The inhibitory effect of flaxseed on the growth and metastasisof estrogen receptor negative human breast cancer xenograftsis attributed to both its lignan and oil components
Article first published online: 22 APR 2005
Copyright © 2005 Wiley-Liss, Inc.
International Journal of Cancer
Volume 116, Issue 5, pages 793–798, 20 September 2005
How to Cite
Wang, L., Chen, J. and Thompson, L. U. (2005), The inhibitory effect of flaxseed on the growth and metastasisof estrogen receptor negative human breast cancer xenograftsis attributed to both its lignan and oil components. Int. J. Cancer, 116: 793–798. doi: 10.1002/ijc.21067
- Issue published online: 20 JUL 2005
- Article first published online: 22 APR 2005
- Manuscript Accepted: 4 FEB 2005
- Manuscript Received: 13 DEC 2004
- Natural Sciences and Engineering Research Council of Canada
- Flax Council, Saskatchewan Flax Development Commission
- Program in Food Safety, University of Toronto
- flaxseed oil;
- breast cancer;
Our previous studies have shown that dietary flaxseed (FS) can reduce the growth and metastasis of human estrogen receptor negative (ER-) breast cancer in nude mice. The aims of our study were to determine (i) whether the tumor inhibitory effect of FS was due to its oil (FO), lignan secoisolariciresinol diglycoside (SDG), or both components, and (ii) whether the effect on tumor growth was related to increased lipid peroxidation. Athymic nude mice were orthotopically injected with ER- breast cancer cells (MDA-MB-435) and 8 weeks later were fed either the basal diet (BD) or BD supplemented with 10% FS, SDG, FO, or combined SDG and FO (SDG + FO) for 6 weeks. The SDG and FO levels were equivalent to the amounts in the 10% FS. Compared to the BD group, the tumor growth rate was significantly lower (p < 0.05) in the FS, FO, and SDG + FO groups, in concordance with decreased cell proliferation and increased apoptosis; however, these did not significantly relate to the lipid peroxidation, indexed as malonaldehyde (MDA), in the primary tumors. Lung metastasis incidence was reduced (16–70%) by all treatments, significantly in the FS and SDG + FO groups. The distant lymph node metastasis was significantly decreased (52%) only in the FO group. Although the total metastasis incidence was lowered (42%) significantly only in the SDG + FO group, all treatment groups did not differ significantly. In conclusion, FS reduced the growth and metastasis of established ER- human breast cancer in part due to its lignan and FO components, and not to lipid peroxidation. © 2005 Wiley-Liss, Inc.
Previous studies have shown that the n-3 fatty acid, α-linolenic acid (ALA), as well as its metabolites, eicosapentaenoic (EPA) and docosahexaenoic acid (DHA), and secoisolariciresinol diglycoside (SDG), the precursor to the mammalian lignans, enterolactone (EL) and enterodiol (ED), can inhibit the various stages of mammary carcinogenesis.1, 2, 3, 4 Flaxseed (FS) contains the highest level of SDG, with values 75–800 times those in other plant foods,5 and is also the richest source of ALA,6 suggesting that it may be protective against mammary cancer.
Dietary supplementation of 5 or 10% FS indeed has been shown to inhibit the initiation and early and late promotion stages of mammary carcinogenesis in carcinogen-treated rats.7, 8, 9 FS at 10% level has also been shown to inhibit the growth of both human estrogen receptor positive (ER+) MCF-7 and ER negative (ER-) MDA-MB- 435 breast cancer xenografts in nude mice.10, 11 Furthermore, spontaneous metastasis has been reduced by feeding mice with 10% FS diet when ER tumors are already established.11, 12 In carcinogen-treated rats, reduction in established tumor growth has been attributed to both the SDG and oil components of FS.8, 9 However, it is not clear which component(s) of FS is responsible for the effect of FS in the growth and, in particular, the metastasis of human ER- breast cancer.
