[CANCER RESEARCH57, 5201-5206, December 1, 1997]

Advances in Brief

Carcinoma-associated Expression of Core 2 @-1,6-.N-AcetylglucosaminyltransferaseGene in Human Colorectal Cancer: Role of O-Glycans in Tumor Progression'

Kazuhisa Shimodaira, Jun Nakayama,2 Naoshi Nakamura, Osamu Hasebe, Tsutomu Katsuyama, and Minoru FukudaSecond Department of Internal Medicine (K. S.] and Department of Laboratory Medicine (K. S., J. N., 7'.K.], Shinshu University School of Medicine, Matsumoto 390, Japan;Department of Internal Medicine, Nagano Municipal Hospital, Nagano 381, Japan (N. N., 0. H.]; and The Glycobiology Program, La Jolla Cancer Research Center. TheBurnha,n institute, La Jolla, California 92037 (M. F.]

Abstract

Recently, it was demonstrated that an increased level of NeuNAca2-3Galpl-4(Fucal-3)G1cNAq3-R (sialyl Lex) and NeuNAca2-3Galfil3(Fuca1-4)GlcNAc@i-R(sialyl Lea)expression on the surface of colorectalcancer cells is positively correlated with progression of the disease. It hasnot been determined, however, which type of glycans, N- or O-glycans, ismore closely associated with progression when cancer cells express thoseollgosaccharides. To address this problem, we have examined expressionof sialyl Le@and slalyl Lex, those oligosaccharides in O-glycans, and core2 fi-1,6-N-acetylglucosaminyltransferase (C2GnT) transcripts in colorec.tat cancerspedmensfrom 46 patientsand comparedthoseresultswithdinicopathological variables. C2GnT is a glycosyltransferase that is responsible for the core 2 branch, which is critical for biosynthesis of sialylLea and sialyl Le5 in O-glycans. Sialyl Le@and sialyl Le@were determinedby immunohWochemistry, and C2GnT transcripts were detected by reverse tramcripflon-PCR. Sialyl Lea or sialyl Le@@ O-glycans was assessedby combining linmunohistochemistry for sialyl Le or sialyl Le@withreverse transcrlption-PCR for C2GnT. Sialyl Lea,detected on cancer cellsin 74% of patients, was well correlated with lymph node metastasis,whereas sialyl Lea@ sialyl Lex j@O-glycans, which were specificallydetected in cancer tissues of 50 and 61% of patients, respectively, wereclosely associated with lymphatic and venous invasion. In addition,C2GnT, which was specifically detected in cancer tissues of 63% ofpatients, was closely correlated with the vessel invasion, as well as depthof tumor Invasion. These results strongly suggest that sialyl Lea@ sialylLex@ O-glycans and C2GnT, expressed in cancer cells, may play hnportent roles in tumor progression through vessel or direct invasion.

Introduction

Specific sets of carbohydrates on the cell surface play critical rolesin many biological events, such as development, differentiation, and

pathological conditions, such as inflammation and oncogenesis (1—3).In particular, among the carbohydrates, sialyl Le'@3and an isomer ofsialyl Le5, sialyl @aare enriched on the cell surface of carcinomacells (4, 5) and could serveas ligandsfor the cell adhesionmoleculeE-selectin, which is exclusively expressed on the cell surface ofactivated endothelial cells (6). During the process of tumor metastasis,the adhesionof carcinomacells to endothelialcells in target organsmaybe crucialfor successfulcolonizationby tumorcells (7). In fact,it has been demonstrated that the highly metastatic colonic carcinoma

Received7/9/97;accepted10/13/97.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with

18U.S.C.Section1734solelyto indicatethisfact.1 This work was supported by Grant-in-Aid (C) 09670222 for Scientific Research from

the Ministryof Education,ScienceandCultureof Japan(to J. N.),a GrantfromEISA]Co., Ltd. (to T. K.), and National Cancer Institute Grams CA 48737 and, in part, CA33000(to M.F.).

2 To whom requests for reprints should be addressed, at Department of Laboratory

Medicine, Shinshu University School of Medicine, Asahi 3-1-1, Matsumoto 390, Japan.Phone: 81-263-37-2801; Fax: 81-263-34-5316.

3 The abbreviations used are: sialyl Lex, NeuNAca2-3Ga1@1-4(Fuca1-3)G1cNAc@-R;sialyl Lea, NeuNAca2-3GaI@1-3(Fuca1-4)GlcNAc@-R; C2GnT, core 2 @-1,6-N-acetylglucosaminyltransferase; RT-PCR, reverse transcñption-PCR.

cell line, KM12-L4, expresses more sialyl Lex moieties on the cellsurface than does a less metastatic counterpart, the KM12-SP cell line(8), and that KM12-L4 cells also adhere more efficiently to activatedendothelialcellsor ChinesehamsterovarycellsexpressingE-selectin(9). More recently, Izumi et al. (10) isolated two sublines showingdistinct expression levels of sialyl Le@from a human colon cancer cellline, KM12-C, and demonstrated that liver metastasis occurred morefrequentlywhena sublinestronglyexpressingsialylLexwas injectedinto the spleen of nude mice (10). In addition, a clinicopathologicalstudy using immunohistochemistry provided evidence that the amountof sialyl Lex on colorectal cancer cells is positively correlated to poorprognosis of the patients (1 1). Similarly, expression of sialyl L&' incancer cells is also associated with poor prognosis (12). These resultsclearly establish that the amount of sialyl Lex and sialyl Le―expressedon the cell surface of carcinoma cells is closely associated withprogression of colorectal cancer.

