Tumor-Microenvironment Interactions: The Fucose-Generating ...€¦ · fucosylation events are...

[CANCER RESEARCH 64, 6571–6578, September 15, 2004]

Tumor-Microenvironment Interactions: The Fucose-Generating FX EnzymeControls Adhesive Properties of Colorectal Cancer Cells

Adi Zipin,1 Mira Israeli-Amit,1 Tsipi Meshel,1 Orit Sagi-Assif,1 Ilana Yron,1 Veronica Lifshitz,1 Eran Bacharach,1

Nechama I. Smorodinsky,1 Ariel Many,2 Peter A. Czernilofsky,3 Donald L. Morton,4 and Isaac P. Witz1

1Department of Cell Research and Immunology, George S. Wise Faculty of Life Sciences, Tel Aviv University, Tel Aviv, Israel; 2Lis Maternity Hospital, Tel Aviv SouraskyMedical Center, Tel Aviv, Israel; 3Faculty of Medicine, University of Vienna, Vienna, Austria; and 4John Wayne Cancer Institute, Santa Monica, California

ABSTRACT

Extravasation of tumor cells is a pivotal step in metastasis formation.This step is initiated by an interaction of extravasating tumor cells withendothelial cells. Among the molecules mediating tumor-endothelium in-teractions are selectins and their fucosylated ligands. In a previous study,we demonstrated that the fucose-generating FX enzyme regulates theexpression of selectin ligands by B and T lymphocytes and by head andneck squamous cell carcinoma cells. It was also shown that the FX enzymeregulated important interaction parameters between these cancer cellsand endothelial cells. The present study was aimed to determine whetherthe FX enzyme controls adhesive interactions between colorectal cancercells and endothelial cells. The results clearly indicate that this is indeedthe case. Overexpressing the FX enzyme by the transfer of FX cDNA tolow FX-expressing colorectal cancer cells resulted in an increased adhe-sive capacity of the transfectants to activated endothelial cells and torecombinant E-selectin. Down-regulating FX levels in colorectal cancercells expressing high levels of endogenous FX by transfection with small-interfering RNA resulted in a down-regulated expression of the selectinligand sialyl Lewis-a and a decrease in the adhesive capacity of thetransfectants to activated endothelial cells and to recombinant E-selectin.These transfection experiments also indicated that manipulating the levelsof the FX enzyme affected global cellular fucosylation and altered theinteraction of colorectal cancer cells with some extracellular matrix com-ponents such as fibronectin. We also found that highly metastatic colo-rectal cancer variants express higher levels of FX and of sialyl Lewis-athan low metastatic variants originating in the same tumors. These resultslead us to hypothesize that the FX enzyme controls the capacity ofcolorectal cancer to extravasate and form metastasis. If this hypothesiswill be confirmed the FX enzyme could become a target molecule formetastasis prevention.

INTRODUCTION

Fucose is a component of many surface-localized and secretedmolecules. It decorates the terminal portions of N-, O-, or lipid-linkedglycans and modifies the core of some N-linked glycans (1). Terminalfucosylated glycans in humans constitute several blood group antigensand function as selectin ligands (2). Fucosylation of these ligandsdetermines their ability to bind to the selectin family of cell adhesionmolecules and therefore controls pivotal steps of selectin-dependentleukocyte and tumor cell adhesion and trafficking (1, 3, 4). In additionto its role in adhesive reactions, fucosylation influences Notch andCripto signaling events (5–9).

The final steps of fucose biosynthesis are mediated by the GDP-D-mannose 4,6 dehydratase generating GDP-mannose-4-keto-6-D-

deoxymannose. This sugar is converted to GDP-L-fucose by the FXenzyme, functioning both as an epimerase and a reductase (10, 11).The GDP-L fucose is then transported to the Golgi. Fucosylation ofmammalian glycans is catalyzed by distinct fucosyltransferases, withcatalytic activities characterized by specificity for specific glycocon-jugate substrates and a requirement for GDP-fucose (12).

The generation of FX knockout mice enabled us to conclude thatfucosylation events are essential for fertility, early growth, and devel-opment, as well as for intercellular adhesion (1). FX knockout resultedin a massive intrauterine mortality. Live-born FX-null mice exhibiteda virtually complete deficiency of cellular fucosylation and a postnatalfailure to thrive. FX (�/�) adults suffer from an extreme neutrophilia,myeloproliferation, and absence of leukocyte selectin ligand expres-sion reminiscent of LAD-II/CDG-IIc (1).

All these studies demonstrate that fucosylation plays an importantrole in development and in adult physiology by influencing at leasttwo different pathways: biosynthesis of selectin ligands and signalingthrough the Notch and Cripto pathways.

Previous studies from our laboratory demonstrated that the FXenzyme is involved in the biosynthesis of sLe-x in activated T or Bcells (13). In head and neck squamous cell carcinomas, the FXenzyme plays a key role in the biosynthesis of selectin ligands such assialyl Lewis-a (sLe-a) and in the interaction of these cancer cells withendothelial cells (14, 15). In lymphocytes, as well as in head and necksquamous cell carcinomas, the FX enzyme is regulated by outside-insignaling (13, 14).

Numerous reports indicated a correlation between a high expressionof selectin ligands by epithelial cancer cells, notably colorectal cancer,and a high rate of metastasis and poor prognosis (16–18). Thefucose-generating FX enzyme may thus be a pivotal element incancer-associated perturbations of differentiation, survival, and pro-liferation, as well as in cancer cell extravasation. It is thereforeimportant to find out if the FX enzyme is involved in controllingselectin ligand expression by colorectal cancer cells and in theirinteraction with endothelial selectin. The present study provides proofthat this is indeed the case.

MATERIALS AND METHODS

Cell Lines. The human colorectal cancer cell lines: 474, 0485, 1086, 1203,044, and 427 were established at the John Wayne Cancer Institute (SantaMonica, CA). The 474 and 1203 cell lines were derived from primary colo-rectal cancer tumors. The 1086 cell line was derived from a colorectal cancerliver metastasis. The 0485 cell line was derived from a lymph node metastasis.The cells were maintained in RPMI 1640 supplemented with 20% FCS, 2mmol/L L-glutamine, 100 units/mL penicillin, 0.1 mg/mL streptomycin, and12.5 units/mL nystatin. All of the medium components were obtained fromBiological Industries (Beit-Haemek, Israel). KM12C, KM12L4, and KM12SMwere kindly provided by Dr. Isaiah J. Fidler (Department of Cell Biology,M. D. Anderson Cancer Center, Houston, TX). The KM12C cell line wasderived from a Duke’s B2 colorectal cancer primary tumor. The KM12L4 andKM12SM variants originated in liver metastases that developed in BALB/cnude mice inoculated with KM12C cells to spleen or cecum, respectively (19).KM12C, KM12L4, and KM12SM cells were maintained in Eagle’s MEMsupplemented with 10% FCS, 2 mmol/L L-glutamine, 100 units/mL penicillin,

Received 12/30/03; revised 5/20/04; accepted 7/20/04.Grant support: The Jacqueline Seroussi Memorial Foundation for Cancer Research,

The Israel Cancer Association, The Fainbarg Family Fund (Orange County, CA), TheFred August and Adele Wolpers Charitable Fund (Clifton, NJ), Arnold and Ruth Feuer-stein (Orange County, CA), and the Pikovsky Fund (Jerusalem, Israel). I. Witz is theincumbent of the David Furman Chair in Immunobiology of Cancer.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Requests for reprints: Isaac P. Witz, Department of Cell Research and ImmunologyGeorge S. Wise Faculty of Life Sciences, Tel Aviv University, 69978 Tel Aviv, Israel.Phone: 972-3-6406979; Fax: 972-3-6422046; E-mail [email protected].

©2004 American Association for Cancer Research.

6571

Cancer Research. on September 24, 2020. © 2004 American Association forcancerres.aacrjournals.org Downloaded from






http://cancerres.aacrjournals.org/






0.1 mg/mL streptomycin, and 12.5 units/mL nystatin, 5 mmol/L sodiumpyruvate, nonessential amino acids, and 2-fold vitamin solution (19). Thecolorectal cancer SW480 and SW620 cells were obtained from American TypeCulture Collection (Manassas, VA). The SW480 cell line was derived from aprimary colorectal cancer, whereas the SW620 line was derived from a lymphnode metastasis of the same patient. SW480 cells were maintained in RPMI1640 supplemented with 10% FCS. FX-cDNA SW480-transfected cells weremaintained in the same medium supplemented with 800 �g/mL G418 (Cal-biochem, San Diego, CA). SW620 cells were maintained in Leibovitz L-15medium supplemented with 15% FCS, 2 mmol/L L-glutamine, 100 units/mLpenicillin, 0.1 mg/mL streptomycin, 12.5 units/mL nystatin, 10 mmol/LHEPES buffer, and 0.075% sodium bicarbonate. The colorectal cancer HT-29cell line was kindly provided by Dr. Livia Theodor (Gastroenterology Depart-ment, The Chaim Sheba Medical Center, Tel Hashomer, Ramat-Gan, Israel).The HT-29 cells were maintained in McCoy’s modified medium (Invitrogen-Life Technologies, Inc., Paisley, Scotland, United Kingdom) supplementedwith 10% FCS, 2 mmol/L L-glutamine, 100 units/mL penicillin, 0.1 mg/mLstreptomycin, and 12.5 units/mL nystatin. All of the cell lines were routinelycultured in humidified air with 5% CO2 at 37°C.