In vitro studies have shown that EL and EPA inhibit the proliferation and migration of endothelial cells, an important process in tumor growth and metastasis.13, 14 ED and EL, alone and in combination, have also been shown to reduce several steps involved in breast cancer cell metastasis, i.e., cell adhesion, migration, and invasion in vitro.15 Yan et al.16 have shown that dietary supplementation of FS inhibits the metastasis of melanoma cells to the lungs in a murine model, and demonstrated that it was in part due to the SDG component of FS.17 However, the effect of SDG and flaxseed oil (FO), alone or in combination, on the growth and metastasis of human ER breast cancer has not yet been investigated.
One mechanism whereby the n-3 fatty acids, such as ALA, DHA, and EPA, reduce tumor growth may be through increased levels of secondary products of lipid peroxidation, which are cytotoxic to the cell, since their combination with an antioxidant (vitamin E) promotes, while combination with prooxidants reduce tumor growth.18, 19 Because the lignans have been shown to have antioxidant effects,20, 21 it is not clear whether the lignans will antagonize the effect of ALA. Since lipid peroxidation in breast tumor upon animal feeding of FS or its components has not yet been studied, it is also of interest to test whether this is one possible mechanism by which FS may exert its effect on tumor growth and metastasis.
Therefore, the objectives of our study were to determine, using the athymic mice model, (i) the effect of 10% FS on mammary tumor growth and metastasis; (ii) whether the SDG, FO, or both components in FS are responsible for the effect; and (iii) whether increased lipid peroxidation is one of the mechanisms whereby FS inhibits the growth and metastasis of ER- human breast cancer.
Material and methods
Cell line and cell culture
The MDA-MB-435 ER- human breast cancer cell line was obtained from Dr. J. Price (M.D. Anderson Center, Houston, TX). This cell line was selected for study because of the cells' ability to metastasize to various organs.22, 23 The cells were maintained in Iscove's modified Dulbecco's medium supplemented with 5% fetal bovine serum, plus penicillin and streptomycin (Sigma, St. Louis, MO). Before injection into nude mice, they were grown in T-150 flasks to 70–90% confluence. The medium was changed 24 hr before cells were harvested by trypsinization. For injection, the cells were resuspended in the medium at a concentration of 2 × 107cells/ml and kept on ice.
Animals and diets
Female athymic nude mice (NCR-nu/nu), 3–4 weeks of age, 24 mice per group, were purchased from Simonsen Laboratories (Gilroy, CA) and maintained in microisolator within a pathogen-free isolation facility with a 12:12 hr light-dark cycle at 22–24°C and 50% humidity. They were fed with sterilized diet and autoclaved water ad libitum. All surgical procedures were conducted under aseptic conditions in a laminar flow hood. Animal care and use conformed with the Guide to the Care and Use of Experimental Animals (Canadian Council on Animal Care, 1984), and the experimental protocol was approved by the University of Toronto Animal Care Committee.
Five different diets were prepared: basal diet (BD) based on AIN93G diet24 modified to have 20% corn oil at the expense of corn starch, and BD supplemented with 10% freshly ground FS (Linott variety, Omega Products, Melfort, SK, Canada), FO (Omega Nutrition, Vancouver, BC, Canada), SDG, or SDG + FO. SDG was prepared from FS as previously described.25 The amounts of SDG (0.2 g/kg) and FO (36.53 g/kg) in the diets were equivalent to those found in the 10% FS diet. Each diet was formulated to be isocaloric and to have the same high-fat content (20% wt/wt) that was provided by corn oil. The FS- and FO-containing diets were corrected for the contribution of FS and FO to fat, carbohydrate, and protein components. All diets were prepared by Dyets, Inc. (Bethlehem, PA) and sterilized by gamma irradiation by Isomedix Corp. (Whitby, Ontario, Canada).
The mice were acclimated for 1 week, while being fed the BD. They were anesthetized with ketamine/xylazine solution, and MDA-MB-435 cells (1 × 106 in 50 μl of medium) were injected into the surgically exposed, right-sided thoracic mammary fat pad. The incision was closed with Vetbound (3M Animal Care Products, St. Paul, MN). The mice were weighed and the inoculation sites were palpated at weekly intervals. The maximum diameters and perpendicular diameter of the palpable mammary fat pad tumors were measured, and the surface areas were calculated using the equation (L/2 × W/2 × 3.14). At week 8, mice were randomized into 5 groups such that the mean body weight and tumor size in each group were similar. They were fed either the BD, FS, SDG, FO, or SDG + FO diets (n = 23/group, except FO group, with n = 22), and their body weights, food intakes, and palpable tumors were monitored weekly for 6 weeks.
