(CANCER RESEARCH 51. 372-380. January 1. 1991)

Characterization of New Pancreatic Cancer-Reactive Monoclonal AntibodiesDirected against Purified Mucin1

Jenny J. L. Ho, Ning Bi, Pei-sha Van, Mei Yuan, Kristine A. Norton, and Young S. Kim2

(iastroinlestinal Research Laboratory, Veterans Administration Medical Center and Department of Medicine, University of California, San Francisco, California 94121

ABSTRACT

Although mucins have been found to be useful in the diagnosis ofpancreatic cancer, the carbohydrate and peptide structures of pancreaticmucins are still not well characterized. Monoclonal antibodies wereproduced using mucins purified from xenografts of a human pancreaticcancer cell line as the immunogen. One of these, Ia3, reacted with almostall pancreatic, gastric, and colorectal carcinomas examined by immuno-peroxidase staining, but with few normal tissues. Ia3 antigen was elevatedin sera of 50.4% of individuals with gastrointestinal tumors, but its levelsdid not correlate with those of CA15-3, CA19-9, or DU-PAN-2. SerumIa3 antigens migrated more slowly in sodium dodecyl sulfate-polyacryl-amide gel electrophoresis than the polymorphic epithelial mucins recognized by DF3 or 1I5D8. Ia3 reacted only with native, and not withpartially deglycosylated, pancreatic cancer xenograft mucins. Periodateor neuraminidase treatment destroyed this reactivity, but protease hadlittle effect. The antigen recognized by another antibody, Nd2, was notdetected in normal pancreatic, colonie, or gastric tissues but was presentin approximately 60% of the pancreatic and gastric carcinomas examined.Nd2 reactivity with native and partially deglycosylated mucin was lostafter pretreatment with protease and ff-mercaptoethanol. We concludethat, while Ia3 reacts with carbohydrates, Nd2 reactivity appears to bedependent on the integrity of the mucin protein core. The antigenicdeterminants of Ia3 and Nd2 are different from those of B72.3, CA19-9,DU-PAN-2, SPan-1, and several breast cancer mucin-directed antibodies.These results suggest that the malignancy-associated structures identifiedby Ia3 and Nd2 may provide new information on the carbohydrate andpeptide structure of pancreatic cancer mucins.

INTRODUCTION

The usefulness of MoAbs' in the serological detection of

malignancy has been demonstrated for a number of tissues,including breast (1), ovary (2), colon (3), and pancreas (4, 5).Often the antigens identified are heavily glycosylated, highmolecular weight glycoproteins known as mucins (6). For example, one mucin-reactive MoAb produced in our laboratory,SPan-1 (7), had a 93% positive rate for pancreatic canceragainst a serum panel (4). Many of these antibodies react withthe carbohydrates of mucins (3, 5, 7). However, MoAbs whichrecognize the protein core of a polymorphic epithelial mucinhave been described recently by several investigators (8-11). Anumber of these identify a similar sequence of amino acids (11),which is revealed in malignant tissue (12). Availability of thesetypes of MoAbs and cDNAs encoding apomucins (8-10, 13)has facilitated the study of the regulation of seemingly tissue-

Received2/27/90;accepted10/15/90.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 study was supported in part by USPHS Grant CA 24321 from the

National Cancer Institute and by the Veterans Administration Medical ResearchService. Y. S. K. is a Medical Investigator of the Veterans Administration.

2To whom requests for reprints should be addressed, at Gl Research Laboratory (151M 2). VA Medical Center, 4150 Clement St., San Francisco. CA 94121.

