(CANCER RESEARCH 50, 4801-4808. August I, 1990]

Analysis of Human Tumor Associated Thomsen-Friedenreich Antigen1

John Samuel, Antoine A. Noujaim, Grant D. MacLean, Mavanur R. Suresh, and B. Michael Longenecker2

Department of Immunology, University of Alberto, Edmonton, Alberta T6G 2H7 [J. S., B. M. L.]; Faculty of Pharmacy and Pharmaceutical Sciences, University ofAlberta, Edmonton, Alberta T6G 2N8 [A. A. N., M. R. S.J; and Department of Medicine, Cross Cancer Institute, Edmonton, Alberta T6G IZ2 [G. D. M.¡,Canada

ABSTRACT

The Thomsen-Friedenreich (TF) antigen is a precursor structure ofMN blood group antigens and is also expressed by about 90% of humancarcinomas. The immunodominant group of TF antigen |/3-galactosyl(l-3)-a-/V-acetylglactosamine] is present in cryptic form in normal RBC andis revealed by neuraminidase treatment. A murine monoclonal antibody(Mah 49H.8) developed against neuraminidase treated human RBC wasreactive against a variety of human tumors. We have characterized thehuman tumor associated TF antigen detected by this antibody from ahuman transitional bladder carcinoma cell line (647V), a human colonadenocarcinoma cell line (LS174T), and a pleural effusion fluid of abreast adenocarcinoma patient (PE 89). A heterologous sandwich radio-immunoassay for TF antigen was developed using Mah 49H.8 as thecatcher and >25I-peanutagglutinin as the probe. Detergent extracts of

647V and LS174T cells, media conditioned by culturing these cells, andPE 89 were shown to contain the antigen by this assay. The specificityof the antigen capture by Mab 49H.8 in this assay was demonstrated byits selective inhibition by nitrophenyl-/î-D-galactoside,phenyl-/3-D-gal-actoside, and a TF hapten. Preliminary studies on TF antigen in serumsamples using this assay showed that about 53.7% of the carcinomasamples contained an antigen concentration greater than 200 units/mlwhereas for 90.9% of the normal samples the antigen concentration wasbelow 200 units/ml. These studies demonstrated that the TF antigen isshed by the tumor cells both in vitro and in vivo. The TF antigen wassensitive to treatment with alkali (0.1 M NaOH for 5 h at 37°C)and

periodate (10 HIMsodium periodate for l h at room temperature), wasresistant to acidic pH (50 HIMacetate buffer, pH 4.5, for 5 h at 37°C),and could be extracted with perchloric acid (0.6 M for l h at 4"( '). The

antigen was shown to be a high molecular weight glycoprotein (M,>1,000,000) by gel filtration chromatography. The density of the antigenwas estimated to be about 1.35 g/ml by cesium chloride density gradientcentrifugaron. The antigen could be isolated from conditioned media bya combination of affinity chromatography and gel filtration with an overallpurification of about 61,432-fold and a final recovery of 53.2%. Imniii-noprecipitation of TF antigen using Mab 49H.8 from a detergent extractof 647V cells with cell surface glycoconjugates labeled with -'H showed

it to be a glycoprotein with an approximate molecular weight of 1,000,000.These studies suggest that the human TF antigen expressed by carcinomacells and secreted into the body fluids is a mucin-like high molecularweight glycoprotein.

INTRODUCTION

The immunodominant group of the TF3 antigen is ß-gal-(l-3)-a-yV-acetylgalactosamine. The TF hapten is expressed incryptic form in normal human RBC, where it is a precursorstructure for the M and N blood group antigens and in normalepithelial cells (1). The TF hapten is revealed by removal of theterminal sialic acid residues by neuraminidase treatment of the

Received 2/7/90; revised 4/6/90.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 inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1This work was supported by a grant from National Cancer Institute of

Canada.2To whom requests for reprints should be addressed.3The abbreviations used are: ABC, avidin-biotin peroxidase complex: HSA,

human serum albumin: Mab, monoclonal antibody; NP-40, Nonidet P-40; PAGE,polyacrylamide gel electrophoresis; PBS. phosphate buffered saline, containing140 mm NaCl, 2.68 mivi KC1, 8.1 min Na;HPO„.and 1.4 mivi KH¡PO4,pH 7.2;PE, pleural effusion; PNA, peanut agglutinin; RT, room temperature; SDS,sodium dodecyl sulfate; SRIA, sandwich radioimmunoassay: TF, Thomsen-Friedenreich.

cryptic TF antigen. Springer et al. have reported that about90% of the human carcinomas express TF hapten in noncrypticform (2). The TF antigen is capable of eliciting both humoraland cell mediated immunity in mice (3-6) and in cancer patients(7, 8) and in animal models can stimulate effective anticancerimmunity (6). In breast cancer patients increased expression ofTF antigen has been correlated with tumor progression andmetastatic spread (9). Studies on the expression of this antigenin bladder carcinomas suggest that it may serve as a prognosticmarker (10-13). Immunohistochemical studies on normal, pre-malignant, and malignant colonie tissues have shown that theexpression of TF antigen is associated with malignant andpremalignant changes in colonie mucosa (14-16). Further, TF

antigen has also been implicated in organotropic metastasis oftumor cells (17).

PNA (18) is the most commonly used anti-TF reagent. Themajor limitation of PNA as an anti-TF reagent is its cross-reactivity with other glycoconjugates having a terminal galactose residue. Development of Mabs against TF antigen offersan approach to the preparation of more specific reagents againstthis important antigen. We have developed anti-TF Mabs usingneuraminidase treated human RBC as the immunogen (3, 4).One of these Mabs (Mab 49H.8) was reactive against a varietyof human and murine tumors. Several recent studies havedemonstrated the usefulness of Mab 49H.8 as a tool for studying the tumor associated expression of TF antigen (16, 19-22).Two of these studies report a greater degree of specificity forMab 49H.8 as an anti-TF reagent as compared to PNA andpolyclonal antibodies (19,20).

The cryptic TF antigen in RBC is known to be glycophorinand has been well characterized (23, 24). A murine cancerassociated TF antigen, epiglycanin, has also been subject todetailed chemical and biochemical investigations (25, 26). However, the molecular characteristics of the TF antigen expressedby human tumor cells have not yet been described. In view ofthe important clinical significance of TF expression in humantumors, we decided to characterize the human tumor associatedTF antigen using Mab 49H.8. We now report the partialcharacterization of a high molecular weight glycoprotein reactive with Mab 49H.8 from a bladder carcinoma cell line (647V),a colon adenocarcinoma cell line (LS174T), and a pleuraleffusion (PE 89) from a breast cancer patient.

