A Comparative Study of Growth-Inhibitory Effects of Isoflavones andTheir Metabolites on Human Breast and Prostate Cancer Cell Lines

Hong Xiang, Galina Schevzov, Peter Gunning, He M. Williams, and Martin Silink

Abstract: The possible growth-inhibitory properties of therecently synthesized novel metabolite 1-(2,4-dihydrobenzoyl)-1-(4-hydroxyphenyl)ethylene (2-de-O-DMA) and six othermetabolites of isoflavones were investigated and comparedwith those of the major isoflavones genistein, daidzein, andglycitein on human breast noncancer and breast and prostatecancer cell lines in vitro. The novel metabolite 2-de-O-DMAwas found to be a more potent inhibitor than genistein on hu-man breast cancer MCF-7, MDA-MB-468, and SK-BR-3 cellsand breast noncancer MCF-10A cells. In prostate cancer celllines, LNCaP and DU145, 2-de-O-DMA elicited a six- tosevenfold more potent inhibition than genistein. Flow cyto-metric analysis showed that 2-de-O-DMA and genisteinblocked cells at the G2/M phase of the cell cycle. Genisteinand 2-de-O-DMA led to apoptosis of a variety of cancer celllines. The rapid response of growth inhibition induced by 2-de-O-DMA compared with genistein strongly suggests thatthe observed antiproliferation effects elicited by this novelmetabolite are mediated via a biological pathway differentfrom that induced by genistein. 2-de-O-DMA, a novel metab-olite of isoflavone, could have a potential role in chemo-preventive and chemotherapeutic treatment of hormonalbreast and prostate cancers.

Introduction

Epidemiological and migrant studies suggest that the in-cidence and mortality of the hormone-dependent breast andprostate cancers are much lower in Asian than in Westerncountries. One of the most important factors to explain thelow risk of hormone-dependent and certain types of cancersis the difference in the diets (1,2). Asian people, who con-sume a diet rich in soy phytoestrogens, have a higher excre-tion of urinary phytoestrogens, especially isoflavones andtheir metabolites (3,4). Isoflavones are one of a major classof phytoestrogens, are structurally similar to estrogens (5),bind to estrogen receptors (ER), and, hence, have estrogenic

and antiestrogenic activities and their own growth-inhibi-tory effects independent of the ER (6–8). Isoflavones andtheir metabolites have been considered to reduce the riskof cancer and to have potent anticarcinogenic activities (9,10). The mechanisms by which these isoflavones exert theiranticancer activities have been attributed to their direct inhibi-tion of protein tyrosine kinase (11), inhibition of DNA-to-poisomerase II (12), inhibition of angiogenesis (13),antiproliferation and cell cycle arrest (14), and induction ofapoptosis (15).

Genistein, daidzein, and glycitein are the main isofla-vones found in soy foods. These three major isoflavones andtheir metabolites dihydrodaidzein, O-demethylangolensin(O-DMA), 6�-hydroxy-O-demethylangolensin (6-OH-O-DMA), dihydrogenistein, 1-(2,4-dihydrobenzoyl)-1-(4-hy-droxyphenyl)ethylene (2-de-O-DMA), and two isomers oftetrahydrodaidzein (cis- and trans-tetrahydrodaidzein) wereidentified from human urinary excretion after soy consump-tion (4).

To understand the potential biological activities of thesemetabolites, we synthesized seven metabolites of isofla-vones, investigated their growth-inhibitory effects, and com-pared them with the isoflavones genistein, daidzein, andglycitein on human breast cancer MCF-7 and MDA-MB-468 cells. The novel metabolite 2-de-O-DMA exhibited apotent growth-inhibitory effect on human breast ER-positiveMCF-7 cells and ER-negative MDA-MB-468 cells. We fur-ther examined this metabolite in other human breast cancerSK-BR-3 (ER-negative), human breast noncancer MCF-10A (ER-negative), and human prostate cancer LNCaP [an-drogen receptor (AR)-positive] and DU145 (AR-negative)cell lines. This study shows that the novel metabolite 2-de-O-DMA is still able to inhibit the proliferation of MCF-10A(ER-negative), SK-BR-3 (ER-negative), LNCaP, andDU145 cells. Our data showing the rapid response elicitedby 2-de-O-DMA compared with genistein suggest that theantiproliferation effects caused by genistein and 2-de-O-DMA may be mediated via different mechanisms.

NUTRITION AND CANCER, 42(2), 224–232Copyright © 2002, Lawrence Erlbaum Associates, Inc.

H. Xiang, H. M. Williams, and M. Silink are affiliated with the Ray Williams Institute of Paediatric Endocrinology and Diabetes and Metabolism and G.Schevzov and P. Gunning with the Oncology Research Unit, The Children’s Hospital at Westmead, Westmead NSW 2145, Sydney, Australia. H. Xiang and M.Silink are also affiliated with the University of Sydney, Sydney, NSW, Australia.

