Dietary lariciresinol attenuates mammary tumor growth and reduces blood vessel density in human MCF-7 breast cancer xenografts and carcinogen-induced mammary tumors in rats

Abstract

Lariciresinol is a dietary lignan that accounts for a significant portion of the total phytoestrogen intake from Western foods. Recent epidemiological studies suggest that high dietary intake of lignans and lariciresinol is associated with reduced breast cancer risk. However, no causal relationship between lariciresinol intake and breast cancer development has been established. In this study, we investigated for the first time the effects and possible mechanisms of action of lariciresinol on hormone responsive mammary cancer in vivo in dimethylbenz[a]anthracene induced mammary cancer in rats, and in human MCF-7 breast cancer xenografts in athymic mice. For tumor bearing rats, lariciresinol (3 or 15 mg/kg of body weight) or vehicle was administered p.o. daily for 9 weeks. For E2-maintained ovariectomized athymic mice bearing orthotopic MCF-7 tumors, control diet (AIN-93G) or lariciresinol containing diet (AIN-93G supplemented with 20 or 100 mg of lariciresinol/kg of diet) was administered for 5 weeks. In both models, lariciresinol administration inhibited the tumor growth and tumor angiogenesis. In MCF-7 cells, enterolactone significantly inhibited the E2-stimulated VEGF secretion. Moreover, in MCF-7 xenografts, lariciresinol administration enhanced tumor cell apoptosis and increased estrogen receptor beta expression. Lariciresinol and its further metabolites secoisolariciresinol, enterodiol and enterolactone were found in serum of both rats and athymic mice confirming a similar lignan metabolism pattern as in humans. These findings indicate conceivable importance of dietary lignan lariciresinol in inhibition of breast cancer development. © 2008 Wiley-Liss, Inc.

Breast cancer is the most prevalent cancer among women in Western countries. Diet and life-style related factors such as overweight and weight gain have been consistently associated with increased breast cancer risk.1 Weight control is linked with consumption of a low-fat fiber-rich diet, e.g., a diet rich in whole grains, vegetables, fruits and berries, with probable importance in breast cancer risk reduction. These fiber-rich foods are among the most important dietary sources of naturally occurring phenolic compounds.2 Some of these compounds are classified as phytoestrogens because of their estrogen-like potency in different in vitro and in vivo assays.3 Thus, phytoestrogens have been postulated to influence the etiology of estrogen-related diseases, such as hormone responsive breast cancer.

In non-soy containing foods, dietary lignans are the main dietary phytoestrogens.4–8 Many of the dietary plant lignans can be transformed to mammalian metabolites, so called enterolignans, by intestinal microbiota.9, 10 In several recent studies, high dietary intake of plant lignans, as well as high serum and urine enterolignan concentrations, have been associated with reduced breast cancer risk11–14 suggesting a biological significance of lignans. However, conflicting epidemiological findings on the association of the enterolignan levels on breast cancer risk exist.15, 16

The latest studies have revealed that lariciresinol is one of the most abundant plant lignans in foods consumed in Europe and North America5, 6 and that a significant portion of the total lignan intake is due to this compound.17 Lariciresinol is bioavailable as it is found in human serum18 and urine.19, 20 Furthermore, it was identified as the main serum lignan in some women consuming their habitual (Western) diet in Finland.18 When ingested, lariciresinol can be metabolized first to secoisolariciresinol, and then further to enterolignans, e.g., enterodiol and enterolactone10, 21, 22 (Fig. 1). Interestingly, all these 3 compounds have been shown to attenuate breast cancer growth in different estrogen-responsive experimental cancer models in vivo.23–27

Figure 1. Metabolism of lariciresinol to other lignans in vivo.

A recent epidemiological study revealed that high lariciresinol intake was inversely associated with breast cancer risk28 suggesting a possibility for biological activity of the compound in women. However, it is not yet known, whether there is a cause-effect relationship between high dietary lariciresinol intake and reduced breast cancer risk or development. According to our knowledge, no experimental studies have been conducted on the possible effects of lariciresinol on mammary cancer or any other cancer. In this study, we investigated for the first time the effects and possible mechanisms of action of lariciresinol on mammary cancer using 2 different hormone responsive breast cancer models: carcinogen induced mammary cancer in rats and MCF-7 human breast cancer xenografts in athymic mice.

DMBA, dimethylbenz[a]anthracene; E2, estradiol; ERα, estrogen receptor alpha; ERβ, estrogen receptor beta; HPLC-MS/MS, high-performance liquid chromatography-tandem mass spectrometry; IGF-1, insulin-like growth factor 1; MVD, microvessel density; PBS, phosphate buffered saline; PR, progesterone receptor; TUNEL, terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling.

Material and methods

Lignans, animals and housing conditions

Lariciresinol was isolated from knots of Pinus cembra as previously described.22 The purity of the product was 97% as determined by gas chromatography. Secoisolariciresinol (purity >95%) was isolated from knots of Araucaria angustifolia.29 Enterolactone (purity >98%) was obtained from VTT, Technical Research Center of Finland (Espoo, Finland) and enterodiol (purity >95%) from Sigma Chemicals (St. Louis, MO).