Sialyl Lex and sialyl @aassociated with glycoproteins, can bepresent in N-glycansor O-glycans(8, 13, 14). However,it has notbeen determined which type of glycans is more closely associated

with the progression of colorectal cancer when carcinoma cells express sialyl @aor sialyl Le5 on their cell surface. As shown in Fig. 1,the core 2 branch is critical for biosynthesis of sialyl Lea or sialyl Le@in O-glycans because both determinants can be formed via this branchstructure (2). C2GnT is a glycosyltransferase that is responsible forcore 2 branch synthesis, and a cDNA encoding this enzyme wasisolated from the human promyelocytic cell line, HL-60, by expression cloning (15). Thus, the presence of C2GnT in carcinoma cells isdirectly correlated with the presence of sialyl Le―or sialyl @XinO-glycans that are present on the carcinoma cells. Here, we combineRT-PCR analysis of C2GnT transcripts with immunohistochemistryfor sialyl @aand sialyl Lex to determine whether sialyl Lex or sialyl

@a@ present in O-glycans in colorectal cancers obtained by endo

scopic biopsy. We also assess the relationship between the expressionof sialyl @aor sialyl L@e5in O-glycans and clinicopathological features exhibited by the patients. Last, we statistically evaluate thecorrelationbetweenthe expressionof C2GnTtranscriptsper se andcinicopathologicalvariablesto determinethe significanceof C2GnTin tumor progression.

Patients and Methods

Patient& Biopsy specimens of carcinoma tissues were obtained using acolonoscope from 46 patients with primary colorectal cancer at Nagano Municipal Hospital (Nagano, Japan) after informed consent was obtained. Carcinomas that were limited to the mucosa were excluded from this study. Inaddition, biopsy specimens of the colorectal mucosa, which was remote fromthe carcinoma and apparently normal at the endoscopic level, were alsoobtained using a colonoscope from the same patients. The patients included 24males and 22 females, ages 28—91years (mean, 67.3 years). Histopathologicalfeatures, including the depth of tumor invasion, vessel invasion, and lymphnode metastasis, were determined at the subsequent surgical operation performed within 4 weeks after biopsy. Biopsy specimens for RT-PCR analysis

5201

on July 7, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

C2GnT GENE IN COLORECTAL CANCER

GalNAcal-Ser/Thr (Tn antigen)'

GaI@1-3Ga1MAca1-SerIThr (core I ,T antigen)

,/ GICNA@f1\

GaI@1-3GalMca1-Ser/Thr (core 2)

G&@1-3/4GIcNAcfl1-3(GaIj@1-4GIcNAc@1)@ (r@O)

I 61L GaI@1-3GalNAca1-Ser/Thr

Fucal NeuM1@ca2-3GaI@1-3GalM@ca1-Ser/Thr

Fig. 1. The biosynthetic pathways of O-glycans. Note thatcore 2 branch formed by C2GnT is critical for the synthesisofboth sialyl Lea and sialyl Le' in O-glycans (based on Refs.2 and 13—15).

Sialyl @a@ si@i LeX

were immediately frozen and kept at —80°Cuntil extraction of total RNA. Inparallel, other tissue specimens of carcinoma tissues, as well as normal mucosafor immunohistochemical staining, were fixed for 48 h at room temperature in20% phosphate-buffered formalin (pH 7.4), embedded in paraffin, and cut into3-pm sections.

RT-PCR. Total cellular RNA was isolated from approximately8 mg ofwet tissue from each biopsy specimen, according to the acid guanidiniumthiocyanate-chioroform extraction method, using an Isogen kit (Nippon-Gene,Tokyo, Japan; Ref. 16). After digestion of the genomic DNA with DNase I,cDNA was then synthesized by reverse transcription of the extracted RNAusing an oligo(dT),5 primer. Using this cDNA as template, C2GnT cDNA,which encodes an enzyme that is critical for O-glycan biosynthesis, wasamplified by PCR. On the basis of the published sequence (15), upstreamand downstream primers used in the first round of PCR were 5'-GTGCTCAGAATGGGGCAGGATGTCACCTGG-3'(nucleotides —214to —185)and 5'-TCACTACTAGGAUCTCCCCAGCAAGCTCC-3' (nucleotides138—167),respectively. These primers were designed to amplify a region fromexon 1 to exon 2 of the C2GnT gene to distinguish amplified cDNA fromamplified genomic DNA (17). Fifty pAof reaction mixture, containing lXPCR buffer, 1.0 mMMgC12,200 @LMdNTP, 1 @.tMeach primer, 1.25unit of Taqpolymerase, and 1.0 @gof the reverse-transcribed RNA, were denatured for 3mm at 98°C,prior to 40 cycles at 95°Cfor 1 mm, 60°Cfor 1 mm, and 72°Cfor 2 mm. In the second round, 1/50 of the amplified cDNA was reamplifiedfor 20 cycles under the same conditions using nested primers. The upstreamand downstream primers for the second round of PCR were 5'-GATGTCACCTGGAATCAGCA-3' (nucleotides —196 to - 177) and 5'-GCAGCAACGT

