CANCER RESEARCH 57. 2916-2921. July 15. 19971

ABSTRACT

Consumption of a plant-based diet can prevent the development andprogression of chronic diseases associated with extensive neovascularization, including solid malignant tumors. In previous studies, we have shownthat the plant-derived isoflavonoid genistein is a potent inhibitor of cellproliferation and in vitroangiogenesis. In the present study, we report thatcertain structurally related flavonoids are more potent inhibitors thangenistein. Indeed, 3-hydroxyflavone, 3',4'-dihydroxyflavone, 2',3'-dihydroxyflavone, fisetin, apigenin, and luteolin inhibited the proliferation ofnormal and tumor cells, as well as in vitro angiogenesis, at half-maximalconcentrations in the low micromolar range. We have previously demonstrated that genistein concentrations in the urine of subjects consuming aplant-based diet Is 30-fold higher than in subjects consuming a traditionalWestern diet. The wider distribution and the more abundant presence offlavonoids in the plant kingdom, together with the present results, suggestthat flavonoids may contribute to the preventive effect of a plant-baseddiet on chronic diseases, including solid tumors.

INTRODUCTION

Angiogenesis, the generation of new capillaries from preexistingvessels, is virtually absent in the healthy adult organism in which it isrestricted to a few conditions including wound healing and the formation of corpus luteum, endometrium, and placenta. These conditions of physiological angiogenesis represent ordered, tightly regubated, and self-limited processes ( I ). However, in certain pathologicalconditions, angiogenesis is dramatically enhanced and loses its selflimiting capacity (2). Although pathological angiogenesis is seen

during the development and progression of many diseases, such asrheumatoid arthritis, psoriasis, and diabetic retinopathy, from a dinical perspective, probably the most important manifestation of pathobogical angiogenesis is that induced by solid tumors (3). Wellvascularized tumors expand both locally and by metastasis, whereasavascular tumors do not grow beyond a diameter of 1—2mm (1, 4).

Dietary factors contribute to about one-third of potentially preventable cancers (5), and the long-known preventive effect of plant-baseddiets on tumorigenesis and other chronic diseases is well documented(6). Breast, prostate, and endometnal cancer belong to a group ofhormone-dependent cancers that, like colon cancer and coronary heartdisease, are among those chronic diseases that have a bower incidencein Asia than in Western countries (7). Immigrants from Asia whomaintain their traditional diet do not increase their risk of these

Received I/7/97; accepted 5/13/97.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance withI8 U.S.C. Section 1734 solely to indicate this fact.

I Work in Heidelberg was supported by grants from Deutsche Forschungsgemeinschaft, Kulemann-Stiftung, and Deutsche Krebshilfe; work in Geneva by Grant 31-43364.95 from the Swiss National Science Foundation; and work in Helsinki by grantsfrom the Finnish Cancer Foundation and the Sigrid Juselius Foundation.

2To whom requests for reprints should be addressed, at Laboratory of BiologicalChemistry. Medical School. University of loannina, 451 10 loannina, Greece. Phone:(30)65 1 28388; Fax: (30)65 1 33442; E-mail: thfotsis@cc.uoi.gr.

3 Present address: Department of Hematology, Oncology and Endocrinology. Children's Hospital. University of Essen, Hufelandstrasse 55. 45 122 Essen. Germany.

diseases (8); however, an increased risk for these diseases accompanies a change toward a Westernized diet (9). These data indicate thatcertain plant-derived dietary groups might contain compounds thatexert antimitotic and antitumorigenic effects, thereby offering anticancer protection to individuals consuming such diets. Identificationand characterization of such compounds might provide us with additional chemotherapeutic agents for pharmacological intervention incancer.

The idea that dietary ingested compounds could modulate proliferation of tumor cells and pathological angiogenesis appeared to us tobe an important possibility meriting further investigation. If dietarycompounds were to inhibit angiogenesis, this could explain, at least inpart, the long-known preventive effect of plant-based diets on tumorigenesis and other chronic diseases, such as inflammation (5). Inprevious studies, we examined this possibility by screening the urineof human subjects consuming a diet rich in plant products for thepresence of antimitotic and antiangiogenic compounds. This work ledto the identification of the isoflavonoid genistein as a potent inhibitorof cell proliferation and in vitro angiogenesis (10, 11). Further studiesshowed that the excretion of genistein in urine of vegetarians is30-fold higher than that of omnivores (12—14).In the present study,we extended these observations by investigating the antimitotic andantiangiogenic effects of flavonoids, a group of compounds that areisomeric to isoflavonoids. Flavonoid aglycones all consist of a benzene ring (A) fused with a six-member ring (C) that in position 2carries a phenyb ring (B) as a substituent (Table I), whereas isoflavonoids carry the B ring in position 3. Flavonoids are more widelydistributed in the plant kingdom (15—17),rendering them a veryattractive target for further studies.

MATERIALS AND METHODS

Materials and Instrumentation. All flavonoids except 2',3'-dihydroxyflavone, 3',4'-dihydroxyflavone, coumarin, and catechin were obtained fromRoth Chemikalien (Karlsruhe, Germany). 2',3'-Dihydroxyflavone and 3',4'-dihydroxyflavone were synthesized as described below. The variousflavonoids were prepared in 10 mM stock solutions. Apigenin, chrysin,3',4'-dihydroxyflavone, 3',4'—dihydroxyflavone, luteolin-7-glucoside, and

3-hydroxyflavone were dissolved in ethanol. All the other compounds weredissolved in ethanol: DMSO (1:1, v/v). DMSO was obtained from Merck

(Darmstadt, Germany). When stored at 4°C,stock solutions of the substancesremained bioactive for more than I month.

