Genistein Inhibits Both Estrogen and Growth Factor- Stimulated Proliferation of Human Breast Cancer Cells1’ 2

Vol. 7, 1345-1351, October 1996 Cell Growth & Differentiation 1345

Genistein Inhibits Both Estrogen and Growth Factor- stimulated Proliferation of Human Breast Cancer Cells1’ 2

Greg Peterson and Stephen Barnes3 (20 gig/mO did not decrease constitutive or EGF- Departments of Pharmacology and Toxicology [S. B.1 and Biochemistry induced tyrosine phosphorylation as determined by and Molecular Genetics [G. P., S. B.] and Comprehensive Cancer Western blothng with antiphosphotyrosine antibodies. Center Mass Spectrometry Shared Facility [S. B.], University of Alabama at Birmingham, Birmingham, Alabama 35294 These data suggest that although genistein inhibits the growth of breast cancer cells in culture, it does so without gross inhibition of PTK activity. Abstract Genistein is a naturally occurring dietary protein Introduction tyrosine kinase (P1K) inhIbitor that is hypothesized to Breast cancer is the major hormone-dependent cancer in be responsible for the lower rate of breast cancer American women. Breast cancer prevention and treatment is observed in Asian women consuming soy. Although the subject of a substantial effort to improve the health of genistein is a potent in vitro PTK inhibitor, its women in the United States. The age-adjusted death rates mechanism of action in vivo is not known. In vivo, from breast cancer are 2-8-fold lower in Asian countries than breast cancer growth is regulated by estrogens and in the United States and Western Europe (1). This difference peptide growth factors, such as epidermal growth in breast cancer incidence has been correlated with differ- factor (EGF), the receptor of which has intrinsic PTK ences in dietary patterns (2). activity. Therefore, genistein may block mammary In some Asian countries, women consume on the average epithelial cell growth by interfering with signal 20-50 times more soy products per capita than Americans transduction events stimulated by estradiol or growth (3, 4). Soy contains significant amounts (1-3 mg/g) of the factors. The effect of genistein, related isoflavones, isoflavone phytoestrogens genistein (5,7,4’-trihydroxyisofla- and other tyrosine kinase inhibitors on fetal bovine vone) and daidzein (7,4’-dihydroxyisoflavone). They are serum-, estradiol-, and EGF-stimulated cell growth and present as their -glucoside conjugates, many of which are signal transduction pathways was examined in five esterified (5-7). Several investigators have suggested that human breast cancer cell lines. Genistein inhibited the soy isoflavones, predominantly genistein, may play a role in growth of these cells by each of the growth stimuli reduction of breast cancer risk (8-10). In addition to the with ICee values ranging from 2.6 to over 20 g/ml. isoflavones, several classes of compounds in soy have been Growth inhibition by genistein was cytostatic and reported to exert anticancer activity (1 1), e.g. , protease in- reversible at ICee concentrations. Related isoflavones hibitors (12) and inositol phosphates (13). were less potent growth inhibitors than genistein, Several mechanisms have been proposed for the effects of whereas the synthetic PTK inhibitor tyrphostin A25 was genistein. Initially, genistein was considered to have estro- an equally potent growth inhibitor. The mechanism of gen agonist/antagonist activity (14) because of its structural genistein growth inhibition in human breast cancer similarity to the physiological estrogens such as E24 and cells did not depend on the presence of functional because of its estrogenic effects when administered to im- estrogen receptor signaling pathways or on inhibition mature mice (1 5). However, other mechanisms have been of EGF-receptor PTK activity. Furthermore, genistein proposed, including PTK inhibition, topoisomerase II inhibition, induction of differentiation, and inhibition of oxidation events (reviewed in Refs. 16-1 8). Most reports have focused Received 4/14/96; revised 7/20/96; accepted 7/26/96. on the PTK-inhibitory action of genistein. Genistein is a po- The costs of publication of this article were defrayed in part by the tent inhibitor of the P11< activity of the EGF-R in vitro with an payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to mdi- IC50 value of 0.7 pg/mI (1 9). The effect of genistein on PTK cate this fact. activity in whole cells, however, is not well documented. 1 These studies were supported in part by Grant 91 B58 from the American Institute for Cancer Research, Grant 5A01 CA-61668 from the National Breast cancer cell growth is partially regulated in vivo Cancer Institute, the Nebraska Soybean Promotion and Utilization Board, through the action of E2 (20). In this classical model, E2 binds and the United Soybean Board. to and activates the cytosolic ER. Activated ER then trans- 2 The data herein were presented in part at the 1994 Annual Meeting of the American Society for Cell Biology, San Francisco, and the 1995 American Association for Cancer Research annual meeting and have appeared in abstract form (Mol. Biol. Cell, 5 (Suppl.): 348a, 1994.) The data were also published as a requirement forthe partial fulfillment of a Ph.D. dissertation 4 The abbreviations used are: E2, estradiol; EGF, epidermal growth factor; from the University of Alabama at Birmingham for T. G. P. EGF-R, EGF receptor; FBS, fetal bovine serum; MEM, Eagle’s modified 3 To whom requests for reprints should be addressed, at Department of essential medium; MU, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazo- Pharmacology and Toxicology, University of Alabama at Birmingham, hum bromide; P1K, protein tyrosmne kinase; ER, estrogen receptor; P1-3-K, Birmingham, Alabama 35294. Phone: (205) 934-71 17; Fax: (205) 934- phosphoinositide-3-kinase; MAP-K, mitogen-activated protein kinase; 8240; E-mail: [email protected]. PVDF, polyvinylidene difluoride. 1346 Genistein Inhibits Breast Cancer Cell Growth

