for research in this area. The low breast cancer mortality rates in Japan and other Asian countries, where soy is commonly consumed, certainly made this hypothesis intriguing, especially because the fat/breast cancer connection began to weaken (43
,44
). Identifying dietary factors that affect breast cancer risk has been a particularly frustrating experience because conflicting data exist concerning the role of dietary fat (45
–48
), fiber (49
) and fruits and vegetables (50
). For this reason, the publication in 1991 of a case-control study conducted in Singapore showing that the consumption of modest amounts of soy was associated with an 50% reduction in premenopausal breast cancer risk (postmenopausal risk was unaffected) did much to further stimulate interest in soy (51
).
Although soy contains a number of putative anticarcinogens (10
), initial excitement over the hypothesized anticancer effects of this legume was based largely on the possibility that isoflavones might exert antiestrogenic effects on breast tissue, as other estrogen agonists/antagonists, such as tamoxifen, were known to do. In 1966, Folman and Pope (52
) were the first to demonstrate antiestrogenic effects of isoflavones, showing that in female mice, subcutaneously injected genistein inhibited estrone stimulation of uterine growth. In 1987, Akiyama et al. (53
,54
) serendipitously discovered that genistein was a specific inhibitor of tyrosine protein kinase, an enzyme frequently overexpressed in cancer cells (55
). This finding led to the widespread use of genistein as a tool for identifying whether certain cellular processes were under the control of a variety of kinases. In fact, genistein affects a variety of cellular molecules that modulate cell death and survival (21
,56
). Naturally, these observations fueled interest in the anticancer effects of isoflavones.
|
Background on isoflavones |
|---|
|
TOPABSTRACTINTRODUCTIONRecent historical perspectiveBackground on isoflavonesEffect of soy and...Soy and breast cancer...Markers of breast cancer...Insights from HRTTamoxifen and soyDoes the form of...Is it a question...ConclusionsLITERATURE CITED |
|---|
). There are also small amounts of a third isoflavone in soybeans, i.e., glycitein (7,4'-dihydroxy-6-methoxyisoflavone) and its glycoside, glycitin. The isoflavone content of raw soybeans is 1.0 mg/g with a range of 0.4–2.4 mg/g (57
–59
). Traditional soy foods (i.e., tofu, miso, natto) typically provide 0.2–0.4 mg/g of fresh weight product and 2–4 mg/g protein. Isolated soy proteins vary in isoflavone content (range 0.5–2.0 mg/g) although the average is 1 mg/g (14
). Native Japanese adults typically consume 30–40 mg (aglycone units) of isoflavones per day, which is roughly equivalent to between 1 and 1.5 servings (U.S. measures) of traditional soy foods (60
–62
).
Serum genistein and isoflavone levels increase in a dose-dependent fashion in response to genistein administration in animals (63
) and soy food consumption in humans (64
,65
) and can easily reach the low micromolar level (66
,67
). Plasma levels in free-living Asians are 500 nmol/L when measured after an overnight fast (68
,69
). Because the half-life of isoflavones is 6–9 h, fasting levels are much lower than postprandial levels (67
,70
–73
). Isoflavones circulate in plasma primarily in the conjugated form, mostly bound to glucuronic acid; <3% circulates in the free form (73
).
Estrogenic and nonestrogenic isoflavone properties.
Isoflavones have a spatial conformation similar to that of mammalian estrogens, bind to estrogen receptors and affect estrogen-regulated gene products (74
,75
). Isoflavones have traditionally been considered to be weak estrogens, possessing between 10-5 and 10-2 of the activity of 17ß-estradiol on a molar basis (74
–76
). It is really not possible, however, to arrive at a general estimate of activity because estrogenicity varies greatly depending on the assay used. Furthermore, isoflavones may bind less tightly than estrogen to serum proteins making them more available to tissues (76
) and they may be tissue selective, exerting quite pronounced estrogen-like effects in some tissues such as the coronary vessels (77
,78
) but not in other tissues such as the endometrium (78
–82
). However, even the lower estimates of estrogenicity suggest that isoflavones have the potential to exert physiologic effects in humans consuming soy foods because serum isoflavone levels will be 100-1000 times higher than endogenous estrogen levels (83
). The consumption of modest amounts (
100 mg) of isolated isoflavones was shown to exert potentially important biological effects in humans, such as enhancing systemic arterial compliance (84
,85
) and reducing urinary levels of 5-hydroxymethyl-2'-deoxyuridine, a marker for oxidative DNA damage (86
).
The higher binding affinity of isoflavones for estrogen receptor-ß (ERß) compared with ER
(87
–89
) and the different tissue distributions of these receptors (90
,91
) are often cited as support for the tissue-selective effects of isoflavones. This may be true, but binding affinity alone does not determine potency because the resulting conformational change in the ligand (isoflavone)-receptor complex varies among ligands regardless of binding affinity (92
).