MATERIALS AND METHODS

Cell Line, Culture, Conditioned Medium, and Extraction. A humanbladder transitional cell carcinoma cell line, 647V (27), was kindlyprovided by Dr. Y. Fradet (Laval University Cancer Center, Hotel-Dieude Quebec, Quebec, Canada). A human colon adenocarcinoma cell line,LS174T (28), was obtained from the American Type Culture Collection, Rockville, MD. A murine mammary adenocarcinoma cell line(Ta3-Ha) was kindly provided by Dr. J. F. Codington, Laboratory forCarbohydrate Research, Massachusetts General Hospital, Boston, MA.The cells were cultured in RPMI 1640 (Gibco Canada, Inc., Burlington,Ontario, Canada) supplemented with fetal bovine serum (5%) andgentamicin (50 units/ml) and the medium was changed every 3 days.The spent (conditioned) medium was collected during each mediumchange, pooled together, centrifuged at 10,000 x g for 30 min to

4801

Research. on October 24, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

HUMAN TUMOR ASSOCIATED TF ANTIGEN

Table l Sandwich radioimmunoassay for TF antigen using '"I-PNA as a probe

cpm bound ±SDSource

ofantigenPBS647V

cell octyl glucosideextract(1:10dilution)Octyl

glucoside buffer (1:10dilution)LS174T cell NP-40 extract ( 1:10 di

lution)NP-40buffer (1:10dilution)647V

conditioned culturemedium(8x)LS174T

conditioned culturemedium(4x)Culture

medium(4x)PE89 (1:2dilution)TF(a)-HSA(10Mg)HSAdOng)Epiglycanin

(10 jxg)Mab

49H.8 ascatcher266

±1912.086±742205

±19,367±57266

±8337,066±219326,035

±613540

±8414,807±19793,997±199351±5122,875±80MOPC-104E

ascatcher256

±101,365±136276

±371,251±60298

±57322±58206

±38153

±5280±58465±32278±25693±197

remove the cell debris, and concentrated about 20-fold (20x) using aHigh-Output Stirred Cell 2000 (Amicon Corporation, Danvers, MA).When confluent, the cells were removed from the flasks by EDTA(0.02%) treatment followed by scraping with a rubber policeman. Thecell suspension (approximately 10s cells) was washed in PBS threetimes and the pellet was extracted with a detergent (NP-40 or octylglucoside) extraction buffer (1% NP-40 or 1% octyl glucoside, 2 mMEDTA, 1 mM phenylmethylsulfonyl fluoride, 2 mM pepstatin, 10 mMo-phenanthroline, and 150 mM NaCl in 50 mM Tris buffer, pH 7.5) for30 min at 4°Cand centrifuged at 30,000 x g; the supernatant was used

for further studies.Immunoperoxidase Staining. Immunoperoxidase staining of cells was

done according to the ABC method (29) using a Vectastain ABCstandard kit (Dimension Laboratories, Inc., Mississauga, Ontario, Canada). The cells were grown on glass coverslips, then air dried, andsequentially incubated with a primary antibody (Mab 49H.8) at RT for15 min, with 0.025% glutaraldehyde at 4C for 5 min, with biotinylatedanti-mouse IgM at RT for 30 min, and with ABC reagent at RT for 20min. After each of the above steps the slides were washed in PBS atRT for 5 min. Finally the slides were incubated with 3,3'-diaminoben-

zidine (Sigma Chemical Co., St. Louis, MO) substrate solution containing 0.5% hydrogen peroxide at RT for 5 min, washed in water,counterstained with hematoxylin, and mounted in Permount. A darkercolor corresponds to positive reaction with the antibody (Fig. 1).

Monoclonal Antibodies, PNA, and Radioiodination. Ascitic fluid containing Mab 49H.8 (isotype, IgM) was prepared as reported previously(3, 4). Mab 49H.8 was purified from the ascitic fluid by affinitychromatography using a Synsorb-T (synthetic TF hapten immobilizedon an inert matrix; supplied by Chembiomed Ltd., Edmonton. Alberta,Canada) column according to a procedure supplied by Chembiomed.Briefly, the ascitic fluid was diluted with equal volume of PBS andpassed through the column; the column was washed with PBS to removeunabsorbed proteins and then eluted with 0.1 Mglycine HCI containing0.15 MNaCl (pH 2.8). The acid eluted fractions containing the antibodywas neutralized to pH 7.5 and shown to contain a single protein bySDS-PAGE. PNA was obtained from E-Y Laboratories, San Mateo,CA. Two murine IgM Mabs, Mab CH4 (isotype, IgM) (30) and MOPC-104E (myeloma protein; Sigma), were used as negative control antibodies. PNA was radioiodinated by the lodo-Gen method (31). lodo-Genwas supplied by Pierce, Rockford, IL. and Na'25I was supplied by

Edmonton Radiopharmaceutical Centre, Edmonton, Alberta, Canada.Protein and Carbohydrate Assays. Protein assays were performed

using a Bio-Rad protein assay kit (Bio-Rad, Richmond, CA) with bovine

serum albumin as a protein standard. Total hexose concentrations ofTF antigen samples were estimated by the phenol sulfuric acid method(32), using galactose as the hexose standard.

SRIA. Breakable strips of microtiter wells (Immulon 2 Removawell;Dynatech Labs., Inc.. Chantilly, VA) were coated with Mab 49H.8 ora control antibody (MOPC-104E) (I ^g /well) by overnight incubationat 4°C.The wells were then serially incubated with (fetal bovine serum,5%) in PBS at 37°Cfor 30 min, with antigen solution at RT for 2 h,and with radioiodinated PNA (2 x 10' cpm/well) at RT for 2 h. The

wells were washed with PBS containing 0.05% Tween 20 after each ofthe above steps. Each well was counted for L25Iactivity using a gamma

counter.Epiglycanin used in SRIA was prepared as reported previously (6).

TF-a-HSA described in Table 1 was obtained from Biomira, Inc.,Edmonton, Alberta, Canada, and had a haptenxarrier molar ratio of10:1.

Inhibition of the antigen binding with the catcher (Mab 49H.8) inthe SRIA by synthetic haptens was done by mixing the antigen solutionwith the haptens before incubation with the antibody coated wells. TheTF hapten used in this study was 2-buten-l-yl-2-acetamido-2-deoxy-3-O-(0-galactopyranosyl)-«-D-galactopyranoside and was supplied byBiomira. Other sugar haptens were supplied by Sigma.

All the assays shown in Tables 1, 2, and 3 and Fig. 2A were done in

Fig. 1. Immunoperoxidase staining of647V cells using Mab 49H.8. A brown colorcorresponds to positive reactivity with theMab.

