Apigenin Acts on the Tumor Cell InvasionProcess and Regulates Protease Production

F. Lindenmeyer, H. Li, S. Menashi, C. Soria, and H. Lu

Abstract: Apigenin is a widely distributed plant flavonoidand was proposed as an antitumor agent. In this study, weinvestigated the apigenin effects on the protease-mediatedinvasiveness in an estrogen-insensitive breast tumor cellline MDA-MB231. The results show that apigenin at22.8–45.5 �M (2.5–10 mg/ml) strongly inhibited, in a dose-dependent manner, tumor cell invasion through Matrigel,cell migration, and cell proliferation. We show that apigenintreatment from 22.8 mM (2.5 mg/ml) led to a partial decreasein urokinase-plasminogen activator expression and to a to-tal inhibition of phorbol 12-myristate 13-acetate-inducedmatrix metalloproteinase-9 secretion. We also demonstratein the apigenin-treated cells a defective adhesion toMatrigel and a G2-M cell cycle arrest. Taken together, ourresults demonstrate that apigenin is a pleiotropic effector af-fecting protease-dependent invasiveness and associatedprocesses and proliferation of tumor cells.

Introduction

The long-known preventive effect of plant-based diets ontumorigenesis and other chronic diseases is well docu-mented (1). Several cancers, including breast cancer, areamong those that have a lower incidence in Asia than inWestern countries (2). These observations have been attrib-uted to the Asian dietary regimen, rich in flavonoid-containing plants, which are thought to be antitumorigenic.The flavonoid family is a group of common phenolic plantpigments, present in significant amounts in various plantslike soya, fruits, and vegetables, as well as in tea, coffee, andwine (3). Yet the molecular mechanisms involved in the ac-tion of these flavonoids are so far not well documented.

Among the flavonoids, apigenin is one of the most effec-tive in inducing cell growth inhibition and is relatively non-toxic and nonmutagenic (4,5). It has been reported thatapigenin has chemopreventive effects (6,7), and it was sug-gested to involve the p53-p21/waf1 response pathway (8).Apigenin arrests rat neuronal cells at the G2-M phase and

promotes their differentiation (3). Moreover, it has anti-proliferative effects on various cell lines (10), includingbreast cancer cells (5). Two cell cycle regulators, p34 (cdc2)kinase and cyclin B1, were shown to be decreased in coloncarcinoma cell lines after apigenin treatment, resulting ingrowth inhibition and reversible arrest of the G2-M phase(11). In HL-60 leukemic cells, treatment with apigenin (60�M) induced apoptosis and caused a rapid induction ofcaspase-3 activity, elevation of reactive oxygen species pro-duction, release of mitochondrial cytochrome c into the cy-tosol, and subsequent induction of pro-caspase-9 processing(12). Recently, apigenin was also shown to inhibit mela-noma B16-BL6 growth and metastatic potential in vivo (13).

It has been suggested that apigenin affects cells by inter-fering with nuclear factor (NF)-�B transcriptional activation(14), which regulates the expression of several genes in-volved in cellular adhesion and invasion, such as intercellu-lar adhesion molecule 1, ELAM-1, vascular cell adhesionmolecule 1 (14), and also type IV collagenases [matrixmetalloproteinase (MMP)-2 and MMP-9] (15,16). Recently,it has been shown that apigenin inhibited mitogen-activatedprotein kinase (MAPK) activity and its downstream onco-gene expression, such as c-jun and c-fos (17). In anotherstudy, the inhibitory effect of apigenin on proliferation ofthyroid carcinoma cells was associated with an inhibition ofboth epidermal growth factor (EGF)-receptor tyrosineautophosphorylation and phosphorylation of its downstreameffector MAPK (18). MAPKs are thought to integrate a vari-ety of mitogenic signals initiated by various cell surfacereceptors and regulate transcription of many genes as inter-mediates in the signal transduction pathway (19,20). Consti-tutively active MAPKs are associated with carcinogenesisand metastasis in human breast cancer (21).

This study investigated the effect of apigenin on cell in-vasion. Cell invasion involves a series of proteolytic eventsinvolving proteolytic enzymes such as urokinase-plasmino-gen activator (u-PA) and MMPs. The concept that these pro-teases are involved in invasion and metastasis, as well astumor angiogenesis, has been well accepted (22). Urokinase

NUTRITION AND CANCER, 39(1), 139–147Copyright © 2001, Lawrence Erlbaum Associates, Inc.

F. Lindenmeyer, S. Menashi, C. Soria, and H. Lu are affiliated with the Institut National de la Santé et de la Recherche Médicale, U553, Bât. INSERM,Institut d’Hématologie, Hôpital Saint-Louis, Université Paris 7, 75475 Paris, France. H. Li and C. Soria are affiliated with the Laboratoire DIFEMA, Groupe deRecherches MERCI, Faculté de Médecine et de Pharmacie de Rouen, Rouen, France.

is a serine protease, and expressions of u-PA and its cell sur-face receptor u-PAR are considered to be poor prognostic in-dexes of breast cancer (23,24). Collagenase type IV andMMP-9, in particular, have also been shown to play a criticalrole in tumor invasion and metastasis (25,26). Consideringthat MAPKs have been reported to regulate the expressionof MMP-9 in breast epithelial cells (27) and that gene ex-pression of proteases, in particular u-PA and MMP-9, is un-der the regulation of NF-�B (16,28), we postulated thatapigenin would affect cell invasion by regulating proteaseproduction. To test this hypothesis, we investigated the ef-fect of apigenin on an estrogen-negative breast tumor cellline (MDA-MB231), which was highly proliferative and in-vasive.