Several investigators have examined the in vitro estrogenic and antiestrogenic potential of isoflavones, especially genistein, and most have concluded that genistein functions primarily as an estrogen agonist (93
–97
), although Miodini et al. (98
) concluded that genistein displays mixed estrogen agonist/antagonist properties. Isoflavones have been shown to down-regulate estrogen receptors, an effect that could lead to reduced estrogenic responses (99
). Interestingly, Pike et al. (92
) found that, in contrast to estradiol, in the ERß-genistein complex the AF-2 helix does not adopt the distinctive agonist position but instead lies in a similar orientation to that induced by estrogen receptor antagonists. No data on the ER-genistein complex was presented in that paper, but it is known that ERß is easier to antagonize than ER (92
).
In contrast to the findings of Pike et al. (92
), An et al. (18
) found that genistein was >1000-fold more potent at triggering transcriptional activity with ERß than ER. This difference is far greater than the 30-fold greater binding affinity of genistein for ERß than ER (100
). These findings indicate that genistein is a potent agonist for ERß and that the divergent transcriptional actions of estrogens and isoflavones result not only from their different binding affinities but from differences in their ability to recruit coregulators and trigger transcriptional functions of ER and ERß (18
). Therefore, determining the role of ERß in breast carcinogenesis appears to be critically important to understanding the effect of soy and isoflavone intake on breast cancer risk. For this reason, experimental models that do not include both ER and ERß and their receptor cofactors in a fashion similar to that found in human breast tissue are likely inappropriate for evaluating the effects of soy and isoflavones.
Isoflavones can exert hormonal and antiestrogenic effects in many ways without direct interaction with the estrogen receptor; for example, isoflavones have been shown in vitro to inhibit the activity of enzymes involved in estrogen metabolism. More specifically, isoflavones inhibit aromatase (101
–103
), a cytochrome P450 enzyme involved in three hydroxylation steps that convert the C19 androgens to aromatic C18 estrogenic steroids (101
) although the 50% inhibitory concentrations (IC50) are quite high and other flavonoids are far superior to the isoflavones in this regard (104
). Isoflavones also inhibit the 17ß-oxidoreduction of estrogens (96
,105
). The 17ß-hydroxysteroid oxidoreductases (also known as 17ß-hydroxysteroid dehydrogenases), which are present in steroidogenic cells as well as some target tissues of estrogen, convert the relatively weak estrogen estrone (E1, oxidized form) into the more potent estrogen, estradiol (E2, reduced form). Importantly, in contrast to aromatase inhibition, inhibition of 17ß-hydroxysteroid oxidoreductases occurs at relatively low micromolar concentrations (105
,106
). Low concentrations of isoflavones also inhibit 3ß-hydroxysteroid dehydrogenase (106
), which converts dehydroepiandrosterone into androstenedione, which can be converted into estrone (107
). Because it is thought that the estrogens, which drive the growth of estrogen-sensitive mammary tumors, are generated locally (108
), the effects of isoflavones on estrogen metabolism at the tissue level may be particularly important.
A recent example of in vivo antiestrogenic activity of soy occurred in adult, surgically postmenopausal female macaques (Macaca fascicularis); the consumption of an isoflavone-rich soy protein isolate inhibited the stimulatory effects of exogenously administered estradiol on mammary gland cell proliferation (109
). Clearly, when evaluating the possible hormone-related effects of isoflavones, it is necessary to look beyond the estrogen receptor, and when evaluating the likely overall biological effects of isoflavones, it is necessary to consider more than hormone-related activities. As noted previously, genistein is thought to be able to influence signal transduction (110
) by inhibiting the activity of many enzymes and influencing cellular factors that control the growth of cells (19
,21
,54
,56
); in some experimental systems, isoflavones demonstrate antioxidant activity (86
,111
,112
).
In vitro anticancer effects.
Genistein inhibits the growth of a wide range of cancer cells in vitro (3
,5
,20
), including both hormone-dependent and hormone-independent breast cancer cells, with IC50 values ranging from 10 to 50 µmol/L (106
,113
–123
). However, at concentrations <10 µmol/L, MCF-7 (an estrogen receptor–positive cell line) cell growth is stimulated by genistein (18
,97
,98
,106
,116
,124
). Genistein does not stimulate the growth of estrogen receptor–negative breast cancer cells in vitro (63
,97
). This biphasic effect is attributed to genistein exerting estrogen-like effects at lower and what may be considered physiologic concentrations but at higher concentrations exerting other non-estrogen receptor–mediated effects, for example, inhibition of the activity of one or more cellular molecules that control cell signaling, growth and death. This situation may be analogous to tamoxifen as noted by Bouker and Hilakivi-Clarke (38
) in that in male rat anterior pituitary cells, low (100 nmol/L) concentrations of tamoxifen increased prolactin secretion but at higher concentrations, estrogen-stimulated prolactin secretion was completed inhibited by tamoxifen (125
).