4802

Research. on October 24, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

HUMAN TUMOR ASSOCIATED TF ANTIGEN

32

28

21

oCD IJ

CLO

200 400 1,000 2,000 4,000

TF Antigen Concentration (ng of hexose/mL)

l2468

Concentration of hapten (inM)10

Fig. 2. A, SRIA for human TF antigen showing the concentration dependenceof the assay. The concentration of the TF antigen is expressed as ng of totalhexoses/ml. B, inhibition of antigen binding with the catcher (Mab 49H.8) by (•)TF hapten. (A) lactose, and (D) A'-acetylgalactosamine in SRIA. Bars, SD.

triplicate for each sample. The assays shown in Fig. IB and Fig. 3/1were done in duplicate.

Serum Samples and Pleural Effusion Fluid. Serum samples of 22normal volunteers were obtained from Red Cross Blood Bank, Edmonton, Alberta, Canada. Serum samples of 84 carcinoma patients wereobtained from Cross Cancer Institute. Edmonton. Alberta, Canada.The serum sample in each type of cancer was random chosen and thepatients were at various stages of disease and therapy. A pleural effusion

fluid (PE 89), obtained from Cross Cancer Institute was from a breastadenocarcinoma patient having extensive metastasis.

TF Antigen Assay of Serum Samples. For the TF antigen assay ofserum samples the sandwich assay was modified to allow 18 h incubation with the serum samples and 4 h incubation with I25I-PNA in order

to increase the sensitivity of the assay. For the quantitation of the TFantigen, a standard curve was generated by using TF-a-HSA (hap-

tenicarrier molar ratio of 26:1, supplied by Biomira) as a synthetic TFstandard using PBS as well as a normal serum sample (1:2 dilutionwith PBS) as assay matrices. One unit of TF antigen was defined as thequantity of antigen showing the same level of binding as 10 ng of TF-a-HSA in a SRIA. The serum sample were assayed at a dilution of 1:2(with PBS).

Chemical Treatment. Antigen solutions were incubated with 0.1 MNaOH, 50 HIM acetate buffer (pH 4.5), or PBS at 37"C for 5 h,neutralized to pH 7.5, and then dialyzed against PBS overnight at 4°C.

Periodate treatment of the antigen solutions was done based on areported procedure (33). Antigen solutions were incubated with 10 mMsodium periodate in 50 mM acetate buffer (pH 4.5) at RT in the darkfor 1 h, neutralized to pH 7.5, incubated with \% glycine in PBS for15 min at RT, and then dialyzed against PBS overnight at 4°C.

Antigen solutions were incubated with 0.6 M perchloric acid at 4°C

for 1 h followed by removal of the precipitate by centrifugation at10.000 x g or 15 min. The supernatants were neutralized to pH 7.5and dialyzed against PBS overnight at 4°C.The TF activity of all the

above samples were evaluated by SRIA.Gel Filtration Chromatography. Tumor cell detergent extracts, con

centrated conditioned media from cell culture (20x), or pleural effusionfluid were fractionated on a Superóse HR-I2 10/30 gel filtrationcolumn (1- x 30-cm column; Pharmacia, Inc., Donai, Quebec. Canada)by fast protein liquid chromatography. The elution was done underisocratic conditions at RT using 0.1 M phosphate buffer (pH 7.5)containing 0.15 M NaCl or 6 M guanidine hydrochloride in 0.1 Mphosphate buffer (pH 7.5) with 0.15 M NaCl at a flow rate of 0.3 ml/min. The eluant fractions (0.6 ml each) were assayed for TF antigen bySRIA either directly or after overnight dialysis (in case of elution by 6M guanidine hydrochloride buffer) against PBS.

Reduction and Alkylation. A high molecular weight fraction enrichedin TF antigen from 647V cell conditioned medium or PE 89 wassubjected to reduction by incubation with dithiothreitol (50 mM) in thepresence of guanidine hydrochloride (6 M) at 4°Cfor 18 h or at RT for

4 h and alkylated by incubation with iodoacetamide (150 mM) in thedark at RT for 1 h, and the excess iodoacetamide was destroyed byaddition of dithiothreitol (200 mM) (34). This solution was furtherfractionated by gel filtration chromatography using guanidine hydro-chloride buffer and the fractions were assayed by SRIA as describedabove.

Affinity Chromatography. An affinity column was prepared by conjugation of Mab 49H.8 with CNBr activated Sepharose 4B beads(Pharmacia) according to the procedure supplied by Pharmacia. Briefly,

Table 2 Inhibition of TF antigen binding with Mab 49H.8 by synthetic carbohydrates in SRIA

% of inhibition ±SDConcentration

Inhibitor(min)Phenyl-ii-D-galactoside510Phenyl-fi-D-glucoside

510p-Nitrophenyl-fi-D-galactoside

510p-Nitrophenyl-o-D-galactoside

5IO647V

conditionedmedium66.60±2.5890.12±5.24-3.00±7.990.58±4.7072.03±2.9794.95±1.0114.64±1.8218.08±1.84LS174T

conditionedmedium86.95±2.3895.91±0.30-7.15±9.17-1.67±1.8783.39±0.9595.84±0.135.75±0.8410.52±1.82PE8945.48±5.469.04±0.54-6.41±7.79-4.61±7.1953.25±1.2378.69±0.57-0.16±6.667.51±1.28Ta3-Ha

conditionedmedium69.40±2.1596.25±0.353.92±7.351.38±0.6458.00±0.1494.15±0.032.41±0.781.21±2.17

4803

Research. on October 24, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

HUMAN TUMOR ASSOCIATED TF ANTIGEN

20

M•oe

ng

5 IO 20Concentration of TF-HSA (fig/ml)

A40 L O >4000

16-0X!>«9'3PH

4-0000

°

08

oo8

°o8 1 1 I 8 •'1*8QOQQQn

CTO OH "nctg» e » oHlililí

A B C D E F GSerum Samples -n

ISFig. 3. A, Standard curve for SRIA using TFa-HSA as the TF standard using

(•)PBS and (A) normal human serum as assay matrices. B, SRIA of a panel ofserum samples: Column A, normal (22); Column B, breast adenocarcinoma (25);Column C, colon adenocarcinoma (18); Column D, lung adenocarcinoma (13);Column E, lung squamous carcinoma (8); Column F, prostate adenocarcinoma(12); Column G, ovarian adenocarcinoma (8).