‘Hand ‘3CNMR4 spectra were recorded on a Varian GEMINI-200 FTspectrometer. Mass spectra were obtained with a JEOL JMS SX1O2 massspectrometer operating at 70 eV. Samples were introduced by a direct inletprobe. The UV spectra were recorded with a CARY SE UV-VIS-NIR spectophotometer. Melting points were determined in open capillary tubes with anElectrothermal apparatus and are uncorrected. TLC was conducted on Merck

4 The abbreviationsusedare: NMR, nuclearmagneticresonance;THF, tetrahydrofuran; LiHMDS, lithium bis(trimethylsilyl)amide; BBCE, brain capillary endothelial; ACE,adrenal cortex endothelial; BAE, bovine aorta endothelial; HUVE, human umbilical veinendothelial; FGF, fibroblast growth factor; bFGF, basic FGF; BME, bovine microvascularendothelial; VEGF, vascular endothelial growth factor.

2916

Flavonoids, Dietary-derived Inhibitors of Cell Proliferation andin Vitro Angiogenesis'

TheodoreFotsis,2MichaelS. Pepper,ErkanAktas,StephenBreit,3Sirpa Rasku,HermanAdlercreutz,Kristiina Wähälä,Roberto Montesano, and Lothar Schweigerer@

Division of Hema:olog@ and Oncologe. Children ‘sHospital, Ruprecht-Karls Universir-v, INF 150, 69120 Heidelberg, Germany (T. F.. E. A., S. B., L SI: Institute of Histology andEmbryology. Departnient of Morphology. Unis'ersitv Medical Center, 1121 Geneva 4, Switzerland (M. S. P., R. MI; Department of C'hemistrv, Organic C'hemistry Laboratory,P. 0. Box 55. University of Helsinki, FIN-IXXJI4Helsinki. Finland (S. R., K. WI: Department of Clinical Chemistry. Meilahti Hospital. University of Helsinki. SF-00290 Helsinki.Finland jH. A.!

on March 3, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

FLAVONOIDS INHIBIT IN VITROANGIOGENESIS

silica gel 60 P254 plates. THF was distilled over CaH2. LiHMDS was titrated

before use with 1-pyreneacetic acid (18).Synthesis of 2',3'-Dlhydroxyflavone and 3',4'-Dihydroxyflavone. 2',3'-

Dihydroxyflavone was synthesized via 2',3'-dimethoxyflavone as follows. A

solution of 0.81 MLiHMDS in THF (38 ml, 30.6 mmol) was added dropwiseto 2-hydroxyacetophenone (0.46 ml, 3.8 mmol) in dry THF (15 ml) underargon at —78°C.The mixture was stirred at —78°Cfor I h and then at —10°Cfor 2 h. Methyl 2,3-dimethoxybenzoate [prepared from 2,3-dimethoxybenzoicacid with methanol and H2S04 (cat.) at room temperature; 0.75 g, 3.8 mmoll

in dry THF (3 ml) was added at —78°Cin one portion. The reaction was

continued at —78°Cfor I h and then at room temperature for 20 h (TLC

monitoring, eluent: hexane:acetone, 7:2). The reaction mixture was poured intoice water (300 ml), acidified with conc. HCI, and extracted with ethyl acetate.

The extract was dried with Na2SO4and evaporated. The crude product wascyclized at room temperature with 30 ml of glacial acetic acid and I ml ofconcentrated H2S04 for 24 h. The precipitated product was filtered, washedwith water, and dried to give 2',3'-dimethoxyflavone in 79% yield. The crudeproduct was recrystallized from cycbohexane, m.p. 89—90°C[lit. (19) 85.5—87.5°C],UV Am,@,,(959'0ethanol) 297 nm (log a 4.21), 247 (4.10). The ‘Hand‘3CNMR spectra were in accord with the structuregiven, and will bepublished elsewhere; m/z 283 (19%), 282 (M+, 100%),267 (4), 162 (19), 121(50). One MBBr3 in CH2CI2(21 ml) was added dropwise to a stirred solutionof 2',3'-dimethoxyflavone (1.20 g, 4.3 mmob) in dry CH2C12(18 ml) underargon at room temperature. After 1 h, the mixture was poured into water (400ml) andrefluxedfor 3 h. CH2C12wasdistilled,andtheprecipitatewasfilteredoff and washed with water. Recrystallization from 95% ethanol gave 0.87 g of2',3'-dihydroxyflavone (81%); m.p., 257—258°C[lit. (20) 246°Cj.‘HNMRand mass spectra were identical with published data (21). UV A,,.,,@(95%ethanol) 307 ian (boga 4.37), 249 (4.10).

3',4'-Dihydroxyflavone was synthesized as follows. Freshly prepared LiH

MDS in THF (0.81 M;90 ml, 80.8 mmol) was added dropwise to 2-hydroxyacetophenone (1.3 ml, 10.1 mmol) in dry THF (40 ml) under argon at —78°C.

The mixture was stirred at —78°Cfor I h and at —10°Cfor 2 h. Methyl3,4-[bis(t-butyldimethylsilyl)benzoatej (22) (4.0 g, 10. 1 mmol) in dry THF (6ml) wasaddedat —78°Cin oneportion.Thereactionwascontinuedfor I h at—78°Cand then for 20 h at room temperature (TLC monitoring, eluent:hexane:acetone, 7:2). The reaction mixture was poured into ice water (500 ml),acidified with conc. HCI, and extracted with ethyl acetate. The organic phasewas dried with Na2SO4and evaporated. The crude product was cyclized anddesilylated at 95—100°Cwith 60 ml of glacial acetic acid and 2 ml of conc.H2S04 for 24 h. The mixture was poured into water (I biter),and the precipitatewas filtered, washed with water, and recrystallized from aqeous ethanol to give2.18 g of 3',4'-dihydroxyflavone (85%), m.p. 247°C[bit.(22) 245—246°C],UVAmax (95°10 EtOH) 346 nm (log a 4.30), 245 (4.17). The ‘Hand ‘3CNMR

spectra were in accord with the structure given, and will be published elsewhere;m/z255 (17%), 254(M+, 100%), 253 (13), 237 (5), 226(14), 134 (17),