Conflicting data exist about the extent to which the addition of estrogen to the medium eliminates the stimulatory effect of genistein in vitro. For example, Wang and Kurzer (126
) and Shao et al. (117
) found that even in the presence of estradiol, genistein stimulated MCF-7 cell growth. In HepG2 human hepatoma cells transfected with rat ER or ERß plus an estrogen-responsive luciferase reporter gene, Casanova et al. (89
) found that set concentrations of genistein (1, 10 and 100 nmol/L) exerted additive effects when combined with estradiol across a complete dose-response range of estradiol and concluded that in vivo, the effect of genistein would be additive to that of endogenous estradiol. In contrast, Zava and Duwe (97
) found that in the presence of high estradiol (0.3 nmol/L) concentrations, genistein at 0.3–10 nmol/L did not further stimulate T47D cell (an estrogen receptor–positive cell line) growth and at slightly higher concentrations (30–300 nmol/L) exerted a modest inhibitory effect even though in the absence of estradiol these concentrations enhanced cell growth. Similarly, Miodini et al. (98
) found that although genistein (<5 µmol/L) modestly stimulated MCF-7 cell growth in the absence of estradiol, at a concentration of 5 µmol/L, genistein completely inhibited the stimulatory effects of 17ß-estradiol (10 nmol/L) on cell growth. Independent of effects on cell growth, genistein in vitro (50% effective concentration 1 µmol/L) inhibited invasion by a highly metastatic subline of BALB/c mammary carcinoma cells (127
), an effect also demonstrated in prostate cancer cells (128
).
The relevance of the in vitro anticancer effects of genistein has been questioned because free genistein concentrations in serum would never reach the low micromolar range, even in people consuming plentiful amounts of soy foods. However, Dalu et al. (129
) found that 0.25 mg/g dietary genistein (Japanese dietary content is 0.05 mg/g) down-regulated epidermal growth factor receptor levels in the dosolateral prostate of Lobund-Wistar rats despite free genistein concentrations in serum and prostate tissue of only 18.4 nmol/L and 17.5 pmol/g, respectively. As a result, Dalu et al. concluded that genistein may be more potent in vivo than in vitro. Note that the relative proportion of free genistein in tissues is higher than in the serum (130
) and that breast tissue isoflavone concentrations are two- to threefold higher than paired serum concentrations (Julie Maubach, Ghent University, unpublished data, 2001). Thus, breast tissue may be exposed to higher levels of biologically active isoflavones than was previously thought.
Finally, reports that genistein may inhibit angiogenesis may be particular germane to the issue of soy consumption by breast cancer survivors. Fotsis et al. (131
) were the first to show angiogenesis inhibition, finding that genistein inhibited endothelial cell proliferation and in vitro angiogenesis with concentrations of 5 and 150 µmol/L giving half-maximal inhibition, respectively. Fotsis et al. (132
) later determined, however, that the concentration of genistein for the half-maximal inhibitory effect on angiogenesis was only 10 µmol/L, the higher value in the previous report being due to the poor solubility of genistein in sodium bicarbonate compared with dimethyl sulfoxide. Genistein has also been shown to shift the proteolytic balance of matrix metalloproteinases and tissue inhibitors of matrix metalloproteinases toward antiproteolysis in media or ascitic fluid conditioned by actively growing Ehrlich cells (133
). The first in vivo inhibition of angiogenesis by genistein was reported by Shao et al. (121
), who found that angiogenesis was inhibited in mice injected subcutaneously with genistein. Others have confirmed antiangiogenesis effects in vitro and well as in vivo (128
,134
–138
).
|
Effect of soy and isoflavones on mammary tumorigenesis in adult animals |
|---|
|
TOPABSTRACTINTRODUCTIONRecent historical perspectiveBackground on isoflavonesEffect of soy and...Soy and breast cancer...Markers of breast cancer...Insights from HRTTamoxifen and soyDoes the form of...Is it a question...ConclusionsLITERATURE CITED |
|---|
Several studies examined the effects of different soy products on the development of chemically induced mammary cancer in adult animals. The data are somewhat inconsistent but generally show that in comparison with control diets, the substitution of soy protein for other protein (generally casein) typically found in a standard laboratory diet modestly (25–50%) reduced tumor incidence or, more commonly, tumor multiplicity (42
,139
–143
). However, not all studies showed protective effects (144
–147
). In one study, although Hakkak et al. (139
) found that soy protein reduced 7,12-dimethylbenz(a)anthracene (DMBA)-induced tumor incidence relative to casein, tumor inhibition was only about half that seen for whey protein. Typically in these animal studies, soy protein isolate was the sole source of protein in the diet (20% by weight) although other forms of soy such as miso (140
,142
,148
) and soy flour (149
) have been shown to reduce tumorigenesis. Barnes et al. (149
) found that alcohol-extracted soy protein was less effective in inhibiting mammary tumor development than unextracted protein, suggesting that isoflavones may have accounted for the anticancer effects.