CNBr activated Sepharose 4B beads (2 g) were swollen in l M HC1 andwashed in a sintered glass filter. The swollen gel (7 ml) was incubatedwith purified Mab 49H.8 (90 mg) in 0.1 M phosphate buffer containing0.5 M NaCl (pH 7.5, 22 ml) at RT, and protein concentration in thebuffer was monitored periodically. After 2 h of incubation about 80%of the antibody was bound to the column. The excess antibody wasremoved at this time, the gel was washed with the coupling buffer, and

the remaining active groups were blocked by incubation with 1Mglycinein coupling buffer (40 ml) at 4°Cfor 18 h, and the gel was then washed

with three cycles of alternating pH, each cycle consisting of 0.1 Macetate buffer (pH 4) containing 0.5 M NaCl followed by 0.1 M Trisbuffer (pH 8) containing 0.5 M NaCl. The gel was packed into a smallglass column (1x15 cm; bed volume, 7 ml). An antigen solution (20ml of 647V conditioned medium 20x or LS174T conditioned medium20x) was passed through the above affinity column (flow rate, 0.2 ml/min), washed with PBS (0.4 ml/min), and then eluted with 0.15 Mglycine HCl (pH 2.8) containing 0.15 M NaCl. The acid eluted fractionswere neutralized to pH 7.5 and an aliquot (25 n\) of each fraction wasanalyzed for TF antigen activity by SRIA.

SDS-PAGE. Electrophoretic analysis of antigen samples was performed by SDS-PAGE (3-15% resolving gel and 3% stacking gel)under reducing conditions (35). For the affinity purified TF sample,after the electrophoresis the gel was cut into 1-cm slices and homogenized with 6 M guanidine HCl in PBS for 30 min. The supernatant wasdialyzed against PBS at 4°Cfor 18 h and an aliquot (100 //') w»s

analyzed by SRIA. For the immunoprecipitated TF sample, the proteinbands in the gel were visualized by fluorography using EN'HANCE

(Du Pont, Inc., Lachine, Quebec, Canada) as the fluor.Cesium Chloride Density Gradient Centrifugation. The density of the

antigen solution (647V conditioned medium 20x, LS174T conditionedmedium 20x, or PE 89) was adjusted to 1.43 g/ml using cesiumchloride. The solution (5 ml) was centrifuged at 150,000 x g for 72 hin a Beckman ultracentrifuge using a SW 50.1 rotor at 10°C.Fractions

(0.45 ml each) were collected from the bottom of the tube; density wasdetermined gravimetrically, dialyzed against PBS overnight at 4°C,and

then assayed for TF antigen by SRIA at 1:4 dilution.Tritium Labeling of Cell Surface Glycoconjugates. Tritium labeling

of cell surface glycoconjugates was done using the galactose oxidase-borohydride method (36). A single cell suspension of 647V cells (5 xIO7)in PBS (0.5 ml) was gently shaken with galactose oxidase (5 units)at 37°Cfor 30 min. The cells were then washed with PBS twice,resuspended in PBS (0.5 ml), and mixed with sodium ['HJborohydride

(0.5 mCi, 20 Ci/mmol) in 0.01 N NaOH (20 n\) at room temperaturefor 30 min. The cells were then washed 3 times in PBS and the pelletwas extracted with NP-40 extraction buffer.

Immunoprecipitation. An aliquot of the above NP-40 extract wasprecleared by overnight incubation with anti-mouse IgM-agarose beads(Sigma) at 4°C.The supernatant was divided into 3 equal portions andincubated with Mab 49H.8, Mab CH4, or normal mouse serum at 4°C

for 4 h. Then each portion was incubated with anti-mouse IgM-agarosebeads at 4°Cwith gentle shaking overnight. The beads were collected

by centrifugation, washed 5 times with 0.05 M Tris HCl buffer (pH 7.5)containing 0.5 M NaCl, and extracted by boiling with Laemmli samplebuffer (35). The samples were analyzed by SDS-PAGE followed byfluorographic detection of the tritium labeled glycoprotein bands.

RESULTS

Detection of TF Antigen in Cell Extracts, Conditioned Media,and Pleural Effusion Fluid by SRIA. The reactivity of Mab49H.8 with 647V cells was first demonstrated by Fradet et al*

in flow cytometry studies. We have further confirmed theseresults by immunoperoxidase staining of these cells with Mab49H.8 (Fig. 1). Similarly the reactivity of LS174T cells withMab 49H.8 was also demonstrated by immunoperoxidase staining.

Non-ionic detergent (NP-40 or octyl glucoside) extracts of647V and LS174T cells, culture media conditioned by thesecells, and a pleura! effusion from a breast cancer patient (PE89) were tested for TF antigen using a heterologous SRIA).Epiglycanin, a murine TF antigen, and TF-«-HSA, a syntheticTF antigen, served as positive control samples, while HSA,fresh culture media, and detergent extraction buffers served asnegative control samples. An isotype matched (IgM) mouse

4 Y. Fradet, unpublished results.

4804

Research. on October 24, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

HUMAN TUMOR ASSOCIATED TF ANTIGEN

Table 3 Recover)' of TF activity after chemical treatment

TreatmentNo

treatmentPBS(37-C for 5h)0.1

M NaOH (37'C for 5h)50

HIMacetate buffer. pH 4.5 (37'C for 5h)10

min periodate in 50 m\i acetate buffer, pH 4.5 (RT for 1h)0.6M perchloric acid (4'C for 1h)E0

0.5-03CM75

0.4-01u=

0.3-2o»

0.2<0.1II/1/.i/;',lì

''--V•*T.-'*»*»**»*«^**«--25?0^~-20

*•ac-15

gm-

1 O2o-5I647Vconditionedmedium100.0097.09±0.740.72±0.6996.25±4.130.00100.57±3.45%

of recovery±LS174Tconditionedmedium100.0097.34±1.1213.90±0.6598.52±1.460.0085.58±1.25geM

l-O00C«•g08^£

O-6Baf

0.40

" O"¿\SDTa3-HaconditionedPE

89medium100.00

100.0095.5996.21±9.89

±2.320.002.33±0.0295.36

93.95±4.00±2.240.00

0.00104.90108.43±4.25

±1.091

\1111t1

\V-I1

111

11

1L

1[

11\¿«kiA^40?O3O*•Oon

oMiogU*

10 V015 20 25Fraction No.

4 eFraction No

CoLOcoMñ

0.8«uZ

°-6SJo,I

0.2-1_

...^,---,n ',iili,'1i

'l'I

11

¡1,'/y»**^^-^•••**••^^»-»05

'o4

X•oC3

oma.