Our results show that the adhesive and invasive behaviorsof apigenin-treated cells were strongly inhibited. Exposureof MDA-MB231 cells to apigenin blocked the constitutiveexpression of u-PA as well as the phorbol 12-myristate 13-acetate (PMA)-induced MMP-9 secretion. We also showthat cell migration and cell proliferation were affected byapigenin, thus demonstrating that apigenin is a pleiotropiceffector on tumor cells.

Material and Methods

Material and Chemicals

All culture products were obtained from GIBCO, plasticsfrom Costar, monoclonal antibodies against u-PA (no. 3689)and u-PAR (no. 3936) from American Diagnostica (Green-wich, CT), Annexin V-fluorescein isothiocyanate kit fromBioproducts (Boehringer Ingelheim, Heidelberg, Germany),and apigenin, dimethyl sulfoxide (DMSO), and bovine se-rum albumin (BSA) from Sigma. Apigenin was dissolved inDMSO and prepared in 90 mM (10 mg/ml) stock solution.Matrigel was obtained from Becton-Dickinson (Bedford,MA), recombinant single-chain urokinase (scu-PA) fromGrunenthal (Aachen, Germany), and recombinant humanbasic fibroblast growth factor (bFGF) from R & D Systems(Abingdon, UK).

Cell Culture and Treatment of Cells

The human breast adenocarcinoma cell line MDA-MB231 was routinely cultured in Dulbecco’s modifiedEagle’s medium (DMEM) supplemented with 2 mM gluta-mine, 100 IU/ml penicillin, 100 �g/ml streptomycin, and10% fetal calf serum (FCS). For experiments, subconfluentcells were washed and incubated with various concentra-tions of apigenin, ranging from 0 to 180 �M (0–20 �g/ml),in serum-free medium for 24 hours. The concentration ofvehicle (DMSO) was adjusted to 0.2% (vol/vol) in all exper-iments, including control. Trypan blue exclusion test re-vealed that after 48 hours of incubation �95% of the cellswere viable in the control experiments, while in the presenceof apigenin the viability was ~90% at the highest concentra-

tion used. After incubation, culture media were collectedand centrifuged to eliminate cell fragments. Tumor cellswere harvested by short exposure to 1 mM EDTA andwashed. To test the role of u-PA in our system, the cell sus-pension was incubated with 40 nM recombinant scu-PA inserum-free medium for 30 minutes at 4°C to saturate the u-PA receptors at the cell surface and washed before use.

Quantification and Activity of u-PA: Enzyme-LinkedImmunosorbent Assay and Zymogram Assays

Cells were incubated with the indicated concentrations ofapigenin in serum-free medium for 24 hours. Culture mediawere then collected and centrifuged before use. The cellswere harvested by short exposure to EDTA and exposed ornot to scu-PA. Volumes of culture media were adjusted tothe same number of cells. The u-PA secreted into serum-freemedia was determined by indirect enzyme-linked immuno-sorbent assay (IMUBIND u-PA ELISA kit) according to themanufacturer’s instructions (American Diagnostica). Zy-mography was used to quantify the cell-bound and the freeu-PA activity. It was performed with casein gel containingplasminogen, as described elsewhere (29). A volume of 5 �lof washed cells (25 × 103) or culture medium was laid downonto the gels. The gels were observed under dark-field illu-mination for the areas of lysis after a five-hour incubation at37°C. Amiloride (Sigma) at 1 mM was included in a controlgel to inhibit u-PA activity to ensure the specificity of u-PA-dependent caseinolysis.

Flow Cytometric Analysis for u-PA, u-PAR, andDNA Content

To examine the cell surface expression of u-PA or u-PAR, cells were treated with apigenin for 24 hours. Thewashed and resuspended cells were incubated with 10 �g/mlof a monoclonal antibody directed against u-PA or u-PARfor 30 minutes at 4°C. We used a fluorescein isothiocyanate-conjugated anti-mouse rabbit immunoglobulin G (BioSys-tem, Compiègne, France) to reveal positive cells. After theywere washed, the cells were fixed with 4% paraformal-dehyde and analyzed in a flow cytometer FACScan (Becton-Dickinson) with the CellQuest software.

For cell cycle analysis, cells were treated with variousconcentrations of apigenin for 48 hours in serum-free me-dium. The cell suspension was washed and incubated withpropidium iodide. After 10 minutes of incubation, the cellsuspension was directly analyzed on a FACScan.

Quantification by Reverse Transcription-PolymeraseChain Reaction

Subconfluent cells were cultured for 24 hours in serum-free medium with various concentrations of apigenin. Thecells were harvested by short exposure to trypsin-EDTA andpelleted. The mRNA were extracted with the SV Total RNA

140 Nutrition and Cancer 2001

Isolation System (Promega, Madison, WI) according tothe manufacturer’s recommendations. The Promega accessreverse transcription-polymerase chain reaction (RT-PCR)system was used in this assay. Sense primer (5�-GGCAGCAATGAACTTCATCAAGTTCC-3 �) andantisense primer (5�-TATTTCACAGTGCTGCCCTCCG-3�)were used for the detection of u-PA mRNA, 5�-CTGGA-GCTGGTGGAGAAAAG-3� and 5�-CATGTCTGATGAG-CCACAGG-3� for u-PAR mRNA, and 5�-ATCTGGC-ACCACACCTTCTACAATGAGCTGCG-3� and 5�-CGTCATACTCCTGCTTGCTGATCCACATCTGC-3� for�-actin mRNA as control. Total RNA (0.5 �g) and 50 pmolof each primer were added to the RT-PCR. The RT was per-formed at 48°C for 50 minutes followed by a cycle reactionof PCR at 94°C (2 min), 60°C (30 s), and 68°C (1 min), then29 cycles at 94°C (1 min), 60°C (30 s), 68°C (1 min), and fi-nally 68°C (7 min). The RT-PCR products were observed on2% agarose gels.