121 (34).Cell Culture. Unless otherwise indicated, dishes and medium used for

tissue culture (23) and human recombinant bFGF (23) were from sourcesmentioned earlier. Human recombinant bFGF used in the in vitro angiogenesis

assay was provided by Dr. P. Sarmientos (Farmitalia Carlo Erba, Milan, Italy).BBCE and ACE cells were kindly provided by D. Gospodarowicz (Univer

sity of California, San Francisco, CA). BAE cells were obtained from Dr. P.Nawroth (Department of Internal Medicine, University of Heidelberg, Heidelberg, Germany). HUVE cells were a gift from Dr. J. Grulich-Henn (Children'sHospital, University of Heidelberg). BBCE, BAE, ACE, and HUVE cells were

maintained in DMEM with low glucose concentration (1000 mg/liter), 10%

newborn calf serum, gbutamine (2 mmol/liter), and antibiotics. Cultures received bFGF (2.5 ng/ml) every other day until they were confluent (24, 25).Cloned BME cells from adrenal cortex were kindly provided by Drs. M. B.Furie and S. C. Silverstein (Columbia University, New York, NY), and weremaintained as described previously (26).

Human fibroblasts were obtained from Dr. A. Habenicht (Department ofInternal Medicine, University of Heidelberg) and maintained in DMEM witha high glucose concentration (4500 mg/liter), 10% fetal bovine serum, and

antibiotics. The same culture conditions were used for human keratinocytes, agift from Prof. Fussening [(German Cancer Research Center), DKFZ, Heidelberg, Germany)]. Tumor cells from human breast adenocarcinoma (MCF-7)were provided by Dr. J. Draetta (European Molecular Biology Laboratory,

Heidelberg, Germany). MCF-7 cells were cultured in MEM with nonessentialamino acids, sodium pyruvate (1 mmol/liter), bovine insulin (10 @xg/liter),10%

fetal bovine serum, and antibiotics. Tumor cells from human neuroblastoma(SH-EP) and their MYCN oncogene stable transfectants (WAC 2) were a giftfrom Prof. M. Schwab (DKFZ) and maintained in RPMI 1640 with 10%newborn calf serum and antibiotics. In the cultures of WAC 2 cells, theantibiotic geneticin (200 @xg/ml) was also added for selection.

Cell Proliferation Assay. Stock cultureswere trypsinized(25), cells wereadjusted to a density of 5 X l0@cells/mI (or 2 X l0@cells/mI in the case ofHUVE cells) in their respective media, and seeded in 1-mI aliquots into12-well cluster dishes. After 16 h, wells received 5-pi aliquots of either solventonly, solvent containing various concentrations of the compounds to be tested,or bFGF (2.5 ng/ml), and these treatments were renewed every other day. Cellswere counted at the times indicated with a Coulter particle counter (25). Unlessotherwise indicated, values of cell densities represent the means of duplicatedeterminations that varied by less than 10% of the mean.

In Vitro Angiogenesis Assay. Three-dimensionalcollagen gels were prepared in 18-mm tissue culture wells as described (26). BME cells (7.5 X l0@

cells in a volume of 0.5 ml of a modification of MEM containing 5% donorcalf serum) were seeded into each well; after 3 days, when the cells reachedconfluence, the medium was changed and donor calf serum was reduced to 2%.

Recombinant human bFGF (10 ng/ml) plus VEGF (30 ng/ml) together with0.1—10 @xmol/biter concentrations of the flavonoids to be tested were then

added. Flavonoids were added 2 h before bFGF and VEGF on the first day oftreatment. Cultures were photographed after 4 days. Invasion was quantifiedfrom three randomly selected fields per well measuring I X 1.4 mm bymeasuring the total length of all cells that penetrated the underlying gel eitheras single cells or in the form of cell cords (27). Unless otherwise stated, results

are from at least three experiments per concentration of each flavonoid (i.e.,three measurements per experiment; nine measurements per concentration ofeach flavonoid).

RESULTS

Several Flavonoids Inhibit bFGF-stimulated Endothelial CellProliferation. To investigate the antiangiogenic effects of flavonoids, a series of flavonoid metabolites was tested with regard totheir effect on the bFGF-stimulated proliferation of BBCE cells. Thelist of the flavonoids was selected to cover a large range of structuresto facilitate the discovery of potential structure-activity relationships(Table 1). To this list, coumarin was added as a control because of itsclosely rebated chemical structure. This initial experiment showed thatseveral flavonoid metabolites could inhibit the in vitro proliferation ofBBCE at half-maximal concentrations in the lower micromolar range(Table 2). At least three flavonoids, 3',4'-dihydroxyflavone, luteolin,and 3-hydroxyflavone, inhibited the bFGF-induced proliferation ofBBCE cells (Table 2; Fig. I) more potently than genistein (5—6p@molfliter; Ref. 10). Therefore, the inhibitory activity of the mostpotent flavonoids was approximately 2—3-foldstronger than that ofthe isomeric genistein. Another group of flavonoids, including apigenm, fisetin, quercetin,and eriodictyol,exhibited inhibitoryactivitycomparable to that of genistein, whereas several flavonoids had mmimal or no effect on the proliferation of endothelial cells. The resultwas essentially the same for the most potent metabolites, when endothebiab cells of different tissues and species were tested. Thus, BAE,ACE, and HUVE cells were all inhibited in a manner similar to BBCEcells (data not shown).