U-1 f10 \fe5 20

Fraction No25

GelSlices Mr(kD)

200

92 5

69

46

3O

Fig. 4. Gel filtration chromatography of 647V conditioned medium (A) andPE 89 (B) using 0.1 M phosphate buffer containing 0.15 M NaCl, pH 7.5, as theeluant. V,,, void volume. The Chromatographie profile was not altered when thesamples were reduced and alkylated and eluted with 6 M guanidine hydrochloridein 0.1 M phosphate buffer containing 0.15 M NaCl, pH 7.5.

myeloma protein (MOPC-104E) was used to further confirmthe specificity of this assay. The results are summarized inTable 1. MOPC-104E did not show any selective reactivity withany of the samples used in this assay demonstrating the specificity of the positive reactivity with Mab 49H.8. While Mab49H.8 showed positive reactivity with epiglycanin and TF-a-HSA, no significant reactivity was observed for negative controlsamples demonstrating the selectivity of the assay for TF antigen. The detergent extracts of 647V and LS174T cells, conditioned culture media from these cells, and PE 89 showed strongpositive reactivity for TF antigen. Fig. 2A shows the concentration dependence of the assay for an affinity purified sample ofTF antigen over a 100-fold range of concentration. At lowconcentrations of the antigen (50-1000 ng of hexose/ml; 20-fold concentration range), a linear relationship between antigenconcentration and cpm bound could be observed. The inhibitionof the antigen binding by the catcher (Mab 49H.8) using synthetic carbohydrate haptens was done to further demonstratethe specificity of the assay (Table 2). The positive reactivity inthis assay could be inhibited by phenyl-ß-D-galactoside and p-

4 8 12 16 2O 24CPM Bound X io"2 B

Fig. 5. A, affinity chromatography of 647V conditioned medium showing TFactivity of each fraction (6 ml). Fractions 1-7 were eluted at a flow rate of 0.2ml/min and fractions 8-30 were eluted at a rate of 0.4 ml/min. Fractions 1-20were eluted at a pH of 7.5 and fractions 21-30 were eluted at a pH of 2.8. B,SDS-PAGE of an affinity purified TF antigen sample on a 3-15% gradientresolving gel with a 3% stacking gel under reducing conditions. After the electro-phoresis the gel was cut into 1-cm slices and proteins were extracted. An aliquotof the extract was assayed for TF antigen activity by SRIA. Arrow, interfacebetween stacking and resolving gels.

nitrophenyl-ß-D-galactoside in a hapten specific manner as expected from previously published results (4), while nonspecifichaptens such as phenyl-ß-o-glucoside and p-nitrophenyl-a-o-galactoside showed little or no inhibition. Similarly, a syntheticTF hapten inhibited the antigen capture by Mab 49H.8 in aconcentration dependent manner, whereas nonspecific sugarssuch as lactose and ¿V-acetylgalactosamine did not show anyspecific inhibitory effect on the assay (Fig. 2B).

TF Antigen Assay of Serum Samples. A panel of 22 normaland 84 carcinoma serum samples were tested for TF antigen bySRIA. For the quantitation of TF antigen, a standard curve wasgenerated using a synthetic TF antigen, TF-«-HSA, as thestandard using PBS and a normal serum sample as matrices(Fig. 3/1). The matrix effect of the serum sample was minimal(less than 10%). The TF level of each serum sample wasestimated in activity units (1 TF unit was defined as that

4805

Research. on October 24, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

HUMAN TUMOR ASSOCIATED TF ANTIGEN

Table 4 Purification of TF antigen from 6471' conditioned medium

Sample647V

conditioned medium (20x)Affinity purifiedGel filtration purifiedTotal

protein(mg)1.500

0.1680.013Total

activity(units)111,000

74,30059,100Units/fig

proteins0.074

4424,546Increase

inpurity(fold)15,973

61,432Recovery

ofantigen

activity(%)66.9

53.2

4 6Fraction

Fig. 6. Cesium chloride density gradient centrifugation profile of 647V conditioned medium showing the TF activity and density of each fraction.

1 2 3 Mr (kD)

- 200

- 116

- 9225

- 66.2

- 45

Fig. 7. Immunoprecipitation of tritium labeled 647V cell extract using (Lane1) normal mouse serum, (Lane 2) control Mah CH4. and (Lane 3) Mah 49H.8.Arrow, interface between stacking and resolving gels.

showing an activity equivalent to 10 ng of TF-a-HSA in SRIA).The results of the serum assay is summarized in Fig. 3Ä.Themajority of the normal serum samples (20 of 22, 90.9%) showedan antigen level below 200 units/ml. Among the carcinomasamples about 53.7% of the samples (breast adenocarcinoma,

48.0%; colon adenocarcinoma, 61.1%; lung adenocarcinoma,53.8%; lung squamous carcinoma, 50.0%; prostate adenocarcinoma, 58.3%; ovary adenocarcinoma, 50.0%) showed activitygreater than 200 units/ml. If this concentration is chosen asthe upper limit of normal values, the specificity and sensitivityof this assay are 90.9 and 53.7%, respectively.

Chemical Treatment. The effect of various chemical treatments on the TF activity of 647V conditioned medium, LS174Tconditioned medium, PE 89, and Ta3-Ha conditioned mediumare shown in Table 3. Incubation with 0.1 M NaOH at 37°Cfor

5 h almost completely abolished the reactivity of all the samples,whereas incubation with PBS or 50 m\i acetate buffer (pH 4.5)at 37°Cfor 5 h resulted in little or no loss of activity. A

periodate treatment of the antigen samples was carried outunder conditions reported to be suitable for the selective destruction of its carbohydrate moieties without significant alteration of the protein epitopes (33). Periodate treatment resultedin 100% loss of reactivity for all the samples. The recovery ofTF activity after perchloric acid extraction for 647V conditioned medium, PE 89, and Ta3-Ha conditioned medium were100%, and that for LS174T conditioned medium was 86%.

Gel Filtration Chromatography. The tumor cell (647V andLS174T) detergent extracts, conditioned media, and PE 89were fractionated on a gel filtration column (Superóse HR-12)using phosphate buffered (0.1 M) saline as the eluant and thefractions were assayed for TF antigen by SRIA. Typical chro-matograms of 647V conditioned medium and PE 89 are shownin Fig. 4. The chromatograms of 647V cell extract, LS174Tcell extract, and LS174T conditioned medium were similar tothose in Fig. 4A. In all the above chromatograms Mab 49H.8reactivity was limited to the void volume fractions as shown bythe SRIA, suggesting that the molecular weight of the TFantigen is greater than 1,000,000. The void volume fractionshaving TF activity from 647V conditioned medium or PE 89were pooled together and subjected to reduction of any disulfidebonds by treatment with dithiothreitol; the thiol groups werealkylated by treatment with iodoacetamide and subjected to gelfiltration chromatography using 6 M guanidine hydrochloridebuffer as the eluant. The eluants were assayed for TF activityby SRIA. The gel filtration chromatograms were similar tothose in Fig. 4 and the TF activity was associated exclusivelywith the void volume fractions.

Affinity Chromatography. The 647V conditioned medium andLS 174T conditioned medium were chromatographed on a Mab49H.8 affinity column. Fig. 5A shows the affinity chromatography profile for 647V conditioned medium. The fractions ofthe conditioned medium passed through the column as well asthe PBS washings contained little or no TF activity. SignificantTF activity was limited to the fractions eluted with acidic pH.The affinity chromatography profile for LS174T conditionedmedium was similar to that in Fig. 5A.