Human Umbilical Vein Endothelial Cells (HUVECs). HUVECs wereeither purchased from Dr. Neomy Lanir (Department of Hematology and BoneMarrow Transplantation, Rambam Medical Center, Haifa, Israel) or preparedin our laboratory from umbilical cords as follows: endothelial cells wereharvested by 0.25 mg/mL collagenase type II (Sigma, Holon, Israel). Cellswere grown to confluence in tissue culture flasks precoated with fibronectin(20 �g/mL; Biological Industries, Beit- Haemek, Israel). Cells were estab-lished as primary cultures in M199 medium supplemented with 20% FCS, 50�g/mL endothelial cell growth factor (Biomedical Technologies, Inc., Stough-ton, MA), heparin (5 units/mL; Laboratoire Choay, Paris, France), and anti-biotics. Cells from the third passage were taken for experiments.

Antibodies. The following antibodies were used in flow cytometry assays:anti-sialyl Lewis-a (sLe-a), anti-Lewis-a (Le-a), and anti-Lewis b (Le-b) werepurchased from Seikagaku America (Falmouth, MA). Anti-sialyl Lewis-x(sLe-x) was purchased from American Type Culture Collection. Anti-Lewis-x(Le-x) and anti-CD24 were purchased from PharMingen (San Diego, CA).Anti-Lewis-y (Le-y) and anti-PSGL-1 were purchased from Serotec (Oxford,United Kingdom). Antihuman VIM2 antibodies were a kind gift from Dr.Walter Knapp (Institute of Immunology, University of Vienna, Vienna, Aus-tria). Anti-CD62E was purchased from Southern Biotechnology Associates(Birmingham, AL). FITC-conjugated goat antimouse IgM and IgG were pur-chased from Jackson Immunosearch Laboratory, Inc. (West Grove, PA). Thefollowing antibodies were used in Western blotting assays: mouse monoclonalantihuman FX antibody (3G10/12) prepared in our laboratory, rabbit poly-clonal antihuman FX antibody 701 (13) also prepared in our laboratory,anti-glyceraldehyde-3-phosphate dehydrogenase antibody (Chemicon Interna-tional, Inc., Temecula, CA), and horseradish peroxidase-conjugated secondarygoat antibody against mouse IgG (Jackson Immunosearch Laboratory, Inc.).

Plasmids. The plasmid pSUPER was kindly provided by Dr. ReuvenAgami (Division of Tumor Biology, The Netherlands Cancer Institute, Ples-manlaan, Amsterdam, the Netherlands). The plasmid pRc/CMV was purchasedfrom Invitrogen BV (Groningen, the Netherlands). The plasmid pEGFP-C waspurchased from Clontech Laboratories, Inc. (Palo Alto, CA).

Construction of FX Small Interfering RNA (siRNA). FX mRNA sup-pression was achieved by using the pSUPER vector. Two gene-specificoligonucleotides were designed as follows: 5�-GATCCCCAGACGCCG-ATCTCACGGATTTCAAGAGAATCCGTGAGATCGGCGTCTTTT-TTGGAAA-3� and 5�-AGCTTTTCCAAAAAAGACGCCGATCTCA-CGGATTCTCTTGAAATCCGTGAGATCGGCGTCTGGG-3�. Regularcharacters represent regions required for the generation of the siRNA aspreviously described (20), and bold characters represent FX-specific comple-mentary sequences. Both oligonucleotides were denatured at 95°C for 4minutes, annealed at 70°C for 10 minutes, and cooled down slowly. Oligonu-cleotides were then phosphorylated by the use of T4 polynucleotide kinase at37°C for 30 minutes. This oligonucleotide mixture was ligated into the pSU-PER vector predigested with BglII and HindIII and pretreated with calfintestinal phosphatase. A mutated FX siRNA that served as control wasgenerated by introducing a point mutation (G to C) at position 24 in the firstFX-specific complementary sequence oligonucleotide shown above. This oli-

gonucleotide was cloned to the pSUPER vector along with the second FX-specific complementary oligonucleotide shown above.

Flow Cytometry. Cells (5 � 105) were incubated for 45 minutes at 4°Cwith primary antibodies directed against the tested selectin ligand. After a washwith cell sorter medium (RPMI 1640 supplemented with 5% FCS and 0.01%sodium azide), the cells were incubated for 45 minutes at 4°C with FITC-conjugated goat antimouse IgG or IgM. After an additional wash, antigenexpression on 5000 live cells was determined using a Becton DickinsonFACSort (Mountain View, CA) and CellQuest software. Baseline staining wasobtained by adding cell sorter medium to the cells instead of primary antibody.Flow cytometry scores for selectin ligand expression were calculated asdescribed previously (13). The scores represent the multiplication of the meanfluorescence by the percent positive cells �10-4. Scores for each selectinligand were divided into three categories: high, medium, and low. For sLe-a,high � 8.2 to 19.7, medium � 5.7 to 7.6, and low � 1.8 to 0.2; for Le-b,high � 8.7 to 11.6, medium � 2.2 to 5.6, and low � 0.1 to 1.3; forsLe-x, high � 7 and medium � 2.3 to 4.9.

RNA Preparation and Northern Blot Analysis. Total RNA was prepared,and Northern blotting was performed as described by Eshel et al. (14) in ourlab.

Western Blotting. Colorectal cancer cells were lysed with Laemmli sam-ple buffer (21). Lysates were boiled for 10 minutes, centrifuged, and appliedon a miniprotean II system (Bio-Rad, Hercules, CA) for SDS-PAGE using a12% slab gel as described by Laemmli (21). Electrophoretic transfer ofproteins from the polyacrylamide gel to nitrocellulose (Schleicher & Schull,Dassel, Germany) was performed by a mini-transblot electrophoretic cell(Bio-Rad) at 250mA for 2 hours. After transfer, the nitrocellulose membranewas incubated at room temperature with 3% BSA in TBS-Tween for 30minutes to block free binding sites on the membrane. The blocked nitrocellu-lose membrane was incubated over night with anti-FX 3G10/12 monoclonalantibody, diluted 1:16 in 1% BSA in TBS-Tween with 0.02% sodium azide orwith rabbit polyclonal antihuman FX antibody 701(2), diluted 1:1000 in 5%milk in TBS-Tween with 0.02% sodium azide, then washed three times for 5minutes with TBS-Tween and incubated for 50 minutes with horseradishperoxidase-conjugated secondary goat antibody against mouse or rabbit IgGdiluted 1:10,000 with 5% milk in TBS-Tween at room temperature. Finally, thenitrocellulose membrane was washed five times for 5 minutes with TBS-Tween. The bands were visualized by chemiluminescence-enhanced chemilu-minescence reaction (Amersham, Buckinghamshire, United Kingdom) andautoradiography by exposure to Kodak XAR5 film (Eastman Kodak Co.,Rochester, NY) for 1 to 5 minutes.

The quantification of protein in the lanes was determined in reference to theamount of glyceraldehyde-3-phosphate dehydrogenase in the lanes. This wasperformed by incubating the membrane with anti-glyceraldehyde-3-phosphatedehydrogenase diluted 1:1000 with 5% milk in TBS-Tween with 0.02%sodium azide.

Lectin Blot. The first steps of blotting were performed as described abovefor Western blotting. After transfer of the proteins to nitrocellulose, themembrane was incubated at room temperature for 60 minutes with 3% BSA inPBS-0.05% Tween- 20 to block free binding sites on the membrane. Theblocked nitrocellulose membrane was incubated for 2 hours with horseradishperoxidase-conjugated Ulex europaeus agglutinin lectin (Sigma), which bindsto L-fucose (22), diluted 1:2000 in PBS containing 3% BSA and 0.1% Tween-20, then washed four times for 10 minutes with PBS containing 0.1% Tween-20. The bands were visualized by the chemiluminescence-enhanced chemilu-minescence reaction (Amersham) and autoradiographed by exposure to KodakX-AR5 film (Eastman Kodak Co.) for 5 seconds.

Transfection. SW620 cells were transfected with pSUPER-FX siRNA(205 cells) or with pSUPER-FX-mutated siRNA oligonucleotides (360 controlcells) by electroporation using the Electro cell manipulator 830 (BTX-Genetronics, San Diego, CA). Thirty micrograms of pSUPER and 3 �g of thepBabe-puro plasmid were used for transfection. Forty-eight hours after electro-poration, the cells were selected with 1 �g/mL puromycin. Individual cloneswere then picked, expanded and analyzed for FX protein levels. To increasetransfection yield, cells were re-transfected using the pSUPER coupled withpcDNA3.1-zeo vectors. In this case, 500 �g/mL zeocin was used as a selectionmarker. The transfected cells were grown in the presence of 25 mmol/L fucose.SW480 cells were cotransfected either with FX-cDNA pRc/CMV vector or

6572

FX ENZYME CONTROLS THE ADHESIVE PROPERTIES OF CRC



with pRc/CMV vector without insert together with pEGFP-C vector by elec-troporation.