Analysis of MMP-9 Activity by Zymography

Detection of MMP-9 secretion to the conditioned mediawas determined by zymography as described elsewhere(30). Briefly, sample volumes were adjusted to the numberof cells and loaded on 10% SDS-polyacrylamide gels co-polymerized with 1 mg/ml gelatin. After electrophoresis, thegels were extensively washed with 2.5% Triton X-100, incu-bated overnight at 37°C in appropriate enzyme buffer (50mM Tris, 2 mM CaCl2, 0.02% NaN3, pH 7.5), and thenstained with 0.1% Coomassie brilliant blue. The MMP-9 ac-tivity was visualized as transparent bands against a bluebackground.

Cell Invasion and Migration Assay

The invasion assay was performed essentially as describedpreviously (31). Briefly, 6.5-mm-diameter polycarbonate fil-ters of 8-�m pore size (Transwell, Costar, Badhoevedorp,The Netherlands) were coated with 20 �g of basement mem-brane Matrigel, dried, and reconstituted at 37°C withDMEM before use. The lower chambers were filled withconditioned medium of fibroblasts as chemoattractant, sup-plemented with 2 mg/ml BSA and 10 ng/ml bFGF. After abrief exposure to EDTA, the cells were collected and thenadded on the upper chamber (105 cells/chamber in DMEMcontaining 0.2 mg/ml BSA and 10% FCS). Apigenin orDMSO (control) was also added to the upper chamber. After24 hours of incubation at 37°C, the filters were removed andstained with a Diff-Quick kit (DADE). The number of cellsthat had spread on the lower surface of the filter was countedwith a microscope. The migration assay was performed asdescribed for the invasion study, except the polycarbonatefilters were not coated with Matrigel and migration time waslimited to five hours. Each invasion or migration experimentwas carried out in duplicate and repeated at least twice.

Cell Adhesion

Adhesion assays of tumor cells to basement membranewere performed in 96-wells plates (Costar) using substratefor cell endogenous phosphatases as previously described(32). The wells were coated overnight at 4°C with Matrigel(20 �g/100 �l) or 1% BSA as control. Nonspecific bindingsites were blocked with 1% BSA for one hour at 37°C andwashed three times with adhesion buffer (10 mM HEPES,140 mM NaCl, 5.6 mM glucose, 5.4 mM KCl, 2 mM CaCl2, 1mM MgCl2, 1 mM MnCl2, pH 7.4). The cells (5 × 104/well inadhesion buffer), pretreated or not by apigenin, were allowedto adhere for 20 minutes at 37°C. Then, nonadhered cellswere removed by washing three times with adhesion buffer.The substrate for phosphatases (p-nitrophenyl phosphate,Sigma) in acetate buffer, 0.1% Triton X-100, pH 5.5, wasadded to each well for one hour of incubation at 37°C. The re-action was stopped with 1 N NaOH, and absorbance wasmeasured at 405 nm. Each point was determined in triplicate.

Cell Proliferation Assay

Stock cultures were detached by trypsin, and cells wereseeded in 96-wells plates at 1.2 × 103 cells/well. Cells weregrown for 24 hours with 10% FCS culture medium, then fur-ther cultured for 30 hours in fresh medium with 10% FCS inthe presence of 0.2% DMSO (control) or various concentra-tions of apigenin. Then, 1 �Ci of [3H]thymidine (Amersham,Courtaboeuf, France) was added to each well, and the cellswere incubated for 18 hours. The cells were extensivelywashed, and cell-incorporated radioactivity was countedwith a beta counter. Each condition was run in triplicate.

Statistical Analysis

All experimental results were analyzed with one-wayanalysis of variance using Microsoft Excel (Microsoft, Red-mond, WA).

Results

Apigenin Regulates Protease Production

Inasmuch as cancer invasiveness is a complex mecha-nism involving pericellular proteolytic activity, we first ex-amined the regulation by apigenin of the production of u-PAand MMP-9. The u-PA secretion into the culture media wasinhibited by apigenin in a dose-dependent manner as demon-strated by enzyme-linked immunsorbent assay measure-ments (Figure 1A). Cell surface-bound u-PA, analyzed byflow cytometry, was also decreased in a dose-dependentmanner after apigenin treatment (Figure 1B, a–d). In con-trast, constitutive u-PAR expression was not affected byapigenin (Figure 1B, e). The decrease in u-PA expressionand the unaltered u-PAR expression were confirmed by

Vol. 39, No. 1 141

zymography and RT-PCR. When the culture media or theliving cells were layered onto a plasminogen-rich casein gel,a clear decrease in secreted and membrane-bound u-PA wasobserved in apigenin-treated cells (Figure 1C). Gel lysis wastotally inhibited in the presence of amiloride (not shown),demonstrating that the lysis was u-PA specific. However,the expression of u-PAR and its ability to bind u-PA werenot altered in these cells, inasmuch as the effect of apigeninwas reversed by preincubation of the treated cells with re-combinant scu-PA (Figure 1C). In addition, the RT-PCR re-sults show that the mRNA level of u-PA was downregulatedby apigenin, although only at 90 �M (10 �g/ml), while themRNA level of u-PAR was not modified (Figure 1D). It ap-peared therefore that reduced u-PA expression was corre-lated with a reduced cell-bound u-PA activity.