TumorCellsAreAlsoa Targetof the AntimitoticEffectsofFlavonoids. Having established the antiproliferative effects of flavonoids on endotheliab cells, we next investigated their antiprobiferative effects on various normal and tumor cells. The rationalewas 2-fold: first, to investigate the possible direct antitumor effectsof flavonoids; and second, to observe whether there is any selectivity in their antimitotic activity. Towards this end, several normaland tumor cells were used, all of human origin. The normal cellsincluded fibroblasts (HFK2) and keratinocytes (HaCaT). The tu

2917

on March 3, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Table 2 Antiprolferative effects offlavonoidsThe half-maximal concentration (@xmol/liter) of the inhibitory effect of the various substances was tested on the in s'itro proliferati

human ceratinocytes (HaCaT), breast cancer adenocarcinoma cells (MCF7), and human neuroblastoma cells (SHEP and WAC 2).onof BBCE cells, human fibroblasts(HFK2),BBCE

HFK2 HaCaT MCF7SHEPWAC2FlavonesFlavone

8.55-Methoxyflavone12.03'.4'-Dihydroxyflavone

1.4 9.5 1.50.75.00.8Chrysin9.0 25.0 20.07.05.0Apigenin6.5 10.0 6.03.54.0Luteolin1.9 7.6 4.06.54.01.1Luteolin-7-glucoside

18.0 6.06.04.0Flavonoles3-Hydroxyflavone

2.7 2.82.12.4Fisetin8.6 20.1 20.010.67.06.6Quercetin5.35 36.2 >50.05.558.06.6Myricetin

28.0 40.0 40.0NE―FlavanonesEriodictyol

7.0 >50.0 15.018.010.0Hesperetin30.0 NE NE40.0>50.0Flavan-3-olCatechin

>50.0 >50.0 NENENECumarinsCumarin

NE NE NENENE‘,

NE, no effect.

FLAVONOIDS INHIBIT IN VITROANGIOGENESIS

Table I Substances testedChemical structures of the various substances tested in vitro for their antiproliferative and anticarcinogenic activities.

o1?XI

Flavonols

Chemical formula

Flavones 3@2@ 4

Flavanones

Name Substitution5 7 2'

Hesperetin

E,iodictyol

3'4'5'Flavone

5-MethoxyflavoneCiysin2',3'-Dihydroxyflavone3'.4'-DihydroxyflavoneApigeninLuteolinLuteolin-7-glucosideH

OCH3OHHHOHOHOHH

HOHHHOHOHOGLCH

HHOHHHHHH

HHOHOHHOHOHH

HHHOHOHOHOHH

HHHHHHH

OH OH H OH OCH3HOH OH H OH OH H

3-HydroxyflavoneFisetinQuercetinMyrisetin

H OH H H H HH OH H OH OH HOH OH H OH OH HOH OH H OH OH OH

CoumarinCatechin-.-- OH

mor cells were also of different tissue origins: MCF7 cells werefrom breast adenocarcinoma, and SHEP and WAC2 cells were ofneuroectodermal origin. WAC2 cells are stable transfectants ofSHEP cells that express the MYCN oncogene and thus had anincreased proliferative potential on the same genetic background asthe SHEP cells (23). 3',4'-Dihydroxyflavone, luteolin, and 3-hydroxyflavone were again among the most potent antiprobiferativemetabolites (Fig. 2; Table 2). As for endothelial cells, several ofthe flavonoids exhibited a potent antimitotic activity, especially ontumor cells, whereas others had minimal or no effect. This infor

mation is particularly useful as it establishes a structure-activityrelationship as discussed below.

Flavonoids Inhibit Angiogenesis in Vitro. To examine whetherflavonoids could inhibit angiogenesis, we used an in vitro experimentab system that recapitulates two essential components of the angiogenic process, namely extracelbular matrix invasion and capillary-liketube formation (26). As shown previously, BME cells seeded on thesurface of collagen gels invade the gels when exposed to bFGF (28)or VEGF (27) and form capillary-like tubes beneath the gel surface.When 10 @.tMconcentrations of the most potent antimitotic flavonoids

2918

on March 3, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

p

10 100

FLAVONOIDS INHIBIT IN VITROANGIOGENESIS

01@

0U

C

5?

I-'5?

5?U

CATECHIN

GENISTEIN

2',3'-DHF

LUTEOLIN

APIGENIN

FISETIN

3',4-DHF

3-HF

CONTROLSubstance, @iM

Fig. 1. Effect of increasing concentrations of flavonoids on the proliferation of BBCEcells. BBCE cells were seeded at a density of 5000 cells/well, and every other day, theyreceived bFGF (2.Sng/ml). In addition, every other day they received 5-@s1aliquots ofsolvent only or of solvent containing the indicated concentrations of 3',4'-dihydroxyflavone (•),luteolin (•),3-hydroxyflavone (0), fisetin (A), or quercetin (0). Cells werecounted after 5 days. Values are expressed as percent of controls (i.e., cells receivingsolvent only).

Fig. 2. Effect of flavonoids on the proliferationof human fibroblasts (HFK2; A) and human breastcancer adenocarcinoma cells (MCF7; B). Cellswere seeded at a density of 10,000 cells/well andreceived 5 @xlof solvent every other day with orwithout the indicated concentrations of 3',4'-dihydroxyflavone (•),luteolin (•),3-hydroxyflavone(0), fisetin (A), or quercetin (0). Cells werecounted after 5 days. Values are expressed as percent of controls (i.e., cells receiving solvent only).

