Flaxseed attenuates the tumor growth stimulating effect of soy protein in ovariectomized athymic mice with MCF-7 human breast cancer xenografts

Abstract

In several epidemiological studies, a phytoestrogen-rich diet containing lignans and isoflavones is associated with reduced breast cancer risk, but experimental findings are controversial. In postmenopausal mammary cancer xenograft model, flaxseed (FS), a rich source of plant lignans, reduced breast cancer growth, while soy protein (SP), a rich source of isoflavones, enhanced it. The intake of phytoestrogens is increasing particularly among postmenopausal women, emphasizing the importance of elucidating their interactive effects on breast cancer. Our study determined the effect of FS and SP diets, alone and in combination, on the established human breast cancer MCF-7 tumor growth in ovariectomized athymic nude mice. Tumor bearing mice were divided into 4 groups and fed for 25 weeks either the basal diet (BD), or BD supplemented with 10% FS, 20% SP or 10% FS and 20% SP. After estrogen deprivation, FS regressed the tumor size similar to that of control. SP initially regressed the tumors but starting at week 13, the tumors regressed significantly less than in control and 43% of the tumors were regrowing until the end of the experiment and were significantly larger in size than in control. The combination of SP with FS reduced the tumor growth similar to that of control, as suggested also by the reduced tumor cell proliferation index. In conclusion, dietary FS did not stimulate the growth of estrogen responsive MCF-7 cancers in ovariectomized mice, while long-term consumption of SP did. Furthermore, FS reduced the tumor growth stimulating effect of SP to the same level as control, suggesting tumor growth attenuating effect of FS. © 2006 Wiley-Liss, Inc.

High dietary intake of phytoestrogens such as isoflavones and lignans has been associated with reduced breast cancer risk in several studies, although some of the studies are controversial.1 Soy protein (SP) is the major dietary source of isoflavones such as genistein, while flaxseed (FS) is one of the richest dietary sources of plant lignans, mainly secoisolariciresinol diglycoside (SDG).2 When ingested, SDG is converted to mammalian lignans enterodiol and enterolactone by the intestinal microbiota.3, 4 The biological effects of lignans have been attributed particularly to enterolactone, which has been suggested to act as an endocrine active compound. In vitro, both soy isoflavones and enterolactone have been demonstrated to interact with estrogen receptor (ER) α and β. Soy isoflavones bind and transactivate both ERs with high affinity5, 6 while enterolactone binds weakly to ERs7 and activate gene expression via ERs at high concentration.7, 8

Isoflavone and lignan-rich diets are reported to have distinct effects on the growth of estrogen responsive tumors in vivo. After estrogen deprivation, SP and its isoflavone genistein have been shown to increase the growth of the MCF-7 tumors in ovariectomized animals,9, 10, 11 while no stimulation of MCF-7 tumor growth has been observed when mice were exposed to FS12 or its mammalian lignan metabolites, enterodiol and enterolactone.11 In addition, FS, SDG and enterolactone have been shown to reduce the growth of the dimethylbenzanthracene (DMBA)-induced hormone responsive mammary tumors in nonovariectomized rats,13, 14, 15 suggesting a protective role of lignan-rich diets and enterolactone in breast cancer.

In soy-containing diets, however, soy isoflavones are not consumed alone but in combination with lignans as many other food items such as whole grains and vegetables contain lignans.2 Furthermore, whole soy beans have also been reported to contain minor amounts of plant lignans, e.g. secoisolariciresinol, but in concentrations approximately 1,000-fold lower when compared to that of FS.2, 16 In addition to lignans, FS is a rich source of n-3 fatty acids. Combining the use of soy with FS may have anticancer effect as combining isoflavones with n-3 fatty acids have been shown to reduce the growth of human breast cancer cells.17 Because the intake of isoflavone-rich soy and lignan-rich FS is increasing particularly among postmenopausal women, it is of interest to determine their putative interactive effects on breast cancer growth. In our study, the long term effect of SP and FS alone and in combination were determined on the established estrogen responsive human breast cancer MCF-7 xenografts in ovariectomized athymic nude mice.