Materials and Methods

Materials

Genistein was purchased from Sigma and glycitein fromPlantech. Daidzein and the seven metabolites of isoflavones,including 2-de-O-DMA, O-DMA, 6-OH-O-DMA, dihydro-genistein, dihydrodaidzein, and cis- and trans-tetrahydro-daidzein, were synthesized at the Ray Williams Institute ofPaediatric Endocrinology, Diabetes, and Metabolism. 4,6-Diamidino-2-phenylindole (DAPI), ribonuclease, propidiumiodide, fetal bovine serum (FBS), insulin, epidermal growthfactor (EGF), cholera toxin, hydrocortisone, and dimethylsulfoxide (DMSO) were purchased from Sigma-Aldrich.The colorimetric 3-(4,5-dimethylthiazol-2-yl)-5-(3-car-boxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium(MTS) assay was obtained from Promega. RPMI 1640, L-glutamine, and donor horse serum were purchased fromTRACE.

Cell Lines and Tissue Culture

MCF-7, MDA-MB-468, SK-BR-3, MCF-10A, andDU145 cell lines were purchased from the American TypeCulture Collection (Manassas, VA). The LNCaP cell linewas kindly provided by the Commonwealth Scientific andIndustrial Research Organization Division of Molecular Sci-ence. MCF-7, MDA-MB-468, SK-BR-3, LNCaP, andDU145 cells were routinely cultured in RPMI 1640 supple-mented with 10% FBS and 2 mM L-glutamine. MCF-10Acells were grown in RPMI 1640 supplemented with 5% do-nor horse serum, 20 ng/ml EGF, 100 ng/ml cholera toxin, 10µg/ml insulin, and 0.5 µg/ml hydrocortisone. MCF-10A is ahuman transformed but nontumorigenic breast epithelial cellline used in this study as a model system of a noncancer celltype. The cells were incubated at 37°C in a humidified atmo-sphere of 5% CO2 in air. Compounds were dissolved inDMSO (final DMSO concentration �0.2%) for all experi-ments. DMSO vehicle controls were used in all studies.

Synthesis of the Isoflavone Daidzein and Metabolites ofIsoflavones

Daidzein was prepared according to the method ofWähälä and Hase (16). Reduction of daidzein using palla-dium-on-charcoal as a catalyst and ammonia formate as ahydrogen donor gave the mixture of dihydrodaidzein andcis- and trans-tetrahydrodaidzein (17,18). The ratio of thesecompounds varied depending on the reaction time. Thecrude product was purified by column chromatography fol-lowed by semipreparative high-performance liquid chroma-tography. Cis- and trans-tetrahydrodaidzein isomers wereisolated successfully and were stable in our hands. We didnot observe autoxidation or degradation of these products asreported by Wähälä and Hase (18). These compounds werecharacterized by 1H and 13C nuclear magnetic resonance

(NMR) and high-resolution mass spectrometry. Hydrogena-tion of genistein (Novogen) using palladium-on-charcoal asa catalyst in acetic acid gave dihydrogenistein exclusively ina reaction in which genistein was not protected (19). O-DMA was readily available by the acylation and thendemethylation of genistein (20). We observed that the reduc-tion of genistein using lithium aluminum hydride yielded 6-OH-O-DMA. The 1H-NMR and gas chromatography-massspectroscopy data of 6-OH-O-DMA agree with data for 6-OH-DMA purchased from Plantech. 6-OH-O-DMA, synthe-sized according to the method of Salakka and Wähälä (21),was the compound used in these studies. 2-de-O-DMA wassynthesized by condensation via introduction of a methylenegroup into the �-position of the 2,4-dimethoxyphenyl-4�-methoxybenzyl ketone followed by demethylation. 2-de-O-DMA was obtained as a white powder: melting point143–145°C, ultraviolet (MeOH) �max (log �) 271 (4.23) and243 (4.08) nm, infrared (CH3CN) �max 3,349 and 1,703 cm–1,1H-NMR (acetone-d6, 200 MHz) � 5.26 (1H, d, J = 5.22 Hz,CH2), 5.88 (1H, d, J = 5.3 Hz, CH2), 6.34 (1H, dd, J =1.34,9.3 Hz, ArH-5), 6.38 (1H, d, J = 1.3 Hz, ArH-3), 6.84(2H, d, J = 8.7 Hz, ArH-3�,5�), 7.30 (2H, d, J = 8.7 Hz, ArH-2�,6�), and 7.55 (1H, d, J = 9.3 Hz), 13C-NMR (acetone-d6, 50MHz) � 103.42, 108.72, 113.45, 113.74, 114.92, 115.21,116.21, 128.81, 136.30, 158.62, 152.81, 166.49, 168.47, and202.58, high-resolution electron ionization mass spectrome-try mass-to-charge ratio 256.07479 (calculated for C15H12O4,256.07355).

In Vitro Growth-Inhibitory Assay

Cytotoxicity of the isoflavones and their metabolites wasestimated using the colorimetric MTS assay, which involvesconversion of a tetrazolium salt to a colored, aqueous solu-ble formazan product made by mitochondrial activity of via-ble cells at 37°C (22). Briefly, cells were trypsinized andplated at a density of 3,000 or 5,000 cells per well onto flat-bottomed 96-well microplates. After 24 h of incubation forcell attachment, the medium was replaced with fresh me-dium containing appropriate concentrations of compoundsand treated for various times (24, 48, 72, and 96 h).Absorbance, as a measure of viable cell number, was read at492 nm on a TECAN SPECTRA Fluor Plus microplatereader.