CCTCAGCAT-3' (nucleotides 1—19),respectively (15). When genomic DNAwas amplified using these primer sets, no PCR product was detected. As aninternal control for the RT-PCR analysis, the (32-microglobulin transcripts

were amplified from the same cDNA samples as described (18). Upstream anddownstream primers for f32-microglobulin gene were 5'-ACCCCCACTGAAAAAGATGA-3' and 5'-ATmCAAAC@TCCATGATG-3', respectively. Amplified products were electrophoresed through 3% agarose gels.

Immunohistochemistry. Deparaffinizedbiopsy specimenswere subjectedtoimmunohistochemical staining for detection of sialyl Lea and sialyl @X@nigtheindirect method as described previously (19). Mouse monoclonal antibodies directed against sialyl Lea and sialyl @Xwere purchased from DAKO (Glostrup,Demnark)and BectonDickinson(San Jose, CA), respectively.Antimouseimmunoglobulin antibody conjugated with horseradish pemxidase (DAKO) was used asthe secondaryantibody,and peroxidaseactivitywas visualizedwith diaminobenzidine-H202 solution. A control experiment was done, omitting the primaryantibody from the stainingprocedure,and no specific stainingwas found. Tissues@mans containing more than 5% positively stained carcinoma cells weredefined as positive.Others were classifiedas negativeaccordingto the criteriaofNakamorietaL (11).

To confirm the presence of sialyl Lea or sialyl Lex in O-glycans, unfixedfrozen sections of carcinoma tissues expressing C2GnT mRNA, as well assialyl Lea and sialyl Lex were prepared from surgical specimens that were

immediately removed from three patients during surgical operation and digested with O-sialoglycoprotein endopeptidase, which specifically cleavesO-glycan-containing peptides from glycoproteins (20), before immunostainingwith anti-sialyl @aor anti-sialyl Lex antibody. Briefly, colorectal cancerspecimens, including normal mucosa were immediately embedded in OCTcompound (Miles, Elkhart, IN) and snap-frozen at —80°C.Frozen tissuespecimens were sectioned at 4 pm using Kryostat CM3000 (Leica, Nussloch,Germany) and then digested with 120 @.&g/miO-sialoglycoprotein endopeptidase from Pasteurella haemolytica (Accurate Chemical, Westbury, NY) at

37°Cfor 4 h, followed by immunostaining for sialyl Lea or sialyl Lex. Theeffect of digestion was evaluated by comparing digested specimens with serial

sections incubated with PBS instead of enzyme.Statistical Analysis of Data. The relationship between the expression

pattern of C2GnT in carcinoma tissue and clinicopathological variables wasstatistically determined by either the@ test or Fisher's exact test usingStatView 4.5 software (Abacus Concepts, Berkeley, CA). The clinicopathological variables examined in this study included tumor size, histologicaldifferentiation, depth of tumor invasion, lymphatic invasion, venous invasion,lymph node metastasis, and tumor stage based on the histopathological tumornode-metastasis classification system (21). Similarly, the relationship betweenthe expression of sialyl Lea or sialyl Lex on carcinoma cells and cinicopathological variables was evaluated by the same tests. A significant difference wasestablished when a P of <0.05 was obtained.

Results and Discussion

Expression of Sialyl Lea and Sialyl @Xm Colorectal Cancer.Immunohistochemical analysis revealed that, among 46 patients cxanüned,sialyl @aand sialyl Lex were detected on the apical surfaceof carcinoma cells in 34 (74%) and 43 (93%) patients, respectively(Fig. 2, A and C). In the normal mucosa, on the other hand, sialyl @aand sialyl Lex were detectable in 12 (26%) and 3 (7%) of the patients,respectively (Fig. 2, A and C, insets).

To determine the correlation between elevated sialyl Lea orsialyl Lex level and progression of colorectal cancer, a statisticalanalysis was performed to compare the expression pattern of sialylLea or sialyl Lex on carcinoma cells with clinicopathologicalvariables. As shown previously (11, 12), the expression of sialylLea and sialyl Lex in the carcinoma cells is associated with progression of colorectal cancer. In particular, statistically significantdifferences were found between expression of sialyl Lea on carcinoma cells and lymph node metastasis (P = 0.02) and tumor stage(P = 0.03). Although statistical significance could not be obtainedin this study, carcinoma progressed more when carcinoma cellsexpressedsialyl Lexthan those lacking sialyl Lex(data not shown).It is possible that examination of a larger number of patients

5202

NeuNAco2-3GaI@1-3GalNAcu1-Ser!Thr

INeUNACa2@\

NeuMccz2-3Gai@1-3Ga1NAca1-Ser/Thr

on July 7, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

C2GnT GENE IN COLORECFALCANCER

(—) (+)‘•.:;_@ ‘ ...@ @‘

@--:.@ ,,..@‘@ ; ..@@

. .@ v―;@'.@.@@ ..,@..v.@ •@.