Adhesion of Colorectal Cancer to HUVECs. Ninety-six-well cultureplates were coated with 50 �L of 20 �g/mL fibronectin per well for 30 minutesat 37°C. After one wash with PBS, 3 � 104 HUVECs in a volume of 100�L/well were cultured on the plate for 16 hours to form a confluent monolayer.Endothelial cells were stimulated for 4 hours with human recombinant IFN-�(100 units/mL) and recombinant tumor necrosis factor � (100 units/mL) andwashed. A total of 2 � 106 colorectal cancer cells were washed twice with PBSand suspended in 1 mL of cold PBS, labeled with the CFDA-SE reagent (250nmol/L; Molecular Probes, Eugene, OR) for 5 minutes, and then washed with2 mL of FCS. After an additional wash with PBS, the labeled cells (1 � 105/100 �L) were added to a stimulated HUVEC monolayer in TBS containing 2mmol/L CaCl2 (for selectin-depended adhesion) and incubated for 30 minutesat 37°C. The total fluorescence signal was that of labeled cells added to thewell before removing the nonadherent cells, which were removed by threewashes with PBS. Total fluorescence and that of adhering cells was measuredby a fluorescent ELISA reader (Bio-Tek FL500; Bio-Tek Instruments, Inc.,Winooski, VT) at wavelength of 490/530. The results are presented as percentadhesion. The number of adherent green fluorescent protein (GFP)-transfectedcells was determined by counting fluorescent cells in several fields with the aidof a fluorescence microscope (magnification �100, FITC filter, OlympusIX70; Olympus, Hamburg, Germany).

Adhesion of Colorectal Cancer to rE-Selectin, Extracellular Matrix(ECM), or Fibronectin. Nontissue culture plates (Nunc-Immuno plateNUNC; Nunc International, Roskilde, Denmark) were coated for 60 minutes atroom temp with 100 �L/well of 2 �g/mL recombinant E-selectin (R&DSystems, Minneapolis, MN) diluted in TBS-Tween containing 2 mmol/LCaCl2. Three wells were not coated for detection of nonspecific binding.Supernatants were aspirated from the wells, and 200 �L of blocking medium(1% BSA in TBS-Tween containing 2 mmol/L CaCl2) were added for at least30 minutes at room temperature. Cells were labeled with CFDA-SE reagent asdescribed above. Labeled cells (1 � 105/100 �L) suspended in 1% BSA inTBS-Tween, 2 mmol/L CaCl2 were added to rE-selectin or uncoated wells for45 minutes at 37°C. The rest of the assay was performed as described above.A similar procedure was performed with plates coated with a mixture of matrixproteins (E-TCMT-F NOVAmed; Jerusalem, Israel) or with fibronectin in thepresence of 5 mmol/L MgCl2.

Statistical Analysis. Significance was calculated using the two-way Stu-dent’s t test.

RESULTS

An Expression Profile of Fucosylated Glycans by ColorectalCancer Cell Lines

Table 1 shows the expression profile of several fucosylated glycansby 12 colorectal cancer cell lines. This profile was derived fromseveral repetitions of flow cytometry experiments. All lines expressedsLe-a as well as Le-b. High levels of sLe-a were expressed by four celllines and medium levels by four lines. Four cell lines expressed low

levels of sLe-a. High levels of Le-b were expressed by cells of twolines (427 and 1203, which also expressed high levels of sLe-a). Fourlines expressed medium levels of Le-b, and six lines expressed lowlevels of this fucosylated glycan. In all, it seems that the expressionpattern of sLe-a by the various cell lines was similar to that of Le-band to some extent to Le-a. The expression levels of sLe-x, a domi-nant selectin ligand on activated lymphocytes, were rather low. Withthe exception of four cell lines, of which, one expressed high and threemedium levels of sLe-x, all other lines expressed either low levels ofthis selectin ligand or did not express it at all.

Taken together, these results indicated that sLe-a is the dominantselectin ligand in all cell lines tested. Because of its relative highexpression by most of the colorectal cancer cell lines assayed, weconsider sLe-a to be the representative selectin ligand of colorectalcancer cells.

In view of the possibility that in colorectal cancer, as in HNSCC(14), the fucose-generating FX enzyme functions as a limiting factorin sLe-a biosynthesis, we studied, in the following series of experi-ments, its contribution to sLe-a expression.

Expression Levels of the FX Enzyme Correlate with Thoseof sLe-a

Fig. 1 shows a regression curve correlating the expression of FXprotein to that of sLe-a in several colorectal cancer cell lines. A

Table 1 Expression levels of selectin ligands by colorectal cancer cell lines

Cell line

Selectin ligand

sLe-a Le-b Le-a sLe-x Le-x Le-y CD24 VIM2 PSGL-1

HT-29 6.5 5.6 4.4 2.3 0.6 0.4 3 0.1 �0.1044 17.1 0.9 1.4 �0.1 �0.1 0.2 0.1 �0.1 �0.11086 7.6 1.3 1.7 �0.1 0.6 �0.1 0.5 0.4 �0.1474 8.2 5.6 0.2 4.7 2.2 0.7 0.5 �0.1 0.80485 5.7 2.5 �0.1 �0.1 1.6 1.1 1.6 �0.1 �0.1427 19.7 11.6 13.5 7 8.2 �0.1 0.6 �0.1 0.21203 9.7 8.7 3.1 4.9 16.3 0.2 0.2 2.7 �0.1SW480 0.2 2.2 0.1 �0.1 4.3 2.8 �0.1 7.5 0.1SW620 5.9 0.1 2.3 �0.1 7.3 2 1.1 0.5 �0.1KM12C 0.2 0.1 0.1 �0.1 �0.1 �0.1 4.2 0.3 �0.1KM12L4 0.7 0.2 0.1 0.1 0.3 0.1 3.4 2 0.1KM12SM 1.8 0.3 0.1 0.9 �0.1 0.2 4.1 0.1 0.4

NOTE. Selectin ligand expression was determined by flow cytometry. Values represent expression scores obtained by the multiplication of mean fluorescence values by the percentof positive cells (�10�4) as described previously (13).

Fig. 1. Correlation between the expression of the FX protein and sLe-a by colorectalcancer cells. Lysates of colorectal cancer cells were assayed for FX protein by Westernblot analysis using anti FX antibodies prepared in our lab. Anti- glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies were used as control. Colorectal cancercells were assayed by flow cytometry for sLe-a expression. The values of sLe-a expressionwere calculated as for Table 1. HT-29 :, 044 ��, 1086 F, 474 Œ, 0485 �, 427 », 1203 �,SW480 ■, and SW620 E.

6573




significant correlation between these two parameters was seen. Asimilar correlation between FX expression levels and other selectinligands was not found (results not shown).

Metastatic Colorectal Cancer Cell Line Variants ExpressHigher Levels of the FX Enzyme and of sLe-a Than theCorresponding Nonmetastatic Variants

To assess the contribution of the FX enzyme to various malig-nancy associated characteristics, we compared cell lines having anidentical genetic background but that differ in their metastaticphenotype. High and low metastatic variants originating from twopatients were used. SW480 and SW620 cells originated in onepatient, and KM12C, KM12SM, and KM12L4 originated fromanother patient.

SW480 and SW620 are colorectal cancer cell lines derived, corre-spondingly, from the primary tumor and from a lymph node metas-tasis of a single patient. This cell pair was recently validated as an

appropriate model to study differences between primary and second-ary tumors (23).

Fig. 2A demonstrates that the metastatic SW620 cells expresshigher levels of the FX enzyme than the nonmetastatic SW480 cells.These cells also express higher levels of sLe-a (Fig. 2B).

The KM12C cell line was established from a primary colorectalcancer, and the KM12SM and KM12L4 lines are metastatic variantsof KM12C derived from nude mouse xenotransplants. These three celllines thus have the same genetic background (19). The KM12SM cellsare more metastatic than the KM12L4 cells (19). Fig. 3A demon-strates that the metastatic KM12SM variant expresses higher levels ofthe FX enzyme than the primary KM12C cells. The more metastaticKM12SM variant expresses twice as much FX enzyme than the lessmetastatic KM12L4 variant in which FX expression was only mar-ginally higher than that by the primary KM12C cells. Fig. 3B showsthat the expression of sLe-a in the three cell lines correlated with theFX expression by these cells.

Fig. 3. Expression of FX protein and of sLe-a by KM12C, KM12L4,and KM12SM cells. Please see legend to Fig. 2 above. A representativeexperiment (of three performed) is presented.

Fig. 2. Expression of FX protein and of sLe-a by SW480 and SW620 cells. A, expression of FX protein. Lysates of SW480 and SW620 cells were assayed for FX expression byWestern blot analysis using anti FX antibodies. Anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibodies were used as control. Values represent the ratio between thesignal of FX protein in each cell and the signal of GAPDH in the same cell. B, expression of sLe-a. SW480 and SW620 cells were assayed by flow cytometry for sLe-a expression.(M � mean fluorescence, %pos � percent of positive cells). A representative experiment (of three performed) is presented.