Several studies directly examined the effects of isolated isoflavones or isoflavone-rich extracts on mammary tumorigenesis (143
,150
–152
). For example, daidzein and genistein given by injection each reduced N-methyl-N-nitrosourea (MNU)-induced mammary carcinogenesis (number of tumors per rat) by 20%, but the results were not significant (151
). Arjmandi et al. found that neither the synthetic isoflavone, ipriflavone, nor genistein had an effect on MNU-induced mammary tumorigenesis although the combination was very inhibitory (152
). More substantial inhibitory effects were found by Ohta et al. (143
). Their work reported that the number of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhiP)-induced mammary tumors per rat was 2.2 ± 0.4 for rats consuming a diet containing a 0.02% isoflavone mixture and 1.5 ± 0.3 for those consuming twice this amount of isoflavones; both values were much lower than the 2.6 ± 0.5 tumors/rat in the control group.
Decreases in tumor incidence were also noted in F344 rats in response to the consumption of diets containing 1.5 or 5% hypocotyls [embryo buds of soybeans; concentrated in isoflavones relative to the whole soybean (153
)] in comparison with the control diet; incidence in these groups was 25, 26 and 37.5%, respectively (150
). It should be noted that in rodents in response to the intake of isoflavones via soy or an isoflavone extract with an isoflavone profile (genistein > daidzein >> glycitein) similar to that found in soybeans, equol is the predominant isoflavonoid found in serum (154
), whereas only about one third of humans actually produce equol (155
–157
). Equol is converted from daidzein by intestinal bacteria (158
), binds to estrogen receptors more tightly than daidzein (88
) and may be more estrogenic (74
) and anticarcinogenic (99
).
The most pronounced inhibitory effects in response to the administration of an isolated isoflavone were reported by Gotoh et al. (140
), who found that a low dose of biochanin A (10 mg/kg) significantly (P < 0.01) reduced tumor multiplicity (tumors per rat; 0.7 vs. 2.2 for control) and that a higher dose (50 mg/kg) also reduced tumor incidence. Biochanin A is a methylated isoflavone that is converted to genistein in vivo (159
) and in MCF-7 cells in vitro (160
). Finally, in contrast to all other studies discussed in this section, Day et al. (161
) recently found that in wild-type but not ER knockout mice, genistein (1 g/kg diet) fed continually beginning at age 3 wk increased the number of DMBA-induced malignant tumors compared with the number in mice not given genistein, although the number of benign tumors, latency and size of the tumors were not affected.
Tumor progression.
The inhibitory effects of soy/isoflavones on the development of chemically induced mammary cancer as discussed above may not be relevant to the issue of soy consumption by breast cancer survivors. Several investigators used models that may more closely resemble the situation faced by breast cancer patients by looking at the effects of soy and isoflavones on tumor progression. For example, Hawrylewicz et al. (162
) first fed rats a casein-based AIN-76 diet and then injected them with the breast carcinogen MNU. When the first tumor grew to 0.3–0.5 cm in diameter (mean latency was 14.7 wk), it was surgically removed and half the animals continued on the casein diet, whereas the other half was fed soy protein isolate. Rats in the soy group developed significantly fewer tumors (53 vs. 128), and histopathological analysis indicated that tumors in the casein-fed rats had considerably more aggressive characteristics than did the tumors in the soy-fed rats. Hawrylewicz et al. (162
) concluded that the inhibition of chemically induced mammary tumor development reported in a previous study (141
) and tumor progression in this study (162
) resulted from the lower methionine content of soy protein. Methionine restriction is known to retard carcinogenesis (163
); it is unlikely, however, that this mechanism could contribute to the low breast cancer mortality in Japan because only 10% of the total protein intake of the Japanese people is derived from soy (164
). This mechanism would not be likely to be relevant to breast cancer patients consuming a mixed diet containing modest amounts of soy protein, although it could be when much of the total protein intake is derived from soy.
The findings of Hawrylewicz et al. (162
), however, are in contrast to those from a study by Charland et al. (165
). In this experiment, female Lewis rats (n = 60) were injected subcutaneously with MAC-33 cells (a mammary tumor cell line known to metastasize spontaneously) and then randomly assigned to receive (intraperitoneally) 18 mg soybean (isoflavone-enriched) extract, 18 mg heat-sterilized (to destroy protease inhibitors) soybean isoflavones or saline five times per week for 30 d. There were significant increases in tumor weight, tumor volume and the tumor-to-carcass ratio in the isoflavone groups relative to the control group. There was also a significant increase in the number of lung metastases (31%; Kruskal-Wallis test) in the rats receiving heat-sterilized soybean isoflavones.