Table 1 Genistein does not interfere with the MIT assay 120 MCF-7 and T47D ER cells were plated as described and grown for 2 days in serum-containing medium. Genistein was added on day 2 with 100 fresh serum-containing medium as described. Cell viability was determined after 4 days by the MIT assay, and cell growth was expressed as a percentage of serum-stimulated cells receiving DMSO vehicle. The averages of at least three separate experiments (n > 12) are reported. 80 0 L . IC50 values, jig/mI Cell line 0 Trypan blue [H]Thymidine MiT assay c) 60 C- MCF-7 7.6 ± 0.84 8.7 ± 0.23 9.7 ± 0.61 C T47D ER 8.7 ± 0.41 10.6 ± 0.74 7.0 ± 1.2 40

20 locates to the nucleus and initiates specific transcriptional events via its interaction with estrogen response elements on DNA. E2 increases the levels of both growth factor receptors 0 and receptor ligands, such as EGF-R and transforming 0 1 5 75 10 15 20 growth factor-a (20, 21). This process establishes an auto- Concentration (.tg/m1) crine growth loop that leads to proliferative mammary epithelial cell growth (21). Establishment of this autocrine path- Fig. 1 . Genistein does not interfere with the mitochondrial reduction of way by other mechanisms (e.g. , overexpression, mutation, MU. Genistein (1-20 jig/mI) was added to serum-stimulated MCF-7 cells. and others) can lead to unregulated mammary epithelial cell Cell viability was determined by the MU assay after 4 days incubation with genistein (dark columns). Eight h prior to the addition of MIT, repli- growth in the absence of E2 (22). cate wells of MCF-7 cells were treated with equal concentrations of Because unregulated PTK activity in E2 and EGF signaling genistein to directly assess the effect of genistein on modulation of pathways is implicated in cellular transformation (23), com- mitochondrial MIT reduction (light columns). Cell growth is expressed as a percentage of serum-stimulated cells receiving DMSO vehicle; columns, pounds such as genistein that inhibit the PTK activity of mean; bars, SE. proteins involved in these pathways (both oncogenes and proto-oncogenes) are candidate chemotherapeutic agents (24-27). Genistein has been reported to inhibit the growth of of 7.6 and 8.7 g/ml by dye exclusion, respectively, and 8.7 a variety of cell lines in vitro stimulated by serum, E2, and and 1 0.6 ig/ml by [3H]thymidmne incorporation, respectively. growth factors (1 7, 1 8), but a systematic study of the effects These values are similar to the IC50 values of 9.4 and 7.0 of genistein on breast cancer cells has not been reported. p.g/ml for MCF-7 and T47D ER cells, respectively, obtained In the present study, the effect of genistein on FBS-, E2-, with the MIT assay (Table 1). Additionally, genistein at con- and EGF-stimulated growth was examined using five hu- centrations up to 20 p.g/ml did not alter MIT mitochondrial man breast cancer cell lines, MCF-7, T47D (both ER and reduction when compared to control cells in an 8-h incuba- ER), BT-20, and ZR-75-l in culture as model systems. To tion period (Fig. 1). Furthermore, neither biochanin A or daid- assess the specificity of genistein growth inhibition, the zein were found to interfere with the MIT assay at IC50 effect of the related isoflavones, biochanin A (5,7-dihy- concentrations (data not shown). Therefore, the MIT assay droxy-4’-methoxyisoflavone), daidzein, daidzin (J3-gluco- was valid for determining growth inhibition by genistein at side of daidzein), and genistmn (p-glucoside of genistein), concentrations under 20 g/ml in the systems studied. and the synthetic PTK inhibitor tyrphostin A25 was also Genistein Inhibits the Growth of Human Breast Cancer examined. To explore potential mechanisms of action, the Cells. Genistein inhibited the serum-stimulated growth of requirement of a functional ER system for growth inhibition the human breast cancer cell lines MCF-7, T47D ER, and and the effect of genistein on the EGF-stimulated tyrosine T47D ER in culture with IC50 values ranging from 7.0 to 9.4 phosphorylation of the EGF-R and other signal transduc- g/ml (Table 2). Growth of the human breast cancer cell lines tion proteins were examined. BT-20 (ER) and ZR-75-l (ER) was less susceptible to inhibition by genistein, with at least 2-fold higher IC50 values Results of 19.7 and 20 .tg/mI, respectively (Table 2). Validation of MTT Assay. It was reported that genistein Growth inhibition in all five cell lines was cytostatic at IC50 interferes with the MIT assay by modulating the activity of concentrations; genistein concentrations over 20 jig/mI were mitochondrial enzymes responsible for reducing the MIT required before cytotoxic effects were observed over a 4-day dye, thereby giving inaccurate data for genistein growth in- incubation period (Table 2). Similar results were reported hibition (28). To confirm the validity of the MIT assay to earlier for MDA-468 cells, a different subclone of MCF-7 monitor growth inhibition by genistein, IC50 values were de- cells, and MCF-7 cells overexpressing the multidrug resist- termined by trypan blue dye exclusion and [3H]thymidine ance protein (29). incorporation for serum-stimulated growth in MCF-7 and Growth inhibition of MCF-7 cells by genistein was com- T47D ER cells and compared to IC50 values obtained with pletely reversible at IC50 concentrations. After 2 days of the MiT assay (Table 1). Genistein inhibited serum-stimu- incubation in the presence of 0-20 g/ml genistein, cell lated growth of MCF-7 and T47D ER cells with IC50 values growth was comparable to that of control cells not exposed Cell Growth & Differentiation 1347

Table 2 Genistein inhibits the serum-smulated growth of human 800 breast cancer cells in culture 700 IC50 values, jig/mI Cell line Serum E2 EGF 600

MCF-7 9.7 ± 0.61 (>20r 2.7 ± 0.47 (9.0) 5.4 ± 0.25 (12.4) Genistein Removed T47D ER 7.0 ± 1.2 (>20) 2.3 ± 0.67 (13.4) 2.8 ± 0.29 (11.7) ‘C T47D ER 9.1 ± 0.16 (>20) ER 3.1 ± 0.13(11.6) ± 14.9 ± 0.61 (>20) BT-20 19.7 0.71 (>20) ER E 400’ ZR-75-1 >20 (>20) ND 13.9 ± 0.46 (>_