ALA, α-linolenic acid; ANOVA, one way analysis of variance; BD, basal diet; DMBA, dimethylbenzanthracene; ER, estrogen receptor; FS, flaxseed; LSD, least significant difference; MNU, methylnitrosourea; SDG, secoisolariciresinol diglycoside; SEM, standard error of mean; SP, soy protein; TUNEL, terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling.

Material and methods

Experimental diets

The basal diet (BD) was AIN-93G formulation18 modified to have high (20%) fat content and corn oil substituted for soy oil (Table I). FS (Linnot variety), containing 0.2 g/kg SDG, was obtained from Omega Products (Melfort, Saskatchewan, Canada). Soy protein isolate (SP) containing isoflavones genistein, daidzein and glycitein (1.61, 1.31 and 0.21 mg/g as aglycones, respectively) was manufactured by Protein Technologies International (St. Louis, MO). FS diet contained 10% of freshly ground FS, SP diet had 20% SP instead of casein and FS + SP diet contained both 10% FS and 20% of SP (Table I). The energy densities were similar in all test diets. The diets were prepared by Dyets (Bethlehem, PA) and sterilized by Co⁶⁰ radiation by Isomedix (Whitby, ON, Canada).

Table I. Composition of Experimental Diets

Component	Group
	Control (g)	FS (g)	SP (g)	FS + SP (g)
Casein	200.0	177.6	0.0	0.0
Corn starch	262.4	249.3	262.4	225.4
Dextrose	132.0	132.0	132.0	132.0
Sucrose	100.0	100.0	100.0	100.0
Cellulose	50.0	22.0	48.47	22.0
Corn oil	200.0	163.5	200.0	163.5
AIN-93G minerals	40.34	40.34	40.34	40.34
AIN-93 VX Vitamins	11.5	11.5	11.5	11.5
L-Cystine	3.47	3.47	3.0	3.47
Methionine	0.0	0.0	2.0	1.53
Choline Bitartrate	0.289	0.289	0.289	0.289
Flaxseed	0.0	100.0	0.0	100.0
Soy protein	0.0	0.0	200.0	200.0

Animals and housing conditions

Animal care and studies were performed according to the Guide to the Care and Use of Experimental Animal.19 All the experimental protocols were approved by the University of Toronto Animal Care Committee. Ovariectomized athymic female mice (Balb/c nu/nu, 4–5 weeks old) were obtained from Charles River Canada (St-Constant, PQ, Canada) and maintained in micro-isolator cages (4 mice per cage) in a pathogen-free facility with 12-hr light–dark cycle at 22–24°C and 50% humidity. Animals were fed the diets and water ad libitum.

MCF-7 cell culture

MCF-7, estrogen responsive human breast cancer cells (The American Type Culture Collection, Manassas, VA), were maintained in Dulbecco's minimum essential medium/F12 supplemented with 10% heat-inactivated fetal bovine serum and 1% antibiotic–antimycotic solution.11 The cells were grown to 70–90% confluence and given fresh medium every 2–3 days, and a day before cell harvest. For cell injections, the cells were trypsinized, resuspended in serum-free medium containing Matrigel (1:1 vol) as previously described,11, 12 and kept on ice. Cell viability after cell injections was confirmed to be at least 85% by trypan blue exclusion assay.

Study design

After 1-week acclimatization with basal diet, mice were injected with MCF-7 cells (5 × 10⁵ cells in 50 μl) orthotopically into 4 sites of mammary fat pads and implanted s.c. with an estradiol pellet (1.7 mg, 60-day release; Innovative Research of America, Sarasota, FL) as previously described.11, 12 After injection, tumors were palpated weekly. The tumor surface area was calculated using the formula (length/2 × width/2) × π. When average tumor area reached ∼35 mm² at week 6, the estradiol pellet was removed and all mice were divided into 4 dietary treatment groups: (1) Control group fed with BD, (2) FS diet group, (3) SP diet group and (4) FS + SP diet group. At the start of the treatment, the average tumor size and body weight were similar in all groups. Food intake, body weights and palpable tumor area were monitored weekly. The estrogen-facilitated growth of the established tumors was verified in mice implanted with a fresh estradiol pellet and fed with BD (positive control). These mice were killed after 7 weeks of treatment because of the high tumor burden. After the start of the treatments, 1 mouse in each of the control, FS and FS + SP groups, died before the end of the experiment due to unknown reason (data not included). All remaining mice in 4 treatment groups (7 in BD, 8 in FS, 10 in SP and 9 in FS + SP) were sacrificed at week 25 by CO₂ asphyxiation, and their body weight, tumor weight and tumor volumes [(length/2 × width/2 × thickness/2) × π] were recorded.