As shown by gelatin zymography, MDA-MB231 cellsconstitutively secreted MMP-9, which was highly stimu-

lated by the tumor promoter PMA (Figure 1E). Only theproenzyme form of MMP-9, with molecular mass of 92 kDa,could be detected in the conditioned media. The addition ofapigenin did not significantly affect the basal secretion ofMMP-9 but completely abolished PMA-induced MMP-9 se-cretion (Figure 1E).

Apigenin Affects Cell Migration and InvasionCapacities

We next examined the effect of apigenin on MDA-MB231 cell invasion through Matrigel, a reconstituted ex-tracellular matrix (ECM). As is shown in Figure 2A, 45.5�M (5 �g/ml) apigenin greatly reduced the invasiveness ofthese cells: 74% inhibition compared with control. When tu-mor cells were saturated with exogenous scu-PA, the inva-sion was increased in apigenin-treated and untreated cells,

142 Nutrition and Cancer 2001

Figure 1. Effects of apigenin on urokinase-plasminogen activator (u-PA) and u-PA receptor (u-PAR) expression and on matrix metalloproteinase-9 (MMP-9)expression. A: quantification by enzyme-linked immunosorbent assay of u-PA secreted by MDA-MB231 cells during 24 h of incubation with apigenin. Valuesare means ± SD of duplicate results from 3 experiments. B: analysis by flow cytometry of MDA-MB231 cells pretreated for 24 h with apigenin at 2.5 (a), 5 (b),10 (c), and 20 (d) �g/ml (22.8, 45.5, 90, and 180 �M) for cell surface expression of u-PA and typical profile for cell-surface expression of u-PAR at 2.5–20�g/ml (22.8–180 �M) apigenin (e). Values are a representative example of 3 independent experiments. C: inhibition by apigenin of u-PA secretion into mediumand of cell surface u-PA binding as demonstrated by zymography. For 1 set of experiments, cells were incubated with recombinant single-chain u-PA (scu-PA)to saturate u-PA binding sites at cell surface before deposit on gel. D: agarose gel electrophoresis of reverse transcription-polymerase chain reaction products ofu-PA and u-PAR after mRNA isolation from control cells treated with dimethyl sulfoxide (1) and cells treated for 24 h with 2.5 (2), 5 (3) and 10 (4) �g/ml (0,22.8, 45.5, and 90 �M) apigenin. E: zymography analysis of MMP-9 secretion on media cultured for 24 h with phorbol 12-myristate 13-acetate (PMA). Vol-umes loaded were adjusted to number of cells. In C and E, typical results are shown for �4 repeated experiments.

although invasion was still reduced in the presence of api-genin. We have also shown that the apigenin-treated cellshave a reduced motility, as demonstrated in a Matrigel-freemigration system. The migration was inhibited by ~15% at22.8 �M (2.5 �g/ml) and 45% at 45.5 �M (5 �g/ml), as shownin Figure 2B. As for invasion, preincubation with scu-PAstimulated cell migration in apigenin-treated and untreatedcells. However, the inhibition by apigenin was still evident atthe same proportions as without scu-PA (Figure 2B).

Apigenin Affects Cell Adhesion to Matrigel

Apigenin has been shown to inhibit expression of adhe-sion proteins in endothelial cells by an action on NF-�Btranscriptional activation (14). Our results show that theability of apigenin-treated cells to adhere to Matrigel was re-duced in a dose-dependent manner, demonstrating 50% inhi-bition at 22.8 �M (2.5 �g/ml) apigenin (Figure 3). However,preincubation with the cells was necessary, inasmuch asapigenin had no effect when added only to the adhesion as-say (not shown). Saturating the cells with u-PA before theadhesion assay did not abrogate this inhibitory effect on celladhesion (not shown).

Apigenin Affects Cell Proliferation and Induces CellCycle Arrest at the G2-M Phase

Analysis of the [3H]thymidine incorporation method hasshown that apigenin markedly reduced the proliferation ofMDA-MB231 cells in a dose-dependent manner (Figure4A). We observed a growth inhibition of 20% at 11.4 �M(1.25 �g/ml) apigenin, reaching 60% inhibition at 90 �M(10 �g/ml). The half-maximal inhibitory concentration was~36.4 �M (4 �g/ml). We next explored using flow cytom-etry analysis of the cell cycle alterations associated with thisgrowth inhibition. After treatment with apigenin for 48hours, the cells were incubated with propidium iodide to an-alyze their DNA content. The results in Figure 4B suggestthat apigenin causes a G2-M cell cycle arrest, since the num-ber of cells increased at the G2-M phase and decreased at theG1 phase.