•@‘@‘:@@ -@. ¼ -- •@ @;

:@@“-ar'.-' ‘@, C @‘k@ . .-@.,-, ,....-@. -V@ .,@ @- . .v@@ .@t.@@@ @;

@ @t@\ f 4)

,__;-‘ . a,.@@ ..,.c..:@&:..@‘@ @e.t@'y,,'-, ,, ,@ •:‘,;@@ @?@ , ..

.. ,-@@ .;,@..@ @.;: .. .@ _ :/.@ ‘ ,@\ /

..-,@‘.@ “/à' @. t

@@ ,..a'@;; . :@.‘@@

@ @.‘,@@ t..@@.,@ •,@t9@/ç@,@ ‘.@?i?@'}@_::. @‘#‘@ ‘ ..@‘.

. .@- .@ . .J-• . @‘

@ . ., @:@

.. ;...@

@ @I. •_ •‘@.•@

?(‘‘@@. @.- .@ @,;:c4.'. tf

‘...@... .

...-

.,

09

.@ . .@.‘— f ....,@ ;. a' .- . - 9_@s., !@@@ , •@@@ -. . ‘ ..‘

,@.@ .@ .- •@;,-@

@C

@@ ,@@@ ::@ @,.@

@ . @-,t

@ \.@@

. , .

5- @. - \

@ S. . ‘.@‘.@

@‘::@ :‘.@

@@p4@•:.:‘

@t•:.1,., ::•@@ .@ .;,.@ .@,.@ ...,@

E1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

C NC N C NC N C N C NC N C N C N C N C N C N C N C NC NC N C N C NC N C N

O—sialoglycoproteinendopeptidase

Sialyl Lea

Sialyl Lex

Patient

C2GnT

$2m

Fig. 2. Expression of sialyl Lea and sialyl Lax and C2GnT transcripts in colorectal cancer and normal mucosa. A—D,immunohistochemistry for sialyl Lea (A and B) and sialyl La'(C and D). In the carcinoma cells from the patient expressing C2GnT transcripts (patient 20 in E), the expressions of sialyl La@(A) and sialyl La' (C) were significantly reduced afterdigestion with O-sialoglycoprotein endopeptidase (B and D). In the normal mucosa, intensity of staining for sialyl@ and sialyl Le' was not changed by preincubation with the enzyme(A—D,insets). Hematoxylin was used for counterstain. Bar, 100 @sm;bar (inset), 200 pm. E, RT-PCR analysis of C2GnT transcripts expression in cancer and normal mucosa from20 patients randomly selected from 46 patients. C2GnT transcripts were specifically expressed in cancer tissues in patients 2—9and patients 15—20.C, cancer tissue; N. normal mucosa;fJ2m, @3@-micmg1obu1in.

lacking sialyl Le5 on carcinoma cells will identify statistically were examined. These results, combined, support the conclusionsignificant differences. In fact, Nakamori et a!. (11) reported that that sialyl Lea and sialyl Lex on carcinoma cells are involved inthe expression of sialyl Lex on colorectal cancer cells was corre- progression of colorectal cancer by mediating E-selectin-dependlated, with statistical significance, with depth of tumor invasion, ent adhesion to endothelial cells (2, 3). However, it has not beenlymphatic invasion, and lymph node metastasis when 132 patients clarified which type of glycans, 0- or N-glycans, contributes more

5203

on July 7, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Table 1 Statistical analysis of the relationship between the expression patternofC2GnTtranscripts in colorectal cancer tissues and clinicopathologicalvariablesNo.

of patients (%) (n =46)C2GnT

transcriptsPositive

Negative(n29)(n17)Tumor

size(mm)307 (24) 7 (41)@ 23

>30 22 (76) 10 (59)Histologicaldifferentiation―Well

20 (69) 11 (65)0.77Moderatelyandpoorly9(31)6(35)Depth

ofinvasion'smand pm 2 (7) 6 (35) o o@'

ss, Se, and si 27 (93) 11(65)LymphaticinvasionNegative

0 (0) 5 (29)@Positive 29 (100) 12 (71)VenousinvasionNegative

3 (10) 7 (41) o 02dPositive 26 (90) 10(59)Lymph

nodemetastasisNegative14(48) 11 (65) 028

Positive 15 (52) 6(35)TumorstageeI

and II 12 (41) 10 (59)@ 25m andIV 17(59) 7(41)a

M@Jy5@ by x2 test or Fisher's exact test.

b Well, well-differentiated adenocarcinoma; moderately, moderately differentiated ad

enocarcinoma;poor,poorlydifferentiatedadenocarcinoma.s@submucosa; pm. muscularis propria; ss, subserosa; se, serosa exposed; si, inva

sive to neighboring tissue.dp <o.os.e

Stage I, pT,_2, pN0, M,@; stage II, pT@4, pN@,, M0; stage II, pT,_4, pN1.3,@ stageIV,

pT,_4, pN,_3, M1.

C2GnT GENE IN COLORECTAL CANCER

significantly to the progression of colorectal cancer when carcinoma cells express sialyl Lea and sialyl Lex on cell surface glycoproteins.