Assessment of tumor growth category

At endpoint, all tumors were categorized as growing and non-growing tumors based on their growth pattern, as previously reported,15 to further clarify the response of each tumor to dietary treatments. Tumors that initially regressed in size in response to estradiol pellet removal but grew back to at least 50% of their original surface area during the dietary treatment period were assessed as growing tumors. Tumors that regressed to non-palpable size were classified as completely regressed tumors. All the remaining tumors (i.e. tumors with surface area reaching less than 50% of the pretreatment area, regressing tumors and tumors with stabilized surface area) were classified as nongrowing tumors.

Ki-67 labeling index and apoptosis

The Ki-67 labeling index used as cell proliferation marker was determined by immunohistochemistry, as previously described.11 The cell apoptosis index was determined by in-situ terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) assay demonstrating the DNA fragmentation (ApopTag Detection Kit, Intergen, Purchase, NY) and was run based on the manufacturer's instructions as previously described.11 All slides were read blind to the treatment groups under a light microscope at 400× magnification. The Ki-67 labeling index was determined by counting the percentage of positive cells over total number of cells (500–1,200) counted from 5–8 fields. The number of apoptotic cancer cells was counted and expressed as apoptotic cell number/mm².

Statistical analyses

Analysis of variance (ANOVA) with general linear model repeated measures procedure was used to determine the difference among treatment groups in palpable tumor regression over treatment time (i.e. regression rate) followed by post-hoc least significant difference (LSD) test. Normally distributed data (animal body weights, food intake, tumor area and tumor weights determined by using Shapiro-Wilk's normality test) were analyzed with ANOVA followed by LSD test. The proportions (frequencies) of tumor with different growth patterns were analyzed with Chi-square (χ²)-test. Non-normally distributed tumor volume data were analyzed with Kruskal–Wallis median test, followed by Mann–Whitney U-test. These analyses were performed by using Statistica software for Windows (Stat Soft, Tulsa, OK). The differences among groups in Ki-67 labeling index and apoptosis were analyzed with one-way ANOVA followed by post-hoc Tukey's test using SigmaStat 2.0 software (San Rafael, CA). The acceptable level of significance was set at p ≤ 0.05 for all analyses. All data are presented as mean ± standard error of mean (SEM).

Results

Body weights and food consumption

No significant differences in food consumption or body weights were measured between the control and dietary treatment groups (FS, SP or FS + SP groups) in ovariectomized mice during the 25-week treatment period (data not shown), suggesting that the dietary treatments had no side effects on the general health condition.

Tumor growth responses

In the first 7 treatment weeks (except week 5), tumors in the mice fed FS diet regressed more than in the control group (p < 0.05), but during the remaining treatment weeks the reduction of the initial tumor area was similar in both groups: 80% in the control and 89% in the FS group (Fig. 1). During the first 10 weeks, SP diet reduced the tumor growth similar to the control, then stopped the tumor regression for 8 weeks, and caused tumor regrowth steadily to the end of the experiment. This resulted in the average tumor area in the SP group that was significantly larger than in the control group (p < 0.05) from treatment week 13 throughout the remaining treatment weeks (Fig. 1). In the SP group, 43% of the tumors regrew slowly until the end of the study (Table II, Fig. 1). A significant portion of the tumors (11%) in SP fed mice were bigger at the time it was killed than at their pretreatment size. Hence, significantly more growing and less completely regressed (p < 0.05) tumors were observed in the SP group when compared to other treatment groups (Table II). However, the combination of SP with FS attenuated the late stage tumor growth stimulating effect of SP to the same level as control (Fig. 1, Table II). Accordingly, at the end of the study, the average tumor weight and volume were 243 and 244% higher in the SP group than the control (p < 0.05), and combination of FS with SP reduced the tumors to the extent similar to the control (Fig. 2).