Cell Cycle Progression Assay

MCF-7, MCF-10A, LNCaP, and DU145 cells wereseeded at low density (1.5 × 105 cells/100-mm dish) andgrown to 50% confluence before addition of various concen-trations of the tested compounds for various times. Adherentcells were harvested by trypsinization and floating cells bycentrifugation (23). Cell cycle progression was analyzed incontrol and treated cells on the basis of the method describedby Bradshaw et al. (24). After the cells were washed twicewith ice-cold phosphate-buffered saline (PBS, pH 7.4), they

Vol. 42, No. 2 225

were suspended in a fluorochrome solution (0.1% Triton X-100, 0.1 mg/ml ribonuclease A, and 0.05 mg/ml propidiumiodide) and incubated for 60 min at 4°C. Flow cytometricDNA ploidy analysis was performed using a Becton Dick-inson FACScan flow cytometer. The Cell-Quest softwarepackage (Becton Dickinson, Mansfield, MA) was used toanalyze data gathered on 20,000 cells for each individualsample (25).

Analysis of Apoptosis

Nuclear morphology was demonstrated by staining cellswith DAPI. Briefly, control and treated cells were harvested3–72 h after treatment. Adherent and floating cells were har-vested together, cytospun onto glass slides using a Cyto-Tekcentrifuge (Miles, Elkart, IN), and air dried. Slides were fixedin 4% paraformaldehyde in PBS for 20 min, permeabilized in�20°C methanol for 20 min, washed twice with 2% FBS inPBS, and incubated with DAPI at 1 µg/ml in PBS for 10 minat 37°C. The staining solution was then removed, and theslides were washed with PBS again, mounted with coverslips,and visualized under fluorescence microscopy.

Terminal Deoxynucleotidyl Transferase-MediateddUTP Nick End Labeling Assay

Control and treated cells were harvested and cytospun ontoslides. TdT-mediated dUTP nick end labeling (TUNEL) as-

says were performed to detect apoptotic nuclei using theIn Situ Cell Death Detection Kit, Fluorescein (BoehringerMannheim) according to the manufacturer’s instructions. Forquantification of apoptotic cells, the percentage of TUNEL-positive cells was assessed in six random fields at ×400 mag-nification.

Results

Effects of Isoflavones and Their Metabolites onGrowth of Breast and Prostate Cancer Cell Lines

Seven metabolites of isoflavones were synthesized to iden-tify potential inhibitors of cancer cell growth. The structuresof these metabolites are shown in Fig. 1. These metaboliteswere first evaluated on human breast cancer MCF-7 (ER-positive) and MDA-MB-468 (ER-negative) cell lines andcompared with the main isoflavones genistein, daidzein, andglycitein. The 50% inhibitory concentrations (IC50) for allcompounds were determined by functional viability assayswith the MTS assay. Table 1 shows the data for the inhibitoryaction of isoflavones and their metabolites. The novel metab-olite 2-de-O-DMA showed potent inhibitory effects on all thecell lines tested (Table 1). The IC50 values of 2-de-O-DMA(15.4 and 6.7 µM) were threefold more potent than the IC50

values of genistein (47 and 22.6 µM) on MCF-7 and MDA-MB-468 cells, respectively. Similarly, in the noncancer MCF-10A cells, 2-de-O-DMA was a more potent growth inhibitor

226 Nutrition and Cancer 2002

Figure 1. Structures of isoflavones and their metabolites compared with 17-estradiol. 2-de-O-DMA, 1-(2,4-dihydroxybenzoyl)-1-(4-hydroxyphenyl)ethyl-ene; O-DMA, O-demethylangolensin; 6-OH-O-DMA, 6�-hydroxy-O-demethylangolensin.

than genistein (Table 1). 2-de-O-DMA also exhibited growth-inhibitory properties in human prostate cancer cell lines andwas six- to sevenfold more potent than genistein in the inhibi-tion of LNCaP and DU145 cells (Table 1).

In MCF-7 cells treated with �40 µM 2-de-O-DMA for 24h, cell growth decreased dramatically (Fig. 2). In contrast,genistein-treated cells showed no such growth inhibition at asimilar time of drug exposure (Fig. 2). However, long-term

Vol. 42, No. 2 227

Table 1. Comparison of Growth-Inhibitory Effect of Isoflavones and Their Metabolites on Human Breast Cancer,Noncancer, and Prostate Cancer Cell Linesa

IC50, µM

Cell Line Genistein Daidzein Glycitein 2-de-O-DMA (M1) M2–M7

Breast cancerMCF-7 [ER(+)] 47 157.4 140.7 15.2 146–155MDA-MB-468 [ER(�)] 22.6 99.1 136.5 6.6 146–155SK-BR-3 [ER(�)] 65.9 157.4 ND 8.6 ND

NoncancerMCF-10A [ER(�)] 28 157.4 ND 19.1 ND

Prostate cancerLNCaP [AR(+)] 26 111.7 ND 4 NDDU145 [AR(�)] 45 157.4 ND 6.2 ND

a: Cells were plated at a density of 3–5 × 103 per well (96-well plate) and allowed to adhere for 24 h before addition of various compounds at 1–155 µM. After96 h of incubation, 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) assays were performed. Valuesare means of �3 representative experiments (n � 12, SD � 7%). IC50, 50% inhibitory concentration; M1–M7, Metabolites 1–7: 2-de-O-DMA, 1-(2,4-dihydroxybenzoyl)-1-(4-hydroxyphenyl)ethylene (M1); O-demethylangolensin (M2), 6�-hydroxy-O-demethylangolensin (M3), dihydrogenistein (M4),dihydrodaidzein (M5), cis-tetrahydrodaidzein (M6), trans-tetrahydrodaidzein (M7); ER, estrogen receptor; AR, androgen receptor; ND, not determined.