Carcinoma-associated Expression of C2GnT Transcripts inHuman Colorectal Cancer. To determine the presence of sialyl Leaand sialyl LeX in O-glycans, it is necessary to define the expressionpattern of C2GnT in colorectal cancer because C2GnT is an enzymethat is required to form a critical branch of core 2 in the biosynthesisof O-glycans, and both sialyl Le@and sialyl LeX in O-glycans aresynthesized via this particular branch (Fig. 1).

The expression of C2GnT in colorectal cancer and normal mucosawas analyzed by RT-PCR. Among 46 patients examined, C2GnTtranscripts were specifically detected in carcinoma tissues from 29patients (63%) but not in the normal mucosa of all of the 46 patients(Fig. 2E and Table 1). This result indicates that sialyl Le@and sialylLe', on the cell surface of carcinoma cells expressing C2GnT, couldbe present in O-glycans. Increased activity of C2GnT has been demonstrated in many biological processes and diseases, such as T-cellactivation, leukemia, and immunodeficiencies such as Wiskott-Aldrich syndrome and AIDS (2). In rodents, Yousefi et al. (22) reportedthat transfection of the T24H-ras gene into rat2 fibroblasts and mammary adenocarcinoma SP1 cells resulted in a 70% increase in C2GnT

activity in both cell lines. The present study provides further evidencethat this enzyme is associated with malignant transformation in thehuman large intestine and suggests that O-glycans having the core 2branch may play important roles in tumor progression.

Expression of Sialyl Lea and Sialyl Lea in O-Glycans in Cobrectal Cancer. Combining the results obtained from immunohistochemistry for sialyl Le@ and sialyl Le'@with those obtained fromRT-PCR analysis of C2GnT, we can differentiate sialyl Le@and sialylLe' in O-glycans from those found in N-glycans and/or in glycolipids.As determined by RT-PCR analysis (Fig. 2E), C2GnT expression incolorectal cancer was detected in 23 (68%) of the 34 patients expressing sialyl Le' and in 28 (58%) of the 43 patients expressing sialyl Lex.These results indicate that sialyl Le@and sialyl LeXcould be present inO-glycans on the cell surface of carcinoma cells of these patients. Inthe patients lacking C2GnT transcripts in cancer tissues, sialyl Le@andsialyl LeX must be present in N-glycans of glycoproteins and/or inglycolipids in the carcinoma cells. In six patients, sialyl Le@was notdetectable in carcinoma cells, despite the presence of C2GnT transcripts. In addition, one patient lacked sialyl LeX in carcinoma cells,but C2GnT was apparently expressed in carcinoma tissue. In thesepatients, it is possible that other glycosyltransferases, such as a-2,3-sialyltransferase or a-l,3/1,4-fucosyltransferases, involved in biosynthesis of sialyl Le@and sialyl Lex, were absent in carcinoma tissues(23,24).

To corroborate the presence of sialyl Le@and sialyl LeX in 0-glycans of carcinoma cells, unfixed frozen sections of the colorectalcancer expressing sialyl Le@and sialyl Lex, together with C2GnTmRNA, were prepared from three patients and subjected to 0-sialo

glycoproteinendopeptidasedigestion(20), prior to immunostainingwith anti-sialyl Le@or anti-sialyl Le' antibodies. After removal of0-glycans with 0-sialoglycoprotein endopeptidase, expression of sialyl Le@and sialyl LeXin carcinoma cells was significantly reduced inall patients examined (Fig. 2, B and D). These results strongly suggestthat sialyl Le―and sialyl LeX were actually expressed in 0-glycans oncarcinoma cells expressing C2GnT and that the enzyme-resistantstaining may be those in N-glycans and/or in glycolipids. In fact, sialylLea and sialyl Le'@in the normal mucosa, in which C2GnT was not

detectable, were not reduced, even after digestion with 0-sialoglycoprotein endopeptidase (Fig. 2, B and D, insets), indicating that sialylLea and sialyl Lex, in normal mucosa, are present in N-glycans and/or

in glycolipids. These findings clearly establish that sialyl Le@and

sialyl LeX present in 0-glycans but not in N-glycans are newlysynthesized in association with malignant transformation of the cobrectal mucosa.

Sialyl@ @dSialyl @XDeterminants Present in O-GlycansAre Involved in Vessel Invasion of Coborectal Cancer Cells. Todetermine whether the expression of sialyl Le@or sialyl LeX in 0-glycans is involved in progression of coborectal cancer, the relationship between the expression of C2GnT in carcinoma tissues andclinicopathobogicalvariables was statisticallyevaluated among patientsexpressingsialylLe@or sialylLeXon carcinomacells.As shownin Table 2, patients expressing C2GnT mRNA in carcinoma tissueshad more advanced clinicopathobogical variables compared to patientslacking the C2GnT transcripts. In particular, lymphatic and venousinvasions were more frequenfly found in patients expressing C2GnTthan in patients lacking C2GnT with statistical significance(P < 0.05). Interestingly, the significant differences found betweenexpression of sialyl Le@and lymph node metastasis or tumor stagecould no longer be obtained (Table 2).