Figure 1. Effect of flaxseed (FS) and soy protein (SP) diet alone and in combination, and basal diet (BD) on palpable tumor growth of mammary MCF-7 breast cancers in ovariectomized athymic nude mice. Mice were implanted with an estradiol (E2) pellet (1.7 mg of E2, 60-day release) and injected with MCF-7 cells into mammary glands in 4 sites. When established tumors reached the average area of 35 mm², the E2 pellet was removed and the mice were divided into 4 treatment groups: control fed basal diet, flaxseed (FS) group fed 10% FS diet, soy protein (SP) group fed 20% SP diet, and FS + SP group fed diet containing both 10% FS and 20% SP. Different letters (a–b) indicate statistical difference at p < 0.05 using general linear model one-way ANOVA repeated measures followed by Fisher's least significant difference (LSD) test. Data are expressed as mean ± SEM.

Figure 2. Effect of flaxseed (FS) and soy protein (SP) diet alone and in combination, and basal diet (BD) on (a) mammary tumor weight and (b) volume of human MCF-7 breast cancers in ovariectomized athymic nude mice. Different letters (a–b) indicate significant difference at p < 0.05 among groups by one-way ANOVA followed by post hoc LSD test (tumor weight data) or by Kruskal–Wallis median test, followed by Mann–Whitney U-test (tumor volume data). Data are expressed as mean ± SEM.

Table II. Classification of the Established MCF-7 Tumor Growth in Ovariectomized Athymic Mice after Long-Term Dietary Treatments

Group	Tumors
	Growing(%)	Non-growing(%)	Completelyregressed (%)
* The numbers of tumors significantly different (p < 0.05) from control group by χ^2-test.
Control	7.1	42.9	50.0
SP	42.9*	40.0	17.1
FS	6.1	33.3	60.6
FS + SP	9.1	63.6*	27.3*

Tumor cell proliferation and apoptosis

The tumor cell proliferation index in the FS group was similar to the control, while that in SP group was significantly higher than control (p < 0.05; Fig. 3a). The combination of SP with FS reduced the proliferation index to similar level as with FS diet alone (Fig. 3a) and increased significantly the number of apoptotic cancer cells compared to FS diet alone or control (p < 0.05; Fig. 3b). Although SP also increased the tumor cell apoptosis compared to the control, this increase (37%) was less than the increase in proliferation index (151%), which is congruent with the measured enhanced tumor growth compared to control (Figs. 1 and 2).

Figure 3. Effect of flaxseed (FS) and soy protein (SP) diet alone and in combination, and basal diet (BD) on (a) Ki-67 labeling and (b) apoptotic indices of MCF-7 human breast cancer xenografts in ovariectomized mice. Different letters (a-c) indicate significant difference at p < 0.05 among groups by one-way ANOVA followed by post hoc tukey's test. Data are expressed as mean ± SEM.

Discussion

Our study demonstrates for the first time that FS attenuates the tumor growth stimulating effect of SP diet in ovariectomized athymic mice with low estrogen levels and increases the proportion of non-growing and completely regressed tumors. The mechanism of FS action was elucidated to occur via inhibition of SP stimulated tumor cell proliferation as well as increased apoptosis. We also show that unlike SP, a long-term (25 weeks) feeding of 10% FS diet does not stimulate the MCF-7 tumor growth, which is in agreement with our earlier work12 showing no growth stimulating effect of FS on MCF-7 tumors after 6–7 weeks of exposure to 10% FS diet.

The tumor growth inhibiting effect of FS has also been demonstrated in non-ovariectomized animals bearing hormone responsive mammary tumors. Dietary FS has been shown to decrease the growth of DMBA-induced13 and methylnitrosourea (MNU)–induced20 hormone responsive mammary cancers in rats. Furthermore, FS consumption (25 g/day) has been shown to decrease the proliferation and increase the apoptosis of newly diagnosed breast cancers in postmenopausal women.21 These studies suggest that dietary consumption of FS does not trigger the classic estrogen-like cancer growth stimulating effect of hormone responsive mammary cancers in animals or in humans, but may inhibit the tumor growth.

