Resveratrol Inhibits Intestinal Tumorigenesis and ModulatesHost-Defense-Related Gene Expression in an Animal

Model of Human Familial Adenomatous Polyposis

Yann Schneider, Benoit Duranton, Francine Gossé,René Schleiffer, Nikolaus Seiler, and Francis Raul

Abstract: We studied the effect of oral administration ofresveratrol, a natural constituent of grapes, on tumorigene-sis in Min mice. Min mice are congenic mice geneticallypredisposed to develop intestinal tumors as a result of a mu-tation of the Apc gene. Resveratrol (0.01% in the drinkingwater containing 0.4% ethanol) was administered for sevenweeks to Min mice starting at five weeks of age. The controlgroup was fed the same diet and received water containing0.4% ethanol. Resveratrol prevented the formation of colontumors and reduced the formation of small intestinal tumorsby 70%. Comparison of the expression of 588 genes in thesmall intestinal mucosa showed that resveratrol downregu-lated genes that are directly involved in cell cycle progres-sion or cell proliferation (cyclins D1 and D2, DP-1transcription factor, and Y-box binding protein). In addi-tion, resveratrol upregulated several genes that are involvedin the recruitment and activation of immune cells (cytotoxicT lymphocyte Ag-4, leukemia inhibitory factor receptor, andmonocyte chemotactic protein 3) and in the inhibition ofthe carcinogenic process and tumor expansion (tumor sus-ceptibility protein TSG101, transforming growth factor-�,inhibin-� A subunit, and desmocollin 2). Our data highlightthe complexity of the events associated with intestinal tu-morigenesis and the multiplicity of the molecular targets ofresveratrol. The high potency and efficacy of resveratrolsupport its use as a chemopreventive agent in the manage-ment of intestinal carcinogenesis.

Introduction

Resveratrol (trans-3,4�,5-trihydroxystilbene) is a naturalantifungal agent found in grapes and a variety of medicinalplants (1). Because of its high concentration in grape skin,significant amounts of resveratrol are present in wines. Ow-ing to the procedure, �0.1 mg/l of resveratrol is found inwhite wines. In contrast, red wines contain up to 8 mg/l (2).

It has been proposed that this high resveratrol content mayexplain in part the apparent ability of moderate consumptionof red wine to reduce the risk of cardiovascular disease (3).More recently, there has been increasing interest in res-veratrol because of reports on its anticarcinogenic effects asassessed by analyzing the major stages (initiation, promo-tion, and progression) of malignant transformation (4). Themolecule has antioxidant and anti-inflammatory properties,and it seems able to inhibit cyclooxygenase and hydroper-oxidase activities (4). Recent reports have demonstrated thatresveratrol is able to inhibit ribonuclease reductase andDNA synthesis (5) and to arrest cell cycle progression at theS/G2 or G1/S phase transition (6,7).

On the basis of the above-reported findings and thepotential relevance of resveratrol as a promising chemopre-ventive component of human diets, we investigated the ef-fects of orally administered resveratrol on the progression ofcarcinogenesis in Min (multiple intestinal neoplasia) micegenetically predisposed to develop intestinal tumors. Min isa mutant allele of the murine Apc (adenomatous polyposiscoli) locus, encoding a nonsense mutation at codon 850. Minmice develop a variety of tumors and lesions similar to thoseseen in humans carrying germline mutations in the Apc geneand developing familial adenomatous polyposis, an inher-ited colon cancer syndrome (8). Apc is the most frequentlymutated gene in sporadic colon tumors in humans. In con-trast to humans, where the colon is more heavily involved,the small intestine is the preferred site of tumor developmentin Min mice.

We report data showing that oral administration of res-veratrol initiates a dramatic decrease in the number of tu-mors in the small intestine and completely suppresses tumorformation in the colon of Min mice. By analyzing differen-tial gene expression patterns in the mucosa of the small in-testine, we were able to show that resveratrol downregulateda panel of genes directly involved in the progression oftumorigenesis and upregulated several genes controlling ge-

NUTRITION AND CANCER, 39(1), 102–107Copyright © 2001, Lawrence Erlbaum Associates, Inc.

The authors are affiliated with the Laboratoire du Contrôle Métabolique et Nutritionnel en Oncologie Digestive de l’Université Louis Pasteur, Institut deRecherche contre les Cancers de l’Appareil Digestif, 67091 Strasbourg-Cedex, France.

nome stability and cellular differentiation and favoring theactivation of antitumoral defense mechanisms.