At 14 weeks after the tumor cell injection, the mice were sacrificed by CO2 asphyxiation. The body and primary mammary tumor weights and volumes were recorded. Primary tumor volume was calculated as length/2 × width/2 × thickness/2 × 3.14. A portion of the primary tumors was fixed in 10% buffered formalin, and the rest was stored at −70°C. The incidence and number of metastatic tumors in lymph nodes (only those in the distant lymph nodes, not those fused with the primary tumor) and other organs, as well as any gross pathologic changes in the organs, were observed. The body weight and major organ weight, e.g., liver, lungs, kidney, uterus, and ovaries, were also recorded. After fixation in 10% buffered formalin, the lungs were observed for nodule number of metastatic tumors under a stereomicroscope.
Cell proliferation assay
The 5 μ sections of formalin-fixed paraffin-embedded primary tumor tissue were deparaffinized and rehydrated. They were then stained for cell proliferation as previously described.10, 11 Briefly, endogenous peroxidase was blocked with aqueous 3% H2O2. The antigen was retrieved by heating sections in 0.01 M citrate buffer at pH 6.0 in a microwave oven. The rabbit anti-human Ki-67 antibody at 5 μg/ml (Santa Cruz Biotechnology, Santa Cruz, CA) was diluted in the Diluent Buffer (Dako, Missisauga, ON, Canada) that blocks nonspecific antigens. The sections were incubated at 4°C overnight. The biotinylated swine anti-rabbit IgG (Dako) was loaded and incubated for 30 min at room temperature. Streptavidin-horseradish peroxidase and AEC substrate chromogen (Dako) were used to demonstrate the antigen followed by a brief counterstaining with hematoxylin. All slides were read blindly under a light microscope at 400× magnification. Over 1,000 cells from at least 5 different fields were counted. Ki-67 labeling index (LI) was calculated as the percentage of positive cells over total cells counted.
Cell apoptosis assay
To demonstrate DNA fragmentation, the in situ terminal deoxynucleolidyl transferase-mediated nick end labeling (TUNEL) assay was utilized with the ApopTag Detection Kit (Intergen, Purchase, NY) according to the manufacturer's protocol.10 Briefly, tumor sections were deparaffinized, rehydrated, and pretreated with proteinase K (20 μg/ml) for 15 min. They were incubated in a reaction mixture containing terminal transferase and digoxigenin dUTP at 37° C for 1 hr. The sections were then washed, followed by incubation with antidigoxigenin antibody coupled to horse-radish peroxidase for 30 min at room temperature. After rinsing with PBS, diaminobenzidine was added for 6 min. The slides were counterstained with methyl green and mounted with Permount. The number of breast carcinoma cells exhibiting positive nuclear immunoreactivity was determined and expressed as a percentage of the total breast carcinoma cells examined under a microscope as described above. The percentage of positively stained cells was expressed as apoptotic index (AI).
Malonaldehyde (MDA) was extracted from the primary tumors using a modification of the methods previously described.26, 27 Briefly, 5 ml of 10% trichloroacetic acid and 250 μl of butylhydroxytoluol were added to 0.5 g of tumor tissue, and then homogenized with a tissue grinder (Fisher Scientific, Ottawa, Canada). The homogenate was heated for 30 min at 90°C, cooled immediately under tap water, and then centrifuged at 671g for 10 min. Then 250 μl of the supernatant was added to 0.25 ml of saturated aqueous thiobarbituric acid solution and heated at 95°C for another 30 min. After cooling, 1 ml of n-butanol was added to the sample, vortexed, and centrifuged again at 1,048g for 10 min. For analysis by high-performance liquid chromatography (HPLC), 100 μl of the extract was diluted with 100 μl of methanol and 800 μl of the HPLC mobile phase. Injection volume was 50 μL. The HPLC (Shimadzu, Mandel Scientific, Guelph, ON, Canada) was fitted with a reverse-phase column (Waters μ-Bondapak C18, 3.9 × 150 mm, particle size = 10 μm), a photodiode array detector (Shimadzu, Mandel Scientific, Guelph, ON, Canada) monitoring at 532 nm, and a mobile phase of 16% acetonitrile, 6% tetrahydrofuran in a 5 mM phosphate buffer at pH 7.0 at a flow rate of 0.6 ml/min. The retention time for MDA was 2.36 min.