An affinity purified TF antigen sample was analyzed by SDS-PAGE, followed by SRIA of extracts of the gel slices (Fig. 5B).The peak TF antigen activity was located near the interface ofstacking and resolving gels, suggesting a high molecular weight(> 1,000,000) for the antigen.

4806

Research. on October 24, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

HUMAN TUMOR ASSOCIATED TF ANTIGEN

Purification of TF Antigen. A combination of affinity chro-matography and gel filtration was used to obtain a highlypurified TF sample from 647V conditioned medium (20x). Theincrease in purification as well as the recovery of each step wasevaluated (Table 4). A purification of 5973-fold was achievedby affinity chromatography as the first step. A subsequent stepof gel filtration resulted in an additional 10.29-fold purification,giving a final 61,432-fold enrichment of the antigen sample.The overall recovery was 53.2%.

Cesium Chloride Density Gradient Centrifugation. The densityof the antigen was estimated by analytical cesium chloridedensity gradient centrifugation of 647V conditioned medium,LS174T conditioned medium, and PE 89. The density gradientprofile of 647V conditioned medium (Fig. 6) showed that themajor peak with TF activity was associated with a density ofapproximately 1.35 g/ml. An additional minor peak with adensity of 1.2 g/ml was also noticed. The density gradientprofiles for LS174T conditioned medium and PE 89 weresimilar to those in Fig. 6 with a peak TF activity at a densityof 1.36 g/ml for the former and 1.35 g/ml for the latter.

Immunoprecipitation of TF Antigen. The carbohydrate residues on the cell surface of 647V cells were labeled with tritiumat position 6 of galactose by the galactose oxidase-borohydride

method for tritium labeling of glycoconjugates. A detergent(NP-40) extract of these radiolabeled cells was used for immu-noprecipitation of the TF antigen by Mab 49H.8, an irrelevantantibody (Mab CH4), and normal mouse serum. SDS-PAGEanalysis of the immunoprecipitated samples showed a highmolecular weight (approximately 1,000,000) band at the top ofthe resolving gel for Mab 49H.8 (Fig. 7, Lane 3). No corresponding protein band was observed for Mab CH4 and mouseserum (Fig. 7, Lanes 1 and 2), although some nonspecificactivity was noticed in the stacking gel of these lanes.

DISCUSSION

In view of the significance of TF antigen expression in bladdercarcinomas (10-13) and colon adenocarcinomas (14-16), wehave chosen a human bladder transitional cell carcinoma cellline (647V) and a human colon adenocarcinoma cell line(LS174T) as model cell lines for the analysis of human tumorassociated TF antigen. Our results show that these cell linesare good sources for the isolation of the human TF antigen andwould serve as convenient in vitro models for studying itsbiosynthesis and secretion. A pleural effusion from a breastcancer patient served as an additional clinical source for thesecreted form of the antigen.

We have developed a heterologous sandwich radioimmuno-assay using Mab 49H.8 as the catcher and '"I-PNA as the

probe. Since this assay is dependent on the epitope requirementsof two different anti-TF reagents, it is likely to reduce background signal due to nonspecific reactivity of either of thereagents. However, the assay is limited to antigens havingmultiple TF epitopes. The specificity of the assay was demonstrated by the inhibition of the antigen binding with the catcherby haptens specific for Mab 49H.8 including a synthetic TFhapten (Table 2; Fig. IB) We have shown that this assay candetect TF antigens from a variety of sources including syntheticTF antigen, TF antigen in cancer cell extracts, and murine andhuman TF antigens secreted by cancer cells in vitro and in vivo.The concentration dependence of the assay (Figs. 2A and J>A)suggests that this SRIA is suitable for quantitation of theantigen. To our knowledge this is the first report of a TFantigen assay suitable for quantitation of the secreted TF anti

gen in body fluids such as serum and pleural effusion fluid.The TF antigen may be purified by a combination of affinity

chromatography and gel filtration (Table 4). The gel filtrationchromatography of conditioned media and pleural effusion fluid(Fig. 4) and SDS-PAGE analysis of affinity purified (Fig. 5B)and immunoprecipitated (Fig. 7) TF antigen samples show thatit is a high molecular weight glycoprotein. Estimation of theexact molecular weight is difficult since the antigen eluted inthe void volume in the gel filtration chromatography and appeared as a band near the interface of stacking and resolvinggel in SDS-PAGE analyses. However these results suggest thatthe molecular weight is in the range of 1,000,000. The densityof the antigen was estimated to be about 1.35 g/ml. The densityof proteins which are not heavily glycosylated is usually lessthan 1.3, whereas that of proteoglycans is greater than 1.6 (37).The periodate oxidation experiment suggested that the epitopedetected by Mab 49H.8 is a carbohydrate. The loss of antigenicactivity on incubation with 0.1 M NaOH suggests that thecarbohydrate epitopes are O-linked to the protein core, typical

of mucin glycoproteins (38). The high molecular weight, presence of multiple O-linked carbohydrate epitopes, and the observed density suggest that the human TF antigen is a mucin-like glycoprotein. The extractability of the antigen by 0.6 Mperchloric acid is consistent with it being a mucin.

Several tumor associated antigens have been characterizedpreviously as mucins (38-42). Tumor associated mucins oftendiffer from their normal counterparts with respect to glycosy-lation. The expression of unmasked TF epitope has been described as an aberrant glycosylation pattern observed in about90% of the carcinomas. However, to our knowledge, none ofthe well characterized tumor associated mucins have been described to express this epitope. Byrd et al. (43) have recentlyreported the characterization of a cancer associated mucin fromLS174T cells. The TF epitope in this mucin appears to becryptic in that the native mucin was unreactive with the anti-TF antibody (Mab 49H.8), whereas the asialomucin was reactive (43). Therefore the molecule we are reporting appears tobe different from that reported by Byrd et al.