These results indicate that, among patients expressing sialyl Le@and sialyl Le' on carcinoma cells, coborectal cancer progresses morethrough vessel invasion rather than lymph node metastasis, whenthese carbohydrates are presented in 0-glycans. Thus, it is suggestedthat these carbohydrate ligands in 0-glycans found on the cell surfaceof carcinoma cells could efficiently facilitate selectin-mediated adhesion between carcinoma cells and endothelial cells. In fact, Mannon eta!. (25) recently demonstrated that treatment of two human coloncarcinoma cell lines, LS18O and T84, with 0-sialoglycoprotein endopeptidaseresultedin reducedbindingto notonlyE-selectinbut alsoP- and L-selectins. In addition, venules distributed along the invasive

margin of colorectal cancer were reported to express E-selectin aswell as P-selectin (26). These results suggest that coborectal carcinomacells may adhere to endothelial cells by binding of not only E-selectinbut also P-selectin. For binding of P-selectin, however, the carbohy

5204

on July 7, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

No. of patients(%)Sialyl

Laa@pesitiVe@røu@(n = 34) SialylLa'-positivegroup(n=43)C2GnT

transcripts C2GnTtranscriptsPositive

(n = 23) Negative (n = 11) P― Positive (n = 28) Negative (n=15)F'Tumor

size(mm)@307(30) 4(36) >099 7(25) 6(40)

>30 16 (70) 7 (64)@ 21 (75) 9 (60)Histologicaldifferentiation―032Well

14 (61) 8 (73)@@ 19(68) 10 (67)Moderately and poorly 9 (39) 3 (27)@ 9 (32) 5(33)>@Depth

ofinvasioncsmand pm 2 (9) 3 (27)@ @o 2 (7) 4 (27)

55' SC, and si 21 (91) 8 (73)@ 26 (93) 1 1(73)16Lymphatic

invasionNegative0 (0) 3 (27)@ 03d 0 (0) 4 (27)

Positive@ 23 (100) 8 (73) • 28 (100) 11(73)VenousinvasionNegative

2 (9) 5 (45)@ 02d 2 (7) 6 (40)Positive 21 (91) 6 (55) • 26 (93) 9 (60)o

01―Lymph

nodemetastasisNegative9 (39) 6 (55) o 47 13 (46) 9 (60)

Positive 14 (61) 5 (45)@ 15 (54) 6 (40)o@toTumor

stageeland!!8(35) 5(45) 071 11(39) 8(53)

ifiandIV 15(65) 6(55)@ 17(61) 7(47)038a

Analyzed by x2 test or Fisher's exact test.

b Well, well-differentiated adenocarcinoma; moderately, moderately differentiated adenocarcinoma; poorly, poorly differentiatedadenocarcinoma.Csm,submucosa; pm, muscularis propria; ss, subserosa; se, serosa exposed; si, invasive to neighboring tissue.

dp <0.05.e

Stage I, pT1_2, pN0, M@@;stage II, pT@4, pN0, M€,@;stage III, pT,_4, pN,_3, M@@;stage N, pT,_4, pN,3, M,.

C2GnT GENE IN COLORECTALCANCER

Table 2 Statistical analysis of the relationship between the expression pattern of C2GnT transcripts in colorectal cancer tissues and clinicopathological variables in patientsexpressing sialyl @aor sialyl Lex on carcinoma cells

metastasis, whereas sialyl Lea and sialyl Lex in 0-glycans oncarcinoma cells are highly correlated with lymphatic invasion aswell as venous invasion. Moreover, the expression of C2GnTtranscripts per se in carcinoma tissue is closely related to not onlyvessel invasion but also depth of tumor invasion. These resultssuggest that oligosaccharides with the core 2 branchings, includingsialyl Lea and sialyl Lex, may play important roles in tumorprogression through direct or vessel invasion and that the presenceof unidentified 0-glycan(s) having the core 2 branch, in addition tosialyl Le@and sialyl Le―,may play a role as ligands for carbohydrate-binding proteins including selectins. Further studies will beimportant in determining whether the expression of C2GnT isclosely related to tumor progression in the absence of siabyl Leaand sialyl Lex and to determine whether the oligosaccharide structure other than sialyl Lea and sialyl Lex are responsible for tumorprogression.

Acknowledgments

We are grateful to Drs. Kendo Kiyosawa and Taiji Akamatsu (ShinshuUniversity School of Medicine) for their encouragement and discussion duringthis study and Drs. Edgar Ong (The Burnham Institute) and Elise Lamar (TheSalk Institute, La Jolla, CA) for critical reading of the manuscript. We alsothank Hirofumi Sakurai and Akiko Kobayashi (Nagano Municipal Hospital)for their excellent technical assistance.

References

1. Feizi, T. Demonstration by monoclonal antibodies that carbohydrate structures ofglycoproteins and glycolipids are onco-developmental antigens. Nature (Land.). 314:53—57,1985.

2. Fukuda, M. Possible roles of tumor-associated carbohydrate antigen. Cancer Res., 56:2237—2244,1996.

3. Hakomori, S. Tumor malignancy defined by aberrant glycosylation and sphingo(glyco)lipid metabolism. Cancer Res., 56: 5309—5318, 1996.