0 10 20 3@l 40 50 60 70 80 90 100

Total additive sprout length% of control

Fig. 3. Effect of flavonoids on in vitro angiogenesis. BME cells were grown toconfluence on three-dimensional collagen gels as described in “Materialsand Methods.―After reaching confluence, they received solvent or recombinant human bFGF (10 ng/ml)plus VEGF (30 ng/ml) with or without 10 @mol/literconcentrations of the substances tobe tested. After 2 days of incubation, the medium was changed, and cells were exposedto the same conditions again. Quantitative analysis was carried out to BME cells treatedfor 5 days as described above. Fields measuring I X I .4 mm were then randomly selectedand photographed, and BME cell invasion was quantified by measuring the total length ofall cells that penetrated the underlying gel either as single cells or in the form ofcell cords.Results are expressed as percent of controls (i.e., cells receiving solvent and recombinanthuman bFGF plus VEGF).

human subjects consuming a diet rich in plant products for thepresence of antiangiogenic and antimitotic compounds (10, 11, 29,30). At least three urine fractions were able to inhibit the proliferationof endothelial and neuroblastoma cells. In one of the fractions, theisoflavonoid genistein was identified as an antimitotic and antiangiogenic compound (10, 29). Another of the active urine fractions contamed compounds with aromatic ortho-hydroxy groups ( 10). Becauseflavonoids, the isomers of isoflavonoids, are a widely distributedgroup of pobyphenolic compounds in the plant kingdom (15—17)andbecause most of them possess aromatic ortho-hydroxy groups, wewere prompted to test their biological activity on proliferation of normaland tumor cells as well as their effects on in vitro angiogenesis.

In the present study, we have established that several flavonoidmetabobites exhibit strong inhibitory activity on bFGF-induced proliferation of endothelial cells in vitro (Fig. 1). To the best of our

were added along with a synergistic combination of bFGF and VEGF,they inhibited the ability of BME cells to invade the gels and generatecapillary-like structures. A quantitative analysis revealed that severalflavonoids inhibited bFGFIVEGF-induced invasion of BME cells to agreater extent than the same concentration of genistein (Fig. 3). Amore extensive evaluation, carried out in a dose-response manner,confirmed these findings and revealed a half-maximal concentrationfor the inhibition of in vitro angiogenesis which closely correlateswith that of the antimitotic effects (Fig. 4). The half-maximal concentration of the inhibitory effect of genistein was around 10 j.LM,avalue that is 15-fold lower than the value that we reported in an earlierpublication (10). We have now been able to confirm that the significantly higher half-maximal concentration observed in our earlierstudies was due to the reduced solubility of genistemn in sodiumbicarbonate as compared to DMSO (data not shown).

DISCUSSION

The presence of phytochemicals in vegetarian diet is thought toexplain, at least in part, the long-known preventive effect of plantbased diets on tumorigenesis and other chronic diseases (5). Toexamine this possibility, in previous studies we screened urine of

A

0.ul 0.1 1

Substance, @tM Substance, @tM

2919

C

0CU

C

5?

1@5?0.rl)

5?U

1°@

L)

10 100

120C‘I

@I1@

ei 20'

@ ‘5.

on March 3, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

FLAVONOIDS INHIBIT IN VITROANGIOGENESIS

show that a nonhydroxybated ring C with oxo function at position 4 anda C2-C3 double bond is required for maximal biological activity. Catechin, which backs both the C4 oxo group and the C2-C3 double bond, iscompletely devoid of antiproliferative activity. Eriodictyob, the flavanonederived from reduction of the C2-C3 double bond of the flavone luteolin,is at least 3-fold weaker in inhibiting all the cells tested. The presence ofa hydroxyl function in position 3, as occurs in flavonols, seems to modifythe antimitotic activity in an unpredictable manner. Thus, 3-hydroxyflavone, the 3-hydroxybated flavonol derived from flavone, has antiproliferative activity which is 4-fold greater. In contrast, quercetin, the flavonolderived from buteolin, is 2.5-fold weaker in its inhibition of cell proliferation. The catechol structure alone is insufficient for inhibition ofproliferation. This is clearly demonstrated by catechin, which despite thepresence of aromatic vicinab hydroxybs at positions 3' and 4', does nothave an antimitotic effect. Also, 3-hydroxyflavone, which has no catecholfunctions, has antiproliferative activity comparable to the catecholic metabolites luteolin and 3',4'-dihydroxyflavone. On the other hand, hesperetin, derived from methybation of the catechob hydroxyl at C-4' oferiodictyol, inhibits BBCE cells with a half-maximal concentration of 30p.mol/liter, which is 4-fold weaker than the antimitotic activity of thecatecholic eriodictyol (7 @molIbiter).

Having established the inhibitory effects of flavonoids on theproliferation of endothebial cells, we next considered it important todetermine whether flavonoids had additional effects on other functions of endothelial cells, which are important for angiogenesis. Angiogenesis is a complex process requiring the coordinated, sequentialinvolvement of a number of cellular events. Indeed, formation of newcapillaries begins with a localized breakdown of the basement mem

brane of the parent vessel, through the finely tuned elaboration ofproteolytic enzymes and their inhibitors (35); this is followed bymigration of endothelial cells and invasion of the surrounding matrix.The in vitro angiogenesis assay used in the present study recapitulatesthese early events. Thus, inhibition of in vitro angiogenesis by flavonoids represents interference with a number of early events duringangiogenesis, including endothebial cell proliferation, proteolytic enzyme production, and endothelial cell migration. We have observedthat the order of potency of the various flavonoids with respect toantimitotic effects and inhibition of in vitro angiogenesis does notstrictly correlate. However, the requirement for a nonhydroxylatedring C with oxo function at position 4 and a C2-C3 double bond is alsovalid for inhibition of in vitro angiogenesis because catechin is onceagain totally inactive.

Flavonoids exhibit several interesting biochemical properties (36).