Materials and Methods

Animals and Diets

The experiments were conducted according to the Na-tional Research Council Guide for the Care and Use of Lab-oratory Animals with the authorization (no. 00573) of theFrench Ministry of Agriculture.

C57BL/6J-ApcMin male mice were five weeks old attime of purchase from Jackson Laboratories (Bar Har-bor, ME). They were housed under standardized conditions:22°C, 60% relative humidity, 12:12-hour light-dark cycle,and 20 air changes per hour. The mice were randomly di-vided into two groups and fed ad libitum a standard pelletteddiet (reference no. 105, UAR, Villemoisson, France) that hasa crude protein content of 22%, a crude fat content of 5%,and a crude fiber content of 4%.

One group (n = 10) of five-week-old mice received drink-ing water containing 0.01% resveratrol (Sigma-Aldrich,Saint Quentin Fallavier, France). This amount of resveratrolis ~10 times the amount found in 1 liter of Pinot Noir redwine. Resveratrol was first dissolved in 0.4 ml of absoluteethanol and added to 100 ml of drinking water. The controlgroup (n = 10) received the same volume of drinking watercontaining 0.4% ethanol.

All mice had free access to the drinking solution between5 PM and 8 AM. The mice consumed 3–4 ml of the drinkingfluid daily. The daily consumption of resveratrol was0.3–0.4 mg/mouse. After seven weeks, the entire colon andthe jejunoileum (from the ligament of Treitz to the cecum)were collected under anesthesia and flushed with ice-cold0.9% NaCl.

Assessment of the Number of Tumors in theJejunoileum and Colon

Seven mice of each group were used for the morphologi-cal studies. The jejunoileum and colon from each mousewere scored for tumors by examination of 5-cm intestinalsections along the entire jejunoileum and colon. Tumorswere counted under a dissecting microscope at ×10 magnifi-cation after the intestine had been cut open and pinned flat.All tumor counts were performed by a single blinded ob-server.

Semiquantitative Monitoring of Gene ExpressionPatterns With cDNA Expression Arrays

Of each treatment group, the mucosa of the jejunoilealsegment of three mice was collected separately by scrapingthe normal-appearing mucosal surface with a sterile glassslide; tumor tissue was not present in the collected samples.

The mucosal samples were weighed in a prechilled steriletube before addition of a denaturating solution (2.7 M guani-dine thiocyanate, 1.3 M ammonium thiocyanate, 0.1 M so-dium acetate, pH 4.0). Total RNA was extracted using theAtlas Pure RNA Isolation Kit (Clontech Laboratories, PaloAlto, CA) following the user’s manual. The purified totalRNA (2 �g/sample) was used for cDNA synthesis by reversetranscription using the Atlas Mouse cDNA Expression Arrayprocedure and [�-32P]dATP (Clontech Laboratories). The At-las Mouse cDNA Expression Array includes 588 mousecDNAs spotted in duplicate on a positively charged nylonmembrane. The array contains coding sequence-specificpolymerase chain reaction products derived from com-pletely sequenced cDNA clones. All genes (cDNA) presentin the commercially available mouse arrays are identifiedon a website (www.clontech.com). These genes representseveral different functional classes, including cell cycle reg-ulators, tumor suppressors, transcription factors, growth fac-tors, chemokines, and cytokines. Plasmid and bacteriophageDNAs are included as negative controls to confirm hybrid-ization specificity, along with several housekeeping cDNAsas positive controls for normalizing mRNA abundance. The32P-labeled cDNA probes prepared from 2 �g of total RNAisolated from the intestinal mucosa of control and resvera-trol-treated mice were hybridized overnight on separateidentical Atlas Mouse cDNA Expression Array membranes.After a high-stringency wash, the hybridization patternswere analyzed and quantified using a PhosphorImager(model 445SI, Molecular Dynamics, Sunnyvale, CA) andImageQuant software. Transcript abundance was normal-ized using a set of housekeeping genes.

Three control and three resveratrol-treated mice were ar-bitrarily paired and compared on separate membranes. Thechanges in gene expression were well reproduced.

The method is semiquantitative. In a given tissue it al-lows the comparison between two populations of mRNAthat reflect changes in gene expression. To exclude technicalartifacts, we considered only changes in mRNA abundanceexceeding 1.6-fold variations relative to controls. Underthese conditions, differences were consistently recorded for18 of 588 genes in samples obtained from resveratrol-treatedand control animals.

Statistics

Values are means ± SE. Statistical differences betweengroups were evaluated by one-way analysis of variance pluscomparison of means. Differences were considered signifi-cant at p � 0.01.