4. Itzkowitz, S. H., Yuan, M., Fukushi, Y., Palekar, A., Phelps, P. C., Shamsuddin,A. M., Tromp, B. F., Hakomori, S., and Kim, Y. S. Lewis'- and sialylated Lewis'-related antigen expression in human malignant and nonmalignant colonic tissues.Cancer Res., 46: 2627—2632,1986.

5. Itzkowitz, S. H., Yuan, M., Fukushi, Y., Lee, H., Shi, Z., Zurawski, ‘/.,Jr.. Hakomori,

drate ligands need to be presented on a specific carrier protein,

PSGL-1 (6). In fact, Handa et al. (27) demonstrated that transfectionof the PSGL-1 gene into a colon cancer cell line, HRT1 8, expressingsialyl LeS and sialyl LeX resulted in P-selectin-mediated adhesionalthough the cells did not adhere to P-selectin before the transfection(27). Thus, it is important in future study to examine whether cobrectal carcinoma cells actually express PSGL-l or PSGL-l-like

molecules.Expression of C2GnT Gene Is Associated with Invasiveness of

Colorectal Cancer. Here, the expression of sialyb Lea or sialyl Le5on carcinoma cells was not correlated with lymphatic or venousinvasion with statistical significance (see above). However, whenthese carbohydrates were presented in 0-glycans, a significantcorrelation could be obtained (Table 2). These results suggest thatthe expression of C2GnT transcripts per se may define tumorprogression. To test this hypothesis, correlations between the expression pattern of C2GnT transcripts and clinicopathobogical vanables were analyzed on 46 patients, irrespective of the expressionof sialyl Lea or sialyl Lex on carcinoma cells. Surprisingly, theexpression of C2GnT transcripts was correlated with not onlyvessel invasion but also depth of invasion (P < 0.05), as shown inTable 1. This result strongly suggests that 0-glycans having core2 branch could be closely related to invasiveness, including depthof tumor invasion and vessel invasion of coborectal cancer cells,because C2GnT is responsible for synthesis of the core 2 branch (2,15). These results also suggest the possibility that carbohydratemoieties other than sialyl Lea or sialyb Lex may also serve asligands for selectins or selectin-like carbohydrate-binding proteins.A statistically significant correlation between C2GnT expressionand various clinicopathological parameters seen in patients lackingsialyl Lea or sialyl Lex on carcinoma cells was not demonstrated(data not shown) because only a small number of patients lackedsialyl Lea (n 12) or sialyl Lex (n = 3) in this study.

In summary, the present study demonstrates that presence ofsialyl Lea on carcinoma cells is well correlated with lymph node

5205

on July 7, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

C2GnT GENE IN COLORECTAL CANCER

S., and Kim, Y. S. Immunohistochemical comparison of Le', monosialosyl Lea (CA19—9),and disialosyl La' antigens in human colorectal and pancreatic tissues. CancerRes.,48:3834—3842,1988.

6. Rosen, S. D., and Bertozzi, C. R. The selectins and their ligands. Curn Opin. CellBiol., 6: 663—673, 1994.

7. Nicolson, G. L. Cancer metastasis: tumor cell and host organ properties important inmetastasis in specific secondary sites. Biochim. Biophys. Acts, 948: 175-224, 1988.

8. Saitoh, 0., Wang, W-C., Lotan, R., and Fukuda, M. Differential glycosylation andcell surface expression of lysosomal membrane glycoproteins in sublines of a humancolon cancer exhibiting distinct metastatic potentials. J. Biol. Chem., 267: 5700—5711,1992.

9. Sawada, R., Tsuboi, S., and Fukuda, M. Differential E-selectin-dependent adhesionefficiency in sublines of a human colon cancer exhibiting distinct metastatic potentials. J. Biol. Chem., 269: 1425—1431,1994.

10. Izumi, Y, Taniuchi, Y., Tsuji, T., Smith, C. W., Nakamori, S., Fidler, I. J., andIrimura, T. Characterization of human colon carcinoma variant cells selected for sialylLa' carbohydrate antigen: liver colonization and adhesion to vascular endothelialcells. Exp. Cell Res., 216: 215—221,1995.

11. Nakamori, S.. Kameyama, M., Imaoka, S., Furukawa, H., Ishikawa, 0., Sasaki, Y.,Kabuto, T., Iwanaga, T., Matsushita, Y., and Irimura, T. Increased expression of sialylLewis' antigen correlates with poor survival in patients with colorectal carcinoma:clinicopathological and immunohistochemical study. Cancer Res., 53: 3632—3637,1993.

12. Shimono, R., Mon. M., Akazawa, K., Adachi, Y., and Sugimachi, K. Immunohistochemical expression of carbohydrate antigen 19-9 in colorectal carcinoma. Am. J.Gastroenterol., 89: 101—105,1994.

13. Fukuda, M. Cell surface carbohydrates in hematopoietic cell differentiation andmalignancy. In: M. Fukuda (ed). Cell Surface Carbohydrate and Cell Development,pp. 127—159.Boca Raton, FL: CRC Press, 1992.