Human breast cancer cells were plated as described and grown for 2 days 300 in serum-containing medium. For E2- and EGF-stimulated growth experiments, cells were quiesced for 48 h in serum-free, phenol red-free me- 200 dium. Genistein was added on day 2 with serum stimulation or on day 4, 15 mm prior to E2 or EGF stimulation as described. Cell viability was determined after 4 days stimulation by the MIT assay, and cell growth 100 was expressed as a percentage of serum-, E2-, or EGF-stimulated cells receiving DMSO vehicle. The averages of at least three separate experi- 0 ments (n > 18) are reported. 0 1 2 3 4 5 6 a Genistein concentrations required before cytotoxic effects are observed are given in parentheses. b ND, not determined. Days Incubation

Fig. 2. Genistein growth inhibition is reversible at cytostatic concentrations. Genistein was added at 0 (U), 10 (#{149}),20 (A), and 50 (0) jig/mI to to genistein after 4 days of growth in genistein-free medium serum-stimulated MCF-7 cells as described and incubated for 4 days. Cell viability was determined by the MIT assay, and cell growth is expressed (Fig. 2). However, cytotoxic concentrations of genistein (over as a percentage (mean ± SE) of serum-stimulated cells receiving DMSO 20 g/ml) caused nonreversible effects on growth after 48 h vehicle. Data points, average of three separate experiments (1 8); hera, SE (within the symbols). incubation (Fig. 2). Similar data were obtained for T47D cells (data not shown). Genistein inhibited E2-stimulated growth of MCF-7 and Effects of Genistein on EGF-stimulated Signal Trans- T47D ER cells with IC50 values of 2.7 and 2.3 p.g/ml, reduction Pathways. Genistein at 10 tg/ml, a concentration spectively (Table 2). Genistein inhibited E2-stimulated growth twice the IC50 value for EGF-stimulated growth, did not in both cell lines in a biphasic manner, with modest growth significantly decrease EGF-R tyrosine phosphorylation in stimulation noted at genistein concentrations less than 10 MGF-7 or T47D cells (8 and 1 2% decrease, respectively, ng/ml (data not shown). Genistein at (l 0 ng/ml) did not compared with DMSO vehicle controls; Fig. 3). A genistein increase cell growth in the presence of 0.1 nM E2 (data not concentration of SO tg/ml (a concentration that caused shown). Genistein inhibited EGF-stimulated growth of each cell death) was required before significant inhibition of of the breast cancer cell lines with IC50 values ranging from EGF-R tyrosine phosphorylation was observed (62 and 2.8 to 14.9 jg/ml (Table 2). As noted for serum-stimulated 46% decrease for MGF-7 and T47D cells, respectively; Fig. 3). In contrast, the specific EGF-R PTK inhibitor tyr- growth, ZR-75-1 and BT-20 cells were 2-4-fold less sensi- phostin A25 significantly inhibited EGF-R tyrosine phos- tive to the inhibitory effects of genistein than MGF-7, T47D phorylation when added to MGF-7 and T47D cells at 6 ER, and T47D ER cells. Genistein inhibited both E2- and g/ml, a concentration near its IC50 value for cell growth EGF-stimulated growth in a cytostatic manner at IC50 con- (Fig. 3). Biochanin A and daidzein also failed to inhibit centrations (Table 2). EGF-R tyrosine phosphorylation at concentrations up to Genistein Specifically Inhibits Human Breast Cancer So j.g/ml in MGF-7 and T47D cells (data not shown). These Cell Growth. The growth-inhibitory effect of genistein data strongly suggest that EGF-R is not the intracellular against breast cancer cells was not due to general cytotoxic target of genistein at cytostatic (IC50) genistein concen- effects of this isoflavone. The related isoflavones daidzein, trations. Similar results were reported for human prostate daidzin, genistin, and biochanin A were generally less potent cancer cells (30). inhibitors of serum-, E2-, and EGF-stimulated growth in hu- EGF-stimulated signal transduction proteins represent an- man breast cancer cells, with IC50 values ranging from 7 to other class of targets for genistein action (24). The activities over 100 ig/ml (Table 3). As with genistein, inhibition of of several signal transduction proteins involved in early EGF- breast cancer cell growth by these isoflavones was cyto- stimulated signal transduction events are regulated by tyrosine static at IC50 concentrations. phosphorylation. Additionally, other cellular proteins undergo The synthetic P1K inhibitor tyrphostin A25 inhibited tyrosine phosphorylation; the effect of phosphorylation remains EGF-stimulated growth of breast cancer cells at similar unknown. These groups of proteins include phospholipase G’y, concentrations to genistein, with IC50 values ranging from P1-3-K, Raf, rasGAP, mammalian son of sevenless, and MAP-K 7.5 to 13.5 g/ml (Table 3). Tyrphostin A25 was not an (25). Therefore, genistein could inhibit cell growth by modulating effective inhibitor of serum-stimulated growth of breast the tyrosine phosphorylation of any or one of these proteins. cancer cells. Similar data was obtained for tyrphostin A47 Genistein at concentrations up to 50 jg/ml did not inhibit (data not shown). the tyrosine phosphorylation of MAP-K or P1-3-K in MGF-7 1348 Genistein Inhibits Breast Cancer Cell Growth

Table 3 Genistein specifically inhibits breast cancer cell growth Growth inhibition of human breast cancer cells by various isoflavones and tyrphostin A25 was carried out as described in the legend to Table 2. Cell viability was determined after 4 days stimulation by the MIT assay and cell growth was expressed as a percentage of serum-, E2-, or EGF-stimulated growth. The averages of at least three separate experiments (n > 1 8) are reported.