Results

Drinking fluid and diet consumption was comparable inboth groups of mice. Initial mean body weight was 25 ± 0.3g; it reached 27 ± 0.6 g in the control group and 29 ± 0.4 g inthe resveratrol-treated group at the end of the seven-week

Vol. 39, No. 1 103

treatment. No toxic effects were observed in the mice receiv-ing resveratrol.

Effect of Resveratrol on Tumor Formation in theJejunoileum and Colon

The number of adenomas was reduced by 70% in Minmice receiving 0.01% resveratrol in the drinking water forseven weeks (Table 1). Only a few tumors were present inthe colon of nontreated Min mice; the colon of mice treatedwith resveratrol was tumor free.

Changes in Gene Expression in the Small IntestinalMucosa After Resveratrol Treatment

The administration of resveratrol led to dramatic changesin the expression of 18 genes as measured by the abundanceof their mRNAs among 588 genes tested simultaneously us-ing the Atlas Mouse cDNA Expression Array procedure(Figure 1). Resveratrol downregulated the expression ofgenes controlling cell cycle progression and transcription.As shown in Figure 2, compared with control Min mice, themice receiving resveratrol exhibited a highly significant de-crease in the expression of 5 genes of a panel of ~180 genesinvolved in the regulation of cell cycle and the expression oftranscription factors and DNA binding proteins: RNA poly-merase 1 termination factor TTF1 (�3-fold), cyclins D1 andD2 (�2.6- and �2-fold, respectively), DP-1 cell cycle regula-tory transcription factor (�2-fold), and YB-1 DNA bindingprotein (�2.6-fold).

In contrast, administration of resveratrol initiated the up-regulation of 13 genes of a panel of ~250 genes involved incontrol of tumor growth, cell differentiation, or control ofthe immune response. Among these genes, TSG101 tumorsusceptibility protein (+1.8-fold), Mas protooncogene (+1.8-fold), MAP kinase (+2-fold), homeobox protein 4.2 (+1.7-fold), monocyte chemoattractant protein 3 (+2.6-fold),leukemia inhibitory factor receptor (+2-fold), glutamate re-ceptor NMDA2B (+5.5-fold), cytotoxic T lymphocyte Ag-4(+5.5-fold), desmocollin 2 (+4-fold), follistatin (+2.3-fold),thrombopoietin (+2-fold), and two genes of the transforminggrowth factor-� (TGF-�) family, inhibin-� A subunit (+2.3-

fold) and TGF-� (+1.7-fold), were most prominently stimu-lated.

Discussion

Several epidemiological studies demonstrated that di-etary factors inhibit the development of many types of can-cers through a chemopreventive mechanism. On the basis ofin vitro and some in vivo studies (3–7,9), the antimutagenicand chemopreventive properties of resveratrol are now in-creasingly recognized. It was recently shown that resveratrolexerts antiproliferative effects on human colonic cancercells (10) and has a protective role in colorectal aberrantcrypt foci formation in rats (11). In this report, we have in-vestigated the effects of oral administration of resveratrol totransgenic Min mice genetically predisposed to develop in-testinal tumors, an animal model of human familial adenom-atous polyposis. When given orally, resveratrol inhibited thecarcinogenic process in small intestine and colon of Minmice, despite the constitutive mutation of the tumor-sup-pressor gene Apc, which is an early and frequent event in theprocess of intestinal carcinogenesis. This indicates that aninitiated carcinogenic process might be suppressed by a che-mopreventive agent present in the diet.

Several cellular and molecular targets that may explainthe chemopreventive effects of resveratrol have been identi-fied. The cell cycle machinery and the pathways of reactiveoxidants are mostly recognized as sites of resveratrol action(5–7,12). By comparing the expression of 588 selected genesin the intestinal mucosa of control Min mice with that inresveratrol-treated Min mice, we were able to show that res-veratrol administration led to major changes in the expressionof 18 genes, with variations exceeding 60% compared withcontrols. Resveratrol downregulated the expression of cyclinsD1 and D2, which are directly involved in cell cycle progres-sion. They are generally stimulated during malignancy (13)and repressed by anticancerous phytochemicals (14). Res-veratrol reduced significantly the expression of transcriptionfactors, including the DP-1 transcription factor, which isstrongly involved in the control of cell proliferation (15).

The present results confirm the antiproliferative proper-ties of resveratrol observed in vitro on several cancer celllines (4–7,10). Resveratrol has also been shown to inhibitDNA synthesis and to arrest the cell division cycle at theS/G2 phase transition in human intestinal cancer cells (10).