Animal care and studies were performed according to the European convention for the protection of Vertebrate Animals used for Experimental and other Scientific Purposes, EU directives and The Law on Statute on Animal Experiments in Finland. All experimental protocols were approved by the Committee for Care and use of Animals in Experiments at the University of Turku and State of Provincial Office of Western Finland. Rats (2–3/cage) were maintained in standard cages and athymic mice (2–5/cage) in micro-isolator cages in isolators. Animals were housed with 12-hour light/dark cycle at 22–24°C and 50% humidity, and they were administered with the diets and water ad libitum.

Effects of lariciresinol on mammary cancers

Carcinogen induced mammary cancer in rats

Fifty-day-old female Sprague-Dawley rats obtained from Harlan Netherlands were induced p.o. with 12.0 mg of 7,12-dimethylbenz[a]anthracene (DMBA) (Sigma Chemical Co., St. Louis, MO) in 1 ml of sesame oil. Starting 7 weeks after induction, rats were fed with lignan-free semipurified C1000 diet (Altromin, Lage, Germany), and at 10 weeks from induction they were divided into control (n = 16), lariciresinol 3 mg/kg body weight (b.w.) (n = 16), and lariciresinol 15 mg/kg b.w. (n = 15) treatment groups, with similar tumor numbers and volumes. The rats were administered p.o. daily with vehicle (10% ethanol in aqueous solution of 50% polyethylene glycol MW 3400) or with lariciresinol dissolved in vehicle for 9 consecutive weeks. The rats were weighed, the food consumption was measured, and the tumors were palpated once a week. The tumor areas were calculated using the formula (length/2 × width/2) × π. The cumulative tumor area per number of animals in each group, and tumor multiplicity (the number of tumors per rat in the treatment group) were followed weekly during the treatment period. In the control and lariciresinol 3 mg/kg b.w. groups, 2 animals had to be sacrificed because of a bleeding tumor before the end of the experiment. The tumor numbers and volumes of these rats were calculated unchanged (i.e., similar as at the time of sacrifice) during the remaining weeks of the experiment. At the end of the experiment (9 weeks after induction), body, liver and uterus weights were recorded. Tumors were weighed, fixed in neutral buffered formalin and embedded in paraffin. The tumors present at the time of sacrifice were confirmed as adenocarcinomas by histopathological evaluation. Blood samples were collected for serum at 2 and 5 hr after the last dose of lariciresinol from a similar separate study with identical lariciresinol treatments. The samples were stored at −20°C until analyzed using HPLC-MS/MS with a method previously described.18

MCF-7 cell culture

MCF-7 human breast cancer cells (The American Type Culture Collection, Manassas, VA), were maintained in Dulbecco's minimum essential medium/F12 (Gibco BRL, Paisley, Scotland) containing 2 times charcoal-stripped heat inactivated FCS (Gibco BRL) and 1% antibiotic–antimycotic solution for 1–2 days prior to the start or the E2 and lignan treatments. Cells were treated with or without 1 nM E2, or with E2 together with 10 μM lariciresinol, secoisolariciresinol, enterodiol or enterolactone. The cells were given fresh treatment mediums every 2–3 days and grown to 80% confluence. For preparation of cell blocks the cultures were trypsinized, centrifuged and resuspended into a mixture of phenol red free Dulbecco's minimum essential medium/F12 and 1% agarose in phosphate-buffered saline (PBS) at 50°C. The solid cell blocks were fixed in formalin and embedded in paraffin.

For vascular endothelial growth factors (VEGF) secretion studies, the previously described method was used.27, 30 Shortly, cells were seeded into 6-well plates (1 million/well) in phenol red free DMEM/F12 medium containing charcoal-stripped heat inactivated FCS. On the next day, phenol red-free DMEM/F12 medium containing 10 g/ml transferrin (Sigma), 1 μg/ml insulin (Sigma), 0.2 mg/ml bovine serum albumin (Sigma), and 10 nM E2 alone or in combination with 1 μM lariciresinol, secoisolariciresinol, enterodiol or enterolactone, or vehicle was added. Fresh mediums were changed daily and after 7 consecutive days the medium was collected for VEGF analyses, and cells measured for protein with the BioRad DC protein Assay (R&D Systems, Minneapolis, MN). The secreted VEGF was quantified from the medium with an immunoassay kit for human VEGF (QuantiGlo, R&D Systems).

Orthotopic MCF-7 xenografts in athymic mice

For MCF-7 xenograft studies, ovariectomized athymic female mice (Balb/c nu/nu, 4–5 weeks old) were obtained from Charles River (Lyon, France). MCF-7 human breast cancer cells were maintained in Dulbecco's minimum essential medium/F12 supplemented with 10% heat-inactivated fetal bovine serum and 1% antibiotic-antimycotic solution. The cells were grown to 70–90% confluence and given fresh medium every 2–3 days, and one day before cell harvest. For cell injections, the cells were trypsinized, resuspended in serum-free medium containing Matrigel (1:1 vol) and kept on ice. After 1-week acclimatization with AIN-93G basal diet (corn oil replaced for soy oil), mice were injected with 0.5 million MCF-7 cells in 50 μl into 4 sites of mammary fat pads and implanted s.c. with an E2 pellet (1.7 mg, 60-day release; Innovative Research of America, Sarasota, FL) as previously described.26, 31 The tumors were palpated weekly and the surface area was calculated using the formula (length/2 × width/2) × π. When the average tumor area reached ∼20 mm², the E2 pellet was changed to a fresh one, and mice were divided into three dietary treatment groups: (i) E2 group (n = 6) fed with basal diet, (ii) 20 mg/kg lariciresinol in basal diet group (n = 8) and (iii) 100 mg/kg lariciresinol in basal diet group (n = 6). Administration of diet containing lariciresinol 20 mg/kg and 100 mg/kg resulted in 3.3–4.2 mg/kg and 17.8–22.9 mg/kg b.w. exposures, respectively. The E2-dependent growth of the established tumors was verified in mice with E2 pellet removed (n = 5, w/o E2 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 mice were sacrificed after 5 weeks of treatment, serum was collected for lignan analysis as described earlier, and body, uterus, liver and tumor weights were recorded.