However, unlike FS diet, the long-term administration of SP diet stimulated the growth of orthotopic MCF-7 xenografts in ovariectomized mice, which concur with the previous studies of Allred and co-workers9, 22 although in this work the tumors grew slower. In our study, however, the energy density was higher than in previous studies9, 22 as high-fat diets (20% corn oil) were fed to mice, while the other studies9, 22 used low fat diet (7% corn oil). The use of high-fat led to lower consumption of SP diet and hence phytoestrogen in our study than in the other studies,9, 22 which may in part explain the observed slower rate of tumor growth. We observed approximately 2-fold increase in MCF-7 cancer cell proliferation index supporting the previous findings of increased tumor proliferation after a long-term use of SP.9 However, in our study, slightly but significantly increased number of apoptotic cancer cells was also observed with SP compared to control. In vitro, increased apoptosis of MCF-7 cells has also been demonstrated when these cells were treated with soy isoflavone aglycones genistein, daidzein and glycitein.23, 24 In our in vivo study, however, the increase in tumor cell apoptosis after SP exposure was smaller (37%) than the increase in proliferation index (151%), as compared to control. Thus the net effect of the combined cell proliferation and apoptosis results was increased proliferation which was reflected in the growth of the MCF-7 tumors in vivo with SP diet.

Several studies with MNU- and DMBA-induced hormone responsive mammary tumors in rats have demonstrated reduced tumor numbers and incidence when SP is administered before or during the tumor initiation phase.25, 26, 27, 28 However, no inhibition of the mammary tumor growth or other growth stimulation parameters by SP or its isoflavones has been documented when the administration is started after tumors were established.29, 30 Instead, dietary genistein at 25 ppm tended to increase the growth of DMBA and MNU induced tumors in ovariectomized rats,30 while a higher level of dietary genistein at 750 ppm significantly increased the weight of ER positive MNU-induced tumors in ovariectomized rats.31 Dietary genistein (1,000 ppm) has also been documented to increase the proportion of malignant mammary adenocarcinomas in mice.32 These findings indicate that isoflavones and isoflavone containing SP stimulate the mammary tumor growth in hormonal milieu low in estrogens. Furthermore, isoflavone containing diet may have a biphasic effect on hormone dependent mammary cancers: it may reduce the risk of breast cancer when consumed prior to tumor induction or cancer occurrence, but when the consumption is started when the hormone responsive tumors are pre-existing, its tumor growth-stimulating effects can not be excluded. As we have shown previously,11 mammalian lignans enterodiol and enterolactone showed no stimulation of MCF-7 tumor growth alone or in combination with genistein in ovariectomized mice. Thus, in this regard, the consumption of isoflavone-rich diet with lignan-rich diet may be more beneficial than consumption of isoflavone-rich diet alone.

In Asian diet, in addition to soy, fish containing high amounts of n-3 fatty acids is commonly consumed, which may have significance in protection against mammary cancer. In mammary cancer cell cultures, the combination of genistein with n-3 fatty acids better reduced the cell growth than either alone.17 Accordingly, the tumor growth inhibiting properties of FS in the present study may in part be due to its oil component. FS oil is a very rich source of α-linolenic acid (ALA, n-3 fatty acid), usually accounting over 50% of the total fatty acids.33 We have shown earlier that FS oil inhibits the growth of established mammary tumors in rats, which may be related to its high ALA content34 or the reduced dietary content of n-6 fatty acids. Furthermore, FS oil in combination with SDG better reduced the growth and metastasis of established ER negative human breast cancer tumors in athymic mice than SDG alone35 showing potent effect of the combination treatments on tumor development.