Figure 2. Effects of genistein, daidzein, and 2-de-O-DMA on growth of human breast cancer MCF-7 and MDA-MB-468 cell lines. Cells were treated withvariable concentrations of isoflavonoids for 24 and 96 h before cell growth was determined by 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium assay. Growth of cells in control medium without isoflavonoids was normalized to 100%. Results are representative of 3 ex-periments; points are averages of 12 wells, which varied by �7%.

exposure (i.e., 96 h) of MCF-7 cells to �20 µM 2-de-O-DMA resulted in detection of nearly no viable cells (con-firmed by trypan blue staining; Fig. 2). As previously reported(26,27), a mild stimulatory effect was induced by 1–2 µMgenistein on MCF-7 (ER-positive) cells after 96 h of expo-sure (Fig. 2). This stimulatory effect displayed by genisteinwas not elicited by 2-de-O-DMA (Fig. 2). Thus 2-de-O-DMA induced a higher degree of cytotoxicity within ashorter exposure time than genistein.

2-de-O-DMA and Genistein Induce G2/M Cell CycleArrest in Breast and Prostate Cancer Cell Lines

To assess the effects of genistein and 2-de-O-DMA oncell cycle progression, flow cytometric analysis was used toexamine the cell cycle distribution of human breast cancerMCF-7, noncancer MCF-10A, and human prostate cancerLNCaP and DU145 cells. Cells were treated with differentconcentrations of these two compounds for various times

and stained with propidium iodide. After 24 h of treatmentwith genistein and 2-de-O-DMA, a two- and a fourfold in-crease in the number of cells arresting in the G2/M phasewere observed for MCF-7 and MCF-10A cells, respectively(P � 0.01 in both cases; Table 2, Fig. 3). A 72-h analysis ofcell cycle progression was not performed on MCF-7,LNCaP, and DU145 cells treated with 39 µM 2-de-O-DMAbecause of extensive cell death. However, in the case of thehuman breast noncancer MCF-10A cells, a 72-h analysisshowed that 74.7% were arrested in the G2/M phase aftertreatment with 20 µM 2-de-O-DMA (5-fold increase, P �

0.01) compared with 37.8% after 148 µM genistein (3-foldincrease, P � 0.01; Table 2, Fig. 3). A higher concentrationof 2-de-O-DMA (39 µM) for a 72-h exposure delayed MCF-10A cells at the S and G2/M phases (Table 2). Genistein and2-de-O-DMA also arrested the cell cycle in LNCaP andDU145 cells at the G2/M phase. Genistein at 74 µM pro-duced a 83.7% G2/M accumulation in DU145 cells (a 3-foldincrease, P � 0.01) and decreased the percentage of cells inthe G1 phase from 51.6% of control to 5.9% of treatedDU145 cells (9-fold decrease, P � 0.01; Table 2).

Effects of Genistein and 2-de-O-DMA onApoptotic Cell Death

To investigate nuclear morphology and determine whether2-de-O-DMA and genistein induced apoptosis in vitro, treatedand untreated MCF-7 and MCF-10A cells were stained withDAPI to identify apoptotic nuclei. Apoptosis induces mor-phological changes, including nuclear condensation, forma-tion of apoptotic bodies, and pycnotic morphology. Thesecharacteristics were clearly visualized in the nuclei of MCF-7cells treated with genistein at 72 h and 2-de-O-DMA at 24 h(Fig. 4). Apoptotic cells were not detected in MCF-10A cellstreated with 37, 74, and 148 µM genistein for 24, 48, and 72 h,but 39 µM 2-de-O-DMA induced apoptosis of MCF-10Acells after 72 h of exposure (Table 3).

TUNEL assay (Boehringer Mannheim) was performed toquantitate the percentage of apoptotic cells (Fig. 4D, Table 3).TUNEL-positive cells were counted in six random fields to-taling �1,000 cells. After treatment of MCF-7 cells with 148µM genistein for 48 and 72 h, 31.7% and 52.2% of the cells,respectively, were observed to be apoptotic (Table 3). At 74µM genistein, no significant number of apoptotic cells wereseen. In contrast, exposure to 78 µM 2-de-O-DMA for only24 h resulted in 31% apoptotic cells. Hence, 2-de-O-DMA in-duced a rate of apoptosis similar to that induced by genisteinbut at a more rapid rate (24 vs. 48 h) and at a lower concentra-tion (78 vs. 156 µM 2-de-O-DMA; Fig. 4D, Table 3).

In the case of the prostate cancer cell lines, no significantinduction of apoptotic cells was observed in the LNCaP cellstreated with genistein or 2-de-O-DMA. In contrast, 48 h ofexposure to 39 µM 2-de-O-DMA or 148 µM genistein sig-nificantly induced apoptosis in the DU145 cells with levelsof 16% and 10%, respectively, compared with control, 1.2%(Table 3).