Inhibition of tyrosine kinases (37, 38) and protein kinase C (39) are,however, of particular importance with regard to cellular processesstudied in this paper. Flavonoids are competitive inhibitors withrespect to the ATP binding (37) site on a variety of enzymes, a regionof considerable homology among kinases. In this context, apigeninwas found to inhibit both protein kinase C and TPA stimulation of theFGF receptor (40). A further important target of flavonoids appears tobe l-phospatidybinositol kinase (41), a key enzyme in signal transduction beading to production of second messengers, inositob 1,4,5-trisphosphate, and diacylgbycerol. Quercetin in breast carcinoma cellsreduced phosphatidylinositol kinase and inositob 1,4,5-trisphosphateconcentration within 60 mm to 5 and 6%, respectively (IC50, 6

@mol/liter).However, no other flavonoids were tested in this study(41). Synthetic analogs of flavonoids are potent inhibitors of cdc2 (42)and cyclin-dependent kinases (43, 44), thereby inhibiting cell cycleprogression. At beast in the case of cdc2, the inhibition was competitive with respect to AlP binding (42). Although inhibition of kinasesappears to be the most probable target, selected flavonoids have beenreported to inhibit the DNA binding or the DNA religation step ofeukaryotic topoisomerase I (45). The strong scavenging of free radi

aCU

1@C

Substance, pM

Fig. 4. Comparison of the effect of 3',4'-dihydroxyflavone, 3-hydroxyflavone, andgenistein on in vitro angiogenesis. BME cells were grown to confluence on threedimensional collagen gels as described in Materials and Methods.― After reachingconfluence, they received solvent or recombinant human bFGF (10 ng/ml) plus VEGF (30ng/ml) or various concentrations of 3'.4'-dihydroxyflavone (•).3-hydroxyflavone (0).and genistein (Y). After 2 days of incubation, the medium was changed, and the cells wereexposed to the same conditions again. Quantitative analysis was carried out to BME cellstreated for 5 days as described above. Fields measuring I X 1.4 mm were then randomlyselected and photographed, and BME cell invasion was quantified by measuring the totallength of all cells that penetrated the underlying gel either as single cells or in the formof cell cords. Results are expressed as percent of controls (i.e., cells receiving solvent andrecombinant human bFGF plus VEGF).

knowledge, there are no previous reports regarding the effects offlavonoids on endothelial cells, the target cells in angiogenesis. Threeflavonoids, 3',4'-dihydroxyflavone, buteolin, and 3-hydroxyflavone,at half-maximal inhibitory concentrations of 1.4, 1.9, and 2.7 mmol/liter, respectively, were more potent than the isoflavonoid genistein(5—6p@mobIbiter).However, this difference does not warrant a concbusion that flavonoids are more potent than isoflavonoids, becauseapigenin, the corresponding flavonoid isomer of genistein, with hydroxybations at positions 5, 7, and 4', exhibits a comparable halfmaximal inhibitory concentration (6.5 @mol/liter).Endothebial cellsfrom aorta, brain, adrenal cortex, and umbilical vein were all inhibitedby flavonoids in a similar fashion, which thus exhibited no specificityfor any endotheliab subtype.

Endothebial cells were not the only target for flavonoids. Flavonoidsinhibited the proliferation of low-density cultures of various normal(HFK2 human fibrobbasts and HaCaT human keratinocytes) and tumor (MCF7 human breast cancer and SHEP and WAC2 humanneuroblastoma-derived) cells. Among the normal cell types, thereappeared to be a greater inhibitory effect on endothebiab cells than onfibroblasts and keratinocytes (Table 2), indicating a certain degree ofselectivity towards endothelial cells. In contrast, the inhibitory effect

of flavonoids on tumor cells was comparable to their effect onendotheliab cells (Table 2). It is possible that the rapid proliferation ofendotheliab cells in response to bFGF stimulation accounts for thesimilar inhibitory profiles seen with endothebial and tumor cells.These results point to a selective targeting of rapidly proliferatingcells, a conclusion that is supported by the results obtained with theisoflavonoid genistein (1 1).

Although antiproliferative effects of flavonoids have been reportedpreviously (3 1—34),most of these studies were confined to the effectsof single or few metabolites on one or two cell types. This has madeit difficult to draw any conclusion regarding a potential structureactivity correlation. In the present study, a more comprehensiveapproach was used in the testing of a set of various flavonoid structureson several normal and tumor cells of different origin. Our results clearly

120@

2920

on March 3, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

FLAVONOIDS INHIBIT IN VITROANGIOGENESIS

cals (36) does not appear to be the mechanism of antimitotic andantiangiogenic activity of flavonoids. Indeed, myricetin and catechin,some of the strongest protectors against single-stranded breaks induced by singlet molecular oxygen (46), are devoid of antimitotic orantiangiogenic activity.

Flavonoids comprise a large group of naturally occurring lowmolecular weight substances that are present in fruits, vegetables,nuts, seeds, and the stems, leaves, flowers, bark, and roots of mostplants, as well as tea, coffee, and wine (15—17).The average Westerndiet is estimated to contain approximately 23 mg/day of quercetin,kaempferol, myricetin, apigenin, and luteobin (47), the total content ofall naturally occurring flavonoids not being exactly known. No informarion is available about the content of flavonoids in the diet ofvegetarians. However, because the concentration of genistein in theurine of vegetarians in 30-fold that of omnivores (12—14), it ispossible that similar values might be also anticipated for flavonoids.Development of sensitive methods for determination of flavonoids inbiological materials, a task that our group has already undertaken, isexpected to provide valuable information in this direction. Considering the biological functions of flavonoids presented in this and previous studies, it is quite possible that flavonoids might contribute tothe preventive effects of vegetarian diet on cancer incidence andmortality. In this respect, animal experiments have already shown thatflavonoids exert inhibitory effects on carcinogenicity (48, 49). Furtherstudies should focus on the in vivo effects of flavonoids on angiogenesis and tumorigenesis models.

ACKNOWLEDGMENTS

We thank C. DiSanza and M. Guisobanfor excellent technical assistance.

REFERENCES

1. Klagsbrun, M., and Folkman, J. Angiogenesis. In: M. B. Spom and A. B. Roberts(eds.), Peptide Growth Factors and Their Receptors II, pp. 549—586.Berlin: SpringerVerlag, 1990.

2. Denekamp, J. Vascular attack as a therapeutic strategy for cancer. Cancer MetastasisRev., 9: 267—282,1990.

3. Folkman, J. Tumor angiogenesis. Adv. Cancer Res., 43: 175—203,1985.4. Folkman, J., and Cotran, R. S. Relation of vascular proliferation to tumor growth. Int.