Vol. 39, No. 1 143

Figure 2. Apigenin inhibits cell invasion and migration. A: invasionthrough Matrigel. MDA-MB231 cells were added in upper chamber withapigenin (5 �g/ml, i.e., 45.5 �M) or dimethyl sulfoxide as control (solidbars). Cells were preincubated with scu-PA to saturate u-PA-binding siteson cell surface (open bars). Assay time was 24 h. Results are expressed astotal number of cells counted in 9 fields for each membrane. Values aremeans ± SD of 2 or 3 separate experiments each carried out in duplicate. B:effect of apigenin on migration of MDA-MB231 cells (0, 2.5, 5, and 10�g/ml, i.e., 0, 22.8, 45.5, and 90 �M apigenin). Cells were allowed to mi-grate through a polycarbonate membrane not coated with Matrigel. Assaytime was 5 h. Results are expressed as total number of cells counted in 9fields for each membrane. Values are means ± SD of 3 separate experimentseach carried out in duplicate. *, Significantly different from untreated con-trol without scu-PA (p � 0.01); **, significantly different from untreatedcontrol with scu-PA (p � 0.05); ***, significantly different from untreatedcontrol with scu-PA (p � 0.01).

Figure 3. Inhibitory effect of apigenin on MDA-MB231 cell adhesion toMatrigel. Pretreated cells were allowed to adhere for 20 min, andabsorbance was measured at 405 nm. Values are means ± SD of 3 separateexperiments each carried out in triplicate.

Apigenin Affects MDA-MB231 Cell Morphology

The MDA-MB231 cells were exposed to various concen-trations (0–180 �M, i.e., 0–20 �g/ml) of apigenin, and thecell morphology was observed after 24 hours of treatment ina serum-free medium (Figure 5). Morphological changeswere clearly observed at 45.5 �M (5 �g/ml) apigenin and be-came very pronounced at 90 and 180 �M (10 and 20 �g/ml),since DMSO alone has no effect. The cells lost the smoothsurface boundary, became irregular, and developed dendrite-like structures with extended arborization. The morphologi-cal changes were reversed on apigenin removal. Inasmuchas a role for MAPKs in the initiation of microtubular reorga-nization has been demonstrated (33), such morphologicalchanges in treated cells could be related to the apigeninMAPK inhibitor activity.

Discussion

The preventive effect of flavonoids on cancer incidencemakes them attractive candidates for the development ofanticancer agents. Apigenin, one of flavonoids, has been de-scribed as a chemopreventive compound (6,7) and as an in-hibitor of certain signal transduction pathways (14,17). Wedemonstrated here that apigenin affects proteolytic activityand invasion capacity of breast tumor cells in addition toother reported aspects, such as inhibition of cell prolifera-tion.

First, apigenin treatment caused a remarkable decrease atthe mRNA and protein levels in the expression of u-PA, awell-recognized activator of pericellular proteolysis in mi-grating cells (34). In particular, we show an important de-crease in cell membrane-bound u-PA activity. Because the

144 Nutrition and Cancer 2001

Figure 4. A: antiproliferative effect of apigenin on MDA-MB231 cells. Cells were treated with 0, 1.25, 2.5, 5, and 10 �g/ml (0, 11.4, 22.8, 45.5, and 90 �M)apigenin for 48 h, and incorporated [3H]thymidine was counted. Values are means ± SD of 2 separate experiments each carried out in triplicate. B: apigenincauses a G2-M cell cycle arrest in tumor cells. DNA content of MDA-MB231 cells was measured by flow cytometry 48 h after treatment with 0, 5, 10, and 20�g/ml (0, 45.5, 90, and 180 �M) apigenin. A representative experiment is shown.

level of u-PAR on the cell surface was not modified byapigenin treatment, as shown by flow cytometry, zymog-raphy, and RT-PCR, the decrease in cell-associated u-PAcould be directly attributed to the reduced u-PA expression. Ithas been demonstrated that MAPK pathways were involvedin the regulation of u-PA gene expression (35). Thus we ob-served an inhibition of u-PA expression that is consistent withapigenin being an MAPK signaling pathway inhibitor.

MMP-9 secretion was also affected by apigenin treat-ment. Under standard culture condition, MDA-MB-231cells secreted a detectable level of MMP-9, which was up-regulated by PMA treatment. Our results show that apigenin,although without effect on the basal secretion of MMP-9, to-tally abolished the response to PMA stimulation.

Our results are in agreement with the observations thatapigenin can block the EGF-induced upregulation of MMP-9 secretion by inhibiting MAPK signaling pathways inbreast epithelial tumor cells (27). A recent study demon-strated that the invasiveness and production of MMP-9 werecorrelated and that they were regulated via protein kinase C,an activator of the MAPK pathway (36). The inhibition ofPMA-induced upregulation of MMP-9 may therefore alsoresult from blockage of the MAPK signaling pathway, inas-much as the tumor promoter PMA activates all three MAPKsubfamilies. In a human squamous cell carcinoma cell line,PMA-enhanced MMP-9 secretion and in vitro invasivenesswere associated with a strong activation of the p38 MAPKand its downstream target, MAPK-activated protein kinase-2. Inhibition of p38 MAPK by a specific inhibitor resulted incomplete reduction of the PMA-induced MMP-9 secretion,but not of u-PA secretion (37). In a recent study, apigeninsignificantly inhibited EGF-stimulated release of prosta-glandin E2, and this inhibition was also p38 dependent (38).