6574




The results presented above established, thus, a correlation betweenthe metastatic phenotype of colorectal cancer and expression levels ofthe FX enzyme and of sLe-a.

The Adhesive Capacity of Colorectal Cancer Cells to ActivatedEndothelial Cells and to E-Selectin Is Linked to the ExpressionLevels of Endogenous FX and sLe-a

Fig. 4A demonstrates that the metastatic SW620 cells expressinghigher levels of FX and sLe-a than SW480 cells (Fig. 2), adhere betterto activated HUVEC (P � 0.03).

To test whether the differential adhesion of SW620 and SW480cells to activated HUVECs reflects their adhesion capacity to E-selectin, we measured the adhesion of these cells to recombinantE-selectin. The results presented in Fig. 4B show that the metastaticSW620 cells adhere significantly better to rE-selectin than SW480cells (P � 0.02), suggesting that the adhesion of the cells is mediatedby a selectin-selectin ligand interaction.

Similar results were obtained when comparing the adhesive capac-ity of primary KM12C cells with that of the metastatic KM12SM cellvariant. The highly metastatic KM12SM cells, expressing high levelsof the FX enzyme and of sLe-a (Fig. 3), adhered better to activatedHUVECs (P � 0.03; Fig. 5A) and to rE-selectin (P � 0.02; Fig. 5B)than the KM12C cells, which express a low metastatic behavior andlow levels of FX and sLe-a.

The FX Enzyme Directly Regulates the Expression of sLe-a andthe Selectin-Mediated Adhesion of Colorectal Cancer Cells

FX cDNA Transfection. The results presented in the previoussections show a positive correlation between expression levels of thefucose-generating FX enzyme, expression levels of the selectin ligandsLe-a, and adhesion of colorectal cancer cells to activated HUVECsand rE-selectin. The next set of experiments was aimed to determineif the FX enzyme directly regulates the expression of sLe-a oncolorectal cancer cells and their adhesion to activated HUVECs andrE-selectin. It was assumed that if fucose is indeed a limiting factor inselectin ligand biosynthesis in colorectal cancer cells and thus of theirE-selectin–dependent adhesion to activated HUVECs, then overex-

pressing the FX enzyme in low FX-expressing cells would result inhigher levels of selectin ligand expression and in an up-regulatedadhesive capacity of FX cDNA-transfected cells. By the same token,down-regulating the levels of FX by siRNA should decrease theexpression levels of selectin ligands. The FX siRNA-transfected cellsshould, as a result, express a down-regulated adhesive capacity. Theexperiments described below demonstrate that this is indeed the case.

We first overexpressed the FX enzyme in SW480 cells, whichexpress very low levels of endogenous FX and sLe-a (Fig. 2). Thesecells were transiently cotransfected with FX and GFP cDNA (in a10:1 ratio). These cells expressed higher levels of FX protein thancontrol cells transfected with GFP alone (Fig. 6A). Fig. 6B shows thatthese transfectants adhered better to activated HUVECs than cellstransfected with GFP cDNA alone (P � 0.001) and shows that theadhesion of the FX-transfectants was reduced to background levels(i.e., adhesion to nonactivated HUVECs) by antibodies directedagainst E-selectin (P � 0.001).

Adhesion experiments in which rE-selectin was used instead ofactivated HUVECs (Fig. 6C) yielded similar results (P � 0.001).Antibodies against E-selectin reduced the adhesion of the FX-transfectants to rE-selectin also in this case (P � 0.001).

FX siRNA Transfection. The next step was to down-regulate theexpression of the FX enzyme in SW620 cells by stably transfectingthem with FX siRNA (205 cells; ref. 20). SW620 cells transfectedwith a mutated FX siRNA sequence in the pSuper vector (360 cells)served as controls. Northern blotting indicated that the FX siRNA-transfected 205 cells expressed lower levels of FX mRNA than thecontrol-transfected 360 cells (Fig. 7A). The FX siRNA-transfected205 cells expressed also lower levels of the FX protein than thecontrol transfected 360 cells (Fig. 7B). It is interesting to note that thedown-regulation of FX protein was more remarkable at the proteinlevel than at the mRNA level. No explanation for this observation isavailable at this stage. Fig. 7C demonstrates that the siRNA-trans-fected 205 cells express less sLe-a than control transfectants.

Fig. 8 shows that the FX siRNA-transfected 205 cells adhered lesswell to rE-selectin than control 360 cells (P � 0.01). In conformitywith the results reported in Fig. 6C, antibodies directed against E-

Fig. 4. Adhesion of SW480 and SW620 cells. A, adhesion to HUVECs. SW480 and SW620 cells were labeled with CFDA-SE reagent and incubated on activated HUVECs for 30minutes as described in Materials and Methods. Nonadherent cells were removed, and the number of adherent cells was determined by a fluorescence ELISA reader in reference toa standard curve. The bars represent mean � SD of values obtained in four independent experiments. �, P � 0.03. B, adhesion to rE-selectin. CFDA-SE–labeled SW480 and SW620cells were incubated on rE-selectin–coated plates for 45 minutes as described in Materials and Methods. Percentage of adhering cells was calculated as the ratio between fluorescencesignal of adhering cells and that obtained from the total number of plated cells. Fluorescence of the cells was determined as described above. The bars represent mean � SD of valuesobtained in three independent experiments. �, P � 0.02.

Fig. 5. Adhesion of KM12C and KM12SM cell lines. A, adhesionto HUVECs. KM12C and KM12SM cells were labeled withCFDA-SE and incubated on activated HUVECs for 30 minutes asdescribed in Materials and Methods. The percentage of adherentKM12C was given a value of 1.0. Normalized adherence values wereobtained by dividing percentage of adhesion of KM12SM cells bypercentage of adhesion of KM12C cells. The bars representmean � SD of values obtained in five independent experiments. �,P � 0.03. B, adhesion to rE-selectin. Please see legend to Fig. 4Babove. The bars represent mean � SD of values obtained in threeindependent experiments. �, P � 0.02. (C � KM12C,SM � KM12SM).

6575




selectin blocked to background levels the adhesion of the control 360cells to rE-selectin (P � 0.001; Fig. 8). The low adhesion of the FXsiRNA-transfected 205 cells to rE-selectin was also blocked to back-ground levels by these antibodies (P � 0.001; Fig. 8). An isotypecontrol antibody used in these experiments had no blocking activitywhatsoever.

The transfection experiments described above provide conclusiveevidence that a FX3selectin ligand3adhesion pathway operates incolorectal cancer. However, the involvement of other fucosylatedadhesion molecules controlled by the fucose-generating FX enzyme inthe adhesion of colorectal cancer cannot be ruled out.

The siRNA transfectants adhered less well also to ECM (P � 0.002;Fig. 9A). To identify the ECM protein to which adhesion was reducedby down-regulating the fucose-generating FX enzyme, we comparedthe adhesion of FX siRNA transfectants (205 cells) and of controltransfectants (360 cells) to collagen, laminin, and fibronectin.Whereas the adhesion of both types of transfectants to collagen and

laminin was similar, the FX siRNA-transfected cells adhered signif-icantly less well to fibronectin than the controls (P � 0.04; Fig. 9B).This finding led us to tentatively conclude that cellular adhesion tofibronectin may be fucose dependent.

We also tested if knocking down FX levels by FX siRNA trans-fection would affect adhesion to uncoated tissue-culture plastic ves-sels under normal culture conditions. Three independent experimentsshowed a reduced survival of SW620 cells in which levels of FX wereknocked down by FX siRNA transfection, as compared with controlcells transfected with mutated FX siRNA. At the time point of 96hours after the initiation of culture, the average number of viable-adherent control cells was 1.96-fold higher than that of the FX siRNAtransfectants (P � 0.02).

The above results suggest that FX knockdown may influence globalfucosylation of colorectal cancer glycomolecules. To test this possi-bility, we evaluated levels of fucoconjugates in FX siRNA transfec-tants (205 cells) and in control 360 cells. Fig. 10 shows a lectin blot

Fig. 6. Adhesive properties of SW480 cells transiently overexpressing FX cDNA. A, expression of FX protein. SW480 cells were cotransfected with FX cDNA and GFP. Controlswere transfected with GFP alone. Lysates were subjected to Western blot analysis as described in the legend for Fig. 2A. B, adhesion to HUVECs and its blocking by antibodies againstE-selectin. The transfected cells were incubated on activated HUVECs, as described in the legend for Fig. 4A. The bars represent mean � SD of values obtained from five independentexperiments. P � 0.001 for the difference between FX cDNA and control transfectants. Similar experiments were performed in the presence of antibodies against E-selectin. Twoindependent experiments were performed. P � 0.001 for the difference between the adhesion of FX cDNA transfectants in the presence or absence of the antibody. C, adhesion torE-selectin and its blocking by antibodies against E-selectin. The transfected cells were incubated on rE-selectin–coated plates, as described in the legend for Fig. 4B. The bars representmean � SD of values obtained by counting several fields in two independent experiments. Adhesion was blocked, in two independent experiments, by anti-E-selectin antibodies (�E-selectin; P � 0.001). The adhesion presented in B and C was determined by counting adherent, GFP-expressing cells under a fluorescence microscope (the average number of cellsper field was obtained by counting several fields). Control � SW480 cells cotransfected with control pRc/CMV vector and GFP; FX cDNA � SW480 cells cotransfected with FX cDNAand GFP; see Materials and Methods.