Values are presented as means ± SEM. Palpable tumor area over treatment time, i.e., growth rate, was compared among groups by the general linear model-repeated measure ANOVA. Metastasis incidence was analyzed by chi-square (χ2) test. All other comparisons were done using 1-way ANOVA followed by Tukey test, or using Kruskal-Wallis 1-way ANOVA on ranks if parameters did not meet normality and equal variance tests. Linear regression analysis was used to determine association between MDA level and tumor size (final tumor weight and volume), growth rate, Ki-67 LI, or AI. The level of significance was set at p < 0.05. SigmaStat (Jandel Corporation, San Rafael, CA) was used for statistical analyses.
Food intake, body weight, and major organ weights
The initial body weight, weight gain, and mean food intake over the 14 weeks did not differ significantly among diet groups (data not shown). There were no significant differences in the weights and gross pathologic appearance of major organs, including liver, kidney, ovaries, and uterus (data not shown), suggesting no potential adverse effect of treatment diets on the major organs, nor estrogenic effect on the hormone-sensitive organs.
Palpable tumor growth, and final tumor volume and weight
The palpable tumor growth rate (mm2/week) was significantly lower in the FS, FO, and SDG + FO groups (p < 0.05) over the treatment period, compared to the BD or SDG groups (Fig. 1). The SDG group did not differ significantly from the BD group. The final tumor volume and weight were not significantly different among groups (Table I), although their patterns were similar to that of the palpable tumor growth (Fig. 1).
|n||Tumorvolume (mm3)||Tumorweight (g)||MDA(μmol/g)|
|BD||23||1.59 ± 0.23||2.08 ± 0.28||68.4 ± 2.2a|
|FS||23||1.25 ± 0.21||1.75 ± 0.29||70.9 ± 2.8a|
|SDG||23||1.69 ± 0.25||2.58 ± 0.34||72.1 ± 2.0a|
|FO||22||1.21 ± 0.19||1.76 ± 0.29||87.0 ± 1.3b|
|SDG+FO||23||1.15 ± 0.14||1.72 ± 0.22||81.2 ± 1.5b|
Compared to the BD group, the incidence of lung metastasis was reduced by 50.1, 30.1, 16.3, and 70.1% in the FS, SDG, FO, and SDG + FO groups, respectively, with the reduction reaching significance in the FS and SDG + FO groups (Fig. 2a). Of the treatment groups, the SDG + FO group had the lowest lung metastasis incidence (Fig. 2a). Regarding distant lymph node metastasis incidence, although all the treatment groups had a lower number than the BD group, only the 52.5% reduction by FO reached significance (Fig. 2b). The incidence of metastasis in other organs, such as liver, bones, kidneys, and abdominal cavity, was reduced in all treatment groups, with significantly none occurring in the FO and SDG + FO groups (Fig. 2c). The total metastasis incidence, which took into account the incidence of metastasis in the lungs, lymph node, and other organs, was lowered significantly in the SDG + FO group by 42.8% (p < 0.05) when compared to the control (Fig. 2d). However, all treatment groups did not differ significantly from each other.
There was no significant difference among groups in the number of metastatic tumors per mouse in the lung (1.48 ± 0.56, 1.26 ± 0.55, 1.0 ± 0.48, 1.24 ± 0.47, and 0.62 ± 0.35 for BD, FS, SDG, FO, and SDG + FO, respectively) and lymph node (0.78 ± 0.21, 0.61 ± 0.25, 0.46 ± 0.15, 0.32 ± 0.15, and 0.39 ± 0.15 for BD, FS, SDG, FO, and SDG + FO, respectively).