Rev. Exp. Pathol., 16: 207—248,1976.5. Adlercreutz, H. Westem diet and westem diseases: some hormonal and biochemical

mechanisms and associations. Scand. J. Clin. Lab. Invest., 50 (Suppl. 201): 3—23,1990.6. Miller, A. B. Diet and cancer: a review. Rev. Oncol., 3: 87—95,1990.7. Rose, D. P., Boyar, A. P., and Wynder, E. L. Intemational comparison of mortality

rates for cancer of the breast, ovary, prostate, and colon, and per capita fat consumption. Cancer (Phila.), 58: 2363—2371, 1986.

8. Kolonel, L. N. Variability in diet and its relation to risk in ethnic and migrant groups.Basic Life Sci., 43: 129—135,1988.

9. Lee, H. P., Courley, L., and Duffy, S. W. Dietary effects on breast cancer risk inSingapore. Lancet, 337: 1197—1200,1991.

10. Fotsis, T., Pepper, M., Adlercreutz, H., Fleischmann, G., Hase, T., Montesano, R., andSchweigerer, L. Genistein, a dietary-derived inhibitor of in vitro angiogenesis. Proc.Nail. Acad. Sci. USA, 90: 2690—2694, 1993.

11. Fotsis, T., Pepper, M., Adlercreutz, H., Hare, T., Montesano, R., and Schweigerer, L.Genistein, a dietary ingested isoflavonoid, inhibits cell proliferation and in vitroangiogenesis. J. Nutr., 125 (Suppl.): 790S—797S, 1995.

12. Adlercreutz, H., Honjo, H., Higashi, A., Fotsis, T., HamBläinen,E., Hasegawa, T., andOkada,H. Urinaryexcretionof lignansandisoflavonoidphytoestrogensin Japanesemen and women consuming traditional Japanese diet. Am. J. Clin. Nutr., 54: 1093—1100, 1991.

13. Adlercreutz, H., Fotsis, T., Bannwart, C., WahSlä,K., Brunow, G., and Hase, T.Isotope dilution gas chromatographic-mass spectrometric method for the determinalion of lignans and isoflavonoids in human urine, including identification of genistein.Clin. Chim. Acta, 199: 263—278,1991.

14. Adlercreutz, H., Van der Wildt, J., Kinzel, J., Attalla, H., Wahala, K., MakelS, T.,Hase, T., and Fotsis, T. Lignan and isoflavonoid conjugates in human urine. J SteroidBiochem. Mol. Biol., 52: 97-103, 1995.

15. Harbome, J. B. Flavonoids. in: L. P. Miller (ed), Phytochemistry, pp. 344—380.NewYork:VanNostrandReinholdCo., 1973.

16. Harbome, J. B., and Marby, T. J. The Flavonoids: Advances in Research. London:Chapman and Hall Ltd., 1982.

17. Middleton, E., Jr. Some biological properties of plant flavonoids. Annals of Allergy,61: 53—57,1988.

18. Kiljunen, H., and Hase, T. Titration of organolithiums and grignards with 1-pyreneacetic acid. J. Org. Chem., 56: 6950—6952, 1991.

19. Hoshino, Y., and Takeno, N. Ortho substituent effect of the side-chain phenyl group

on the dehydrogenation of flavanones with 2.3-dichloro-5.6-dicyano-p-benzoquinone.Bull. Chem. Soc. Jpn., 60: 4468—4470. 1987.

20. Arcoleo, A., Bellino, A., Casinovi, C., and Venturella, P. Flavones of the veratroleseries. Ann. Chim. (Rome), 520: 75—70,1957.

21. Wollenweber, E., Mann, K., limura, M., Tanala, T., and Mizuno, M. 8,2a-Dihydroxyflavone from primula pulverulenta. Phytochemistry, 27: 1483—1486,1988.

22. Nagarathman. D., and Cushman. M. A short and facile synthetic route to hydmxylatedflavones.Newsynthesesofapigenai,Iiicin,andluteolmJ. Org.Chem.,56:4884-4887, 1991.

23. Schweigerer, L., Breit, S., Wenzel, A., Tsunamoto, K.. Ludwig, R., and Schwab, M.Augmented MYCN expression advances the malignant phenotype of human neuroblastoma cells: evidence for induction of autocrine growth activity. Cancer Res.. 50:4411—4416,1990.

24. Schweigerer, L., Neufeld, G., Friedman, J., Abraham, J. A., Fiddes, J. C.. andGospodarowicz, D. Capillary endothelial cells express basic fibroblast growth factor.a mitogen that promotes their own growth. Nature (Lond.). 325: 257—259.1987.

25. Schweigerer, L., Neufeld, G., Mergia, A., Abraham, J. A., Fiddes, J. C., andGospodarowicz, D. Basic fibroblast growth factor in human rhabdomyosarcoma cells:implications for the proliferation and neovascularization of myoblast-denved tumors.Proc. NatI. Acad. Sci. USA, 84: 842—846, 1987.

26. Montesano, R., and Orci, L. Tumor-promoting phorbol esters induce angiogenesis invitro. Cell, 42: 469—477, 1985.

27. Pepper. M. S.. Ferrara, N., Orci, L., and Montesano, R. Potent synergism betweenvascular endothelial growth factor and basic fibroblast growth factor in the inductionof angiogenesis in vitro. Biochem. Biophys. Res. Commun.. 189: 824—83 1, 1992.

28. Montesano, R., Vassalli, J-D., Baird, A., Guillemin. R., and Orci, L. Basic fibroblastgrowth factor induces angiogenesis in vitro. Proc. NatI. Acad. Sci. USA, 83: 7297—7301, 1986.

29. Schweigerer, L., Christeleit, K., Fleischmann, G., Adlercreutz, H., WShBlS, K.. Hase,T., Schwab, M., Ludwig, R., and Fotsis, T. Identification in human urine of a naturalgrowth inhibitor for cells derived from solid pediatric tumors. Eur. J. Clin. Invest., 22:260—264, 1992.