. IC50 values, jig/mI Cell line Daidzein Daidzin Genistin Biochanin A Tyrphostin A25

Serum stimulation MCF-7 20.0 ± 0.43 >100 >100 7.0 ± 0.79 >20 T47D ER 21.5 ± 0.67 >100 >100 12.8 ± 0.38 >20 ZR-75-1 >50 ND ND >20 >20 BT-20 >50 ND ND >20 >20

EGF stimulation MCF-7 10.2 ± 0.48 >50 >50 8.8 ± 0.24 7.5 ± 0.61 T47D ER 21.5 ± 0.67 >50 >50 9.0 ± 0.25 9.2 ± 0.42 ZR-75-1 >50 ND ND >20 13.5 ± 0.81 BT-20 >50 ND ND >20 9.8 ± 0.34

Estrogen stimulation MCF-7 25.0 ± 0.36 ND ND 13.0 ± 0.43 ND

a ND, not determined.

120’ A - - - - ..#{248}-85 100 . #{149} S5

A B C 0 E F

20 B 0 10 hi1l1iii11. ..- 44 Genistein Concentration (.tg/mI)

Fig. 3. Genistein does not inhibit EGF-R tyrosine phosphorylation. MCF-7 (light columns) and T47D ER cells (dark columns) were plated in A B C D E F 1 00-mm plates and grown to 80% confluence in FBS-containing medium. Cells were quiesced for 48 h in serum-free MEM + ITS-BSA. Medium was Fig. 4. Genistein does not inhibit the tyrosine phosphorylation of P1-3-K removed and replaced with fresh MEM + ITS-BSA. Genistein was added or MAP-K. MCF-7 cells were plated in 100-mm plates and grown to 80% in 1 00% DMSO (0-50 jig/mI; DMSO final concentration, 1 %, v/v) 1 5 mm confluence in FBS-containing medium. Cells were quiesced for 48 h in prior to EGF stimulation. Cells were stimulated with 50 ng/ml EGF for 1 serum-free MEM + ITS-BSA. Medium was removed and replaced with mm and lysed with NP4O buffer. EGF-R was immunoprecipitated from cell fresh MEM + ITS-BSA. Genistein was added as described in the legend lysates (100 jig) with B1 D8 monoclonal anti-EGF-R antibody (1 jig). to Fig. 3. Cells were stimulated with 50 ng/mI EGF for 1 mm for P1-3-K and Proteins were separated by SDS-PAGE and transferred to PVDF. EGF-R for 15 mm for MAP-K and lysed with NP4O buffer. Cell lysates (100 jig) was detected with 4G1 0 monoclonal antiphosphotyrosmne antibody and were immunoprecipitated with 4G10 monoclonal antiphosphotyrosine an- visualized with alkaline phosphatase. Proteins were quantitated using the tibody (0.5 jig). Proteins were separated by SDS-PAGE and transferred to Bio-Rad molecular imaging system. Results are the average of four ex- PVDF. Proteins were detected with anti-Pl-3-K (1:10,000; A) or anti- penments and are presented as a percentage (mean ± standard error) of MAP-K antibody (1 :1 0,000; B) and visualized with alkaline phosphatase. EGF-stimulated EGF-R tyrosine phosphorylation in control cells receiving LaneA, DMSO; Lane B, EGF + DMSO; Lane C, EGF + genistein (1 jig/mI); DMSO vehicle. Lane 0, EGF + genistein (5 jig/mI); Lane E, EGF + genistein (1 0 jig/mI); Lane F, EGF + genistein (50 jig/mI).

cells (Fig. 4). However, these proteins were constitutively tyrosine phosphorylated and did not show increased tyrosine Furthermore, genistein did not consistently decrease the phosphorylation in response to EGF. In similar experiments, constitutive or EGF-stimulated tyrosine phosphorylation of genistein at concentrations up to 20 g/ml did not inhibit the any protein in MCF-7 or T47D cells as observed by Western tyrosine phosphorylation of phospholipase C’y, mammalian blotting with an antiphosphotyrosine antibody (data not son of sevenless, or Raf in MCF-7 cells (data not shown). shown). Cell Growth & Differentiation 1349