Immunohistochemical determination of apoptosis, proliferation and steroid hormone receptor expression indices

The expression indices were determined by counting the positively stained nuclei out of 1,000. The expression indices were determined from cultured MCF-7 cells and orthotopic MCF-7 xenografts.

Expression of ERα, ERβ, and PR in the MCF-7 cell cultures and tumors

Eighty percent confluent MCF-7 cells were trypsinized and centrifuged at 130 g for 5 min. The cell pellet was suspended into 10% formalin, incubated at room temperature for 20 min, centrifuged at 200g, resuspended into phenol red-free Dulbecco's minimum essential medium/F12 (Gibco), and gently mixed with 1% agarose in PBS (50°C). The solid cell blocks were fixed in formalin over night and embedded in paraffin.

Five micrometer sections of the MCF-7 tumor samples or paraffin-embedded cultured cells were used for determining the expression indices of ERα, ERβ, and PR. Monoclonal anti-human ERα primary antibody (mouse anti-human estrogen receptor, clone 1D5, 1:200 dilution; DAKO A/S, Glostrup, Denmark), monoclonal mouse anti-human ERβ clone PPG5/10 (1:200 dilution, Serotec Scandinavia, Hamar, Norway), and rabbit anti-human PR-antibody code A0098 (1:400 dilution rabbit anti-human, DAKO A/S) were used for ERα, ERβ and PR, respectively. Deparaffinized, rehydrated sections were rinsed with water and treated with 10 mM sodium citrate buffer (pH 6.0) in a microwave oven for 15 min for antigen retrieval. Slides were allowed to cool down and rinsed with PBS. Nonspecific endogenous peroxidase activity was blocked with 1% hydrogen peroxide in PBS for 20 min at room temperature. The slides were incubated with primary antibodies diluted into PBS containing 3% bovine serum albumin and 0.05% Tween overnight at +4°C. On the next day, the slides were rinsed with PBS and incubated with the secondary antibody (anti-mouse HRP conjugated Dako Envision®+ System for ERα and ERβ and anti-rabbit HRP conjugated Dako Envision®+ System for PR) for 30 min at room temperature, washed with PBS, treated with 3-3′diaminobenzidine, rinsed in distilled water and counterstained with hematoxylin.

Apoptosis and proliferation of MCF-7 tumors

The TUNEL (terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling) method was used to determine apoptotic cells (i.e., cells with fragmented DNA) from non-necrotic tumors. The analysis was performed using ApopTag Peroxidase in situ Apoptosis Detection Kit (Chemicon International, Hampshire, UK) according to manufacturer's instructions for deparaffinized, rehydrated 5 μm sections pretreated as described above for antigen retrieval. The Ki-67 labeling index was determined as a marker of proliferation. The pretreated 5 μm sections were incubated with monoclonal mouse anti-human Ki-67 antibody clone MIB-1 (1:600 dilution, DakoCytomation, Corpinteria, CA) and incubated with the secondary anti-mouse HRP Conjugated Dako Envision®+ System as described earlier.

Analysis of vessel density in DMBA-induced mammary tumors and in MCF-7 xenografts

DMBA-induced mammary gland tumors in rats have different spontaneous growth patterns, growing and nongrowing (stabilized and regressing). The tumors were allocated into these categories as previously described.25 DMBA-induced rat mammary carcinomas have also distinct histologies; poorly differentiated, well differentiated, atrophic and secretory type,32 majority of the tumors being well differentiated. For analysis of tumor vessels numbers, tumors that were established during the treatment period were selected. The analyzed tumors were further selected to have similar growth pattern, comparable sizes and well differentiated histological type. The weights of the analyzed tumors were 1.09 ± 0.52 g in control group and 1.28 ± 0.80 g and 1.27 ± 0.88 g in lariciresinol 3 mg/kg and 100 mg/kg groups, respectively. The tumors (7 in control group, 6 in lariciresinol 3 mg/kg b.w. group, and 5 in lariciresinol 15 mg/kg b.w. group) were stained with an automated Lab Vision Autostainer (Thermo Fisher Scientific, Fremont, CA) from 5 μm rehydrated tumor sections treated with in T10E1 buffer (pH 9.0) in a microwave over at 1,000 W for 3.5 min followed by 15 min at 180 W for antigen retrieval, and incubated with 3% bovine serum albumin. Sections were incubated with polyclonal rabbit primary antibody to von Willebrand Factor VIII (code ab6994, Abcam, Cambridge, UK) for 30 min and endogenous peroxidase blocked with 3% H₂O₂. HRP Conjugated Dako Envision®+ System was used as secondary antitibody followed by diaminobenzidine (Dakocytomation) for visualization. The vessels were counted with Olympus BX51 microsope at 200× magnification from 10 to 15 randomly selected fields of each tumor. The counted tumor areas were similar between the treatment groups, 13.5 ± 2.8 mm² in control group, 12.0 ± 2.7 in lariciresinol 3 mg/kg groups and 13.3 ± 2.3 in lariciresinol 15 mg/kg group.