Undoubtedly, Hsieh and colleagues (116
) from the University of Illinois conducted the animal study most widely cited in support of breast cancer patients avoiding soy. They found that dietary genistein (750 µg/g) stimulated the growth of subcutaneously implanted MCF-7 cells (cells were initially stimulated by estrogen) in ovariectomized athymic mice although the effects were much less pronounced than those of estrogen. In a later study by these investigators using a similar experimental design, isolated soy protein containing various amounts of genistein (15, 150 and 300 µg/g) increased tumor growth in a dose-dependent manner and to a similar extent as when isolated genistein was administered at these same doses (166
). Genistein however, does not appear to stimulate the growth of estrogen receptor–negative mammary cancer cells in vivo (63
).
Although the genistein concentration of the Asian diet (50 µg/g) is much lower than the genistein concentration used by Hsieh et al. (116
) (750 µg/g) and Day et al. (161
) (1000 µg/g), serum genistein levels in response to this amount of genistein in rodents are 1–2 µmol/L (63
,116
,166
), which is similar to that resulting from the consumption of fairly modest amounts (1–2 servings) of soy foods by humans (65
) and, as noted previously, not dramatically lower than that found in free-living Asians (68
,69
).
In contrast to the research from Helferich and colleagues (116
,166
), Shao et al. (121
), using very low doses (0.1–0.5 mg/kg body weight) of genistein, injected subcutaneously and Zhou et al. (Jin-Rong Zhou, Harvard Medical School, Boston, unpublished data, 2001), using dietary genistin (the glycoside form of genistein), observed tumor inhibition in intact nude mice fed estrogen pellets and orthotopically implanted with MCF-7 cells. The dietary genistin concentration (1400 µg/g) used by Shao et al. is equivalent to 900 µg genistein/g, which is very similar to the 750 µg genistein/g used by Helferich and colleagues (116
). Although still speculative, these four rodent studies collectively suggest that in a low-estrogen environment, perhaps as exists in a postmenopausal women (and possibly breast cancer patients as a result of chemotherapy-induced menopause), genistein is estrogenic and has a proliferative effect on breast tissue, but in a high estrogen environment such as exists in a premenopausal women, it has an antiproliferative and possibly antiestrogenic effect (116,121,166; Jin-Rong Zhou, 2001). This conclusion agrees with findings by Makela et al. (167
) who showed that diets containing soy increased uterine weight in immature outbred Han-NMRI mice not given diethylstilbestrol but decreased uterine weight in mice fed diethylstilbestrol.
Effects on the breast may differ from effects on the uterus; furthermore, in the study by Shao et al. (121
), genistein also inhibited the in vivo growth of estrogen receptor–negative mammary cancer cells (MDA-MB-231). Therefore, tumor inhibition in the case of both cell lines (MCF-7 and MDA-MB-231 cells) may have been the result of an estrogen receptor–independent mechanism. More importantly, it is unclear which hormonal environment is more representative of breast cancer patients or postmenopausal women, i.e., that which occurs in ovariectomized mice or in intact mice fed estrogen pellets. Plasma levels of estradiol in ovariectomized athymic mice are 99–139 pmol/L, similar to the levels seen in postmenopausal women (168
). Although postmenopausal serum estradiol levels (55 pmol/L) (83
,169
) are 10% of the level in premenopausal women (169
), some data indicate that the concentration of estradiol in breast cancer tissue of postmenopausal women is similar to that in premenopausal women (170
) and that breast tissue estradiol concentrations are higher than paired serum concentrations (171
). Consequently, identifying a high vs. low estrogen environment may not be as simple as identifying pre-vs. postmenopausal women, especially in the case of breast cancer patients.
Effects of early exposure.
Perhaps the most intriguing animal data come from the laboratory of Lamartiniere and colleagues (172
) at the University of Alabama. They showed that in rats, neonatal exposure to pharmacologic amounts of genistein given by injection on d 2, 4 and 6 of life (173
,174
) and during the prepubertal period given on 16, 18, and 20 d postpartum (175
) and via the diet from conception to 21 d postpartum at physiologic concentrations (25 mg/g body weight) (176
) reduces the later development of DMBA-induced (administered on postpartum d 50) mammary cancer by 50%. In the dietary study, total genistein concentrations in the serum and mammary glands of 21-d-old offspring were only 1.8 µmol/L and 370 nmol/g, respectively. Hilakivi-Clarke et al. (177
) also found that prepubertal genistein (1 mg/kg body weight) administration reduces DMBA-induced mammary cancer in rats treated between postnatal d 7 and 20. Genistein exposure significantly reduced tumor multiplicity in this study, and 60% of the tumors in the genistein group were not malignant, whereas all of the tumors analyzed for histopathology in the vehicle group were adenocarcinomas.