Discussion vone structure. The related isoflavone daidzein, which differs Genistein inhibited the growth of the five human breast cancer only in the absence of a hydroxyl group at the 5 position, was cell lines stimulated by FBS, E, and EGF. lGso values for a much weaker inhibitor of serum-, E2-, and EGF-stimulated growth inhibition of MCF-7, T47D ER, and T47D ER cells by cell growth. The exception to this trend was biochanin A, each growth stimuli were similar. However, BT-20 and ZR-75-1 which was a slightly better inhibitor of serum-stimulated cells were more resistant to the growth inhibition of genistein. MGF-7 cell growth than genistein. However, biochanin A is This resistance was unrelated to ER status, as BT-20 cells are metabolized to genistein in all of the breast cancer cell lines ER and ZR-75-1 cells are ER. ZR-75-1 cells were previously tested (37)#{149}5 reported to have similarly high IGso values for several flavone The effect of genistein on EGF-stimulated tyrosine phos- and isoflavone compounds (31). A potential rationale for resist- phorylation was also examined. The first step in EGF-stim- ance of BT-20 and ZR-75-1 cells to genistein growth inhibition ulated signal transduction is activation of the EGF-R itself has come from preliminary experiments in which genistein is through autophosphorylation of the cytoplasmic tail (23). If extensively metabolized in these cells.5 genistein inhibited autophosphorylation, then receptor acti- Previous investigators have suggested that genistein may vation, signal transduction, and subsequent mitogenesis inhibit cell growth by an antiestrogenic mechanism through would be blocked. Because genistein was shown to inhibit competition with E2 for occupancy of the ER (8, 9, 14). We EGF-R autophosphorylation in vitro through competitive in- previously reported that genistein inhibited the serum-stim- hibition of ATP (1 9), blocking receptor activation seemed to ulated growth of MGF-7 (ER) and MDA-468 (ER-) cells with provide a plausible mechanism for genistein action. similar IC50 values (29). However, MDA-468 cells overex- However, genistein did not block EGF-R tyrosine phospho- press EGF-R, which could make these cells hypersensitive to rylation induced by EGF in MGF-7 orT47D ER cells. In parallel genistein growth inhibition. To minimize the differences be- experiments, the specific EGF-R P11< inhibitor, tyrphostin A25 tween ER and ER cells, two closely related cell lines, T47D (38), blocked EGF-R tyrosine phosphorylation in MGF-7 and ER and a T47D ER- subclone, were used. Once again, T47D ER cells. These data indicated that activation of EGF-R genistein inhibited the growth of ER and ER- T47D cells was not inhibited by genistein at concentrations that inhibited with similar IC50 values, suggesting that the ER is not nec- cell growth by 50%. We conclude, therefore, that EGF-R was essary for growth inhibition by genistein. competent to transmit mitogenic signals in the presence of Genistein inhibited E2-stimulated growth of MGF-7 and genistein. Several other investigators have reported similar data T47D ER cells with lG50 values below 3 pg/mI. However, at (reviewed in Ref. 17). In addition, genistein at IGso concentra- low concentrations (1 0 ng/mI), genistein caused a small in- tions did not decrease overall levels of EGF-R phosphorylation crease in cell growth. A larger growth stimulus of MGF-7 cells assessed using cell labeling with 32P.6 by genistein was reported under cell culture conditions In addition, genistein did not inhibit the tyrosine phospho- where every effort has been made to remove estrogens (32); rylation of MAP-K or P1-3-K. These two proteins are on however, genistein inhibited cell growth at concentrations divergent signaling pathways and rely on tyrosine phospho- greater than 1 pg/mI (32). Because there are circulating rylation for kinase activity and intracellular localization, re- estrogens both pre- and postmenopausally, it is unlikely that spectively. Furthermore, genistein did not detectably de- levels of genistein produced by dietary intake are estrogenic. crease the tyrosine phosphorylation of any protein in whole- Genistein could block E-mediated events through several cell lysates of MGF-7 and T47D cells stimulated with EGF. mechanisms. Serine and tyrosine phosphorylations are neces- These data failed to provide evidence that genistein sary for full E2-induced transcnption (33) and receptor dimer- growth inhibition involves blocking the tyrosine phosphoryl- ization (34), respectively. Additionally, signal transduction path- ation of EGF-R or other proteins involved in EGF signaling. ways stimulated by protein kinase A and G, neurotransmitters, This is contrary to its reported role as a P1K inhibitor in vitro and PTh-linked growth factor receptors can modulate the (1 9). Several explanations could account for the difference phosphorylation and activation of the ER. Genistein could mod- between the in vitro data and cell culture experiments. ulate E2 action (indirectly or directly) by altering these phospho- Genistein may selectively inhibit single sites of tyrosine phos- rylation patterns and changing the activity of the E2-receptor phorylation of the EGF-R or other signal transduction pro- (35). Alternatively, genistein may inhibit ER-independent tyro- teins, which may not be observed by assessing the overall sine phosphorylation events stimulated by E2 (36). levels of tyrosine phosphorylation. These changes may be Estrogens stimulate cell growth by inducing the production very important to the signaling process, as some src-homol- of autocrine/paracrine growth regulators, such as EGF and ogy 2-containing proteins bind only to specific tyrosine transforming growth factor-a, which directly induce cell pro- phosphorylated residues (23). Inhibition of a subset of tyro- liferation (1 9). Genistein may block E2-stimulated growth by sine phosphorylation events would allow specificity to genis- inhibiting a step in this autocrine/paracrine growth loop. In tein action. Tryptic mapping methods and in vitro kinase support of this hypothesis, the IC50 values for E2-stimulated assays could provide answers to these questions. Alterna- growth are similar to those for EGF-stimulated growth. tively, cellular metabolism or sequestration may reduce the Growth inhibition of breast cancer cell lines by genistein intracellular genistein concentrations to levels below that was not the result of nonspecific cytotoxicity related to isofla- needed to inhibit tyrosine phosphorylation in whole cells.

5 T. G. Peterson, G-P. Ji, M. Kirk, and S. Barnes, unpublished data. 6 T. G. Peterson and S. Barnes, unpublished data. 1350 Genistein Inhibits Breast Cancer Cell Growth

In summary, genistein inhibited the serum-, E2-, and EGF- week before plating. Cells were maintained in MEM with 2.5% E2-free FBS stimulated growth of human breast cancer cells in a cyto- for the duration of the experiment. Isoflavones were added as described static and reversible manner, with lG50 values from 2 to >20 previously 15 mm prior to E2 addition. E2 was added in 1 00% ethanol (final concentration, 0.1 % v/v) to give a concentration of 0.1 nri. Cells were p.g/mI. Growth inhibition was specific, as closely related incubated at 37CC with medium changes every 2 days. After 6 days, isoflavones were equal or less potent growth inhibitors. The growth was assayed by the MiT assay as described below. mechanism of growth inhibition did not depend on the pres- MTTAssay. The lC values for genistein were determined by the MIT ence of a functional ER system or involve gross inhibition of assay as described previously (29, 30). Briefly, the MIT assay is a color- imetric assay that is based on the ability of living but not dead cells to tyrosine phosphorylation. reduce a tetrazolium-based compound to a blue formazan product (40). The formazan crystals are solubilized in DMSO, and the absorbance is measured at 540 nm. The absorbance at 540 nm is proportional to the Materials and Methods number of viable cells. The lC values obtained with the MIT assay were Materials. A genistein concentrate (40-50% genistemn by weight) was a compared with the lC values obtained by counting viable cells using