It may be therefore that the inhibition of the PMA-inducedMMP-9 secretion by apigenin observed in our study is alsomediated by a p38-dependent pathway. Moreover, MMP-9gene expression was regulated by transcription factors bind-ing the activator protein-1 (AP-1) binding site (the TRE ele-ment) (39), and MAPKs have been suggested to be a criticalmediator in AP-1-associated changes (40). This is supportedby the fact that u-PA and MMP-9 gene expression was de-pendent on NF-�B (15,16,28), that NF-�B interacts withtranscription factors such as AP-1 (41), and that MAPK sig-naling pathways regulate NF-�B activity (19,20). Althoughapigenin was shown to inhibit some steps in the signaltransduction pathway, an apigenin receptor has not beenidentified.

Apigenin reduced tumor cell invasiveness through Ma-trigel, and this seems to correlate with the decrease in prote-ase production. The anti-invasive effects of genistein, anisoconformer of apigenin, was also shown to be mediatedthrough a downregulation of MMP-9 and an upregulation ofTIMP-1 (42).

However, the reduced invasiveness and migration ofMDA-MB 231 cells induced by apigenin was not simply dueto insufficient protease-dependent matrix degradation. Al-though cell surface saturation by exogenous scu-PA did in-crease the invasion capacity of the apigenin-treated cells, itwas still reduced compared with that of scu-PA-saturatedcontrol cells.

In addition, scu-PA saturation could still stimulate cellmigration in the absence of matrix regardless of apigenintreatment. It has already been shown that u-PA was able topromote cell motility independently of its capacity to inducematrix proteolysis (43), and u-PA was also shown to beinvolved in signal transduction pathways leading to modifi-cations of cytoskeleton organization. Apigenin appearstherefore to affect MDA-MB231 cell invasion by inhibitingu-PA-related matrix proteolysis and cell motility.

Indeed, apigenin also altered MDA-MB231 adhesive be-havior in ECM proteins, and this inhibitory effect on cell ad-hesion seems not to be related to the u-PA system, inasmuchas it was not influenced by saturating the cells with scu-PA(not shown). Nevertheless, u-PA was found to be implicatedin cellular adhesion and migration (44). The reason for thisreduced adhesion is unclear, and further study is needed toidentify the cell surface adhesive molecules and the matrixproteins involved.

Our results reveal that apigenin, which inhibits the ex-pression of proteases involved in degradation of the ECM,may also regulate other cellular activities associated with theinvasive phenotype, including cell migration and cell-matrixadhesion. The decrease in invasiveness may also be due tomodification of the ultrastructure of these cells, which wasnot restored by u-PA. In addition, apigenin treatment re-sulted in a dose-dependent tumor cell growth inhibition, aneffect that was also reported for other cancer cell lines(10,45), and this antiproliferative action was correlated tocell cycle arrest in the G2-M phase (9). However, this effecton cell growth was not specific to apigenin, since other

Vol. 39, No. 1 145

Figure 5. Morphological appearance of MDA-MB231 tumor cells in se-rum-free medium after 24 h of treatment with apigenin: 0.2% dimethylsulfoxide as control (a) and 1.25, 2.5, 5, 10, and 20 �g/ml (11.4, 22.8, 45.5,90, and 180 �M) apigenin. Scale bar, 20 �m.

flavonoids have also been shown to have antiproliferativeeffects (46).

In conclusion, our findings indicate that apigenin inhibitstumor cell invasiveness by downregulating protease expres-sion, cell migration, cell adhesion to the ECM, and cell pro-liferation. This effect is not tumor cell specific, sinceapigenin also inhibited the growth of endothelial cells andangiogenesis (47,48). Our work supports a direct anticancereffect of apigenin, as a pleiotropic effector, in addition to itseffect on angiogenesis. Considering the broad type ofapigenin activities in our results, we suggest that interferingwith the MAPK signaling pathway and inhibiting the ex-pression of NF-�B-regulated genes represent a promisingway to control growth and invasiveness of breast tumors.

Acknowledgments and Notes

The authors thank E. Savariau, R. Nancel, and B. Boursin for technicalassistance (photography). Financial support was provided by the InstitutNational de la Santé et de la Recherche Médicale, the Centre National dela Recherche Scientifique, the Institut d’Hématologie, the Université ParisVII, the Institut Gustave Roussy, the Association pour la Recherche sur leCancer, the Ligue Nationale contre le Cancer, and la Fondation de France.Address correspondence to H. Lu, INSERM U553, Institut d’Hématol-ogie, Hôpital Saint-Louis, Université Paris VII, 1 Ave Claude Vallefaux,75475, Paris, France. Phone: 33 (1) 53 72 40 26. FAX: 33 (1) 53 72 40 27.E-mail: helu@chu-stlouis.fr.

Submitted 21 June 2000; accepted in final form 1 November 2000.

References

1. Adlercreutz, H: Western diet and Western diseases: some hormonaland biochemical mechanisms and associations. Scand Clin Lab InvestSuppl 201, 3–23, 1990.

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

3. Middleton, E: The flavonoids. Trends Pharmacol Sci 277, 375–381,1984.

4. Birt, DF, Mitchell, D, Gold, B, Pour, P, Pinch, HC, et al.: Inhibition ofultraviolet light-induced skin carcinogenesis in SKH-1 mice by api-genin, a plant flavonoid. Anticancer Res 17, 85–91, 1997.

5. Wang, C, and Kurzer, MS: Phytoestrogen concentration determines ef-fects on DNA synthesis in human breast cancer cells. Nutr Cancer 28,236–247, 1997.