30. Fotsis, T., Pepper, M. S., Aktas, E., Joussen, A., Kruse, F., Adlercreutz, H., WtthElS,K., Hare, T., Montesano, R., and Schweigerer. L. Inhibitors of angiogenesis in humanurine. In: M. E. Maragoudakis (ed). Molecular, Cellular, and Clinical Aspects ofAngiogenesis, pp. 213—227.New York: Plenum Press, 1996.

31 . Kandaswami, C., Perkins, E., Soloniuk, D. S., Drzewiecki, G., and Middleton, E. J.Antiproliferative effects of citrus flavonoids on a human squamous cell carcinoma invitro. Cancer Lett., 56: 147—152,1991.

32. Hirano, T., Abe, K., Gotoh, M., and Oka, K. Citrus flavone tangeretin inhibitsleukemic HL-60 cell growth partially through induction of apoptosis with lesscytotoxicity on normal lymphocytes. Br. J. Cancer. 72: 1380—1388, 1995.

33. Yit, C. C., and Das, N. P. Cytotoxic effect of butein on human colon adenocarcinomacell proliferation. Cancer Lett., 82: 65—72,1994.

34. Huang, H. C., Wang, H. R., and Hsieh, L. M. Antiproliferative effect of baicalein, aflavonoid from a Chinese herb, on vascular smooth muscle cell. Eur. J. Pharmacol.,251: 91—93,1994.

35. Pepper. M. S., and Montesano, R. Proteolytic balance and capillary morphogenesis.Cell Differ. Dcv., 32: 319—328, 1990.

36. Havsteen, B. Flavonoids, a class of natural products of high pharmacological potency.Biochem.Pharmacol.,32: 1141—1148,1983.

37. Graziani, Y., Erikson, E., and Erikson, R. L. The effect of quercetin on the phosphorylation activity of the Rous sarcoma virus transforming gene product in vitro and ini@ivo.Eur. J. Biochem., 135: 583—589,1983.

38. Cunningham, B. D. M., Threadgill, M. D., Groundwater, P. W., Dale, I. L., andHickman, J. A. Synthesis and biological evaluation of a series of flavones designedas inhibitors of protein tyrosine kinases. Anticancer Drug Des., 7: 365—384,1992.

39. Femola, P. C.. Cody, V.. and Middleton, E. J. Protein kinase C inhibition by plantflavonoids. Biochem. Pharmacol., 38: 1617—1624,1989.

40. Huang, Y. T., Kuo, M. L., Liu, J. Y., Huang. S. Y., and Lin, J. K. Inhibitions ofprotein kinase C and proto-oncogene expressions in NIH 3T3 cells by apigenin. Eur.J. Cancer, 32A: 146—151,1996.

41. Singhal, R. L., Yeh, Y. A., Praja, N., Olab, E., Sledge, G. W. J., and Weber, G.Quercetin down-regulates signal transduction in human breast carcinoma cells. Biochem. Biophys. Res. Commun., 208: 425—431, 1995.

42. Losiewicz, M. D., Carlson, B. A., Kaur, G., Sausville, E. A., and Worland, P. J. Potentinhibition of CDC2 kinase activity by the flavonoid L86—8275. Biochem. Biophys.Res. Commun., 201: 589—595, 1994.

43. Carlson, B. A., Bubay, M. M., Sausville, E. A., Brizuela, L., and Worland, P. J.Flavopiridol induces G, arrst with inhibition ofcyclin-dependent kinase (CDK) 2 andCDK4 in human breast carcinoma cells. Cancer Res.. 56: 2473—2478.1996.

44. Dc Azevedo, W. F.. Mueller-Dieckmann, H. J., Schulze-Gahman, U., Worland, P. J.,Dausville, E., and Kim, S. H. Structural basis for specificity and potency of aflavonoid inhibitor of human CDK2, a cell cycle kinase. Proc. NatI. Acad. Sci. USA,93: 2735—2740,1996.

45. Bocge, F., Straub, T., Kehr, A., Boesenberg, C., Christiansen. K., Andersen. A..Jakob, F., and Kohrle, J. Selected novel flavones inhibit the DNA binding or the DNAreligation step of eukaryotic topoisomerase I. J. Biol. Chem., 271: 2262—2270, 1996.

46. Devasagayam, T. P., Subramanian, M., Singh, B. B., Ramanathan, R., and Das, N. P.Protection of plasmid pBR322 DNA by flavonoids against single-stranded breaksinduced by singlet molecular oxygen. J. Photochem. Photobiol., 30: 97—103,1995.

47. Hertog, M. G., HoIlman, P. C., Katan, M. B., and Kromhout, D. Intake of potentiallyanticarcinogenic flavonoids and their determinants in adults in the Netherlands. Nutr.Cancer, 20: 21—29,1993.

48. Das, A., Wang, J. H., and Lien, E. J. Carcinogenicity. mutagenicity and cancerpreventing activities of flavonoids: a structure-system-activity relationship (SSAR)analysis. in: E. Jucker (ed), Progress in Drug Research, pp. 133—166.Basel: Birkhaeuser Verlag, 1995.

49. Elangovan, V., Sekar. N., and Govindasamy, S. Chemopreventive potential of dietarybioflavonoids against 20-methylcholanthrene-induced tumorigenesis. Cancer Lett.,87: 107—113,1994.

2921

on March 3, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1997;57:2916-2921. Cancer Res   Theodore Fotsis, Michael S. Pepper, Erkan Aktas, et al.  

Angiogenesisin VitroFlavonoids, Dietary-derived Inhibitors of Cell Proliferation and

  Updated version

  http://cancerres.aacrjournals.org/content/57/14/2916

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/57/14/2916To request permission to re-use all or part of this article, use this link

on March 3, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from