228 Nutrition and Cancer 2002

Table 2. Cell Cycle Analysis of Human Breast CancerMCF-7, Breast Noncancer MCF-10A, and HumanProstate Cancer LNCaP and DU145 Cells Treated WithGenistein or 2-de-O-DMAa

24 h 72 h

G1 S G2/M G1 S G2/M

Breast

MCF-7Control 47.2 26.9 26.2 72.2 17.3 10.5Genistein (37 µM) 49.6 24.5 25.9 67.6 21.5 10.9Genistein (74 µM) 42.7 27.3 30.0 56.3 18.3 25.4Genistein (148 µM) 37.3 18.8 44.2 47.9 12.9 39.22-de-O-DMA (39 µM) 37.7 19.2 43.3

MCF-10AControl 71.6 16.8 12.1 80.7 5.9 13.7Genistein (37 µM) 58.4 18.3 23.3 66.2 6.8 27.0Genistein (74 µM) 43.3 9.5 47.4 48.2 9.8 45.5Genistein (148 µM) 52.7 9.1 38.5 59.1 3.3 37.82-de-O-DMA (20 µM) 38.9 8.3 52.8 18.2 7.1 74.72-de-O-DMA (39 µM) 54.0 26.3 20.3 45.5 21.6 33.6

Prostate

LNCaPControl 49.9 24.6 25.5Genistein (37 µM) 37.6 22.5 39.9Geinstein (74 µM) 30.9 10.3 58.82-de-O-DMA (39 µM) 36.0 20.5 43.5

DU145Control 51.6 18.9 29.5Genistein (37 µM) 26.4 17.8 55.8Genistein (74 µM) 5.9 10.4 83.72-de-O-DMA (39 µM) 33.2 19.2 47.6

a: Treated cells were harvested and fixed, and their DNA contents werestudied by propidium iodide staining followed by flow cytometry. Con-trol cells were treated with 0.2% dimethyl sulfoxide. Values representaverage of 3 separate experiments. See Results for statistical signifi-cance.

Discussion

Phytoestrogens are derived from plants. Two majorchemical types of phytoestrogens that have been identifiedare isoflavones (genistein, daidzein, and glycitein) and lig-nans (28). Isoflavones are plant diphenols found enriched insoy, which have been proposed to contribute an importantpart of the anticancer effects found in soy. Their estrogen ag-onist, antagonist, and ER-independent anticancer propertieshave attracted many researchers to identify the metabolitesof isoflavones from human urine and evaluate their bio-logical activities (4,29–31). Isoflavones are metabolized toisoflavonoid metabolites by the intestinal bacteria. The bio-logical effects of metabolites of isoflavones could be differ-ent from the primary isoflavones. For example, equol and O-DMA, two metabolites of daidzein, are more potent thangenistein and daidzein as antioxidants (32).

In the present study, we compared the growth-inhibitoryeffects of three major isoflavones, genistein, daidzein, andglycitein, and seven metabolites of isoflavones in hormonalcancer cell lines. We have found that the novel metabolite 2-de-O-DMA is able to inhibit the proliferation of breast andprostate cancer cells at three- to eightfold lower concentra-tions than genistein. This high degree of potency displayedby 2-de-O-DMA has not been demonstrated by any otherisoflavone metabolites. Booth and co-workers (39) tested a

total of seven novel isoflavones on normal and transformedintestinal epithelial cell lines and found that none was a morepotent inhibitor than genistein. Furthermore, the time takento induce an antiproliferative effect by 2-de-O-DMA ismuch shorter than that seen with genistein. 2-de-O-DMA in-duced an inhibitory effect on ER-positive and ER-negativecells, and the stimulatory effect displayed by low levels ofgenistein was absent. The inhibitory and stimulatory effectsof genistein observed in this study are due in part to thedose-dependent estrogenic activities, as previously reported(26,33). Our results of genistein in prostate cancer cells areconsistent with those of Peterson and Barnes (34) and Zhouet al. (40), who reported that genistein inhibited human pros-tate cancer LNCaP and DU145 cells. However, 2-de-O-DMA is a more efficient inhibitor of cell growth thangenistein in these prostate cancer cell lines. In the case ofLNCaP cells, an IC50 of 50 µM was observed for genistein(40), whereas an IC50 of only 4 µM was observed for 2-de-O-DMA. The other six metabolites and two major isoflavones,daidzein and glycitein, show weaker or no inhibitory effectson the human breast cancer cell lines at 1–155 µM. The pres-ent data demonstrate that ER and AR are not essential forisoflavones and their metabolites to inhibit the growth ofhuman breast and prostate cancer cells. The finding that 2-de-O-DMA acts as an inhibitor of cell growth on hormonalcancer cell lines in vitro supports the evidence of cancer-

Vol. 42, No. 2 229

Figure 3. Effects of genistein and 2-de-O-DMA on cell cycle distribution of human breast cancer MCF-7 and breast noncancer MCF-10A cells. Cells weretreated with genistein (148 µM for MCF-7 and 74 µM for MCF-10A) and 2-de-O-DMA (39 µM for MCF-7 and 20 µM for MCF-10A) for 24 h. Treated cellswere harvested and fixed, and their DNA contents were studied by propidium iodide staining followed by flow cytometry.

protective effects of dietary soy products that are attributednot only to the primary isoflavones but also to some of theirmetabolites (30).