gift of Protein Technologies International (St. Louis, MO). FBS, tissue trypan blue dye exclusion and by tritiated thymidine incorporation into culture media, supplements, and antibiotics were obtained from Life DNA. Technologies, Inc. (Gaithersburg, MD) or Upstate Biotechnology (Lake TrItiated Thymidine Incorporation. MCF-7 cells were plated in 24- Placid, NY). MiT and biochanin A were from Sigma Chemical Co. (St. well plates at 5 x iO cells/well and grown to 100% confluency. Cells Louis, MO). Prestained molecular weight standards and SDS-PAGE were quiesced for 2 days in serum-free MEM plus ITS-BSA. The medium chemicals were from Bio-Aad (Richmond, CA). Immobilon PVDF mem- was replaced by MEM plus 5% FBS, and genistein was added to cells in branes were from Millipore (Bedford, MA). Alkaline phosphatase sub- as described above for serum stimulation. Cells were incubated for 8 h, strates (5-bromo-4-chloro-3-indolyl phosphate and nitro blue tetrazolium) and [HJthymidmne was added at a concentration of 0.2 jiCVwell. lncuba- were from Research Organics (Cleveland, OH). Pansorbin and Sansorbin tion was continued for 12-24 h. Cells were washed three times with cells were from CalBiochem (La Jolla, CA). 4G10 monoclonal antiphos- ice-cold PBS (pH 7.6) and twice with ice-cold 70% methanol. Cells were photyrosmne antibody, anti-PI-3-K polyclonal antibody, and anti-MAP-K incubated with 750 jil of 0.1 N NaOH at 37CC for 30 mm; 100-jil aliquots polyclonal antibody were from Upstate Biotechnology (Lake Placid, NY). were taken, and their radioactivity was determined by liquid scintillation B1D8 monoclonal anti-EGF-R antibody was the gift of Dr. Jeff Kudlow counting. Values were expressed as a percentage of serum-stimulated (University of Alabama at Birmingham). Biotmnylated goat antimouse lgG, control cells treated with DMSO. goat antirabbit IgG, and streptavidin-alkalmne phosphatase conjugate were Cell Lysis. MCF-7 and T47D ER cells were grown in 100-mm dishes from Klrkegaard and Perry Laboratories (Gaithersburg, MD). All other to 80-100% confluence in FBS-containing medium. Cells were then qui- materials were of the highest quality available. esced for 2 days as described for the EGF-stimulated growth assays. Cell Culture. MCF-7, T47D ER, and T47D ER (39)cells were the gift Isoflavones were added as described 15 mm prior to EGF addition. EGF of Dr. Craig Jordan (Northwestern University, Chicago, IL). Additional was added at 50 ng/ml, and cells were incubated at 37CC for the indicated MCF-7 cells, BT-20 cells, and ZR-75-1 cells were obtained from the times. The medium was aspirated thoroughly, and cells were lysed im- American Type Culture Collection (Rockville, MD). MCF-7, T47D EA, and mediately in NP4O buffer (1 % NP4O, 150 mi NaCI, 50 mi Tris-HCI, pH T47D ER- cells were maintained in Eagles MEM, buffered with HEPES 7.5, 1 jig/mI aprotinin, 1 jig/mI leupeptin, 1 jig/mI pepstatin, 1 mi phen- with 5% (v/v) FBS and antibiotic (100 units/mI penicillin and 100 jig/mI ylmethylsulfonyl fluoride, 1 mM sodium pyrophosphate, 5 m EDIA and 1 streptomycmn). BT-20 and ZA-75-1 cells were maintained in RPMI plus 7% mM Na3VOJ. Subsequent steps were performed at 4#{176}C.The plates were (v/v) FBS and antibiotics as above. All cells were cultured as monolayers rocked for 20 mm, and cell lysates were scraped into microcentrifuge (passed every 6-8 days) in a 95% air:5% CO2 water-saturated atmo- tubes. Supematant was passed through a 21 gauge needle to shear DNA sphere. and centrifuged for 10 mm at 10,000 x g to remove insoluble material. Isoflavene Isolation. Genistein and daidzein were isolated from soy Protein concentrations were determined by the method of Lowry et al. molasses as described by Peterson and Barnes (29). In brief, genistin and (41). Samples were placed at 4#{176}Cuntil further use (samples could be daidzin were purified by fractional crystallization and then subjected to stored at 4#{176}Cfor up to 3 weeks). acid hydrolysis. Genistein and daidzein were recovered by ether extrac- Immunoprecipitation. Samples (100 jig protein) were preabsorbed tion and recrystallized twice from 80% aqueous ethanol. Both were >98% with 50 jil Pansorbin cells bound to 200 jil mouse serum. After 30 mm pure as jud9ed by their UV absorbance at 262 nm and by reverse-phase incubation at room temperature, cells were centrifuged for 5 mm at high-performance liquid chromatography. 10,000 x g at 4#{176}C,and supernatants were removed and used for the Serum StImulatIon. MCF-7, T47D ER and ER, BT-20, and ZA-75-1 immunoprecipitation. Lysates were boiled for 5 mm and cooled on ice. cells were plated into 96-welltissue culture clusters at densities of 2 x 10 B1 D8 monoclonal anti-EGF-R antibody (1 jig) or 4G10 monoclonal an- to 4 x 1 o cells/well in 100 jil of FBS-supplemented medium. After tiphosphotyrosine antibody (0.5 jig) was added and samples were incu- plating, the cells were allowed to attach for 2 days in FBS-supplernented bated for 4 h at 4#{176}Cwith gentle rotation. Pansorbin cells were added (25 medium. On day 2, medium was aspirated and 199 jil fresh FBS-supple- i.i.I) to collect immune complexes. Immune complexes were washed three mented medium were added. Isoflavones were added in 100% DMSO (1 .0 times in lysis buffer. SDS-PAGE buffer was added, and the immune pA volume, 0.5% v/v DMSO) and incubated for 4 days; control wells complexes were boiled for 5 mm and spun at 1 0,000 x g for another 5 received 1 .0 jil DMSO. After 4 days, plates were developed by the MiT mm. The supematants were separated by SDS-PAGE as recommended assay as described below. by Bio-Rad, with the exception that no SDS was added in the running or EGF Stimulation. MCF-7, T47D EA and EA, BT-20, and ZR-75-1 stacking gels. cells were plated at 5 x 1O cells/well in 96-well plates. Cells were Western Blotting. Proteins were transferred to a PVDF membrane at incubated for 2 days as above. On day 2, cells were quiesced in MEM or 75 V constant voltage for 1 h. Completion of transfer was judged by APMI supplemented with ITS-BSA medium (5 jig/mI insulin, transferrmn, 5 Coomassie blue staining of the gel after transfer. PVDF membrane was