Fig. 7. Expression of FX mRNA, FX proteinand of sLe-a by SW620 cells stably transfectedwith FX siRNA. A, FX mRNA. Expression wasdetermined by Northern blot analysis. Values rep-resent the ratio between the signal of FX mRNA inthe cells and the signal of 18S rRNA in the samecell sample. B, FX protein. Expression was deter-mined as in Fig. 2A. C, sLe-a. Expression wasdetermined as in Fig. 2B. A representative experi-ment (of three performed) is presented (360cells � control cells; 205 cells � FX siRNA trans-fectants). M � mean fluorescence; %pos � %positive cells.

6576




of extracts derived from control or FX siRNA-transfected cells de-veloped with the Ulex europaeus agglutinin 1 agglutinin, a fucose-specific lectin (22). A clear reduction in global fucosylation is indi-cated in the FX siRNA transfectants.

DISCUSSION

The attachment of selectin ligand-expressing tumor cells to selec-tin-expressing endothelial cells is crucial event in the initial step oftumor cell-endothelium interactions, including extravasation (24–29).Therefore, understanding the regulation of selectin ligand synthesis intumor cells is crucial for the development of metastasis-targetingmanipulations.

The proper functioning of selectin ligands requires their fucosyla-tion (1, 30–34), mainly by the fucose-generating FX enzyme (1, 13,14, 33, 34). The present study documents a functional relationshipbetween expression levels of the FX enzyme and those of the selectinligand sLe-a by colorectal cancer cells. This conclusion was based ona direct and positive correlation between expression levels of FX andthose of sLe-a by several colorectal cancer cell lines. Furthermoredown-regulating FX expression by FX siRNA transfection decreasedsLe-a expression.

We also documented a functional axis linking FX and sLe-a ex-pression to the capacity of colorectal cancer cells to adhere to E-selectin. It was thus found that the FX enzyme is a limiting factor forthe capacity of at least those colorectal cancer cells used in this studyto adhere to endothelium. This conclusion is supported both bycorrelative evidence, as well as by direct evidence provided by trans-fection experiments.

It is assumed that highly metastatic tumor variants extravasate moreefficiently than cells with a low metastatic phenotype (18, 26, 29).This implies that the former variants would adhere better to endothe-

lial cells than the latter ones. Indeed, the highly metastatic variantsSW620 (23) and KM12SM (19) adhered better to endothelial cellsthan the corresponding low metastatic variants SW480 and KM12C.Each of these variant pairs originated from a single patient. The highlymetastatic variants expressed also higher levels of the FX enzyme andof sLe-a than the variants expressing a low metastatic phenotype. Itwas interesting to note that KM12SM, the most highly metastaticvariant of the KM12C primary tumor (19), expressed higher levels ofthe FX enzyme and of sLe-a than KM12L4, the less metastatic variantfrom the same tumor. It seems, therefore, that the degree of malig-nancy correlates positively with expression of the fucose-generatingFX enzyme and of its selectin ligand product sLea.

Taken together, these results lead us to hypothesize that the FXenzyme controls, by regulating selectin ligand biosynthesis, the inter-action of at least certain colorectal cancer cells with endothelium andthus the capacity to extravasate and form metastasis. If this hypothesiswill be confirmed by testing additional high and low metastatic cellpairs, the FX enzyme could become a target molecule for preventionof metastasis.

It was also demonstrated that the FX siRNA-mediated decrease ofFX expression by colorectal cancer cells down-regulated the ability ofthe siRNA-transfected cells to bind to the ECM protein fibronectin.This raises the possibility that fucose is also involved in the interac-

Fig. 10. Expression of fucosylated proteins by SW620 cells stably transfected with FXsiRNA. Lysates of control (360) and FX siRNA (205) transfected cells were assayed forthe expression of fucosylated proteins by Western blot analysis using horseradish perox-idase-conjugated Ulex europaeus agglutinin 1. Expression of FX protein in the two cellpopulations was determined as in Fig. 2A. Glyceraldehyde-3-phosphate dehydrogenase(GAPDH) served as loading control.

Fig. 8. Adhesion of SW620 cells transfected with FX siRNA to rE-selectin. 205 and360 cells were labeled with CFDA-SE reagent and incubated on rE-selectin in thepresence of antibodies against E-selectin (� E-selectin) or of irrelevant antibodies of thesame isotype (iso.con). For details, please see legend to Fig. 4B. The bars representmean � SD values obtained in four independent replicates (�, P � 0.01). 360 cells � con-trol cells; 205 cells � FX siRNA transfectants.

Fig. 9. Adhesion of SW620 cells transfected with FX siRNA toECM. 205 and control 360 cells were labeled with CFDA-SE reagentand incubated on ECM (A) or on fibronectin (B). For details, pleasesee legend to Fig. 8. The bars represent mean � SD of valuesobtained in three independent experiments. �, P � 0.04; ��,P � 0.002. 360 cells � control cells; 205 cells � FX siRNAtransfectants.

6577




tion between tumor cells and ECM. Indeed it was demonstrated thatfucosylation of integrins, e.g., �3�1, is essential for a correct assemblyof the integrin � and � subunits and for the binding to fibronectin (35).

Fucose participates not only in controlling adhesive properties butalso in other functions such as modifying signal transduction events.For example, Fringe, a fucose-specific glycosyltransferase initiateselongation of O-linked fucose residues attached to epidermal growthfactor-like sequence repeats of Notch (6). This glycosylation modu-lates Notch-mediating signaling (5). Another example is the O-fucosemodification of Cripto. This modification is essential for Nodal-dependent signaling (7). Because both Notch as well as Nodal-medi-ated signaling bear importance with respect to tumor progression (9,36–39), it would be of interest to compare these signaling pathwaysin colorectal cancer variants expressing high or low levels of FX.

Taken together, the results of the present study, as well as those ofother studies cited above, suggest that altering expression levels of theFX enzyme, thereby altering global fucosylation of colorectal cancercells, could have far reaching effects on the survival and phenotype ofthese cells.

The FX enzyme supplies 90% of the cellular fucose while the restis supplied by the salvage pathway (1, 40). In this pathway, extracel-lular fucose is taken up by the cell, phosphorylated by a fucose-kinase,and subsequently converted to GDP L-fucose by GDP-fucose-pyro-phosphorylase. The GDP L-fucose generated by the salvage pathwayundergoes an identical biosynthetic route as the GDP L-fucose gen-erated by the FX enzyme (34, 40). To fully assess the role of the FXenzyme in regulating various fucose-dependent reactions in colorectalcancer cells such as adhesion, Notch, or Cripto-mediated signaling, itwould be important to determine the relative contribution of thesalvage pathway to the cellular fucose supply.

REFERENCES

1. Smith PL, Myers JT, Rogers CE, et al. Conditional control of selectin ligandexpression and global fucosylation events in mice with a targeted mutation at the FXlocus. J Cell Biol 2002;158:801–15.

2. Sakamoto S, Watanabe T, Tokumaru T, Takagi H, Nakazato H, Lloyd KO. Expres-sion of Lewis-a, Lewis-b, Lewis-x, Lewis-y, siayl-Lewis-a, and sialyl-Lewis-x bloodgroup antigens in human gastric carcinoma and in normal gastric tissue. Cancer Res1989;49:745–52.

3. Homeister JW, Thall AD, Petryniak B, et al. The alpha(1,3)fucosyltransferasesFucT-IV and FucT-VII exert collaborative control over selectin-dependent leukocyterecruitment and lymphocyte homing. Immunity 2001;15:115–26.

4. Carlow DA, Corbel SY, Williams MJ, Ziltener HJ. IL-2, -4, and -15 differentiallyregulate O-glycan branching and P-selectin ligand formation in activated CD8 T cells.J Immunol 2001;167:6841–8.

5. Bruckner K, Perez L, Clausen H, Cohen S. Glycosyltransferase activity of Fringemodulates Notch-Delta interactions. Nature (Lond.) 2000;406:411–5.

6. Moloney DJ, Panin VM, Johnston SH, et al. Fringe is a glycosyltransferase thatmodifies Notch. Nature (Lond.) 2000;406:369–75.

7. Haltiwanger RS. Regulation of signal transduction pathways in development byglycosylation. Curr Opin Struct Biol 2002;12:593–8.

8. Yan YT, Liu JJ, Luo Y, et al. Dual roles of Cripto as a ligand and coreceptor in thenodal signaling pathway. Mol Cell Biol 2002;22:4439–49.

9. Adamson ED, Minchiotti G, Salomon DS. Cripto: a tumor growth factor and more.J Cell Physiol 2002;190:267–78.

10. Sullivan FX, Kumar R, Kriz R, et al. Molecular cloning of human GDP-mannose4,6-dehydratase and reconstitution of GDP-fucose biosynthesis in vitro. J Biol Chem1998;273:8193–202.