Cell cycle analysis suggests that the mechanisms of ac-tion of genistein and 2-de-O-DMA are not identical. As pre-viously reported (15,33,35), the growth-inhibitory action ofgenistein on MCF-7, LNCaP, and DU145 cells was accom-panied by G2/M phase arrest. G2/M phase arrest is also in-duced by genistein in the human nonbreast cancer cell lineMCF-10A. 2-de-O-DMA was similarly seen to arrest cells atthe G2/M phase. The G2/M arrest induced by genistein and 2-de-O-DMA was coupled with a parallel depletion of the G0/G1 phase. Interestingly, in the human nonbreast cancer cellline MCF-10A, a G2/M phase arrest was induced by 2-de-O-DMA at a low concentration (20 µM); however, at a higherconcentration (39 µM), an S-G2/M phase arrest was evident.An S-G2/M arrest by 2-de-O-DMA was not apparent in theother cell lines tested and was not seen with genistein.

The apoptosis induced by genistein has been previouslydemonstrated in a number of human cell lines (15,35–38),and this, together with the induced G2/M phase arrest, islikely to be responsible for the observed inhibition in cell

proliferation. Even though genistein and 2-de-O-DMA in-duced a G2/M phase arrest and cell death via apoptosis, theresponse of 2-de-O-DMA on apoptosis was more rapid thanthat seen with genistein. Together, these findings stronglysuggest that the mechanism of action of 2-de-O-DMA is inpart distinct from that of genistein.

In conclusion, although the present results do not revealthe mechanism(s) responsible for the S- and G2/M phase ar-rest of 2-de-O-DMA, our in vitro results demonstrate thatthe novel metabolite of isoflavone 2-de-O-DMA is a potentinhibitor of hormonal cancer proliferation. Further in vitrostudies coupled with in vivo studies are needed to establishthe biologically significant concentration of 2-de-O-DMA.As previously reported in the literature, the in vivo concen-tration of the isoflavone genistein in humans consuming adiet rich in soy is between 276 nM and 4 µM (41,42). Theselevels are low compared with the in vitro concentrations ofgenistein needed to inhibit cell growth, 22–66 µM, as re-ported in our study and by others (41). In the case of ournovel metabolite 2-de-O-DMA, our in vitro results demon-strate that relatively low concentrations, 4–20 µM, can act asan effective inducer of cell growth inhibition. Hence, we

230 Nutrition and Cancer 2002

Figure 4. Evaluation of apoptosis of MCF-7 cells treated with 148 µM genistein for 72 h (B) and 39 µM 2-de-O-DMA for 24 h (C and D). A: control cellstreated with 0.2% DMSO for 72. A–C: 4,6-diamidino-2-phenylindole stain demonstrating nuclear morphology. D: TdT-mediated nick end labeling assay show-ing apoptotic nuclei. Scale bars, 20 µM (A–C) and 40 µM (D).

propose that this novel compound may contribute to the pre-vention of breast and prostate cancers.

Acknowledgments and Notes

The authors thank Christine Smyth for assistance with flow cytometrydata. Contributions of Dr. G. Joannou in leading the Metabolic MassSpectroscopy Unit, stimulating interest in phytoestrogen research, andhelping start the Ph.D. research project are warmly acknowledged. Thiswork was supported by unrestricted educational grants from Dr. R. R.Williams and HIH Insurance Co. and forms part of the research under-taken by H. Xiang for a Ph.D. thesis. Address correspondence to M.Silink, Ray Williams Institute of Paediatric Endocrinology, Diabetes &Metabolism, The Children’s Hospital at Westmead, Locked Bag 4001

Westmead, Westmead NSW 2145, Sydney, Australia. Phone: (+612)98453172. FAX: (+612) 98453170. E-mail: MartinS@chw.edu.au.

Submitted 16 May 2001; accepted in final form 14 January 2002.

References

1. Rose DP, Boyar AP, and Wynder EL: International comparison ofmortality rates for cancer of the breast, ovary, prostate, and colon, andper capita food consumption. Cancer 58, 2363–2371, 1986.

2. Shimizu H, Ross RK, Bernstein L, Yatani R, Henderson BE, et al.:Cancer of the prostate and breast among Japanese and white immi-grants in Los Angeles County. Br J Cancer 63, 963–966, 1991.

3. Adlercreutz H, Honjo H, Higashi A, Fotsis T, Hamalainen E, et al.:Urinary excretion of lignans and isoflavonoid phytoestrogens in Japa-nese men and women consuming a traditional Japanese diet. Am J ClinNutr 54, 1093–1100, 1991.

4. Kelly GE, Nelson C, Waring MA, Joannou GE, and Reeder AY: Me-tabolites of dietary (soya) isoflavones in human urine. Clin Chim Acta223, 9–22, 1993.

5. Setchell KDR and Adlercreutz H: Mammalian lignans and phyto-estrogens. Recent studies on their formation, metabolism and biologi-cal role in health and disease. In Role of the Gut Flora in Toxicity andCancer, Rowland IR (ed). London: Academic, 1988, pp 315–345.