ng/mI selenium , and 500 jig/mI BSA) plus antibiotics as above. Quies- blocked for 30 mm at room temperature with 5% powdered milk in TTBS cence was continued for 2 days at 37#{176}C.Cell medium was replaced with (20 m Iris, pH 7.6, 0.5% Tween-20, 0.45% NaCI). Blots were then fresh quiescence medium, and isoflavones were added as described incubated with primary antibodies (4Gb, 1 :5000; anti-MAP-K, 1:10,000; above. After 1 5 mm incubation, EGF was added at a concentration of 50 and anti-PI-3-K, 1 :10,000) in TIBS + 0.5% BSA and 0.01 % NaN3 for 4 h ng/mI for MCF-7, T47D ER, and T47D ER- and at 75 ng/mI for BT-20 and at room temperature or overnight. Blots were washed 3 times (5 mm) with ZR-75-1 cells. The optimal amounts of EGF were determined to give TTBS and incubated with secondary antibody (1/3000 dilution of biotyni- maximal growth over a 4-day incubation. Cells were incubated for 4 days lated goat antimouse IgG or goat antirabbit IgG) for 30 mm at room at 37#{176}Cand growth was assayed by the MIT assay as described below. temperature. Blots were washed 3 times as above and incubated with Estrogen Stimulation. MCF-7 and T47D ER cells were plated as streptavidin-alkaline phosphatase conjugate (1/3000 dillution) for 30 mm described for growth factor stimulation, with the exception that cells were at room temperature. Blots were washed 5 times as above and developed grown in phenol red-free, charcoal-stripped FBS (E2-free) medium for 1 by alkaline phosphatase (42). Cell Growth & Differentiation 1351