Preparations and procedures for the formalin fixed paraffin embedded MCF-7 tumor sections were performed as described earlier.33 Briefly, similar size tumors (average 150 mm³) were cut in 5 μm sections. After rehydration and antigen retrieval the slides were stained for microvessel density (MVD; CD34, PharMingen) in a 1:50 dilution, developed with 3-3′diaminobenzidine and lightly counterstained with hematoxylin.33 Images of the sections were acquired on Olympus BX-60 microscope at 200× magnification and digitally analyzed and differentially quantified by morphometric analysis.34, 35

Statistical analyses

The analyses were performed by using Statistica software for Windows (Stat Soft, Tulsa, OK). Normally distributed data, determined by using Shapiro-Wilk's normality test, were analyzed with one-way ANOVA followed by LSD post hoc test. Non-normally distributed data were analyzed with Kruskal-Wallis median test, followed by Mann-Whitney U-test. The level of significance was set at p ≤ 0.05 for all analyses. The data are expressed average ± SEM for normally distributed data and as median ± SEM for non-normally distributed data.

Results

Effects of orally administered lariciresinol on growth of DMBA-induced mammary gland carcinoma in rats

Nine-week administration of lariciresinol to rats starting 10 weeks after tumor induction by DMBA reduced the growth of mammary gland tumors measured as a cumulative tumor area (Fig. 2), but did not affect their b.w., uterus or liver weights (Table I). The growth inhibitory effect of laciriciresinol was more pronounced in tumors that developed during the treatment period (Fig. 2a) than those established already before the start of the treatments (Fig. 2b). However, no reduction in tumor multiplicity (2.56, 3.06 and 2.40), incidence (75%, 100% and 87%) or proportion of the nongrowing tumors (46.9%, 41.4% and 50%) was observed in control, lariciresinol 3 mg/kg b.w. and lariciresinol 15 mg/kg b.w. treatment groups, respectively.

Figure 2. Cumulative area of DMBA-induced mammary gland tumors established during lariciresinol administration (a) and tumors already established before the start of the lariciresinol administration (b) in female rats. The Sprague Dawley rats were administered p.o. with a single 12 mg dose of 7,12-dimethylbenzanthracene (DMBA). Eleven weeks after induction, the rats were allocated into 3 groups with similar tumor burden and administered daily for 9 consecutive weeks with vehicle (control, n = 16) or with lariciresinol 3 mg/kg b.w. (n = 16) or 15 mg/kg b.w. (n = 15).

Table I. Uterus and Mammary Tumor Weights of Lariciresinol-Treated Rats and Athymic Mice at the End of the Experiment

	b.w. (g)	Uterus (mg/100 g b.w.)	Tumors (mg)
Sprague Dawley rats with DMBA-induced mammary gland tumors were administered daily p.o. with 3 or 15 mg/kg body weight (b.w.) of lariciresinol for 9 weeks and E2 pelleted balb/c nu/nu athymic athymic mice with orthotopic MCF-7 human breast cancer tumors with dietary lariciresinol (20 mg/kg or 100 mg/kg in AIN-93G diet) for 5 weeks. The different letters (a,b) indicate statistically significant differences at p < 0.05 using Kruskal-Wallis median test, followed by Mann-Whitney U-test. The values are expressed as median ± SEM.
DMBA-induced rats
Control	289 ± 5	171 ± 14	398 ± 123
Lariciresinol 3 mg/kg b.w.	294 ± 6	172 ± 10	218 ± 685
Lariciresinol 15 mg/kg b.w.	293 ± 7	153 ± 10	306 ± 313
Athymic nude mice with MCF-7 human breast cancer xenografts
E2	20.8 ± 0.5	422 ± 70	214 ± 27^a
E2 + 20 mg/kg lariciresinol in diet	21.5 ± 0.5	408 ± 40	191 ± 16^a
E2 + 100 mg/kg lariciresinol in diet	21.1 ± 0.5	408 ± 41	125 ± 17^b

Effects of dietary lariciresinol on orthotopic MCF-7 human breast cancer xenograft growth, apoptosis and proliferation

The control group without E2 had significantly (p < 0.05) higher b.w. and smaller uterine weights compared to E2 maintained group (25.4 ± 0.8 g vs. 20.8 ± 0.5 g and 94 ± 13 vs. 422 ± 70 mg/100g b.w., respectively). Similarly to rats, no significant differences in b.w., uterus, or liver weight were observed in E2-maintained mice administered with a control diet or diet containing lariciresinol 20 or 100 mg/kg (Table II). Removal of E2 rapidly reduced the tumor sizes (Fig. 3) confirming the E2 responsiveness of the tumors. Intake of 100 mg/kg lariciresinol containing diet for 5 weeks reduced significantly MCF-7 tumor area in E2-maintained mice (Fig. 3). Accordingly, at the end of the experiment, the tumor weights were significantly lower in mice administered with 100 mg/kg lariciresinol containing diet compared to mice having a control diet (Table I). Lariciresinol diet increased the number of apoptotic TUNEL-positive tumor cells but did not affect the number of cells expressing the proliferation marker Ki-67 (Fig. 4).