11. Ohyama C, Smith PL, Angata K, Fukuda MN, Lowe JB, Fukuda M. Molecularcloning and expression of GDP-D-mannose-4,6-dehydratase, a key enzyme for fucosemetabolism defective in Lec13 cells. J Biol Chem 1998;273:14582–7.

12. Lopez-Ferrer A, de Bolos C, Barranco C, et al. Role of fucosyltransferases in theassociation between apomucin and Lewis antigen expression in normal and malignantgastric epithelium. Gut 2000;47:349–56.

13. Eshel R, Besser M, Zanin A, Sagi-Assif O, Witz IP. The FX enzyme is a functionalcomponent of lymphocyte activation. Cell Immunol 2001;213:141–8.

14. Eshel R, Zanin A, Sagi-Assif O, et al. The GPI-linked Ly-6 antigen E48 regulatesexpression levels of the FX enzyme and of E-selectin ligands on head and necksquamous carcinoma cells. J Biol Chem 2000;275:12833–40.

15. Eshel R, Zanin A, Kapon D, et al. Human Ly-6 antigen E48 (Ly-6D) regulatesimportant interaction parameters between endothelial cells and head-and-neck squa-mous carcinoma cells. Int J Cancer 2002;98:803–10.

16. Izumi Y, Taniuchi Y, Tsuji T, et al. Characterization of human colon carcinomavariant cells selected for sialyl Lex carbohydrate antigen: liver colonization andadhesion to vascular endothelial cells. Exp Cell Res 1995;216:215–21.

17. Shirahama T, Ikoma M, Muramatsu T, et al. The binding site for fucose-bindingproteins of Lotus tetragonolobus is a prognostic marker for transitional cell carcinomaof the human urinary bladder. Cancer (Phila.) 1993;72:1329–34.

18. Thurin M, Kieber-Emmons T. SA-Lea and tumor metastasis: the old prediction andrecent findings. Hybrid Hybridomics 2002;21:111–6.

19. Morikawa K, Walker SM, Nakajima M, Pathak S, Jessup JM, Fidler IJ. Influence oforgan environment on the growth, selection, and metastasis of human colon carci-noma cells in nude mice. Cancer Res 1988;48:6863–71.

20. Brummelkamp TR, Bernards R, Agami R. A system for stable expression of shortinterfering RNAs in mammalian cells. Science (Wash. DC) 2002;296:550–3.

21. Laemmli UK. Cleavage of structural proteins during the assembly of the head ofbacteriophage T4. Nature (Lond.) 1970;227:680–5.

22. Lerrer B, Gilboa-Garber N. Differential staining of Western blots of avian egg whiteglycoproteins using diverse lectins. Electrophoresis 2002;23:8–14.

23. Hewitt RE, McMarlin A, Kleiner D, et al. Validation of a model of colon cancerprogression. J Pathol 2000;192:446–54.

24. Brandley BK, Swiedler SJ, Robbins PW. Carbohydrate ligands of the LEC celladhesion molecules. Cell 1990;63:861–3.

25. Mannori G, Crottet P, Cecconi O, et al. Differential colon cancer cell adhesion to E-,P-, and L-selectin: role of mucin-type glycoproteins. Cancer Res 1995;55:4425–31.

26. Voura EB, Sandig M, Siu CH. Cell-cell interactions during transendothelial migrationof tumor cells. Microsc Res Tech 1998;43:265–75.

27. Kim YJ, Borsig L, Han HL, Varki NM, Varki A. Distinct selectin ligands on coloncarcinoma mucins can mediate pathological interactions among platelets, leukocytes,and endothelium. Am J Pathol 1999;155:461–72.

28. Burdick MM, McCaffery JM, Kim YS, Bochner BS, Konstantopoulos K. Coloncarcinoma cell glycolipids, integrins, and other glycoproteins mediate adhesion toHUVECs under flow. Am J Physiol Cell Physiol 2003;284:C977–87.

29. Laferriere J, Houle F, Huot J. Regulation of the metastatic process by E-selectin andstress-activated protein kinase-2/p38. Ann NY Acad Sci 2002;973:562–72.

30. Varki A. Selectin ligands. Proc Natl Acad Sci USA 1994;91:7390–7.31. Renkonen R, Mattila P, Majuri ML, et al. In vitro experimental studies of sialyl Lewis

x and sialyl Lewis a on endothelial and carcinoma cells: crucial glycans on selectinligands. Glycoconj J 1997;14:593–600.

32. Karsan A, Cornejo CJ, Winn RK, et al. Leukocyte adhesion deficiency type II is ageneralized defect of de novo GDP-fucose biosynthesis. Endothelial cell fucosylationis not required for neutrophil rolling on human nonlymphoid endothelium. J ClinInvestig 1998;101:2438–45.

33. Listinsky JJ, Listinsky CM, Alapati V, Siegal GP. Cell surface fucose ablation as atherapeutic strategy for malignant neoplasms. Adv Anat Pathol 2001;8:330–7.

34. Noda K, Miyoshi E, Gu J, et al. Relationship between elevated FX expression andincreased production of GDP-L-fucose, a common donor substrate for fucosylation inhuman hepatocellular carcinoma and hepatoma cell lines. Cancer Res 2003;63:6282–9.

35. Litynska A, Pochec E, Hoja-Lukowicz D, et al. The structure of the oligosaccharidesof alpha3beta1 integrin from human ureter epithelium (HCV29) cell line. ActaBiochim Pol 2002;49:491–500.

36. Allenspach EJ, Maillard I, Aster JC, Pear WS. Notch signaling in cancer. Cancer BiolTher 2002;1:466–76.

37. Weijzen S, Rizzo P, Braid M, et al. Activation of Notch-1 signaling maintains theneoplastic phenotype in human Ras-transformed cells. Nat Med 2002;8:979–86.

38. Maillard I, Pear WS. Notch and cancer: best to avoid the ups and downs. Cancer Cell2003;3:203–5.

39. Adkins HB, Bianco C, Schiffer SG, et al. Antibody blockade of the Cripto CFCdomain suppresses tumor cell growth in vivo. J Clin Investig 2003;112:575–87.

40. Marquardt T, Luhn K, Srikrishna G, Freeze HH, Harms E, Vestweber D. Correctionof leukocyte adhesion deficiency type II with oral fucose. Blood 1999;94:3976–85.

6578




[CANCER RESEARCH 64, 8130–8133, November 1, 2004]

Corrections

FX Enzyme Controls the Adhesive Properties of CRC

In the article on how FX enzyme controls the adhesive properties ofCRC in the September 15, 2004 issue of Cancer Research (1), someof the labeling in Figures 7 and 10 was incorrect. The correctedfigures are below.

1. Zipin A, Israeli-Amit M, Meshel T, Sagi-Assif O, Yron I, Lifshitz V, Bacharach E,Smorodinsky NI, Many A, Czernilofsky PA, Morton DL, Witz IP: Tumor-microenviron-ment interactions: the fucose-generating FX enzyme controls adhesive properties of colo-rectal cancer cells. Cancer Res 2004;64:6571–8.

Chromosome 11q LOH in Human Breast Cancer

In the article on chromosome 11q LOH in human breast cancer inthe September 1, 1994 issue of Cancer Research (1), the name of oneof the contributing authors was misspelled. The correct spelling isRobert Winqvist.

1. Hampton GM, Mannermaa A, Winqvist R, Alavaikko M, Blanco G, Taskinen PJ,Kiviniemi H, Newsham I, Cavenee WK, Evans GA: Loss of heterozygosity in sporadichuman breast carcinoma: a common region between 11q22 and 11q23.3 Cancer Res1994;54:4586–9.

p110� Isoform of PI3 Kinase in Tumor Endothelium

In the article on p110� Isoform of PI3 Kinase in Tumor Endothe-lium in the July 15, 2004 issue of Cancer Research (1), the name ofone of the contributing authors, Jeffrey Brousal, was missing. Thecorrect list of authors should read: Ling Geng, Jiahuai Tan, EricHimmelfarb, Aaron Schueneman, Ken Niermann, Jeffrey Brousal,Allie Fu, Kyle Cuneo, Edward A. Kesicki, Jennifer Treiberg, Joel S.Hayflick, and Dennis E. Hallahan. Dr. Brousal’s affiliation is theDepartment of Radiation Oncology, Vanderbilt University School ofMedicine, Nashville, Tennessee.

1. Geng L, Tan J, Himmelfarb E, Schueneman A, Niermann K, Brousal J, Fu A, Cuneo K,Kesicki EA, Treiberg J, Hayflick JS, Hallahan DE. A specific antagonist of the p110�catalytic component of phosphatidylinositol 3�-kinase, IC486068, enhances radiation-induced tumor vascular destruction. Cancer Res 2004;64:4893–9.

Glioblastoma-Founding Human Neural Precursors

In the article on glioblastoma-founding human neural precursors inthe October 1, 2004 issue of Cancer Research (1), the e-mail addressof R. Galli should have been included in the requests for reprintssection. Dr. Galli’s e-mail address is [email protected].

1. Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, De Vitis S, Fiocco R, Foroni C, DimecoF, Vescovi A. Isolation and characterization of tumorigenic, stem-like neural precursorsfrom human glioblastoma. Cancer Res 2004;64:7011–21.