14. Maemura. K., and Fukuda, M. Poly-N-acetyllactosaminyl O-glycans attached toleukosialin. The presence of sialyl Lax structures in O-glycans. J. Biol. Chem., 267:24379—24386,1992.

15. Bierhuizen, M. F. A., and Fukuda, M. Expression cloning of a cDNA encodingUDP-GIcNAc: Galj3l—3GalNAc-R (GlcNAc to Ga1NAc) f31—6GlcNActransferaseby gene transfer into CHO cells expressing polyoma large tumor antigen. Proc. NatI.Acad.Sci.USA,89:9326—9330,1992.

16. Chomczynsk, P., and Sacchi, N. Single-step method of RNA isolation by acidguanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem., 162: 156—159, 1987.

17. Bierhuizen, M. F. A., Maemura, K., Kudo, S., and Fukuda, M. Genomic organization

of core 2 and I branching (3-1,6-N-acetylglucosaminyltransferases. Implication forevolution of the @-l,6-N-acetylglucosaminyltransferase gene family. Glycobiology,5:417—425,1995.

18. Noonan, K. E., Beck, C., Holzmayer, T. A., Chin, J. E., Wunder, J. S., Andrulis, I. L.,Gazdar, A. F., Willman, C. L., Griffith, B., Von Hoff, D. D., and Roninson, I. B.

Quantitative analysis of MDR1 (multidrug resistance) gene expression in humantumors by polymerase chain reaction. Proc. Natl. Acad. Sci. USA, 87: 7160—7164,1990.

19. Nakayama, J., Ota, H., Katsuyama, T., Kanai, M., Taki, T., Hirabayashi, Y., andSuzuki, Y. Immunohistochemical demonstration of glycoconjugates bearing the type2 chain-backbone structure in human fetal, normal and neoplastic gastrointestinaltract. Histochemistry, 97: 303—310, 1992.

20. Abdullah, K. M., Udoh, E. A., Shewen, P. E., and Mellors, A. A neutral glycoproteaseof P. haemolytica Al specifically cleaves O-sialoglycoproteins. Infect. Immun., 60:56—62, 1992.

21. Colon and rectum. In: B. Spiessl, 0. H. Beahrs, P. Hermanek, R. V. P. Hutter, 0.Scheibe, L. H. Sobin, and G. Wagner (eds.). Illustrated guide to the TNM/pTNMclassification of malignant tumours, Ed. 3, pp. 82—89.Berlin: Springer-Verlag, 1989.

22. Yousefi, S., Higgins, E., Daoling, Z., Krtiger-Pollex, A., Hindsgaul, 0., and Dennis,J. W. Increased UDP-GlcNAc. Gal@l—3GalNAc-R(GlcNAc to Ga1NAc) (3-1,6-N-acetylgiucosaminyltransferase activity in metastatic murine tumor cell lines. Controlof polylactosamine synthesis. J. Biol. Chem., 266: 1772—1782,1991.

23. Kitagawa, H., and Paulson, J. C. Differential expression of five sialyltransferase genesin human tissues. J. Biol. Chem., 269: 17872—17878, 1994.

24. Yago, K., Zenita, K., Ginya, H., Sawada, M., Ohmori, K., Okuma. M., Kannagi, R.,and Lowe, J. B. Expression of a-(l,3)-fucosyltransferases which synthesize sialyl La'and sialyl Lee, the carbohydrate ligands for E- and P-selectins, in human malignantcell lines. Cancer Res., 53: 5559—5565, 1993.

25. Mannon, G., Crottet, P., Cecconi, 0., Hanasaki, K., Aruffo, A., Nelson, R. M., Varki,A., and Bevilacqua, M. P. Differential colon cancer cell adhesion to E-, P-, andL-selectin: role of mucin-type glycoproteins. Cancer Res., 55: 4425—4431, 1995.

26. Suzuki, Y., Ohtani, H., Mizoi, T., Takeha, S., Shiiba, K., Matsuno, S., and Nagura, H.Cell adhesion molecule expression by vascular endothelial cells as an immuno/inflammatory reaction in human colon carcinoma. Jpn. J. Cancer Res., 86: 585—593,1995.

27. Handa, K., White, T., Ito, K., Fang, H., Wang, S., and Hakomori, S. P-selectindependent adhesion of human cancer cells: requirement for co-expression of a‘PSGL-l-like'core protein and the glycosylation process for sialosyl-Le' or sialosylLea. Int. J. Oncol., 6: 773—781, 1995.

5206

on July 7, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1997;57:5201-5206. Cancer Res   Kazuhisa Shimodaira, Jun Nakayama, Naoshi Nakamura, et al.  

-Glycans in Tumor ProgressionOCancer: Role of -Acetylglucosaminyltransferase Gene in Human Colorectal

N-1,6-βCarcinoma-associated Expression of Core 2

  Updated version

  http://cancerres.aacrjournals.org/content/57/23/5201

Access the most recent version of this article at:

   

   

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

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

  Permissions

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

.http://cancerres.aacrjournals.org/content/57/23/5201To request permission to re-use all or part of this article, use this link

on July 7, 2020. © 1997 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from