Figure 3. Inhibition of estradiol (E2) -stimulated growth of orthotopic MCF-7 xenografts by dietary lariciresinol in athymic nude mice. Ovariectomized balb/c nu/nu athymic nude mice implanted with E2 pellets (1.7 mg, 60-day release) were injected with MCF-7 cells in matrigel (1:1 vol) into four sites of mammary glands (inguinal and thoracic). When the tumor area reached ∼20 mm², mice were allocated into groups with similar starting conditions; control (n = 5) fed with AIN-93G basal diet, or basal diet containing lariciresinol 20 mg/kg (n = 8) or 100 mg/kg (n = 6). From a group of mice (n = 5), the E2 pellet was removed to verify the E2 response of the established tumors (w/o E2-group). The tumor sizes were measured once per week. The non-normally distributed data are expressed as medians. Different letters (a–c) indicate statistically significant difference among the treatment groups at p < 0.05 determined by Kruskall-Wallis followed by Mann-Whitney U-test.

Figure 4. Effect of dietary lariciresinol on (a) Ki-67 labeling and (b) apoptotic indices of MCF-7 human breast cancer xenografts in ovariectomized E2-supplemented mice. The numbers of analyzed tumors were 19 for E2, 21 for E2 + 20 mg/kg lariciresinol, and 16 for E2 + 100 mg/kg lariciresinol in diet groups. Different letters (a–c) indicate statistically significant differences among the treatment groups at p < 0.05 determined by Kruskall-Wallis followed by Mann-Whitney U-test. Data are expressed as median ± SEM.

Table II. Receptor Expression Indices of MCF-7 Human Breast Cancer Tumors and Cultured Cells

	ERα	ERβ	PR
The tumors were established in ovariectomized athymic Balb/c nu/nu mice implanted with E2 pellet (1.7 mg/60-day release) by injecting 0.5 million MCF-7 cells into 4 sites of mammary fat pads. Mice with ∼20 mm^2 tumors were given a fresh E2 pellet, after which they were divided into three groups receiving control or lariciresinol 20 mg/kg or 100 mg/kg containing AIN-93G based diets. The tumor samples were collected after a 5-week dietary regimen. MCF-7 cells treated with or without 1nM E2 or E2 and the lignans were grown to 80% confluent cultures and then trypsinized, centrifuged and resuspended into a mixture of phenol red-free Dulbecco's minimum essential medium/F12 medium and 1% agarose in PBS at 50°C. The solid cell blocks were fixed in formalin and embedded in paraffin. Immunohistochemical staining was used to reveal cells expressing the receptors in tumors and cell cultures, and the positive nuclei more than 1000 cells were counted to determine the expression indices (values given as the number of positive nuclei per 1000 cells). The different letters (a–c) indicate statistical difference of normally distributed data at p < 0.05 using ANOVA followed by LSD post hoc test. The values are expressed as mean ± SEM.
MCF-7 orthotopic tumors
w/o E2	669 ± 76^a	170 ± 68^a	23 ± 9^a
E2	472 ± 35^b	145 ± 67^a	746 ± 26^b
E2 + 20 mg/kg lariciresinol in diet	597 ± 23^b	176 ± 82^a	856 ± 30^c
E2 + 100 mg/kg lariciresinol in diet	557 ± 30^b	337 ± 37^b	855 ± 33^c
MCF-7 cell cultures
w/o E2	107 ± 27	82 ± 9^a	19 ± 5^a
E2	0	75 ± 6^a	75 ± 21^b,c
E2 + 10 μM lariciresinol	0	157 ± 10^b	99 ± 5^c
E2 + 10 μM secoisolariciresinol	0	197 ± 51^b	46 ± 11^a,b
E2 + 10 μM enterodiol	0	111 ± 26^a	109 ± 17^c
E2 + 10 μM enterolactone	0	133 ± 20^a,b	102 ± 23^c

ERα, ERβ and PR expression of the MCF-7 tumors and cultured cells

There were significant differences in the expression indices of ERα, ERβ, and PR in MCF-7 human breast cancer cells grown either in vitro or othotopically as tumors in vivo in nude mice. The expression indices of all 3 receptor types were lower in cultured cells than in the tumors. Approximately half of the cancer cells in the MCF-7 tumor tissue of mice implanted with E2 pellet expressed ERα, while no ERα expressing nuclei were observed in E2 maintained cells in culture (Table II). ERα-positive nuclei were detected only in cells cultured without E2, and still at significantly lower frequency than in the xenografts. Also the expression index of PR was ∼10-fold higher in the MCF-7 tumors than in the cell cultures (Table II) in the presence of E2. Both E2-maintained and -deprived tumors and cell cultures expressed ERβ, and the expression indices were ∼2-fold higher in the tumors (Table II). However, the presence or absence of E2 did not significantly affect on ERβ expression in cell cultures or tumors.