6. Kuo, ML, Lee, KC, and Lin, JK: Genotoxicities of nitropyrenes andtheir modulation by apigenin, tannic acid, ellagic acid and indole-3-carbinol in the Salmonella and CHO systems. Mutat Res 270, 87–95,1992.

7. Chaumontet, C, Bex, V, Gaillard-Sanchez, I, Seillan-Heberden, C,Suschetet, M et al.: Apigenin and tangeretin enhance gap junctionalintercellular communication in rat liver epithelial cells. Carcinogene-sis 15, 2325–2330, 1994.

8. MacVean, M, Xiao, H, Isobe, K, and Pelling, JC: Increase in wild-typep53 stability and transactivational activity by the chemopreventiveagent apigenin in keratinocytes. Carcinogenesis 21, 633–639, 2000.

9. Sato, F, Matsukawa, Y, Matsumoto, K, Nishino, H, Sakai, T, et al.:Apigenin induces morphological differentiation and G2-M arrest in ratneuronal cells. Biochem Biophys Res Commun 204, 578–584, 1994.

10. Fotsis, T, Pepper, MS, Aktas, E, Breit, S, Rasku, S, et al.: Flavonoids,dietary-derived inhibitors of cell proliferation and in vitro angiogene-sis. Cancer Res 57, 2916–2921, 1997.

11. Wang, W, Heideman, L, Chung, CS, Pelling, JC, Koehler, KJ, et al.:Cell-cycle arrest at G2/M and growth inhibition by apigenin in humancolon carcinoma cell lines. Mol Carcinog 28, 102–110, 2000.

12. Wang, IK, Lin-Shiau, SY, and Lin, JK: Induction of apoptosis byapigenin and related flavonoids through cytochrome c release and acti-vation of caspase-9 and caspase-3 in leukaemia HL-60 cells. Eur JCancer 35, 1517–1525, 1999.

13. Caitagirone, S, Rossi, C, Poggi, A, Ranelletti, FO, Natali, PG, et al.:Flavonoids apigenin and quercitin inhibit melanoma growth and meta-static potential. Int J Cancer 87, 595–600, 2000.

14. Gerritsen, ME, Carley, WW, Ranges, GE, Shen, CP, Phan, SA, et al.:Flavonoids inhibit cytokine-induced endothelial cell adhesion proteingene expression. Am J Pathol 147, 278–292, 1995.

15. Wilhelm, SM, Collier, IE, Marmer, BL, Eisen, AZ, Grant, GA, et al.:SV40-transformed human lung fibroblasts secrete a 92-kDa type IVcollagenase, which is identical to that secreted by normal humanmacrophages. J Biol Chem 264, 17213–17221, 1989.

16. Sato, H, and Seiki, M: Regulatory mechanism of 92 kDa type IV colla-genase gene expression, which is associated with invasiveness of tu-mor cells. Oncogene 8, 395–405, 1993.

17. Kuo, ML, and Yang, NC: Reversion of v-H-ras-transformed NIH 3T3cells by apigenin through inhibiting mitogen-activated protein kinaseand its downstream oncogene. Biochem Biophys Res Commun 212,767–775, 1995.

18. Yin, F, Giuliano, AE, and Van Herle, AJ: Signal pathways involved inapigenin inhibition of growth and induction of apoptosis of humananaplastic thyroid cancer cells (ARO). Anticancer Res 19, 4297–4303,1999.

19. Davis, RJ: The mitogen-activated protein kinase signal transductionpathway. J Biol Chem 268, 14553–14556, 1993.

20. Cobb, MH, and Goldsmith, EJ: How MAP kinases are regulated. J BiolChem 270, 14843–14846, 1995.

21. Sivaraman, VS, Wang, H, Nuovo, GJ, and Malbon, CC: Hyperex-pression of mitogen-activated protein kinase in human breast cancer. JClin Invest 99, 1478–1483, 1997.

22. Mignatti, P, and Rifkin, DB: Plasminogen activators and angiogenesis.Curr Top Microbiol Immunol 213, 33–50, 1996.

23. Schmitt, M, Harbeck, N, Thomssen, C, Wilhelm, O, Magdolen, V, etal.: Clinical impact of the plasminogen activation system in tumor in-vasion and metastasis: prognostic relevance and target for therapy.Thromb Haemost 78, 285–296, 1997.

24. Andreasen, PA, Kjoller, L, Christensen, L, and Duffy, MJ: The uroki-nase-type plasminogen activator system in cancer metastasis: a review.Int J Cancer 72, 1–22, 1997.

25. Stetler-Stevenson, WG: Type IV collagenases in tumor invasion andmetastasis. Cancer Metastasis Rev 9, 289–303, 1990.

26. Bae, SN, Arand, G, Azzam, H, Pavasant, P, Torri, J, et al.: Molecularand cellular analysis of basement membrane invasion by human breastcancer cells in Matrigel-based in vitro assays. Breast Cancer Res Treat24, 241–255, 1993.

27. Reddy, KB, Krueger, JS, Kondapaka, SB, and Diglio, CA: Mitogen-activated protein kinase (MAPK) regulates the expression of progela-tinase B (MMP-9) in breast epithelial cells. Int J Cancer 82, 268–273,1999.

28. Wang, W, Abbruzzese, JL, Evans, DB, and Chiao, P: Overexpressionof urokinase-type plasminogen activator in pancreatic adenocarcinomais regulated by constitutively activated RelA. Oncogene 18, 4554–4563, 1999.