The proposed mechanism for the protective effect of early genistein exposure is from the stimulation of cell proliferation in the mammary gland resulting in enhancement of mammary gland maturation. Early genistein exposure results in fewer terminal end buds and more lobules, and lower rates of cell proliferation in the terminal end buds. Lobules are the most differentiated terminal ductal structures, whereas terminal end buds are the least differentiated terminal ductal structures and the most susceptible to carcinogenesis (178
). In the study by Hilakivi-Clarke et al. (177
), a higher number of differentiated alveolar buds and lower number of terminal ducts were present in the DMBA-treated mammary glands of the genistein-exposed rats. The genistein findings are consistent with the protective effects of early exposure to estrogen and progesterone in rats (179
–181
) and of early but not late pregnancy in women (182
), although recent findings from Guzman et al. (181
) suggest that the protective effects of early estrogen exposure are not simply due to an overall enhancement of differentiation. A particularly fascinating finding from the Larmartiniere group is that when genistein is given to adult rats, mammary carcinogenesis is reduced only if these animals were also exposed to genistein early in life (Coral Lamartiniere, University of Alabama, Birmingham, unpublished data, 2001). The timing effect may not be straightforward because Hilakivi-Clarke et al. (183
–185
) found that the administration of genistein to dams increased DMBA-induced mammary carcinogenesis in the offspring.
|
Soy and breast cancer risk: epidemiology |
|---|
|
TOPABSTRACTINTRODUCTIONRecent historical perspectiveBackground on isoflavonesEffect of soy and...Soy and breast cancer...Markers of breast cancer...Insights from HRTTamoxifen and soyDoes the form of...Is it a question...ConclusionsLITERATURE CITED |
|---|
–189
). Whether soy contributes to this better survival is unknown because no studies looking specifically at the effect of soy consumption on breast cancer survival have been published, although Huang et al. (190
) recently reported that tofu consumption was associated with enhanced survival among gastric cancer patients. However, not only are the etiologies of breast and stomach cancer different, but in this study, dietary assessment was conducted before cancer diagnosis and likely reflected lifelong consumption. Thus, these observations shed little light on the possible effect of soy exposure on breast cancer survival among Western patients who did not consume soy before diagnosis, especially considering the work of Lamartiniere et al. in which adult genistein exposure reduced breast cancer risk only in animals previously exposed to genistein during early life (Coral Lamartineiere, unpublished data, 2001).
As noted previously, the low Asian breast cancer mortality rates provided much of the basis for initial enthusiasm about the possible anticancer effects of soy (191
). However, on the basis of national household survey dietary data for 1980–1985, Nagata (164
) did not find that soy or isoflavone intake was related to breast cancer mortality in 47 prefectures in Japan. The existing case-control and prospective studies are also generally not supportive of the hypothesis that adult consumption of soy reduces postmenopausal breast cancer risk and are only modestly supportive in regard to premenopausal breast cancer (192
–196
). In fact, no epidemiologic studies conducted in Asia were identified that found that adult consumption of soy reduces breast cancer risk in postmenopausal women (51
,197
,198
). Interestingly, Wu et al. (199
) found that among Asian Americans, tofu consumption protected against both pre- and postmenopausal breast cancer although intake was quite modest, which is consistent with data showing that Asian immigrants consume less soy than do native Asians (200
–202
). In a follow-up analysis of data, protective effects were found only in non-U.S.-born Asians, not in U.S.-born Asians (203
). One explanation for this observation is that among women born in Japan but not in the United States, soy consumption reflects a more traditional lifestyle that protects against breast cancer rather than soy intake per se being protective.
Data for premenopausal breast cancer risk are inconsistent, with some studies reporting no association (198
), others reporting only modestly reduced risks (197
) and some finding quite pronounced protective effects (51
). None of the Asian studies found that soy intake was associated with an increased breast cancer risk, however. Typically, in Asian epidemiologic studies, consuming soy at least once per day is classified as the highest intake level, whereas less than once per week is the lowest (4
,5
,51
,196
,197
,203
).
Several case-control studies conducted in Western countries found that soy intake (based on dietary data or urinary isoflavone excretion) protects against breast cancer; not unexpectedly, however, intake was extremely low in these populations. Consequently, the biological relevance of the inverse correlation between soy intake and breast cancer risk observed in these studies is of questionable significance (204
–206
). The inverse relationships noted in these studies may indicate that typical Western women who eat any soy lead a lifestyle that is protective against breast cancer.