3The abbreviations used are: MoAb, monoclonal antibody; ABTS. 2,2'-azino-bis(3-ethyl-benzthiazoline-6-sulfonic acid): BSA. bovine serum albumin: EIA.enzyme immunoassay; HRP. horseradish peroxidase; SDS. sodium dodecyl sulfate: PAGE, polyacrylamide gel electrophoresis; PBS. phosphate-buffered saline;cDNA. complementary DNA; ABC. avidin-biotin complex; GI, gastrointestinal:MRP. mucin repeat peptide.

specific gene expression and structures in normal and canceroustissues (9, 12, 14, 15). New MoAbs which identify additionalcancer-associated epitopes in both the carbohydrate and proteinmoieties could enhance our growing understanding of the roleof these clinically significant glycoproteins in the pathogenesisof pancreatic cancer. In the present study we describe two newMoAbs, Ia3 and Nd2, produced against mucins purified fromxenografts of the human pancreatic cancer cell line SW1990(16), which react with structures distinct from those identifiedby a large number of previously characterized mucin-associatedantibodies. Our results suggest that they may be useful for boththe diagnosis of GI malignancies and the characterization ofthe peptide and carbohydrate structures of pancreatic cancermucins.

MATERIALS AND METHODS

Purification and Deglycosylation of SW1990 Mucins. Mucins werepurified from SW1990 xenografts and deglycosylated as previouslydescribed (16, 17).

Immunization and Production of Hybridomas. The initial immunization was with 100 ng protein of purified mucins in Freund's complete

adjuvant, followed by injections of mucins in incomplete adjuvant 3more times at 2-week intervals. Production of hybridomas and asciticfluids was as previously reported (18). Myelomas used were P3 X 63-Ag8.653 for Ia3 and SP2/0-Agl4 for Nd2 (American Type CultureCollection).

Screening of Hybridomas. Purified mucin (10 /ig hexose/ml PBS)was adsorbed to polystyrene microtiter plates (Costar, Cambridge, MA)overnight at 4°C,and ¡mmunoreactivitywas detected with a /i-galacto-

sidase EIA (4). RBC ghosts (Sigma Chemical Co., St. Louis, MO) wereused to test for blood group immunoreactivity. Isotyping was performedwith a kit from Boehringer-Mannheim (Indianapolis, IN).

Other Antibodies. Sources of antibodies were as follows: hybridomasB72.3 and NS 19-9, American Type Culture Collection; hybridomaSPan-1, developed in our laboratory (4); DU-PAN-2 ascitic fluid, Dr.Richard Metzgar (Duke University Medical Center, Durham, NC);HMFG-1, HMFG-2, and SM-3 culture media. Dr. Joyce Taylor-Papadimitriou (Imperial Cancer Research Fund Laboratories, London,UK); 115D8 and I39H2 ascitic fluids. Dr. John Milkens (The Netherlands Cancer Institute. Amsterdam, Netherlands); DF3 ascitic fluid.Dr. D. W. Kufe (Harvard Medical School, Boston, MA); and polyclonalanti-MRP, developed in our laboratory (13).

Sources of Human Tissues. Human sera were from the Mayo-National Cancer Institute Serum Diagnostic Bank (Dr. V. L. W. Go),National Cancer Institute Diagnosis Serum Bank (Dr. R. L. Aamodt),and the Osaka City University Hospital (Osaka, Japan). Blood fromhealthy donors (which had been determined to be human immunodeficiency virus and hepatitis B negative) was purchased from the IrwinMemorial Blood Bank (San Francisco). Samples were stored at —40°C

until used. Tissues were obtained from surgical specimens and autopsiesfrom the United States, Japan, and China, according to the establishedguidelines of the nations and institutions involved. All tissues werefixed in formalin and embedded in paraffin.

Immunohistochemistry. Immunoperoxidase staining was by the ABCmethod (18). Briefly, endogenous peroxidase activity was blocked with1% H2O2and then the tissue was incubated sequentially with hybridomamedium diluted in 5% normal rabbit serum, biotinylated-rabbit anti-mouse IgG+IgA+IgM (Zymed Laboratories, South San Francisco, CA)

372

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

MONOCLONAL ANTIBODIES REACTIVE WITH PANCREATIC CANCER MUC1N

diluted 1:100 in 1% normal rabbit serum, ABC reagent (Vector Laboratories, Burlingame, CA), and 3.3'-diaminobenzidine (Sigma) in PBS

containing 0.1% H2O2 and 0.03% CoCl2. The counterstain was 1%methyl green. Pure nonspecific myeloma IgG and IgM were used asnegative controls. A specimen was considered positive if at least 5% ofthe optical field (xlO objective) was stained.