In the present study we show that resveratrol repressedthe expression of the Y-box binding protein, which binds toinverted CCAAT box sequences that are present in the pro-moter region of many genes. Interestingly, it was reported re-cently that Y-box binding protein is overexpressed in humancancer cells resistant to chemotherapy (16). By downregu-lating the expression of this transcription factor, resveratrolmay lower the resistance of cancer cells to anticancer effec-tors, such as cytokines and chemokines generated throughthe stimulation of the immune response. In this regard, ourresults show that resveratrol upregulated the expression in

104 Nutrition and Cancer 2001

Table 1. Effect of Resveratrol on Development ofTumors in Small Intestine and Colon of Min Micea–c

Number of Tumors

Groups Small intestine Colon

Control 30 ± 4 4 ± 0.5(21–45) (2–6)

Resveratrol 9 ± 1* 0(2–11)

a: Resveratrol (0.01%) was administered orally in drinking water.b: Values are means ± SE of 7 animals/group; ranges are in parentheses.c: Statistical significance is as follows: *, p � 0.01.

Vol. 39, No. 1 105

Figure 1. Differential gene expression in normal-appearing intestinal mucosa of a control and a resveratrol-treated Min mouse. 32P-labeled cDNA probes wereprepared in parallel from 2 �g of total RNA isolated from intestinal mucosa of control and resveratrol-treated Min mice (3 pairs). Probes were hybridized toseparate and identical Atlas Mouse Array membranes, and hybridization patterns were analyzed by phosphorimaging. A representative pair of cDNA patternsis shown from a control (top panel) and a resveratrol-treated mouse (bottom panel). Important differences were consistently recorded for 18 of 588 genes ana-lyzed between samples obtained from resveratrol-treated and control animals. Ratio of intensities of hybridization signals of ubiquitin (Std 1) to glyceraldehyde3-phosphate dehydrogenase (Std 2) was used for normalization and for comparison of transcript abundance in both membranes. A: oncogenes, tumorsuppressors, and cell cycle regulators. 1n, TSG101 tumor susceptibility protein (GenBank access no. U52945); 2j, RNA polymerase 1 termination factor TTF-1(X83974); 5l, Mas protooncogene (X67735); 6f, cyclin D1 (S78355); 6g, cyclin D2 (M83749). B: stress response, intracellular transducers, and ion channelsand transport. 5m, mitogen-activated protein kinase, p38 (U10871). D: transcription factors and DNA-binding proteins. 2g, DP-1 cell cycle regulatory tran-scription factor (X72310); 4e, homeobox protein 4.2 (J03770); 7j, YB-1 DNA binding protein (X57621). E: receptors, cell surface antigens, and cell adhesion.1k, monocyte chemoattractant protein 3 (S71251); 1l, leukemia inhibitory factor receptor (D26177); 5j, glutamate receptor (D10651); 6k, CTLA-4 (X05719);6l, desmocollin 2 (L33779). F: cell-to-cell communication, cytokines, chemokines, cytoskeleton, and protein turnover. 1l, follistatin (Z29532); 2h, inhibin-� Asubunit (X69619); 4e, thrombopoietin (L34169); 4f, transforming growth factor-� (M13177). Housekeeping genes: Std1, ubiquitin; Std 2, glyceraldehyde 3-phosphate dehydrogenase; Std3, �-actin; Std 4, ribosomal S19 protein.

the intestinal mucosa of 3 genes of a panel of 22 genes thatare directly involved in the recruitment and activation of im-mune cells, such as T lymphocytes and macrophages. Thusresveratrol stimulated the expression of cytotoxic T lympho-cyte Ag-4, which is an important T cell regulatory molecule,playing an important role in the differentiation of T cells invivo (17), of leukemia inhibitory factor receptor, a moleculeimplicated in growth arrest and macrophage differentiation(18), and of the monocyte chemotactic protein-3, which hasbeen shown to activate monocytes, dendritic cells, lympho-cytes, and natural killer cells (19). These events stronglysuggest that the increased recruitment of active immunecells by resveratrol may be an important clue in the chemo-preventive effects of this molecule, by decreasing the rate ofneoplastic development and by reducing tumorigenicity.This hypothesis is supported by the recent observation thatresveratrol seems to be an important cofactor in anti-inflammatory and anticancer nonspecific immune reactions(20). Furthermore, resveratrol has been shown to interferewith arachidonate metabolism, by reducing the levels ofprostanoids as a result of inhibition of cyclooxygenase, andprostaglandin H synthase activities (4,21).