NF-�B in Squamous Cell Carcinoma

In the article on NF-�B in squamous cell carcinoma in the Sep-tember 15, 2004 issue of Cancer Research (1), the entries in Tables 1and 2 indicating NF-�B modulated genes should have been boldfaced.The corrected Tables 1 and 2 are reproduced below.

1. Loercher A, Lee TL, Ricker JL, Howard A, Geoghegen J, Chen Z, Sunwoo JB,Sitcheran R, Chuang EY, Mitchell JB, Baldwin AS Jr, Van Waes C: Nuclear factor-�Bis an important modulator of the altered gene expression profile and malignantphenotype in squamous cell carcinoma. Cancer Res 2004;64:6511–23.

Fig. 7. Expression of FX mRNA, FX protein and of sLe-a by SW620 cells stablytransfected with FX siRNA. A, FX mRNA. Expression was determined by Northern blotanalysis. Values represent the ratio between the signal of FX mRNA in the cells and thesignal of 18S rRNA in the same cell sample. B, FX protein. Expression was determinedas in Fig. 2A. C, sLe-a. Expression was determined as in Fig. 2B. A representativeexperiment (of three performed) is presented (360 cells � control cells; 205 cells � FXsiRNA transfectants). M � mean fluorescence; %pos � % positive cells.

Fig. 10. Expression of fucosylated proteins by SW620 cells stably transfected with FXsiRNA. Lysates of control (360) and FX siRNA (205) transfected cells were assayed forthe expression of fucosylated proteins by Western blot analysis using horseradish perox-idase-conjugated Ulex europaeus agglutinin 1. Expression of FX protein in the two cellpopulations was determined as in Fig. 2A. Glyceraldehyde-3-phosphate dehydrogenase(GAPDH) served as loading control.

8130

Table 1 Selected list of genes increased with tumor progression

FunctionGene Symbol Clone ID NF-�B Association

Fold change

LY-2/Ker LY-2/IkB-aM

Cell cycle/growthCyclin D1* Ccnd1 H3084D05 Target gene 3.351 �2.266Cyclin D2* Ccnd2 H3152D01 Target gene 2.957 �2.05Growth arrest specific 5* Gas5 H3113A12 6.304 �7.813Milk fat globule-EGF factor 8* Mfge8 H3126F11 Target gene 2.594 �3.525Protein phosphatase 3*† Ppp3cb H3065C08 Inhibitor of NF-�B 3.117 �2.231Proliferating cell nuclear antigen* Pcna H3021F12 Target gene 5.775 �2.811

ApoptosisBaculoviral IAP repeat* Birc2 H3074A02 Target gene 9.154 �6.013Bcl-2 related ovarian killer*† Bok1 H3081D02 1.994 �3.461Immediate early response 3* Ier3 H3057B07 Target gene 2.865 �3.851Transformation related protein*† Trp53 H3142D07 Target gene 2.9 �4.142Uchrp* Uchrp IMAGE:605056 2.188 �4.353

Inflammation/angiogenesisColony stimulating factor 1* Csf1 H3057D05 Target gene 14.639 �5.574Complement component 3*† C3 H3054A08 Target gene 5.145 �8.303FGF receptor* Fgfr4 IMAGE:406823 Target gene 1.982 �1.991Gro 1 oncogene* Gro1 H3051F10 Target gene 12.394 �4.094Histocompatibility 2-L* H2-L H3096A12 Target gene 2.342 �3.431Histocompatibility 2-D* H2-D H3141B11 Target gene 2.968 �3.927Interferon receptor* Ifnar H3118F09 Inhibitor of NF-�B 4.5 �3.054Lymphocyte antigen complex* Ly6e H3027D05 Inducer of NF-�B 4.029 �2.613

MetastasisIntegrin � 3*† Itga3 H3137A03 Inducer of NF-�B 15.597 �3.917Laminin � 5*† Lama5 H3002G01 Target gene 2.272 �2.967Laminin receptor 1 Lamr1 H3075G08 2.169 �1.619Plasminogen activator, tissue Plat H3080H11 Target gene 2.667 �2.487Procollagen type 5 � 2 Col5a2 H3156E09 Target gene 5.103 �2.094Syndecan 1* Sdc1 H3013F05 Target gene 2.948 �2.034

MetabolismATPase H� transport*† Atp6b H3120H04 Inhibitor of NF-�B 2.247 �2.858Branched chain ketoacid dehyd* Bckdk H3136B09 2.455 �3.042Choline kinase*† Chk H3088E07 Inducer of NF-�B 6.662 �3.252Cytochrome p450* Cyp1b1 J0216F07 Target gene 22.289 �4.286Glutathione-S-transferase* Gstm1 H3133A06 Target gene 3.037 �3.061Low density lipoprotein receptor* Ldlr H3014C04 Target gene 2.689 �3.226Mannose-6-phosphate receptor† M6pr H3092C05 Inhibitor of NF-�B 6.303 �3.612Potassium intermediate* Kcnn4 H3054H04 2.778 �2.197Solute carrier family 12*† Slc12a2 H3077B02 2.551 �2.145

Stress responseHeat shock protein, 70 kDa* Hspa5 H3032A08 Activates NF-�B 14.303 �6.369Heat shock protein 84 kDa* Hsp84 H3042G07 Activates NF-�B 3.437 �7.198Heat shock protein 86 kDa* Hsp86 H3023G01 Activates NF-�B 2.61 �2.34Heat shock protein cognate 70* Hsc70 H3133H01 Binds NF-�B 3.42 �10.229Superoxide dismutase* Sod1 H3130B11 Target gene 4.784 �2.145

Signal transductionAXL receptortyrosine kinase*† Axl H3152F05 Inhibitor of NF-�B 2.459 0.938CD97 (EGF-TM7)*† Cd97 H3032G06 2.283 �1.352Interleukin-1 receptor associated* Il1rak H3042E08 Activates NF-�B 1.999 �1.296Frizzled 7 homolog Fzd7 H3031A03 2.717 �0.843Growth arrest & DNA damage specific* Gadd45g H3054C02 Activates NF-�B 3.144 �1.407Growth factor receptor bound* Grb2 H3153D02 Activates NF-�B 2.483 �2.341N-myc downstream regulated* Ndr2 G0110H06 Inhibits p50 2.248 �2.039P13 kinase regulatory* Pik3r1 H3067B08 Activates NF-�B 4.206 �3.539Protein tyrosine phosphatase*† Ptpn13 H3118G02 Inhibitor of NF-�B 2.844 �9.071Ras p21 protein activator 3*† Rasa3 H3054E01 Activates NF-�B 3.223 �2.71Ras-related C3* Rac1 IMAGE:477981 2.042 �3.875Transferrin receptor* Trfr H3059G03 2.216 �1.145

Nuclear proteins/transcription factorsActivating transcription factor† Atf2 J0221F08 1.606 �2.696Breast cancer, early onset† Brca2 H3069F08 2.216 �0.604Butyrate response factor* Brf2 H3015E08 2.072 �2.008High mobility group AT* Hmga1 H3029B11 Target gene 2.816 �2.591Jerky* Jrk H3119F06 2.849 �3.329Myelocytomatosis oncogene* Myc H3089H11 Target gene 2.224 �2.42Nuclear factor �B p105* Nfkb1 H3072E09 2.919 �1.443Sex comb on midleg-like 1† Scml1 H3113B01 2.355 �1.154Yes-associated protein 65 kDa Yap H3089H07 2.072 �1.746

RNA processingDEAD box protein 3* Ddx3 H3018F11 2.976 �2.055DJ-1 protein† DJ-1 H3150D06 3.155 �4.611FGF inducible 14 Fin14 H3018G01 2.358 �3.355Nuclear ribonuclease Hnrpa1 H3111H11 4.926 �2.785RNA polymerase 1-1* Rpo1-1 H3049D09 2.703 �2.328

Protein synthesis/modificationERO1 like*† Ero1l H3126B01 2.529 �2.456Ribosomal protein L27a Rpl27a H3009B05 2.183 1.369

8131

Table 1 Continued

FunctionGene Symbol Clone ID NF-�B Association

Fold change

LY-2/Ker LY-2/IkB-aM

Ribosomal protein L8 Rpl8 H3141F09 2.778 1.005Ribosomal protein S18* Rps18 H3006C11 2.487 �2.424Ubiquitin activating enzyme E1 Ube1x H3022E03 Phosphorylates �B 5.313 �3.139Ubiquitin B Ubb H3138A08 Labels �B 8.144 �1.171Ubiquitin conjugating enzyme E2 Ube2h H3057B09 Phosphorylates �B 3.333 �1.199Ubiquitin conjugating enzymeE3†