Primary tumor cell proliferation and apoptosis
The cell proliferation (Ki-67 LI; Fig. 3a) in the primary tumors was significantly lower (p < 0.05) in the FO (26.5%) and SDG + FO groups (30.7%) than in the BD group. Tumor cell proliferation in the FS group did not differ significantly from all the other groups. The SDG and BD groups did not differ significantly.
The AI was significantly increased in the FS, FO, and SDG + FO groups by 65.1, 60.0, and 49.2%, respectively, compared to the BD group (p < 0.05) (Fig. 3b). These groups also had a higher AI than the SDG group alone, which in turn had an AI similar to that of the BD group.
Primary tumor MDA
Compared to the BD group, the primary tumor MDA levels were significantly higher, 27.2 and 18.7%, respectively, in the FO and SDG + FO groups (p < 0.05) (Table I). The MDA levels in these 2 groups were also significantly higher (p < 0.05) than those detected in the FS and SDG groups, both of which were similar to those of the BD group. There was no significant relationship between tumor MDA levels and the tumor size, cell proliferation or apoptosis at the time of sacrifice, or the tumor growth rates among groups.
Our study has shown that 10% FS can significantly reduce the tumor growth rate and metastasis incidence of established human ER- breast cancer xenografts in nude mice, in agreement with our earlier observations.11, 12 The tumor growth rate was similar in the FS, FO, and SDG + FO groups, suggesting that FO, alone and in combination with SDG, contributed to the inhibitory effect induced by the FS diet. The total metastasis incidence did not differ significantly among the diet treatment groups, again indicating that the effect of FS was in part due to both its FO and SDG components. However, the fact that the lung and total metastasis incidences were significantly the lowest in the SDG + FO group, when compared to the control, suggests that the FO and SDG complement, rather than antagonize, each other's antimetastatic activity.
SDG, the plant lignan present in very high concentration in FS, is not absorbed, but is metabolized to the mammalian lignans, ED and EL, by the bacterial flora in the colon.3, 4, 5 Therefore, the anticancer effect induced by SDG is primarily through its metabolites ED and EL, which are absorbed and partly interact with n-3 fatty acids in the body.
Our previous study has shown that 10% FS downregulates the insulin growth factor-1 (IGF-I) and epidermal growth factor receptor (EGFR) expressions in MDA-MB-435 tumors.11 It also decreases the tumor expression of vascular endothelial growth factor (VEGF),12 suggesting reduction in angiogenesis. These in turn related to a lower tumor cell proliferation and enhanced apoptosis, which led to slow tumor growth and inhibition of tumor metastasis. The decreased cell proliferation and increased apoptosis induced by FS were also observed in the ER + breast cancer xenografts in nude mice10 and breast carcinomas of postmenopausal patients.28 These effects of FS, i.e., lower tumor growth, decreased cell proliferation, and increased apoptosis, were again observed in our study. Furthermore, a greater reduction of Ki-67 LI and induction of apoptosis were observed in the FO and SDG + FO groups, suggesting that both components contributed to the effects of FS, in agreement with the lower tumor growth rate.
Because the mice fed the diets that contain FO, i.e., the FS, FO, and SDG + FO groups, had significantly lower tumor growth rate than those fed without it, i.e., the BD and SDG groups, it was hypothesized that the higher amount of ALA in these diets may have increased lipid peroxidation in the tumors, which then led to their reduced growth. Other studies have shown that enhancement of lipid peroxidation by n-3 fatty acids can reduce tumor growth.18, 19, 29 Hence, tumor MDA, a product of lipid peroxidation, was measured in our study. The tumor MDA levels were indeed found to be higher in the FO and SDG + FO groups. However, the MDA level in the FS group was lower than that in the FO and SDG + FO groups, although the FS diet contained the same amount of oil as the FO diet. Perhaps other compounds with antioxidant properties in FS, including the lignans, reduced the oxidation of FO. However, SDG alone did not decrease the tumor MDA level and tumor growth, perhaps because the dose was low. ED and EL have been shown to have antioxidant effects in vitro at 10–100 μM,21 and reduce the oxidized DNA bases in colon cells at 100 μM,30 while ingestion of SDG equivalent to the levels in the 5% FS diet resulted in only about a 1 μM concentration of plasma mammalian lignans.31 On the other hand, the lack of effect of SDG in reducing tumor growth may also be related to its antioxidant activity. Cognault et al.19 have shown that the rat mammary tumor growth was inhibited by 15% FO, but was increased when an antioxidant (vitamin E) was combined with 15% FO and decreased when proxidants were added to the diet. A recent study also showed that the MMTVPyVT transgenic mice have higher tumor growth rate, increased incidence of lung metastasis, and decreased apoptosis of mammary tumor when fed a diet supplemented with vitamins E and A, compared to the diet with antioxidant depletion.32 Nonetheless, in our study, no significant relationship was observed between the tumor MDA levels and the tumor growth rate or final tumor size, indicating that the difference in lipid oxidation may not be the main mechanism involved in the reduction of tumor growth.