Dietary lariciresinol increased the expression indices of ERβ and PR in the MCF-7 tumor tissue (Table II). Accordingly in cell cultures, treatment with lariciresinol or its further metabolite secoisolariciresinol increased the ERβ expression, while the enterolignans, enterodiol and enterolactone increased PR expression. No changes in ERα expression indices were measured in vivo or in vitro by lariciresinol or its further metabolites (Table II).

Vascularization of lariciresinol administered DMBA-induced mammary gland tumors in rats and MCF-7 xenografts in athymic mice

Administration of lariciresinol to mammary tumor bearing rats and mice reduced the tumor vascularization (p = 0.04). In DMBA-induced tumors, the average number of vessels in well differentiated growing tumors established during the treatment period were 30.9 ± 5.0, 25.9 ± 5.3 and 21.4 ± 1.3 vessels per tumor section area in control, lariciresinol 3, and 15 mg/kg b.w. groups, respectively. Accordingly, significantly reduced microvessel densities of the MCF-7 tumors were detected in mice administered with the lariciresinol diets compared to tumors from mice administered with the control diet (Fig. 5).

Figure 5. Dietary lariciresinol reduces microvessel density in MCF-7 tumors. (a) Immunohistochemical analyses of representative images of the means are shown. Tumor cross-sections were stained for microvessel density with anti-CD34 antibody (left panel) and digitized (right panel). Original magnification is ×200. (b) After binarization of the images from CD34-staining, microvessel density was estimated by scoring the total number of black pixels per field, as described earlier.34 All results are expressed as mean pixel counts per image ± SEM from 15 to 17 images of 3–5 tumors per group. Different letters (a–b) indicate statistically significant differences among the treatment groups at p < 0.05 determined by Kruskall-Wallis followed by Mann-Whitney U-test.

Estrogen-induced VEGF secretion in MCF-7 cells

7-Day estrogen treatment of the MCF-7 cells significantly increased the VEGF secretion into culture medium (Fig. 6). Enterolactone was able to significantly inhibit this E2 stimulated VEGF secretion, while other lignans showed smaller, statistically nonsignificant inhibitory effects (Fig. 6).

Figure 6. Secreted VEGF in MCF-7 cell cultures. The cells were cultured with 10 nM E2 alone or together with lignans at 1 μM concentration. The cells were cultured for 7 consecutive days in serum-free culture medium which was changed daily. VEGF was measured from media with ELISA. The different letters (a–c) indicate statistical difference in normally distributed data at p < 0.05 using ANOVA followed by LSD post hoc test. The values are expressed as average ± SEM.

Serum lignan concentrations of rats and athymic mice

Both daily administration of lariciresinol to rats by gavage (3 or 15 mg/kg b.w) or to athymic mice through the diet increased the serum enterolignan, i.e., enterodiol and enterolactone concentrations, compared to control animals (Table III). In addition, a significant increase in serum plant lignan concentrations, i.e., lariciresinol and secoisolariciresinol, was observed. Administration of a higher lariciresinol dose resulted in higher serum concentrations of both plant and enterolignans.

Table III. Serum Lignan Concentrations (nM) of Lariciresinol Administered DMBA-Induced Mammary Gland Tumors Bearing Female Rats and MCF-7 Xenografts Bearing Athymic Mice1

DMBA-induced rats	Time after the last dose	Lariciresinol	Secoisolariciresinol	Enterodiol	Enterolactone
1 The median serum lignan concentration (nM) is presented with the range of detected concentrations in parenthesis. Samples with no detectable lignans are marked as nd. The samples were analyzed by HPLC-MS/MS using a method previously described in Ref. 16, Sprague Drawley rats with DMBA-induced mammary tumors were administered daily p.o. with 3 or 15 mg/kg b.w. of lariciresinol for 9 weeks. The serum samples were collected at the time of sacrifice either 2 or 5h after the last administered lariciresinol dose. The serum samples of ovariectomized E2 pelleted balb/c nu/nu athymic mice were collected after 5-week administration of lariciresinol (20 mg/kg) in AIN-93G diet. Because of dietary administration of lariciresinol to athymic mice, the time from the last dose is not available, n/a.
Control	2 h	9.7 (8.6–13.4)	0.0 (nd)	0.0 (0–18.3)	0.0 (nd)
Lariciresinol 3 mg/kg b.w.	2 h	3382 (782–4427)	0.0 (0–17.2)	15.7 (0–156)	167 (22.5–182)
Lariciresinol 15 mg/kg b.w.	2 h	13531 (9342–15799)	39.3 (22.4–41.4)	339 (140–470)	625 (381–1546)
Control	5 h	6 (4.3–7.2)	0.0 (nd)	0.0 (nd)	0.0 (nd)
Lariciresinol 3 mg/kg b.w.	5 h	1086 (565–3.594)	0.0 (0–30.8)	192 (39.9–632)	206 (41.1–317)
Lariciresinol 15 mg/kg b.w.	5 h	4597 (3183–11374)	57.1 (19.9–207)	3612 (473–5470)	808 (409–1211)
Athymic nude mice with MCF-7 human breast cancer xenografts
E2	n/a	0.0 (0–12.1)	1.2 (0–4.0)	2.3 (0–3.4)	4.9 (0–10.1)
E2 + 20 ppm lariciresinol	n/a	0.0 (0–49.0)	22.5 (7.7–38.6)	23.6 (13.3–57.5)	17.5 (0.1–55.6)
E2 + 100 ppm lariciresinol	n/a	10.5 (0–542)	30.8 (27.4–42.2)	68.1 (50.8–257)	39.8 (25.4–125)

The serum lignan, especially lariciresinol, concentration in rats administered with lariciresinol by gavage varied depending on the duration since the last administered dose (Table III). Higher serum lariciresinol concentrations were detected at 2 hr post gavage than at 5 hr post gavage. In addition to time-dependent differences in serum plant lignan concentrations, significant interindividual variations in enterolignan concentrations were observed within the treatment groups (Table III).