Ube3a H3102B01 Phosphorylates �B 4.878 �5.096

Ubiquitin specific protease 9† Usp9x H3139F12 3.769 �3.175Structural proteins

Alpha tropomyosin* Tpm1 H3120G06 Binds p65 2.421 �4.36Cadherin* Cdh3 H3018F05 Inflammatory 2.328 �2.706Capping protein � 2 Cappa2 H3085F12 14.293 �7.458Dystroglycan 1* Dag1 H3008B05 Activates NF-�B 2.593 �4.738Epithelial protein lost*† Eplin H3153C05 Activates NF-�B 2.066 �2.971Fascin homolog 1* Fscn1 H3006D08 2.315 �2.581Four and a half LIM domains* Fhl2 H3033C07 2.169 �1.189Keratin complex 1 acidic* Krt1-18 H3021B02 Target gene 3.105 �5.517Keratin complex 2 basic* Krt2-8 H3031C01 2.094 �5.568PDZ and LIM domain 1* Pdlim1 J0824B03 5.7 �3.775Protocadherin 7*† Pcdh7 H3067F12 Inflammatory 6.556 �1.026Thymopoietin* Tmpo H3096B08 9.939 �8.153

OtherGlobin inducing factor† Gbif H3053F12 2.003 1.002Metallothionein 2 Mt2 H3013D11 Inhibits I�B degradation 2.145 �3.029Next to the Brca1*† Nbr1 H3061D04 2.367 �1.888RAN binding protein* Ranbp9 H3013A10 Accumulates I�B� 2.87 �3.189Repeat family 3 gene* Llrep3 H3107F07 3.556 �7.704Ring finger protein 19* Rnf19 H3153A08 Activates NF-�B 2.148 �4.402Sema domain, immunoglobin*† Sema3f H3134D09 2.191 �1.667Suppressor of Lec15† Supl15h H3090D12 2.001 �1.383TGF � inducible transcript* Tgfb1i1 H3122H01 3.068 �2.974

NOTE. Total number of genes regulated by NF-�B � 105/167. Total number of genes previously associated with NF-�B � 67/167.* Genes containing �B site in promoter region.† Genes containing ACTACAG motif in coding sequence.

8132

Table 2 Selected list of genes decreased with tumor progression

FunctionGene Symbol Clone ID NF-�B association

Fold change

LY2/Ker LY2/IkB-aM

Cell cycle/growthC-src tyrosine kinase* Csk L0237H04 �2.693 2.358Calmodulin Calm H3006H05 Activates NF-�B via IKK �2.554 2.42Cell division cycle homolog 25a Cdc25a H3050E04 �2.341 2.269Cell division cycle homolog 45 Cdc45l H3003E07 �2.032 2.029Cyclin C Ccnc C0117F09 �2.739 2.418Cyclin E2* Ccne2 C0186A01 �2.309 2.155Cyclin dependent kinase 4 Cdk4 H3147D06 Target gene of NF-�B �2.734 2.05Cyclin dependent kinase inhibit* Cdkn1c H3097D03 �5.208 2.913Platelet derived growth factor* Pdgfa H3146C02 �2.15 2.906

ApoptosisATP binding cassette* Abcd3 H3143E03 NF-�B site in promoter �2.89 2.535Bcl2/adenovirus E1B Bnip3 H3103B07 Transient inhibitor of NF-�B �2.364 2.436Fas associating w/death domain Fadd H3095D08 Inducer of NF-�B �3.597 2.975

Inflammation/angiogenesisCoagulation factor III*† F3 H3014G02 Activates NF-�B via IKK �4.672 3.858Interleukin 17 receptor Il17r H3008A03 Activates NF-�B via MAPK �3.021 3.048Interleukin 2 receptor Il2ra J0052C08 NF-�B site in promoter �2.262 2.207Lymphocyte antigen 6 complex Ly6 H3115A08 Inducer of NF-�B �5.617 3.311Prothymosin � Ptmb4 H3143A02 �21.739 8.858

MetastasisA disintegrin/MMP Adamts1 H3034B07 �2.695 2.736Cadherin 1* Cdh1 H3076B06 Associated with inflammation �2.597 2.313Kangai 1† Kai1 H3154D02 Target gene of NF-�B �2.816 2.139Lipocalin 2 Lcn2 H3083G02 Associated with inflammation �2.506 11.235Procollagen type 1 � Col1a2 H3125D01 �4.901 2.886Procollagen type II � Col2a1 H3026G09 �6.896 2.139Procollagen type III � Col3a1 H3005D11 Inducer of NF-�B �3.205 4.268Secreted acidic C-rich Sparc H3026D08 �8.928 2.872Tissue inhibitor of MMPs Timp3 H3031E01 Two NF-�B sites in promoter �3.3 3.289

MetabolismATPase, type 11A† Atp11a H3097B05 �2.888 2.734ATP synthase H� transport Atp5j2 H3118C01 Inhibitor of NF-�B �2.191 2.111Glutathione peroxidase* Gpx3 J0088G08 Inhibitor of NF-�B upregulates IkBa normal half-life �2.604 4.065Lipopolysaccharide binding*† Lbp H3086G08 Activates NF-�B via MAPK �2.977 2.103Phosphoprotein enriched† Pea15 H3014G07 Inducer of NF-�B �2.424 1.159Pyruvate dehydrogenase*† Pdha1 H3068G07 NF-�B site in promoter �3.021 1.643Sterol carrier protein 2*† Scp2 H3122F12 �2.412 1.488

Stress responseCrystallin � 2 Crya2 H3143B04 Inhibitor NF-�B �3.921 3.479

Signal transductionAdenylate kinase† Ak2 H3052D11 �3.755 3.546Max dimerization protein 4*† Mad4 H3131B07 �2.008 2.131MAD homolog 4 Madh4 H3128C04 �2.765 2.114NF-�B enhancer inhibitor* Nfkbia H3026A08 �1.433 4.611NIK-related kinase Nrk H3008B02 Activates NF-�B �2.244 4.859Phosphoglycerate kinase* Pgk1 H3023D06 �2.659 2.061Protein tyrosine phosphatase 4 Ptp4a2 H3088F03 �2.118 1.615Rho-associated coiled-coil Rock1 H3069C09 �2.808 3.183TNF receptor associated factor Traf1 H3015E06 NF-�B dependent �2.011 1.243

Nuclear proteins/transcription factorsCbp/p300 interacting transactivation Cited4 H3076H08 NF-�B co-activator �2.178 2.745High mobility group box 1 Hmgb1 H3126A05 Binds p50 subunit �3.104 1.739Jun oncogene Jun H3058C09 NF-�B co-activator �2.906 2.057Ras-related C3 Rac1 H3018C09 Inducer of NF-�B �2.004 3.329

RNA processingNuclear protein 220† Np220 H3029A07 �2.259 1.827RNA polymerase II* Rpo2-3 H3055H08 Coactivator of p65 �2.639 4.501

Protein synthesis/modificationEukaryotic translation 4g2 Eif4g2 H3113E10 �4.698 2.222Nedd4 WW-binding protein 4* N4wbp4 H3062G06 �6.966 2.814

Structural proteinsAlpha 2 glycoprotein 1*† Azgp1 IMAGE:521249 �2.013 1.398Beta spectrin 2*† Spnb2 H3010G09 �3.781 1.015Catenin beta* Catnb H3031E05 Regulated by IKK �3.998 3.061Fibronectin† Fn1 H3116A10 NF-�B site in promoter �6.201 4.723Catenin � 1*† Catna1 H3018E08 �1.996 1.723Tenascin C Tnc L0062E01 NF-�B site in promoter �2.557 1.956

OtherDeleted in polyposis Dp1 J0420H06 �2.473 2.137Insulin-like growth factor r2 Igfr2 H3148G08 �2.765 2.494Ninjurin 1* Ninj1 H3072B10 �2.579 4.878Rabaptin 5*† Rab5ep H3002C01 �2.739 1.333Topoisomerase II �*† Top2a H3139A05 Inducer of NF-�B �2.427 2.893Tumor differentially expressed† Tde11 H3014H10 �2.593 1.18WW domain binding 5* Wbp5 H3127H02 �2.087 3.313Zinc finger protein 68*† Zfp68 H3058F07 �4.122 5.885

NOTE. Number of genes regulated by NF-�B � 47/141. Number of genes previously associated with NF-�B � 39/141.* Genes containing kB site in promoter region.† Genes containing ACTACAG motif in coding sequence.

8133

2004;64:6571-6578. Cancer Res Adi Zipin, Mira Israeli-Amit, Tsipi Meshel, et al. of Colorectal Cancer Cells

PropertiesFucose-Generating FX Enzyme Controls Adhesive Tumor-Microenvironment Interactions: The

Updated version

http://cancerres.aacrjournals.org/content/64/18/6571

Access the most recent version of this article at:

Cited articles

http://cancerres.aacrjournals.org/content/64/18/6571.full#ref-list-1

This article cites 39 articles, 14 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/64/18/6571.full#related-urls

This article has been cited by 4 HighWire-hosted articles. Access the articles at:

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

SubscriptionsReprints and

[email protected] at

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

Permissions

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

.http://cancerres.aacrjournals.org/content/64/18/6571To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/content/64/18/6571.full#ref-list-1

http://cancerres.aacrjournals.org/content/64/18/6571.full#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Tumor-Microenvironment Interactions: The Fucose-Generating ...€¦ · fucosylation events are...

Documents

Transcript of Tumor-Microenvironment Interactions: The Fucose-Generating ...€¦ · fucosylation events are...