In contrast to these relatively unimpressive epidemiologic findings, Shu et al. (207
) found that women who consumed tofu during their teenage years (13–15 y) were less likely to develop premenopausal and postmenopausal breast cancer as adults. The odds ratio for girls in the highest (11 g soy protein/d) vs. the lowest quintile of intake (<2.2 g soy protein/d) was 0.51 [95% confidence interval, 0.40–0.65; trend test, P < 0.001]. These data are extremely noteworthy for two important reasons: first, they are consistent with the animal work by Lamartiniere’s group showing the protective effects of early soy exposure (172
), and second, they are consistent with migration data showing that for breast cancer in contrast to prostate cancer, early life events are particularly important (208
).
|
Markers of breast cancer risk: clinical studies |
|---|
|
TOPABSTRACTINTRODUCTIONRecent historical perspectiveBackground on isoflavonesEffect of soy and...Soy and breast cancer...Markers of breast cancer...Insights from HRTTamoxifen and soyDoes the form of...Is it a question...ConclusionsLITERATURE CITED |
|---|
–211
), whereas several other studies found no changes (212
–215
). Wu et al. (216
) found that in premenopausal women, the serum luteal phase estradiol levels decreased in response to soy but only in Asian, not Caucasian, women. Duncan et al. noted few hormonal changes in premenopausal (79
) or postmenopausal women (217
) who were fed soy protein containing various amounts of isoflavones. Other studies suggest that soy consumption favorably alters estrogen metabolism on the basis of the urinary production of potentially carcinogenic estrogen metabolites [16-(OH) estrone, 4-(OH) estrone, and 4-(OH) estradiol] (218
,219
) although Martini et al. (215
) found no effect of soy protein on estrogen levels or metabolism in premenopausal women. Furthermore, there are conflicting data regarding the relative importance of the 16()-hydroxylation vs. the 2-hydroxylation of estrogen on breast cancer risk (220
).
Early in vitro work suggested that isoflavones increase sex hormone binding globulin (SHBG) levels (221
,222
). In support of these data, Pino et al. (223
) recently found that in response to soy milk consumption (20 g soy protein/d), serum isoflavone levels in postmenopausal women were positively correlated with increases in SHBG levels, especially in those with low initial SHBG values. However, most feeding studies have not shown that soy or isoflavone consumption significantly increases SHBG (79
,80
,215
,224
–226
) and, in fact, some studies have reported decreases (212
,227
,228
). Similarly, although a few studies reported increases in menstrual cycle length (209
,210
,228
), several others have not (79
,215
,216
,229
). Finally, one of the more potentially important findings was reported by Lu et al. (229
), who found that soy decreased serum progesterone levels in premenopausal women by 45% (P < 0.0001) compared with levels during the home (usual) diet period; the women in this study consumed 1080 mL soy milk/d, which provided 154 mg isoflavones.
The next study has largely been responsible for raising concerns about soy consumption by breast cancer patients. Petrakis et al. (212
) found that daily soy consumption (38 g soy protein isolate, 80 mg isoflavones) over 5 mo was associated with an increase in breast nipple aspirate fluid (NAF) secretion and breast cell hyperplasia in premenopausal women. Previous epidemiologic research by these investigators showed that nonlactating women who can secrete breast fluid are at an increased breast cancer risk compared with those who cannot (230
) and that abnormal NAF cytology is associated with a further increased breast cancer risk (231
) although the increased risk may be confined to women with a previous history of breast biopsy (232
). Interestingly, Chinese-born premenopausal women in the United States are about twice as likely to secrete NAF than Chinese women born in China (233
). Petrakis et al. (212
) found no effect of soy consumption on NAF secretion in postmenopausal women overall; surprisingly, however, there was a significant increase in NAF secretion in postmenopausal women (n = 4) receiving estrogen replacement therapy.
On the basis of the above findings, Petrakis et al. concluded that soy exerts estrogenic effects on breast tissue. However, this study did not include a control group and fluid secretion increased in women even after soy feeding was discontinued. Bouker and Hilakivi-Clarke (38
) noted that women were eligible for this study only if they were secretors of NAF, which means that they were at an increased risk of breast cancer and perhaps particularly sensitive to soy. Thus, in theory, the findings may not be applicable to all women, whereas they could be particularly germane to breast cancer patients.