Discussion

A recent epidemiological study indicates an inverse association between high dietary lariciresinol intake and breast cancer risk.28 However, according to our knowledge no experimental data exist concerning the effects of dietary lariciresinol on mammary gland cancer. In this study, we show for the first time that intake of lariciresinol inhibits the growth of hormone responsive mammary gland cancers in vivo. This effect was confirmed by using 2 hormone responsive mammary cancer models in 2 different species: DMBA-induced hormone responsive mammary gland cancer in rats and in orthotopic human MCF-7 breast cancer xenografts in athymic mice.

DMBA-induced mammary gland tumors are adenocarcinomas with different histological characteristics similar to majority of the human cancers.32 The model can be considered to give an overall estimation of the intervention effects on tumors with multiple histological characteristics at their different stages of development and growth. Previous studies have shown that dietary lignans secoisolariciresinol diglucoside,23, 24 7-hydroxymatairesinol36 and their metabolite enterolactone25 inhibit DMBA-induced mammary cancer growth when the lignan administration was started after part of the tumors were already established. Secoisolariciresinol diglycoside has also been reported to inhibit mammary tumorigenesis when its administration was started prior to formation of palpable tumors.23 However, 7-hydroxymatairesinol or enterolactone could not significantly reduce mammary tumor incidence or multiplicity if their administration was started 9 weeks after the tumor induction.25, 36 Similarly, in this study, lariciresinol administration that was started 10 weeks after the induction did not inhibit development of mammary tumors. These findings suggest that lariciresinol or its metabolite enterolactone cannot reverse or inhibit the DMBA-induced mutations or alterations in the mammary gland epithelium or stroma that lead to tumor development. Instead, we found that the tumors in lariciresinol treated rats grew slower than those in control rats, and that this growth inhibition was more pronounced in those tumors that developed during the lignan administration. This indicates that the lignans can better attenuate the growth of mammary cancers at their early stages of development. This may at least in part be because of their better accessibility of compounds into small size tumors than into large size tumors. This has recently been demonstrated in E2-maintained orthotopic human MCF-7 breast cancer xenografts in athymic mice administered with secoisolariciresinol diglucoside.37 These results in rats and in mice suggest possible direct effects of lignans on mammary tumor tissue.

Thus far, it was unclear whether athymic mice are able to convert plant lignans into mammalian lignans, due to possible defects in their intestinal microbiota caused by the immunodeficiency and aseptic housing environment. In this study, we demonstrate for the first time that the microbiota of athymic female mice administered with dietary lignan lariciresinol can produce enterolignans, similar to human microbiota.10 However, both serum plant lignan and enterolignan concentrations in athymic mice were significantly lower than those in the DMBA-administered rats after a similar dose of lariciresinol. This may be due to less efficient absorption or metabolism of the plant lignans in athymic mice compared to conventional mice and rats. Interestingly, in this study, we also found that the serum lignan concentrations in athymic mice were comparable to those reported in humans consuming lignan-rich diets38, 39 rather than those found in rats.

High variations in serum plant lignan concentrations were observed both in rats and in athymic mice. In rat serum, the lariciresinol concentrations were very high at 2 and 5 hr post gavage ranging from 782 to 15,799 nM. However, in serum samples collected from athymic mice fed with lariciresinol containing diet for 5 weeks, lariciresinol was present in only some of the samples and the concentration varied from 0 to 542 nM. These findings suggest that lariciresinol is mainly absorbed in the stomach and upper parts of the intestine and is efficiently and rapidly excreted into urine.

Lignan-rich flaxseed has been reported to attenuate E2-stimulated growth of human MCF-7 breast cancer xenografts in athymic mice,40 and similar attenuation of E2-enhanced MCF-7 tumor growth was recently demonstrated with the enterolignans enterodiol and enterolactone.27 Here, we demonstrated that intake of lariciresinol, a precursor of secoisolariciresinol and enterolignans, was able to delay E2-stimulated growth of the MCF-7 tumors, in line with previous reports. The mechanistic studies of the tumor growth inhibition of dietary lignans have focused on enterolignans and especially on enterolactone. However, the results especially on breast cancer cell proliferation in vivo and in vitro have been conflicting. In several studies in vitro, enterolactone has been shown to stimulate the proliferation of estrogen-sensitive MCF-7 cells.41–46 This stimulatory effect could be abolished with antiestrogens tamoxifen41, 45 and ICI-181,780.46 In the presence of E2, enterolactone, enterodiol and the plant lignans secoisolariciresinol and lariciresinol have an additive effect on cell proliferation47 suggesting that these compounds could act as ER agonists in breast cancer cells. Conversely, no tumor proliferation or growth stimulating properties of enterolactone, enterodiol or their plant precursors have been reported in human MCF-7 breast cancer xenografts grown in ovariectomized mice.26, 48 Furthermore, enterolignans have shown not to stimulate cell proliferation in E2-supplemented MCF-7 tumors in mice27 or in rats with mammary cancers.25 In E2-deprived or antiestrogen treated MCF-7 tumors, ERα expression has been reported to be upregulated while E2 supplementation reduces the expression levels. In the present study, no changes were observed in ERα expression levels of E2-stimulated MCF-7 tumors in response to several weeks' of lariciresinol administration. Interestingly, the relative binding affinity of enterolactone to both ERα and ERβ is very low,25, 49, 50 suggesting that enterolignans do not act as classical ER agonists in breast tumors in vivo. However, in vitro and in vivo, enterolactone has been shown to induce estrogen responsive reporter gene expression in tissue specific manner through ERs, and to promote some estrogen responsive genes and morphological changes in estrogen target tissues (e.g., uterus) in female mice.50 Thus, the ERα or ERβ mediated effects of the dietary lignans and their metabolites acting systemically in normal endocrine target tissues cannot be excluded.