29. Lu, H, Yeh, P, Guitton, JD, Mabilat, C, Desanlis, F, et al.: Blockage ofthe urokinase receptor on the cell surface: construction and character-ization of a hybrid protein consisting of the N-terminal fragment of hu-man urokinase and human albumin. FEBS Lett 356, 56–59, 1994.

30. Belloc, C, Lu, H, Soria, C, Fridman, R, Legrand, Y, et al.: The effect ofplatelets on invasiveness and protease production of human mammarytumor cells. Int J Cancer 60, 413–417, 1995.

146 Nutrition and Cancer 2001

31. Albini, A, Iwamoto, Y, Kleinman, HK, Martin, GR, Aaronson, SA, etal.: A rapid in vitro assay for quantitating the invasive potential of tu-mor cells. Cancer Res 47, 3239–3245, 1987.

32. Trochon, V, Mabilat-Pragnon, C, Bertrand, P, Legrand, Y, Soria, J, etal.: Hyaluronectin blocks the stimulatory effect of hyaluronan-derivedfragments on endothelial cells during angiogenesis in vitro. FEBS Lett418, 6–10, 1997.

33. Gotoh, Y, Nishida, E, Matsuda, S, Shiina, N, Kosako, H, et al.: In vitroeffects on microtubule dynamics of purified Xenopus M phase-acti-vated MAP kinase. Nature 349, 251–254, 1991.

34. Chapman, HA: Plasminogen activators, integrins, and the coordinatedregulation of cell adhesion and migration. Curr Opin Cell Biol 9,714–724, 1997.

35. Cirillo, G, Casalino, L, Vallone, D, Caracciolo, A, de Cesare, D, et al.:Role of distinct mitogen-activated protein kinase pathways and coop-eration between Ets-2, ATF-2, and Jun family members in humanurokinase-type plasminogen activator gene induction by interleukin-1and tetradecanoyl phorbol acetate. Mol Cell Biol 19, 6240–6252, 1999.

36. Hirata, M, Tsuchida, A, Iwao, T, Sasaki, T, Matsubara, K, et al.:Cholecystokinin regulates the invasiveness of human pancreatic cancercell lines via protein kinase C pathway. Int J Oncol 14, 1129–1135,1999.

37. Simon, C, Goepfert, H, and Boyd, D: Inhibition of the p38 mitogen-activated protein kinase by SB 203580 blocks PMA-induced Mr92,000 type IV collagenase secretion and in vitro invasion. Cancer Res58, 1135–1139, 1998.

38. Choudhury, QG, McKay, DT, Flower, RJ, and Croxtall, JD: Investiga-tion into the involvement of phospholipases A2 and MAP kinases inmodulation of AA release and cell growth in A549 cells. Br J Pharma-col 131, 255–265, 2000.

39. Himelstein, BP, Lee, EJ, Sato, H, Seiki, M, Muschel, RJ, et al.: Tumorcell contact-mediated transcriptional activation of the fibroblast matrixmetalloproteinase-9 gene: involvement of multiple transcription fac-

tors including Ets and an alternating purine-pyrimidine repeat. ClinExp Metastasis 16, 169–177, 1998.

40. Guo, Z, Du, X, and Iacovitti, L: Regulation of tyrosine hydroxylasegene expression during transdifferentiation of striatal neurons: changesin transcription factors binding the AP-1 site. J Neurosci 18, 8163–8174, 1998.

41. Stein, B, Baldwin, JR, Ballard, DW, Greene, WC, Angel, P, et al.:Cross-coupling of the NF-�B p65 and Fos/Jun transcription factorsproduces potentiated biological function. EMBO J 12, 3879–3891,1993.

42. Shao, ZM, Wu, J, Shen, ZZ, and Barsky, SH: Genistein exerts multiplesuppressive effects on human breast carcinoma cells. Cancer Res 58,4851–4857, 1998.

43. Waltz, DA, Natkin, LR, Fujita, RM, Wei, Y, Chapman, HA, et al.:Plasmin and plasminogen activator inhibitor type 1 promote cellularmotility by regulating the interaction between the urokinase receptorand vitronectin. J Clin Invest 100, 58–67, 1997.

44. Conese, M, and Blasi, F: The urokinase/urokinase-receptor system andcancer invasion. Baillieres Clin Haematol 8, 365–389, 1995.

45. Hirano, T, Oka, K, and Akiba, M: Antiproliferative effects of syntheticand naturally occurring flavonoids on tumor cells of the human breastcarcinoma cell line, ZR-75-1. Res Commun Chem Pathol Pharmacol64, 69–78, 1989.

46. Monti, E, and Sinha, BK: Antiproliferative effect of genistein andadriamycin against estrogen-dependent and -independent humanbreast carcinoma cell lines. Anticancer Res 14, 1221–1226, 1994.

47. Liu, D, Pearlman, E, Diaconu, E, Guo, K, Mori, H, et al.: Expression ofhyaluronidase by tumor cells induces angiogenesis in vivo. Proc NatlAcad Sci USA 93, 7832–7837, 1996.

48. Trochon, V, Blot, E, Cymbalista, F, Engelmann, C, Tang, RP, et al.:Apigenin inhibits endothelial cell proliferation in G2-M phase whereasit stimulates smooth muscle cells by inhibiting p21 and p27 expression.Int J Cancer 85, 691–696, 2000.

Vol. 39, No. 1 147