The antimetastatic effect induced by FS may be attributed to both the lignan and oil components. Although the SDG group did not significantly differ from the control in metastasis incidence, the SDG + FO group showed a lower lung, other, and total incidence of metastasis, suggesting that SDG played a role in the inhibition of metastasis as well. EL has been shown in vitro to inhibit endothelial cell proliferation to reduce angiogenesis.13 SDG has also been shown to reduce the IGF-I expression.33 Expression of IGF-I and EGFR, both of which are mitogens involved in the development of breast cancer and play crucial roles in the tumor growth and metastasis, can also be downregulated by the 10% FS diet.11 All of these implicate the role of lignans in the signal transduction pathway of breast cancer progression involved in the process of metastasis. Moreover, we have shown in vitro that the mammalian metabolites of SDG (EL and ED) at the 1–5 μM level can significantly reduce the various steps of metastasis, such as tumor adhesion, invasion, and migration of the MDA-MB-435 human breast cancer cells, and their combination induced an even greater effect.15 This provides a potential cellular mechanism by which FS exerts its antimetastatic effect. The weaker effect of SDG when given alone than when given in combination with FO or in FS in our study may also be due to other components in FS that may have complemented the effect of FO.
ALA, the major fatty acid in FS, may have played an important role in the inhibition of tumor growth and metastasis. Besides the lipid peroxidation mentioned above, ALA can be metabolized to EPA and DHA, which reduce the production of prostaglandin E2 and 12-hydroxyeicosatetraenoic acid, both of which are known to stimulate breast cancer growth and metastasis.1, 34, 35 The diets supplemented with fish oil rich in EPA and DHA significantly reduced the tumor growth and metastasis in studies with MDA-MB-435 cancer,23, 34 similar to our study. Both EPA and DHA exert their inhibitory effect on tumor growth and metastasis through reduction of cell proliferation,35 induction of apoptosis,36 and inhibition of angiogenesis37 and signal transduction pathway.38 Some of these actions are similar to those elicited by SDG discussed above, suggesting that the actions of FO and SDG are agonistic. Moreover, a diet containing 4% EPA suppressed the development of lung metastases in nude mice from MDA-MB-435 xenografts, which is associated with reduced levels of 92 kDa type IV collagenase gelatinolytic activity in the primary tumor,39 while dietary linoleic acid enhances the enzyme activity and tumor cell invasiveness.40 Thus, in combination, SDG and FO complemented each other and triggered a greater inhibitory effect, rather than antagonistic action, on the tumor growth and metastasis. At 10% level, dietary FS, as well as its equivalent amount of SDG and FO, caused no weight or gross pathologic changes in major organs of the mice, suggesting that it may be safe for the treatment of estrogen-independent breast cancer.
In conclusion, feeding 10% dietary FS to athymic mice with established ER- human breast cancer xenografts reduced the tumor growth and metastasis. The effect of FS can be partly attributed to both its oil and lignan components. The reduced primary tumor growth was due, in part, to the reduced tumor cell proliferation and increased apoptosis, but cannot be related to an increase in tumor lipid peroxidation. The inhibitory effect on the tumor growth and metastasis induced by FS and its components seen in our study is promising, and confirmation of these results in clinical trials would be of interest in the future.
- 6Dietary sources and metabolism of α-linolenic acid. In: ThompsonLU, CunnaneSC, eds. Flaxseed in human nutrition, 2nd ed. Champaign, IL: AOCS Press, 2003. 63–91..