In the present study, we observed a slight increase of PR expression index in MCF-7 tumors in the lariciresinol administered mice. PR expression has been generally used as a marker for ERα transactivation and its expression along with ERα is used as an indicator of better prognosis of breast cancer, and a criterion for endocrine therapy. However, in human breast tumors, the amount of ERα and PR show only weak correlation suggesting that other PR regulators than ERα are involved.51 Here, the increase in PR expression did not correlate with any of the other determined markers (e.g., ERα, ERβ, proliferation or apoptosis). Thus, PR expression in MCF-7 xenografts may merely be a marker of responsiveness to endocrine therapy rather than a direct indicator of tumor cell proliferation and growth.

Majority of the human ER-positive breast cancers express ERβ along with ERα52–55 but there is controversy about the expression of ERβ in MCF-7 cells and tumors. In our hands, both the MCF-7 cells in two-dimensional cultures and in orthotopic xenografts expressed ERβ, and similar to ERα and PR, the expression levels were higher in tumors than in cell cultures. A suggested explanation for the differences between the receptor expression levels in vitro and in vivo is that MCF-7 cells in two-dimensional cultures represent different cell populations with distinct expression patterns than those in three-dimensional cultures56 or in the xenograft tumors.57 Moreover, recent studies indicate that only a subpopulation of the cultured MCF-7 cells with specific gene and protein expression profiles is able to form tumors in vivo.58 These distinct expression patterns of the MCF-7 cells in cultures and tumors are likely to affect treatment responses and may also explain discrepancies between the obtained results with lignans in vitro and in vivo.

We observed increased ERβ expression levels in MCF-7 tumors in E2-supplemented ovariectomized mice fed with a diet containing 100 mg/kg lariciresinol, i.e., in tumors showing reduced growth rate. High ERβ expression in MCF-7 cell cultures has been shown to inhibit E2-dependent59 and independent proliferation.60 Enterolignans have been documented to have cytostatic properties on cell cycle and proliferation in rat hepatomas61 and human colon cancer xenografts,62 but we did not find any effect of lariciresinol on expression of tumor proliferation marker Ki-67 in MCF-7 xenografts. Instead, dietary lariciresinol increased apoptotic index of the tumor cells, which concurs with previous findings on enterolignans showing increased apoptosis of the orthotopic MCF-7 tumor cells in ovariectomized athymic mice in the absence of E2.26

Tumor vasculatization is a prerequisite for cancer growth and progression. In this study, administration of lariciresinol inhibited tumor angiogenesis measured as reduced vessel number and density both in mammary tumors bearing rats and in athymic mice. This effect of lariciresinol is likely to occur at least in part through enterolignans formed from lariciresinol in vivo. Previously, both enterodiol and enterolactone have been shown to inhibit E2-stimulated MCF-7 tumor growth through inhibition of angiogenesis by blocking E2-induced secretion of VEGF into extracellular space of the cancer cells.27 Soluble and freely diffusible VEGF is considered as a biologically active angiogenic protein in extracellular space of the cancer cells.63 Accordingly in this study, enterolactone significantly inhibited E2-stimulated VEGF secretion of MCF-7 cells. These findings indicate that VEGF signaling pathway is one putative target for the lignan effects in inhibition breast cancer angiogenesis.

In summary, lariciresinol, one of the main dietary lignans, inhibited mammary cancer growth in 2 different in vivo models for hormone responsive breast cancer; carcinogen-induced rat mammary tumors and human MCF-7 breast cancer xenografts. These findings support the recent epidemiological data on the favorable role of the dietary lignan lariciresinol in breast cancer risk reduction.28 We also demonstrate that inhibition of tumor angiogenesis, enhanced tumor cell apoptosis and increased ERβ expression are possible mechanisms that account for the reduced MCF-7 human breast cancer growth in lariciresinol treated animals. This study further supports the importance of diet with specific components in inhibition of breast cancer development.

Acknowledgements

We thank Dr. Johanna Junnila, Ms. Terhi Sivonen, Ms. Erica Nyman, Ms. Teija Hurmerinta, and Ms. Liisi Kortela for their help with the tumor sample processing and immunohistochemistry.