Purification of Monoclonal Antibodies. Purification of B72.3, NS19-9, and Nd2 (IgGIs) was by affinity chromatography on Protein Acolumns (MAPS II kit; Bio-Rad, Richmond, CA) and that of DU-PAN-2, Ia3, and SPan-1 (IgMs) was by molecular sieving on SephacrylS-300 (4).

Labeling with HRP. Ia3 and Nd2 were labeled as previously described(19).

Immunoassays. CA15-3 (Centocor, Malvern, PA), CA19-9 (Fujire-bio, Inc., Tokyo, Japan), and DU-PAN-2 (Kyowa Medex Co.. Ltd.,Tokyo, Japan) immunoassays were performed according to manufacturers' instructions, and Ia3 and Nd2 HRP assays were performed as

previously described (19). Sandwich immunoassays were used for thescreening of sera (1 Ia3 unit was equivalent to 1 ng SW1990 mucinhexose standard). Optimal concentrations of capture antibody weredetermined by checkerboard titration. All washes were with PBS/0.05%Tween-20. Alternatively, submaximal concentrations of antigen wereadsorbed to plates without capture antibody. In inhibition studies, thesugar (7V-acetyl-glucosamine, jV-acetyl-galactosamine, fucose, or galactose) was incubated overnight with submaximal concentrations of labeled Ia3 or Nd2 at 4°C.

Nitrocellulose Membrane Nd2 Immunoassays. Five //I of a 1:10 dilution of serum (in PBS) were incubated overnight with nitrocellulosemembrane (Bio-Rad). The blocking solution was 1% BSA in PBS, andHRP-labeled Nd2 was diluted in 1% BSA. Washes were with PBS/0.05% Nonidet P-40. Each 1-cm x 1-cm square containing a serumsample was incubated with 0.5 ml of ABTS (2 m.M ABTS in 0.1 Mcitrate, pH 4.0, 0.03% H2O2). Alternatively, the nitrocellulose wasimmunostained with Nd2, as described below for Western blots usingradiolabeled Protein A. After the final wash, individual squares werecounted in a Beckman Gamma 7000 counter.

Treatments of Mucins. Mucins were reduced and alkylated (20) orheated prior to adsorption to microtiter plates. Otherwise, submaximalconcentrations of untreated SW1990 mucin were adsorbed first andthen incubated with 10% formalin (1 h), bovine pancreatic trypsin (5h; Sigma), Clostridium perfringens neuraminidase (2 h. type X; Sigma),or periodate (21).

PAGE and Western Blotting. Electrophoretic and immunostainingprocedures were as reported earlier (7). Immunoreactivity was detectedwith radioiodinated Protein A and autoradiography.

Glycolipid Isolation. The membrane fraction of SW1990 xenografts(16) was used as a source of glycolipids (22, 23).

Glycolipid Detection. Glycolipids were detected by EIA on microtiterplates (24) or by immunostaining (25) of high performance thin layerchromatography plates (J.T. Baker Chemical Co., Phillipsburg, NJ)after separation with chloroform/methanol/0.02% CaCl2 in H2O (50/40/10). ABC reagents (Vector) and biotinylated second antibodies(Zymed) were used. NS19-9 was the positive control.

RESULTS

Reactivities of Ia3 and Nd2 with Tissues. Ia3 (IgM) and Nd2(IgGl) were selected because they did not react with RBC andbecause they could discriminate between normal and malignantGI tissues (Table 1).