6. Shutt DA and Cox RI: Steroid and phytoestrogen binding to sheeputerine receptor in vitro. J Endocrinol 52, 299–310, 1972.

7. Martin PM, Horwitz KB, Ryan DS, and McGuire WL: Phytoestrogeninteraction with estrogen receptors in human breast cancer cells. Endo-crinology 103, 1860–1867, 1978.

8. Peterson G and Barnes S: Genistein inhibits both estrogen- and growthfactor-stimulated proliferation of human breast cancer. Cell GrowthDiffer 7, 1345–1351, 1996.

9. Messina MJ, Persky V, Barnes S, and Setchell KDR: Soy intake andcancer risk: a review of the in vitro and in vivo data. Nutr Cancer 21,113–131, 1994.

10. Wei H, Bowen R, Cai Q, Barnes S, and Wang Y: Antioxidant andantipromotional effects of the soybean isoflavone genistein. Proc SocExp Biol Med 208, 124–130, 1995.

11. Akiyama T, Ishida J, Nakagawa S, Ogawara R, Watanabe SI, et al.:Genistein, a specific inhibitor of tyrosine-specific protein kinases. JBiol Chem 262, 5592–5595, 1987.

12. Markovits J, Markovitz J, Linassier C, Fosse P, Couprie J, et al.: Inhib-itory effects of the tyrosine kinase inhibitor genistein on mammalianDNA topoisomerase II. Cancer Res 49, 5111–5117, 1989.

13. Fotsis T, Pepper M, Adlercreutz H, Hase T, Montesano R, et al.:Genistein, a dietary ingested isoflavone, inhibits cell proliferation andin vitro angiogenesis. Proc Natl Acad Sci USA 90, 2690–2694, 1993.

14. Matsukawa Y, Marui N, Sakai T, Satomi Y, and Yoshida M: Genisteinarrests cell cycle progression at G2-M. Cancer Res 53, 1328–1331,1993.

15. Constantinou AI, Kamath N, and Murley JS: Genistein inactivates bcl-2, delays the G2/M phase of the cell cycle, and induces apoptosis of hu-man breast adenocarcinoma MCF-7 cells. Eur J Cancer 34, 1927–1934, 1998.

16. Wähälä K and Hase T: Expedient synthesis of polyhydroxyiso-flavones. J Chem Soc Perkin Trans 1, 3005–3008, 1991.

17. Krishnarnurty HG and Sathyanarayana S: Catalytic transfer hydroge-nation, a chemo-selective reduction of isoflavones to isoflavanones.Synth Commun 16, 1657–1663, 1986.

18. Wähälä K and Hase TA: Hydrogen transfer reduction of isoflavones.Heterocycles 28, 183–186, 1989.

19. Lamberton JA, Suares H, and Watson KG: Catalytic hydrogenation ofisoflavones, the preparation of (±)equol and related isoflavans. Aust JChem 31, 455–457, 1978.

20. Bannwart C, Adlercreutz H, Fotsis T, Wahala K, Hase T, et al.: Identi-fication of O-desmethylangolensin, a metabolite of daidzein, and of

Vol. 42, No. 2 231

Table 3. Quantification of Apoptosis Induced byGenistein and 2-de-O-DMAa,b

Apoptotic Cells, %

24 h 48 h 72 h

Breast

MCF-7Control 0 0 0.6 ± 0.8Genistein (37 µM) 0 0 0.7 ± 0.3Genistein (74 µM) 0 1.7 ± 0.6 5.3 ± 1.5*Genistein (148 µM) 5 ± 2.3* 31.7 ± 2.5* 52.2 ± 4.5*2-de-O-DMA (39 µM) 10 ± 1.7* 16 ± 2.3*2-de-O-DMA (78 µM) 31 ± 2.9* 32 ± 2.1*

MCF-10AControl 0 0 1.2 ± 2.3Genistein (37 µM) 0 0 0Genistein (74 µM) 0 0 0Genistein (148 µM) 0 0 1.3 ± 1.82-de-O-DMA (20 µM) 0 0 15 ± 2.7*2-de-O-DMA (39 µM) 0 2.1 ± 1.6 23 ± 4.5*2-de-O-DMA (78 µM) 12 ± 3.5* ND ND

Prostate

LNCaPControl 0 0Genistein (37 µM) 0 0Genistein (74 µM) 0 0Genistein (148 µM) 0 1.3 ± 1.82-de-O-DMA (39 µM) 0 02-de-O-DMA (78 µM) 0 02-de-O-DMA (156 µM) 0 0.6 ± 2.7

DU145Control 0 1.2 ± 1.5Genistein (37 µM) 0 0Genistein (74 µM) 0 0Genistein (148 µM) 0 16 ± 2.7*2-de-O-DMA (39 µM) 0 7.2 ± 2.2*2-de-O-DMA (78 µM) 0 10 ± 1.8*2-de-O-DMA (156 µM) 12.8 ± 3.2* ND

a: Values (means ± SE of �3 independent experiments) represent percent-age of apoptotic cells after they were treated with different isoflavonoidsat various times. Apoptotic cells were determined by TdT-mediateddUTP nick end labeling assay (see Materials and Methods). Numberof apoptotic cells among 1,000 nuclei was counted under a fluorescentmicroscope. ND, not done.

b: Statistical significance is as follows: *, increased significantly com-pared with control (P � 0.05).

metairesinol, one likely plant precursor of the animal lignan entero-lactone, in human urine. Finn Chem Lett 11, 120–125, 1984.