Understanding the cellular and molecular mechanisms of action of lignans as well as isoflavones, alone and in combination, are of critical importance. In many in vitro studies with MCF-7 cells, soy isoflavones have been shown to increase the proliferation of the cells in concentration dependent manner6, 36, 37 and to inhibit apoptosis,38 which agrees with the findings in MCF-7 xenograft model in vivo showing increased tumor growth via increased cell proliferation after consumption of isoflavone-rich diets.9, 11, 22 The mammalian lignan metabolites of plant lignans in FS, enterodiol and especially enterolactone, have also been shown to increase the MCF-7 cell proliferation.39, 40 However, these in vitro findings of mammalian lignans are not in agreement with the findings in vivo as no hormone responsive mammary tumor growth stimulating effect of FS12, 13, 20 or enterolactone11, 15 has been demonstrated. This suggests that in the case of FS lignans or its mammalian metabolites, the in vitro proliferation assays alone may poorly predict the lignan responses in vivo possibly due to distinct mechanisms dominating in regulation of proliferation of MCF-7 cells in vitro and in vivo.

Lignans have distinct effects in different estrogen responsive tissues of ovariectomized animals when administered alone or in combination with isoflavones. In ovariectomized athymic mice with established MCF-7 tumors, no stimulation of the tumor growth or increase in uterine weight was observed with a daily dose of 10 mg/kg mammalian lignans enterodiol or enterolactone.11 Accordingly, the combined lignans and isoflavone genistein did not stimulate the tumor growth induced by genistein alone.11 In uterus, however, the combination of the lignans with genistein in a total dose of 10 mg/kg increased the uterine weight similar to 10 mg/kg dose of genistein alone, suggesting either a growth stimulating effect of genistein at even low concentration or an additive stimulating effect of lignans and isoflavones on the uterus. No such interactive effects were measured in bone mineral density or biomechanical strength.41 These findings indicate that lignans may have pleiotropic effects in hormone responsive organs depending on the target tissue and hormonal status of the individual and act both as antiestrogen-like or estrogen-like compounds. They may also affect steroidogenesis by inhibiting aromatase,15, 42, 43 17β-hydroxysteroid dehydrogenase type 2,43 and 5α-reductase enzymes44 as has been demonstrated in in-vitro studies. Lignans may as well act via other non-hormone related mechanisms e.g. by inhibiting angiogenesis,45 tyrosine kinases via IGF-1 receptor and c-erbB2/HER2/neu mediated pathways,46 COX-2 expression47 or DNA topoisomerase II.48

Ovariectomized athymic mice with MCF-7 cells has limitations as an animal model to determine the effect of hormonally active compounds on postmenopausal breast cancer. While circulating estrogen levels are low in postmenopausal women, breast tumors still grow because of estrogen synthesis that takes place in the tumors and peripheral tissues. In ovariectomized mice, however, the synthesis of estrogen precursors (i.e. androgens) by adrenal cortex as well as aromatase activity of peripheral tissues are very low; therefore, tumors don't grow unless estrogen or compounds with estrogenic activity are provided. Intact MCF-7 cells have a low aromatase activity,43 but it is not sufficient to produce enough estrogens even with supplemented androgens to maintain the MCF-7 tumor growth in vivo.49 Xenografts of MCF-7 cells stably transfected with the human aromatase gene (MCF-7Ca) in ovariectomized, immune-suppressed mice with androgen supplement has been suggested to be a better model to simulate the postmenopausal breast cancer patient,49 in whom the major source of the hormone is nonovarian tissue. Nonetheless, ovariectomized athymic mice with MCF-7 tumors has been used as an animal model in the early studies of tamoxifen as a breast cancer drug50 with results that translated well to human situation.51 Also, this model has been useful in determining whether compounds or dietary components have estrogenic effects.52, 53

In conclusion, FS did not stimulate the growth of estrogen responsive human breast cancer MCF-7 xenografts in mice with low circulating estrogen, whereas a long-term administration of SP diet alone increased the tumor growth. Combination of FS with SP attenuated the tumor growth similar to control and FS alone by increasing apoptosis, but primarily via inhibition of tumor cell proliferation. The FS effects may be partly due to its plant lignans or their mammalian metabolites as well as the altered dietary fatty acids composition due to FS oil.

Acknowledgements

The authors thank Linda Lee, Jaskaren Mann and Agnes Forgacs for technical assistance. The study was supported by Natural Sciences and Engineering Research Council of Canada Grant A9995 (L.U. Thompson) and by postdoctoral fellowship from National Technology Agency of Finland, TEKES Grant 40285/02 (N.M. Saarinen).