Pancreas. Ia3 reactivity in normal pancreatic tissue (23%,Table 1) was principally with the apical membranes of ductsand ductules but not with islets or acini. Ia3 antigenic activitywas also expressed in apical membranes and occasionally inapical cytoplasm of pancreatitis tissue (44%). Ia3 had about thesame percentage of reactivity with poorly differentiated (80%,Fig. la) as with well differentiated tumors (87.5%, Fig. \b). Inpoorly differentiated tumors, Ia3 reacted diffusely with thecytoplasm. In well differentiated cancers (Fig. \b), there was

Table 1 lai and Nd2 reactivity with normal and malignant gastrointestinaltissues

No. positive/total (%)

lai Nd2

PancreasCancerNormalPancreatitis

StomachCancerNormal

ColonCancerNormal

LiverCancerNormal

38/42 (90.5)6/26(23.1)4/9 (44.4)

25/31 (80.6)4/16(25.0)

20/22 (90.9)3/23(13.0)

0/2 (0)0/10(0)

26/41 (63.4)0/36 (0)2/11 (18.2)

14/24(58.3)0/10(0)

9/37 (24.3)0/23 (0)

0/3 (0)0/13(0)

little cytoplasmic staining and staining was mainly of the apicalmembrane, with some reactivity with luminal contents.

No Nd2 reactivity was observed with normal pancreas (Table1. Fig. 2a). However, Nd2 reactivity with 18% of the pancreatitis samples was characterized by weak staining of the apicalmembranes of ducts and ductules and, occasionally, of apicalcytoplasm. A much higher percentage of the pancreatic cancerspecimens (63%) had Nd2 reactivity, with a clear differencebetween well (85.7%, Fig. 2b) and poorly differentiated cancers(20%). In moderate to well differentiated pancreatic cancertissue, almost all of the cancer tissue, but not the surroundingconnective tissue, was stained with Nd2 (Fig. 2b). The patternconsisted of strong staining of luminal contents and strong anddiffuse apical cytoplasmic staining.

Stomach. In stomach, Ia3 had approximately the same levelof reactivity with cancer (81%) and normal tissue (25%) as inpancreas. Fig. \c shows an example of intestinal metaplasia ofthe stomach, a premalignant condition adjacent to cancer tissue.The vacuoles of goblet cells were stained by Ia3. There wasstrong Ia3 staining of luminal contents and apical membranesof cancer cells in well differentiated stomach cancer (Fig. !</).Generally, staining of cytoplasm was weak, although some cellsexhibited strong staining of the apical cytoplasm.

The Nd2 reactivity with stomach cancer was approximatelythe same (58%) as in pancreas (Table 1). No reactivity withnormal stomach was observed. Fig. 2c shows an example ofmoderately differentiated gastric cancer in the form of a gland-like structure. There was a strong Nd2 reactivity with both theluminal contents and the entire plasma membrane of cancercells but no cytoplasmic staining.

Colon. The positive rate of Ia3 reactivity with normal andmalignant colonie tissue was similar to that found in pancreasand stomach (91% for cancer and 13% for normal tissue). Innormal colon, Ia3 reacted occasionally with the luminal contents but not with the cytoplasm, vacuoles of goblet cells, orcolumnar cells (Fig. \e). Fig. If shows an example of transitional mucosa adjacent to cancerous tissue. As with intestinalmetaplasia, Ia3 reacted with vacuoles of goblet cells and luminalcontents. Columnar cells immediately adjacent to positive goblet cells were negative. Ia3 reactivity in well differentiatedcolonie cancer tissue was predominantly with luminal contentsand apical membranes surrounding the lumen of the gland-likestructures but not with the cytoplasm (Fig. \g).