21. Salakka A and Wähälä K: Synthesis of �-methydeoxybenzoins. JChem Soc Perkin Trans 1, 2601–2604, 1999.

22. Malich G, Markovic B, and Winder C: The sensitivity and specificityof the MTS tetrazolium assay for detecting the in vitro cytotoxicity of20 chemicals using human cell lines. Toxicology 124, 179–192, 1997.

23. Marcelli M, Cunningham GR, Haidacher SJ, Padayatty SJ, Sturgis L,et al.: Caspase-7 is activated during lovastatin-induced apoptosis of theprostate cancer cell line LNCaP. Cancer Res 58, 76–83, 1998.

24. Bradshaw TD, Wrigley S, Shi DF, Schultz RJ, Paull KD, et al.: 2-(4-Aminophenyl)benzothiazoles: novel agents with selective profiles ofin vitro anti-tumour activity. Br J Cancer 77, 745–752, 1998.

25. Mutoh M, Lung FD, Long YQ, Roller PP, Sikorski RS, et al.: Ap21Waf1/Cip1 carboxyl-terminal peptide exhibited cyclin-dependentkinase-inhibitory activity and cytotoxicity when introduced into hu-man cells. Cancer Res 59, 3480–3488, 1999.

26. Zava DT and Duwe G: Estrogenic and antiproliferative properties ofgenistein and other isoflavonoids in human breast cancer cells in vitro.Nutr Cancer 27, 31–40, 1997.

27. Verma SP and Goldin BR: Effect of soy-derived isoflavonoids on theinduced growth of MCF-7 cells by estrogenic environmental chemi-cals. Nutr Cancer 30, 232–239, 1998.

28. Tham DM, Gardiner CD, and Haskell WL: Potential health benefits ofdietary phytoestrogens: a review of the clinical, epidemiological, andmechanistic evidence. J Clin Endocrinol Metab 83, 2223–2235, 1998.

29. Chang YC and Nair MG: Metabolites of daidzein and genistein andtheir biological activities. J Nat Prod 58, 1901–1905, 1995.

30. Peterson TG, Ji GP, Kirk M, Coward L, Falany CN, et al.: Metabolismof the isoflavones genistein and biochanin A in human breast cancercell lines. Am J Clin Nutr 68 Suppl, 1505S–1511S, 1998.

31. Franke AA, Custer LJ, Wang W, and Shi CY: HPLC analysis ofisoflavonoids and other phenolic agents from foods and from humanfluids. Proc Soc Exp Biol Med 217, 263–273, 1998.

32. Hodgson JM, Croft KD, and Puddey IB: Soybean isoflavonoids andtheir metabolic products inhibit in vitro lipoprotein oxidation in serum.J Nutr Biochem 7, 664–669, 1996.

33. Pagliacci MC, Smachia M, Migliorati SG, Grignani MF, Riccardi C, etal.: Growth-inhibitory effects of the natural phytoestrogen genistein inMCF-7 human breast cancer cells. Eur J Cancer 11, 1675–1682, 1994.

34. Peterson G and Barnes S: Genistein and biochanin A inhibit the growthof human prostate cancer cells but not epidermal growth factor recep-tor tyrosine autophosphorylation. Prostate 22, 335–345, 1993.

35. Davi JN, Sigh B, Bhuiyan M, and Sarkar FH: Genistein-inducedupregulation of p21WAF1, downregulation of cyclin B, and induction ofapoptosis in prostate cancer cells. Nutr Cancer 32, 123–131, 1998.

36. Lian F, Bhuiyan M, Li Y, Wall N, Kraut M, et al.: Genistein-inducedG2-M arrest, p21WAF1 upregulation, and apoptosis in a non-small-celllung cancer cell line. Nutr Cancer 31, 184–191, 1998.

37. McCabe MJ and Orrenius S: Genistein induces apoptosis in immaturehuman thymocytes by inhibiting topoisomerase II. Biochem BiophysRes Commun 194, 944–950, 1993.

38. Zhou JR, Mukherjee P, Gugger ET, Tanaka T, Blackburn GL, andClinton SK: Inhibition of murine bladder tumourigenesis by soy iso-flavones via alterations in the cell cycle, apoptosis, and angiogenesis.Cancer Res 58, 5231–5238, 1998.

39. Booth C, Hargreaves DF, Hadfield JA, McGown AT, and Potten CS:Isoflavones inhibit intestinal epithelial cell proliferation and induceapoptosis in vitro. Br J Cancer 80, 1550–1557, 1999.

40. Zhou JR, Gugger ET, Tanaka T, Guo Y, Blackburn GL, et al.: Soybeanphytochemicals inhibit the growth of transplantable human prostatecarcinoma and tumor angiogenesis in mice. J Nutr 129, 1628–1635,1999.

41. Barnes S: Effects of genistein on in vitro and in vivo models of cancer.J Nutr 125, 777S–783S, 1995.

42. Adlercreutz H, Markkanen H, and Watanabe S: Plasma concentrationof phytoestrogens in Japanese men. Lancet 342, 1209–1210, 1993.

232 Nutrition and Cancer 2002