References 23. Cantley, L C., Auger, K. A., Carpenter, C., Duckworth, B., Kapeller, A., and Soltoff, S. Oncogenes and signal transduction. Cell, 64: 281-302, 1 . Parker, S. L, Tong, T., Bolden, S., and Wingo, P. A. Cancer Statistics. CA Cancer J. Clin., 65: 5-27, 1996. 1991. 2. Dunn, J. E. Cancer epidemiology in populations ofthe United States- 24. Powis, G. Recent advances in the development of anticancer drugs with emphasis on Hawaii and California-and Japan. Cancer Res., 35: that act against signalling pathways. Tumori, 80: 69-87, 1994. 3240-3245, 1975. 25. Levitski, A. Signal transduction therapy. Eur. J. Biochem., 226: 1-13, 3. Messina, M. J., Persky, V., Setchell, K. D. A., and Barnes, S. Soy intake 1994. and cancer risk: a review of the in vitro and in vivo data. Nutr. Cancer, 21: 26. Arnold, S. F., and Notides, A. C. An antiestrogen: a phosphotyrosyl 113-131, 1994. peptide that blocks dimerization of the human estrogen receptor. Proc. 4. Jones, A. E., Price, K. A., and Fenwick, G. A. Development of a high NatI. Acad. Sci. USA, 92: 7475-7479, 1995. performance liquid chromatographic method for the analysis of isofla- 27. Levitski, A. Tyrphostmns: tyrosine kinase blockers as novel antiprolif- vones. J. Sci. Food Agric., 46: 357-364, 1989. erative agents and dissectors of signal transduction. FASEB J., 6: 3275- 5. Barnes, S., Kirk, M., and Coward, L Isoflavones and their conjugates 3282, 1992. in soy foods; extraction conditions and analysis by HPLC-mass spectrometry. J. Agric. Food Chem., 42: 2466-2474, 1994. 28. Pagliacci, M. C., Spinozzi, F., Migliorati, G., Fumi, G., Smacchia, M., Grignani, F., Riccardi, C., and Nicoletti, I. Genistein inhibits tumor growth 6. Kudou, S., Fleury, Y., Welti, D., Magnolato, D., Uchida, T., Kitamura, K., in vitro but enhances mitochondrial reduction of tetrazolium salts: a further and Okuba, K. Malonyl isoflavone glycosides in soybean seeds (Glycmne pitfall of the MIT assay for evaluating cell growth and survival. Eur. J. max MERRILL). Agric Biol. Chem., 55: 2227-2233, 1991. Cancer, 29A: 1573-1577, 1993. 7. Farmakalidis, E., and Murphy, P. A. Isolation of 6”-o-acetyldaidzemn 29. Peterson, G., and Barnes, S. Genistein inhibition of the growth of and 6”-o-acetylgenistemn from toasted defatted soy flakes. J. Agric. Food human breast cancer cells: independence from estrogen receptors and Chem., 33: 385-389, 1985. multi-drug resistance gene product. Biochem. Biophys. Res. Commun., 8. Setchell, K. D. A., Borriello, S. P., Hulme, P., Kirk, D. N., and Axelson, 179: 661-667, 1991. M. Nonestradiol estrogens of dietary origin: possible roles in hormone dependent diseases. Am. J. Clin. Nutr., 40: 569-578, 1984. 30. Peterson, G., and Barnes, S. Genistein and biochanin A inhibit the growth of human prostate cancer cells in culture but not epidermal growth 9. Barnes, S., Grubbs, C., Setchell, K. D. A., and Carlson, J. Soybeans factor receptor tyrosine phosphorylation. Prostate, 22: 335-345, 1993. inhibit mammary tumors in models of breast cancer. In: M. W. Pariza, H-U. Aeschbacher, J. S. Felton, and S. Sato (eds.), Mutagens and Carcinogens 31 . Hirano, I., Oka, K., and Akiba, M. Antiproliferative effects of synthetic in the Diet, pp. 239-253. New York: Wiley-Uss, 1990. and naturally occurring flavonoids on tumor cells of the human breast 10. Adlercreutz, H. Western diet and Western disease: some hormonal carcinoma cell line. ZR-75-1 . Res. Commun. Chem. Pathol. Pharmacol., 64: and biochemical mechanisms of action. Scand. J. Clin. Lab. Invest., 50: 69-78, 1989. 3-23, 1990. 32. Martin, P. M., Horwltz, K. B., Ryan, D. S., and McGuire, W. L Phy- toestrogen interaction with estrogen receptors in human breast cancer 11 . Messina, M., and Barnes, S. Workshop report from the Division of Cancer Etiology, National Cancer Institute, NIH. The role of soy in reduc- cells. Endocrinology, 103: 1860-1867, 1978. ing the risk of certain cancers. J. NatI. Cancer Inst., 83: 541-546, 1991. 33. Le Goff, P., Montano, M. M., Schodin, D. J., and Katzenellenbogen, B. 12. Kennedy, A. A. The evidence for soybean products as cancer pre- S. Phosphorylation of the human estrogen receptor. Identification of hor- ventive agents. J. Nutr., 125 (Suppl.): 733s-743s, 1995. mone-regulated sites and examination of their influence on transcriptional 13. Shamsuddin, A. M. Inositol phosphates have novel anti-cancer func- activity. J. Biol. Chem., 269: 4458-4466, 1994. tion. J. Nub’., 125 (Suppl.): 725s-732s, 1995. 34. Arnold, S. F., Obourn, J. D., Jaffe, H., and Notides, A. C. Phospho- 14. Barnes, S., Peterson, G., Grubbs, C., and Setchell, K. Potential role of rylation of the human estrogen receptor on tyrosine 537 in vivo and by src dietary isoflavones in the prevention of cancer. In: M. M. Jacobs (ed), Diet family tyrosine kinases in vitro. Mol. Endocrinol, 9: 24-33, 1995. and Cancer: Markers, Prevention, and Treatment, pp. 1 35-1 47. New York: 35. Katzenellenbogen, B. S., Montano, M. M., Le Goff, P., Schodin, D. J., Plenum Press, 1994. Kraus, W. L, Bhardwaj, B., and Fujimoto, N. Antiestrogens: mechanisms 15. Cheng, E., Story, C. D., Yoder, L, Hale, W. H., and Burroughs, W. and actions in target cells. J. Steroid Biochem. Molec. Biol., 53: 387-393, Estrogenic activity of isoflavone derivatives extracted and prepared from 1995. soybean oil meal. Science (Washington DC), 164-165, 1956. 118: 36. Migliaccio, A., Pagano, M., and Auricchio, F. Immediate and transient 16. Barnes, S., and Peterson, I. G. Biochemical targets of the isoflavone stimulation of protein tyrosine phosphorylation by estrogens in MCF-7 genistein in tumor cell lines. Proc. Soc. Exp. Bid. Med., 208: 103-108, cells. Oncogene, 8: 2183-2189, 1994. 1995. 37. Peterson, I. G., Coward, L, Kirk, M., Falany, C. N., and Barnes, S. 1 7. Peterson, G. Evaluation of the biochemical targets of genistein in The role of metabolism in mammary epithelial cell growth inhibition by the tumor cells. J. Nutr., 125 (Suppl.): 784s-789s, 1995. isoflavones genistemn and biochanin A. Carcmnogenesis (Lond.), in press, 1 8. Constantinou, A., and Huberman, E. Genistein as an inducer of tumor 1996. cell differentiation: possible mechanisms of action. Proc. Soc. Exp. Bid. 38. Levitski, A., and Gilion, C. Tyrphostmns as molecular tools and poten- Med., 125: 109-115, 1995. tial antiprollferative drugs. Trends Pharmacol. Scm., 127: 171-176, 1991. 19. Akiyama, I., Ishida, J., Nakagawa, S., Ogawara, H., Watanabe, S-I., 39. Murphy, C. 5, Meismer, L F., Wu, S. K., and Jordan, V. C. Short and Itoh, N., Shibuya, M., and Fukami, Y. Genistein, a specific inhibitor of long term estrogen deprivation of T47D human breast cancer cells in tyrosmne-specific protein kinases. J. Biol. Chem., 262: 5592-5595, 1987. culture. Eur. J. Cancer Clin. Oncol., 25: 1777-1788, 1982. 20. Dickson, A. B., and Uppman, M. E. Estrogenic regulation of growth and polypeptide growth factor secretion in human breast carcinoma. 40. Carmichael, J., DeGraff, W. G., Gazdar, A. F., Minna, J. D., and Endocr. Rev., 8: 29-43, 1987. Mitchell, J. B. Evaluation of a tetrazolium-based semiautomated colori- metric assay: assessment of chemosensitivity testing. Cancer Res., 47: 21 . Chrysogelos, S. A., Dickson, A. B. EGF receptor expression, regula- 936-942, 1987. tion, and function in breast cancer. Breast Can. Aes. Treat., 29: 29-40, 1994. 41 . Lowry, 0. H., Rosebrough, N. J., Farr, A. L, and Randall, A. J. Protein measurements with folin phenol reagent. J. Biol. Chem., 193: 265-275, 22. Pietras, A. J., Arboleda, J., Reese, D. M., Wongvipat, N., Pegram, M. D., Ramos, L, Gorman, C. M., Parker, M. G., Sliwkowski, M. X., and 1951. Slamon, D. J. HER-2 tyrosine kinase pathway targets estrogen receptor 42. Harlow, E., and Lane, 0. Cell staining.In: E. Harlow and D. Lane (eds.), and promotes hormone-independent growth in human breast cancer Antibodies: A Laboratory Manual, pp. 407-408. Cold Spring Harbor, NY: cells. Oncogene, 10: 2435-2446, 1995. Cold Spring Harbor Laboratory, 1988.