Nd2 did not react with normal colon. However, its reactivitywith cancerous colonie tissues was somewhat lower (24%) than

373

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

MONOCLONAL ANTIBODIES REACTIVE WITH PANCREATIC CANCER MUCIN

*

Fig. 1. Immunopcroxidase histochemical reactivity of lai. a. poorly differentiated pancreatic cancer, showing cytoplasmic staining, x 248. ft. well differentiatedpancreatic cancer, stained principally in the apical membranes. Occasional staining of luminal contents was also observed, x 122. c. intestinal metaplasia of thestomach, showing staining of the vacuoles of goblet cells, x 188. a, well differentiated stomach cancer, with strong staining of luminal contents and apical membranes.X 165. e, lack of Ia.1 reactivity with columnar cells and vacuoles of normal colonie goblet cells. X 220. as compared to f. transitional mucosa adjacent to coloniecancer tissue, x 290. where Ia3 stained both the vacuoles and the luminal contents. #, well differentiated colonie cancer tissue stained with Ia3. where reactivity waswith luminal contents and apical membranes surrounding the lumens, x 155.

that in either pancreatic or stomach cancers. In cancer tissue,the staining was of luminal contents and apical membranes butlittle of cytoplasm (Fig. 2d).

Liver. No reactivity was observed with either normal or cancerous liver by either Ia3 or Nd2.

Non-GI Tissues. Ia3 reactivity in breast cancer was usuallywith the luminal contents, apical membrane, and apical cytoplasm. Nd2 reactivity with malignant breast tissue was lower

than that of Ia3 (14.3% as compared to 67% (Table 2). In onecase of infiltrative duct breast cancer, the pattern of reactivitywas strong staining of both apical membrane and cytoplasm.

Both Ia3 and Nd2 reacted with normal bronchial glands(Table 2). Ia3 had some reactivity with the endothelium ofcapillaries of the glomeruli of normal kidney but none with thetubules. Nd2 did not react with normal kidney.

Neither Ia3 nor Nd2 reacted with the following normal374

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

MONOCLONAL ANTIBODIES REACTIVE WITH PANCREATIC CANCER MUCIN

Fig. 2. Immunoperoxidase histochemical reactivity of Nd2. a, normal pancreas, x 390, showing no reactivity with ductal cells or acini, b, well differentiatedpancreatic cancer tissue, X 244, showing intense staining of luminal contents and apical cytoplasm, c. moderately differentiated gastric cancer tissue, x 195, showingintense staining of luminal contents and cellular membranes, d. moderate to well differentiated cancer tissue immediately adjacent to seemingly normal colonie tissue,x 82. Nd2 staining of the cancer tissue was characterized by staining of luminal contents and the apical membrane.

tissues (n = 5 each): adrenal gland, brain, muscle, nerves, ovary,skin, testis, thyroid, and uterus.

Detection of Ia3 and Nd2 in Serum. The absorbances of bothIa3 and Nd2 EIAs related well to the concentration of antigen(Fig. 3). A consistent elevation of Nd2 in sera from patientswith malignancies could not be demonstrated by EIA, immuno-

staining of sera spotted onto nitrocellulose membrane witheither HRP-Nd2 or Nd2 followed by second antibody and

radioiodinated Protein A, or immunostaining of Western blotsafter electrophoresis. Both denaturing and nondenaturing elec-trophoretic conditions (see below) were tried.

Table 3 and Fig. 4 show the results of a screening of humansera with the Ia3 EIA. Using the mean +2 SD (1078 units/ml)as a cut-off point, 7.7% of normal individuals had levels abovethis value. When individuals with GI malignancies were examined, 43.5% of the patients with pancreatic cancer, 62.5% with

375

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

MONOCLONAL ANTIBODIES REACTIVE WITH PANCREATIC CANCER MUCIN

Table 2 lai and Nd2 reactivity with normal and malignant nongastrointestinal Ia3 (Units/ml)

No. positive/total(%)Ia3

Nd2Cancer

Normal CancerNormalLung8/8(100) 7/10(70)° 1/9(11.1) 8/13(61.5)°

Kidney 0/12(0) 9/10(90) 0/13(0) 0/13(0)Bladder 6/6(100) 0/4(0) 0/8(0) 0/3(0)Breast 10/15(66.7) ND* 2/14(14.3)ND"