The successful radiolabeling of the TF antigen by the tritia-tion of the cell surface glycoproteins and its subsequent immu-noprecipitation suggest that the antigen is expressed on the cellsurface. The detection of the antigen in culture media conditioned by the cell lines as well as in the pleural effusion of abreast cancer patient suggests that the TF antigen is also shedby cancer cells. Our preliminary' studies on a panel of serum

samples showed increased levels of TF antigen in 53.7% ofcarcinoma patients as compared to the normal controls (Fig.3Ä). It should be pointed out, however, that the cancer serawere from a variety of cancer patients in various stages of theirdiseases, including primary cancer, various stages of metastaticdisease, and those in apparent remission. No attempt to correlate TF antigen level with disease status was done in thispreliminary study. We are currently evaluating the new SRIAas a possible novel monitoring assay in cancer patients. Wolfet al. (9) have used immunohistological studies to correlateincreased expression of TF antigen in breast cancer tissues withtumor progression and metastatic spread. The SRIA describednow makes it possible to evaluate any such potential correlationbetween serum TF levels and cancer progression. The secretionof various cancer associated mucins into serum have beenreported previously (38, 44, 45). The biological significance ofthe secretion of the cancer associated mucins or their effect onthe immune system is not yet clearly understood. We haverecently demonstrated the immunosuppressive effect of a mu-

4807

Research. on October 24, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

HUMAN TUMOR ASSOCIATED TF ANTIGEN

riñeTF antigen, epiglycanin, on a delayed type hypersensitivityresponse in a syngeneic adenocarcinoma model.5 We postulate

that the human TF mucin may also have a similar effect.We are currently purifying larger quantities of the TF antigen

from conditioned media for detailed structural analysis of thecore protein and the carbohydrate groups. The biological effectsof this antigen including immunosuppression will also be investigated.

ACKNOWLEDGMENTS

We are grateful to T. Tan, V. Pazderka, and G. Stanczyk-Brzezinskafor their technical assistance in these investigations.

REFERENCES

1. Springer, G. F., and Desai. P. R. Human blood-group MN and precursorspecificities: structural and biological aspects. Carbohydr. Res.. 40: 183-192.1975.

2. Springer, G. F. T and Tn, general carcinoma autoantigens. Science (Wash.DC), 224: 1198-1206, 1984.

3. Rahman. A. F. R., and Longenecker, B. M. A monoclonal antibody specificfor the Thomsen-Friedenreich cryptic T antigen. J. Immunol., 129: 2021-2024, 1982.

4. Longenecker. B. M.. Rahman. A. F. R.. Leigh. J. B.. Purser, R. A., Greenberg.A. H.. Willans. D. J.. Keller, O., Petrik. P. K.. Thay. T. Y., Suresh, M. R.,and Noujaim. A. A. Monoclonal antibody against a cryptic carbohydrateantigen of murine and human lymphocytes. I. Antigen expression in non-cryptic or unsubstituted form on certain murine lymphomas, on a spontaneous murine mammary carcinoma, and on several human adenocarcinomas.Int. J. Cancer, 33: 123-129, 1984.

5. Longenecker, B. M., Willans, D. J., MacLean, G. D., Selvaraj, S., Suresh,M. R., and Noujaim. A. A. Monoclonal antibodies to synthetic tumor-associated glycoconjugates in the study of expression of Thomsen-Frieden-reich-like and Tn-like antigens on human cancers. J. Nail. Cancer lust.. 78:489-496, 1987.

6. Henningson, C. M.. Selvaraj. S.. MacLean. G. D., Suresh, M. R., Noujaim.A. A., and Longenecker, B. M. T cell recognition of a tumor-associatedglycoprotein and its synthetic carbohydrate epitopes: stimulation of antican-cer T cell immunity in vivo. Cancer Immunol. Immunother.. 25: 231-241.1987.

7. Springer. G. F.. Murthy, S. M., Desai, P. R.. and Scanion, E. F. Breastcancer patient's cell-mediated immune response to Thomsen-Friedenreich(T) antigen. Cancer (Phila.). 45: 2949-2954. 1980.

8. Springer, G. F., Murthy, S. M., Desai. P. R.. Fry. W. A.. Tegtmeyer, H., andScanion, E. F. Patients' immune response to breast and lung carcinomaassociated Thomsen-Friedenreich (T) specificity. Klin. Wochenschr., 60:121-131, 1982.

9. Wolf, M. F.. Ludwig. A.. Fritz. P., and Schumacher. K. Increased expressionof Thomsen-Friedenreich antigens during tumor progression in breast cancerpatients. Tumor Biol., 9: 190-194, 1988.

10. Ohoka. H.. Shinomiya, H., Yokoyama, M., Ochi. K., Takeuchi, M., and1 isunii. S. Thomsen-Friedenreich antigen in bladder tumors as detected byspecific antibody: a possible marker of recurrence. Urol. Res., 13: 47-50,1985.

11. Summers, J. L., Coon, J. S., Ward, R. M.. Falor, W. H., Miller. A. W.. Ill,and Weinstein. R. S. Prognosis in carcinoma of the urinary bladder basedupon tissue blood group ABH and Thomsen-Friedenreich antigen status andkaryotype of the initial tumor. Cancer Res., 48: 934-939, 1983.

12. Coon, J. S., Weinstein, R. S., and Summers. J. L. Blood group precursor T-antigen expression in human urinary bladder carcinoma. Am. J. Clin. Pathol.,77:692-699, 1982.

13. Limas, C., and Lange, P. T-antigen in normal and neoplastic urothelium.Cancer (Phila.), 5«:1236-1245, 1986.

14. Boland. C. R., Montgomery. C. K.. and Kim, Y. S. Alterations in humancolonie mucin occurring with cellular differentiation and malignant transformation. Proc. Nati. Acad. Sci. USA, 79: 2051-2055. 1982.

15. Boland, C. R., Montgomery. C. K., and Kim, Y. S. A cancer-associatedmucin alteration in benign colonie polyps. Gastroenterology, 82: 664-672,1982.

16. Itzkowitz, S. H., Yuan, M., Montgomery, C. K., Kjeldsen, T., Takahashi, H.K., Bigbee. W. L., and Kim. Y. S. Expression of Tn, sialosyl-Tn. and Tantigens in human colon cancer. Cancer Res., 49: 197-204. 1989.

17. Springer, G. F., Cheingsong-Popov. R., Schirrmacher. V., Desai, P. R., andTegtmeyer. H. Proposed molecular basis of murine tumor cell-hepatocyteinteraction. J. Biol. Chem., 258: 5702-5706. 1983.

18. Lotan, R., Skutelsky, E.. Danon, D., and Sharon, N. The purification,composition, and specificity of the anti-T lectin from peanut (Arachis hypo-gaea). J. Biol. Chem., 250: 8518-8523, 1975.

19. Yuan, M.. Itzkowitz, S. H.. Boland. C. R., Kim. Y. D.. Tornita, J. T.. Palekar.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45

* P. Y. S. Fung and B. M. Longenecker, manuscript submitted for publication.