Another study that has raised concern in breast cancer survivors examined the effects of feeding textured vegetable protein (60 g, 45 mg isoflavones/d) for 2 wk on breast cell proliferation in premenopausal women with benign or malignant breast disease (234
). Increased cell proliferation is generally considered to be a marker for increased cancer risk (235
–237
) although this is not always the case (238
). A preliminary analysis of this study, based on biopsies from only about half of the study subjects (n = 48) indicated that soy consumption markedly increased breast cell proliferation (239
); in the final analysis, however, which included all 84 subjects, no such effect was found (234
). Nevertheless, soy did appear to exert a weak estrogen-like effect on breast tissue because biopsies indicated that the expression of two proteins, pS2 and apolipoprotein D, found in breast cells, were up-regulated and down-regulated, respectively, which is similar to the effects of estrogen. However, there was no effect of soy supplementation on estrogen and progesterone receptor status, apoptosis, mitosis or Bcl-2 expression. The investigators concluded that soy exerted a weak estrogenic effect on breast tissue but also that the long-term implications of this effect were unclear because soy did not increase cell proliferation and because the study lasted only 2 wk. It is certainly possible that 2 wk is not sufficiently long for any antiestrogenic and even antiproliferative effects of soy to become evident.
Recently, two laboratories examined the effect of isoflavone supplements on breast tissue density. Mammographic density appears to be an excellent short-term marker of the effect on the breast of potential preventive interventions for breast cancer (240
–242
). Differences in the parenchymal pattern of the breast on mammography reflect differences in the amounts of stromal, epithelial and fat tissue present in the breast. Extensive areas of mammographically dense breast tissue are strongly associated with an increased risk of breast cancer. Tamoxifen use and low fat diets have been shown to decrease density, whereas HRT has been shown to increase it (242
).
In one study, women (n = 175) aged 49–65 y with Wolfe’s P2 or DY mammographic patterns were randomly assigned to receive either an isoflavone tablet (40 mg isoflavones derived from red clover) or placebo for 1 y (Charlotte Atkinson, Fred Hutchinson Cancer Research Center, Seattle, unpublished data, 2001). After 1 y, breast tissue density in postmenopausal women decreased significantly in the isoflavone group and nonsignificantly in the placebo group, but differences between the groups were not significant. There were also no differences between the isoflavone and placebo groups for pre- and perimenopausal women. However, when subjects were divided into tertiles of age, in women who were 56–65 y old and received isoflavones, breast tissue density significantly decreased (P < 0.05), whereas in women of this age who received a placebo, it increased, although not significantly so. In the second study, premenopausal women (n = 34) were randomly assigned to receive either 100 mg isoflavones (derived from soy) or placebo for 1 y (Gertraud Maskarinec, Cancer Research Center of Hawaii, unpublished data, 2001). No differences in breast tissue density were noted between groups. Therefore, although direct comparisons are needed, isoflavones appear to exert effects on breast tissue density opposite to those of HRT (243
).
|
Insights from HRT |
|---|
|
TOPABSTRACTINTRODUCTIONRecent historical perspectiveBackground on isoflavonesEffect of soy and...Soy and breast cancer...Markers of breast cancer...Insights from HRTTamoxifen and soyDoes the form of...Is it a question...ConclusionsLITERATURE CITED |
|---|
,245
). It has generally been accepted that higher endogenous estrogen levels (246
) and greater lifelong exposure to estrogen increase breast cancer risk, which is presumably why earlier age at menses and later age at menopause (247
) and higher bone mineral density (248
–252
) are considered to be risk factors for breast cancer and why obesity (because of the greater peripheral synthesis of estrogen and lower SHBG levels) appears to decrease survival of breast cancer patients (253
,254
). As recently as 1997, a consensus development conference recommended that estrogen replacement should be avoided, if possible, by patients with a history of breast cancer (255
).
This position is not without controversy because some data suggest that estrogen is not contraindicated for such patients (37
,256
,257
); it does not appear that estrogen increases disease-specific mortality because breast tumors in HRT users tend to be smaller, better differentiated and more likely to be of a favorable histological type (245
,258
–260
). For example, in a small study in Spain, women with breast cancer (n = 121) who were undergoing HRT at time of diagnosis were found to have lower tumoral stages, a lower degree of affected axillary lymph node dissemination and a greater percentage of well-differentiated tumors than women (n = 121) who were not using HRT (261
). Recently, O’Meara et al. (262
) found that in a group of 2755 women diagnosed with invasive breast cancer, hormone use was associated with a significant decrease in breast cancer recurrence and mortality, although the number of women (n = 174) who used hormones was small and the follow-up period was not long. However, after a number of factors were adjusted for, the relative risk of contralateral cancer was actually increased by one third in HRT users.
Even the view that estrogen therapy increases breast cancer risk in healthy women is now being challenged. Recent observational data indicate that estrogen per se increases breast cancer risk only very slightly, whereas the combination of estrogen plus progestogen may increase breast cancer risk two- to threefold over a woman’s lifetime (263
–268
). These epidemiologic observations are consistent with the increased breast cell proliferation that occurs during the luteal phase of the menstrual cycle when progesterone levels are elevated (269
,270
), with increased mammary and breast cell proliferation (271