The chemopreventive effects of resveratrol may also berelated to the inhibition of tumor invasiveness through theactivation of desmocollin 2, which is one of the desmosomalcomponents involved in suppression of tumor spreading(22), and also through the upregulation of a tumor suscepti-bility gene (TSG101), the inactivation of which in mice re-sults in genome instability, cell transformation, and theability to form metastatic tumors in nude mice (23). Another

target for resveratrol seems to be genes of the TGF-� family:TGF-�, the inhibin-� A subunit, and follistatin, which arewell-known growth factors, play an important role in thecontrol of the progression of colorectal cancer (24).

Resveratrol, a phytoalexin, has major biological func-tions in plants and grapes, particularly protection againstfungal infections. Our study highlights the complexity andmultiplicity of the molecular targets of the chemopreventiveactions of resveratrol. When given orally to animals predis-posed to intestinal carcinogenesis, it is able to inhibit tumori-genesis and to induce in the mouse intestinal mucosa cellsimportant changes in the expression of ~3% of the 588screened genes involved in the control of genome stability,cell growth, and differentiation, by favoring the downregu-lation of genes known to be implicated in the progressionof carcinogenesis and by activating genes involved in thehost-defense mechanisms against tumorigenesis and tumorspreading.

The use of a powerful tool such as the simultaneous com-parison of differential gene expression has allowed us to dis-cover new aspects of the chemopreventive properties ofresveratrol. This type of global approach should becomemore general as a preliminary search of genetic targets ofpotential chemopreventive agents.

Acknowledgments and Notes

This work was supported by a grant from Ligue Nationale contre leCancer (Comités Départementaux du Bas-Rhin et du Haut-Rhin, France).Address correspondence to Dr. Francis Raul, IRCAD, 1, place de

106 Nutrition and Cancer 2001

Figure 2. Variations of mRNA expression in intestinal mucosa of Min mice after treatment with resveratrol. For each cDNA array, signals from duplicatespots for each gene were averaged and background was subtracted from negative control. Ratio of intensities of hybridization signals of ubiquitin (Std 1) toglyceraldehyde 3-phosphate dehydrogenase (Std 2) was used for normalization and comparison of transcript abundance. Ratio of normalized signals for con-trols was considered as 1 and used as reference. RNAs expressed higher in resveratrol-treated animals than in controls are assigned a positive value, and thoseexpressed higher in controls than in resveratrol-treated mice are assigned a negative value. Three separate series of comparison were realized using in parallelRNA isolated from 1 control and 1 resveratrol-treated mouse, leading to good reproducibility. Each column corresponds to mean ± SE (n = 3). For further de-tails see Figure 1 legend. TGF�, transforming growth factor-�; MAP, mitogen-activated protein; LIF, leukemia inhibitory factor; HOX 4.2, homeobox protein4.2; CTLA-4, cytotoxic T lymphocyte Ag-4; YB1, Y-box DNA-binding protein; TSG, tumor susceptibility gene.

l’hôpital, BP 426, 67091 Strasbourg cedex, France. E-mail: francis.raul@ircad.u-strasbg.fr.

Submitted 19 October 2000; accepted in final form 8 January 2001.

References

1. Langcake, P: The production of resveratrol by Vitis vinifera and othermembers of the Vitaceae as a response to infection or injury. PhysiolPlant Pathol 9, 77–86, 1976.

2. Siemann, EH, and Creasy, LL: Concentration of phytoalexin resvera-trol in wine. Am J Enol Vitic 43, 49–52, 1992.

3. Soleas, GJ, Diamandis, EP, and Goldberg, DM: Resveratrol: a mole-cule whose time has come? And gone? Clin Biochem 30, 91–113,1997.

4. Jang, M, Cai, L, Udeani, GO, Slowing, KV, Thomas, CF, et al.: Cancerchemopreventive activity of resveratrol, a natural product derived fromgrapes. Science 275, 218–220, 1997.

5. Fontecave, M, Lepoivre, M, Elleingand, E, Gerez, C, and Guittet, O:Resveratrol, a remarkable inhibitor of ribonucleotide reductase. FEBSLett 421, 277–279, 1998.

6. Della Ragione, F, Cucciolla, V, Boriello, A, Della Pietra, V, Racioppi,L, et al.: Resveratrol arrests the cell division cycle at S/G2 phase transi-tion. Biochem Biophys Res Commun 250, 53–58, 1998.