4808

A.. Bennington, J. L., Trump. B. F.. and Kim. Y. S. Comparison of T-antigen expression in normal, premalignant, and malignant human colonietissue using lectin and antibody immunohistochemistry. Cancer Res., 46:4841-4847, 1986.Wolf. M. F.. Koerner, U., and Schumacher. K. Specificity of reagents directedto the Thomsen-Friedenreich antigen and their capacity to bind to the surfaceof human carcinoma cell lines. Cancer Res.. 46: 1779-1782, 1986.Ravn. V.. and Jensen. H. Thomsen-Friedenreich-related antigen in humancndometrium. An immunohistochemical study employing the monoclonalantibody 49H.8. A preliminary report. APMIS, 96: 552-558, 1988.Juhl, B. R., Hartzen. S. H., and Hainau. B. Thomsen-Friedenreich-relatedantigen in non-neoplastic ureter urothelium and transitional cell tumours ofthe urinary bladder. An immunohistochemical study employing the monoclonal antibody 49H.8. Acta Pathol. Microbiol. Immunol. Scand. A, 95: 83-91.1987.Marchesi. V. T.. Furthmayr, H., and Tornita. M. The red cell membrane.Annu. Rev. Biochem.. 45: 667-698. 1976.Prohaska. R., Koerner, T. A. W., Armitage. I. M., and Furthmayr. H.Chemical and carbon-13 nuclear magnetic resonance studies of blood groupM and N active sialoglycopeptides from human glycophorin A. J. Biol.Chem.. 256:5781-5791. 1981.Codington, J. F., Linsley. K. B.. Jeanloz, R. W., Irimura, T., and Osawa, T.Immunochemical and chemical investigations of the structure of glycoproteinfragments obtained from epiglycanin. a glycoprotein at the surface of theTA3-Ha cancer cell. Carbohydr. Res.. 40: 171-182, 1975.Van den Eijnden, D. H.. Evans. N. A.. Codington. J. F., Reinhold, V., Silber,C., and Jeanloz. R. W. Chemical structure of epiglycanin. the major glycoprotein of the TA3-Ha ascites cell. J. Biol. Chem., 254: 12153-12159, 1979.Coon, J. S., Watkins, J. R., Pauli, B. U., and Weinstein, R. S. Flow cytometricanalysis of heterogeneity in blood group-related antigen expression in ahuman urinary bladder carcinoma cell line. 647V. Cancer Res.. 45: 3014-3021. 1985.Tom. B. H., Rutzky. L. P.. Jakstys. M. M., Oyasu, R.. Kaye, C. I., andKahan. B. D. Human colonie adenocarcinoma cells. I. Establishment anddescription of a new line. In Vitro (Rockville), 12: 180-191. 1976.Hsu. S-M.. Raine. L., and Fanger. H. Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques. A comparison between ABCand unlabeled antibody (PAP) procedures. J. Histochem. Cytochem., 29:577-580. 1981.Mosmann. T. R.. Gallatin. M., and Longenecker, B. M. Alteration ofapparent specificity of monoclonal (hybridoma) antibodies recognizing polymorphic histocompatibility and blood group determinants. J. Immunol.,/25: 1152-1156, 1980.Fraker. P. J., and Speck. J. C.. Jr. Protein and cell membrane iodinationswith a sparingly soluble chloroamide, l,3,4,6-tetrachloro-3a,6a-diphenylgly-coluril. Biochem. Biophys. Res. Commun., «0:849-857, 1978.Dubois. M.. Gilles. K. A., Hamilton, J. K.. Rebers, P. A., and Smith. F.Colorimetrie method for determination of sugars and related substances.Anal. Biochem., 2«:350-356, 1956.Woodward. M. P., Young. W. W., Jr.. and Bloodgood. R. A. Detection ofmonoclonal antibodies specific for carbohydrate epitopes using periodateoxidation. J. Immunol. Methods, 78: 143-153. 1985.Chase, K. V., Flux. M., and Sachdev, G. P. Comparison of physicochemicalproperties of purified mucus glycoproteins isolated from respiratory secretions of cystic fibrosis and asthmatic patients. Biochemistry, 24: 7334-7341.1985.Laemmli, U. K. Cleavage of structural proteins during the assembly of thehead of bacteriophage T4. Nature (Lond.). 227: 680-685, 1970.Gahmberg. C. G. Tritium labeling of cell-surface glycoproteins and glyco-lipids using galactose oxidase. Methods Enzymol.. 50: 204-206.1978.Hascall, V. C., and Kimura, J. H. Proteoglycans: isolation and characterization. Methods Enzymol., «2:769-800. 1982.Magnani. J. L., Steplewski, /., Koprowski, H., and Ginsburg, V. Identification of the gastrointestinal and pancreatic cancer-associated antigen detectedby monoclonal antibody 19-9 in the sera of patients as a mucin. Cancer Res.,«.•5489-5492,1983.Burchell. J., Durbin. H., and Taylor-Papadimitriou, J. Complexity of anti-genie determinants recognized by monoclonal antibodies HMFG-1 andHMFG-2 in normal and malignant human mammary epithelial cells. J.Immunol.. 131: 508-513, 1983.Lan, M.S., Finn, O. J., Fernstein. P. D.. and Metzgar. R. S. Isolation andproperties of a human pancreatic adenocarcinoma-associatcd antigen DU-PAN-2. Cancer Res.. 45: 305-310, 1985.Motti, S., Aguanno. S.. Canevari, S., Diotti. A., Orlandi. R.. Sonnino, S..and Colnaghi, M. I. Biochemical analysis of human ovarian cancer-associatedantigens defined by murine monoclonal antibodies. Cancer Res., 45: 826-832, 1985.Johnson, V. G.. Schlom, J., Paterson, A. J., Bcnnet, J., Magnani, J. L., andColcher. D. Analysis of a tumor-associated glycoprotein (TAG-72) identifiedby monoclonal antibody B72.3. Cancer Res.. 46: 850-857, 1986.Byrd. J. C.. Nardelli, J., Siddiqui. B.. and Kim. Y. S. Isolation and characterization of colon cancer mucin from xenografts of LS174T cells. CancerRes.. 48: 6678-6685. 1988.Linsley, P. S.. Brown, J. P., Magnani. J. L., and Horn. D. Monoclonalantibodies reactive with mucin glycoproteins found in sera from breast cancerpatients. Cancer Res., 48: 2138-2148, 1988.Paterson, A. J., Schlom. J.. Sears, H. F., Bennet. J., and Colcher, D. Aradioimmunoassay for the detection of of a human tumor associated glycoprotein (TAG-72) using monoclonal antibody B72.3. Int. J. Cancer, 37: 659-666. 1986.

Research. on October 24, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

1990;50:4801-4808. Cancer Res   John Samuel, Antoine A. Noujaim, Grant D. MacLean, et al.   AntigenAnalysis of Human Tumor Associated Thomsen-Friedenreich

  Updated version

  http://cancerres.aacrjournals.org/content/50/15/4801

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/50/15/4801To request permission to re-use all or part of this article, use this link

Research. on October 24, 2020. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from