7. Hsieh, T, and Wu, JM: Differential effects on growth, cell cycle arrest,and induction of apoptosis by resveratrol in human prostate cancer celllines. Exp Cell Res 249, 109–115, 1999.

8. Moser, AR, Pitot, HC, and Dove, WF: A dominant mutation that pre-disposes to multiple intestinal neoplasia in the mouse. Science 247,322–324, 1990.

9. Carbo, N, Costelli, P, Baccino, FM, Lopez-Sorriano, FJ, and Argiles,JM: Resveratrol, a natural product present in wine, decreases tumorgrowth in a rat tumour model. Biochem Biophys Res Commun 254,739–743, 1999.

10. Schneider, Y, Vincent, F, Duranton, B, Badolo, L, Gossé, F, et al.:Anti-proliferative effect of resveratrol, a natural component of grapesand wine, on human colonic cancer cells. Cancer Lett 158, 85–91, 2000.

11. Tessitore, L, Davit, A, Sarotto, I, and Caderni, G: Resveratrol de-presses the growth of colorectal aberrant crypt foci by affecting baxand p21 expression. Carcinogenesis 21, 1619–1622, 2000.

12. Jang, M, and Pezzuto, JM: Effects of resveratrol on 12-O-tetradec-anoylphorbol-13-acetate-induced oxidative events and gene expres-sion in mouse skin. Cancer Lett 134, 81–89, 1998.

13. Suzuki, R, Kuroda, H, Komatsu, H, Hosokawa, Y, Kagami, Y, et al.:Selective usage of D-type cyclins in lymphoid malignancies. Leukemia13, 1335–1342, 1999.

14. Carlson, B, Lahusen, T, Singh, S, Loaisa-Perez, A, Worland, PJ, et al.:Down-regulation of cyclin D1 by transcriptional repression in MCF-7human breast carcinoma cells induced by flavopiridol. Cancer Res 59,4634–4641, 1999.

15. Choubey, D, and Gutterman, JU: Inhibition of E2F/DP-1-stimulatedtranscription by p202. Oncogene 15, 291–301, 1997.

16. Ohga, T, Uchiumi, T, Koike, K, Wada, M, Kuwano, M, et al.: Directinvolvement of the Y-box binding protein YB-1 in genotoxic stress-induced activation of the human multidrug resistance 1 gene. J BiolChem 273, 5997–6000, 1998.

17. Walunas, TL, and Bluestone, JA: CTLA-4 regulates tolerance induc-tion and T cell differentiation in vivo. J Immunol 160, 3855–3860,1998.

18. Tomida, M, Heike, T, and Yokota, T: Cytoplasmic domains of the leu-kemia inhibitory factor receptor required for STAT3 activation, differ-entiation, and growth arrest of myeloid leukemic cells. Blood 93,1934–1941, 1999.

19. Fioretti, F, Fradelizi, D, Stoppacciaro, A, Ramponi, S, Ruco, L, et al.:Reduced tumorigenicity and augmented leukocyte infiltration aftermonocyte chemotactic protein-3 (MCP-3) gene transfer: perivascularaccumulation of dendritic cells in peritumoral tissue and neutrophil re-cruitment within tumor. J Immunol 161, 342–346, 1998.

20. Bertelli, AEE, Ferrara, F, Diana, G, Fulgenzi, A, Corsi, M, et al.:Resveratrol, a natural stilbene in grapes and wine, enhances intraphag-ocytosis in human promonocytes: a co-factor in anti-inflammatory andanticancer chemopreventive activity. Int J Tissue React XXI, 93–104,1999.

21. Subbaramaiah, K, Chung, WL, Michaluart, P, Telang, N, Tanabe, T, etal.: Resveratrol inhibits cyclooxygenase-2 transcription and activity inphorbol ester-treated human mammary epithelial cells. J Biol Chem273, 21875–21882, 1998.

22. Tselepis, C, Chidgey, M, North, A, and Garrod, D: Desmosomaladhesion inhibits invasive behaviour. Proc Natl Acad Sci USA 95,8064–8069, 1998.

23. Xie, W, Li, L, and Cohen, SN: Cell cycle-dependent subcellular local-ization of the TSG101 protein and mitotic and nuclear abnormalitiesassociated with TSG101 deficiency. Proc Natl Acad Sci USA 95,1595–1600, 1998.

24. Dumont, N: Genetic and epigenetic contribution to colorectal cancer.APMIS 107, 711–722, 1999.

Vol. 39, No. 1 107