Reactivity was wit*ND, notdetermin2.0-IIt]

CJ3i.o-in

ri

ll bronchial glands of the lung,ed.0*•0.5I

•0.4•0.3

S•0.2

g>

,, fCNORMAL

PANCREATITISPANCREATIC

CANCERCOLORECTAL,

CANCERGASTRIC

CANCERHEPATOMALUNG

CANCERLYMPHOMABREAST

,CANCERBLADDER

,CANCERKIDNEY

CANCERFig.

4. SerologieO, five cases; venie•

100O 20OO 30OO 4OOO 3OUU>5OOO11 1 1 JJ«HH

*•i *:-M*i •-i.»4*1»»•«•

0•l"f

•<m

*•<+•1-4|

. 1 .«gH

.8.••al

screening with sandwich EIA using Ia3. ••individual cases;al line, mean of normal + 2 SD ( 1078 units/ml).

0 2 4 6 8 10 12 14

ANTIGEN CONCENTRATION

Fig. 3. Standard cunes of Ia3 and Nd2 sandwich El As. Native SW1990 mucinwas used as antigen. The highest concentration of antigen used washexose. Antigen was serially diluted 2-fold to construct the curves.

Table 3 lai reactivity with sera

SW1990 xenograft mucin

Reactivity (units/ml)

n Mean SD Median Range i Positive

Normal 78 467 311Pancreatitis 16 341 218

Total normal/benign 94

350 78-1.300 7.7°250 90-700 0

6.4

CancersGastrointestinalPancreasStomachColonLiverTotal

GINongastrointestinalLungLymphomaBreastBladderKidneyTotal

non-GI4640301913518182220IS933,6746,1753,9945,7251,5871.2983491341108,6347,97710,1058.0654,6224,671326119887001,5006121,250425161200909090-50.0008-20,0000-50,0000-20,00019-20.00019-20.0000-1,2500-50010-31243.562.536.763.250.416.75.64.5005.4

Reactivity(% ofcontrol)Treatment100'C,

15minReduction

andalkylation10%

formalinNeuraminidase0.004

units/ml0.02units/ml0.1units/ml2.0

units/mlPeriodateO.I

mM1.0mM5.0mM20.0

mMIa3in

no90.475.913.73.2453.5000Nd275090.492.567.858.170.19277.655.760.2

SD (1078 units/ml).

gastric cancers, 36.7% with colonie cancers, and 63.2% withliver cancers had levels above the cut-off. None of the 16individuals with pancreatitis had elevated levels. When non-GImalignancies were examined, only lung cancer patients as agroup had a higher positive rate than normal (16.7%). Patientswith lymphomas, breast cancer, bladder cancer, and kidneycancer had positive rates of 5.6% or less. Ia3 had a highersensitivity for GI (50.4% positive rate) than non-GI malignancies (5.4% positive rate).

Characteristics of the Ia3 and Nd2 Epitopes. Ia3 reactivitywith mucins was sensitive to low concentrations of either periodate or neuraminidase (Table 4). This suggests that Ia3 reactsprincipally with carbohydrates, requiring sialic acid for binding.Accordingly, such protein-denaturing conditions as the reduction of disulfide bonds and heat did not affect Ia3 reactivity

" Mean of two determinations.

(Table 4). In contrast, reduction of disulfide bonds in mucincompletely abolished Nd2 reactivity (Table 4). Nd2 reactivitywas unaffected by concentrations of neuraminidase which completely abolished Ia3 reactivity (Fig. 5). At higher concentrations of either neuraminidase or periodate (Table 4), somereduction in Nd2 reactivity was observed, although the maximum reduction produced was approximately 30-40%. This wasrelatively unaltered over an increase of 100-fold in neuraminidase or of 4-fold in periodate concentrations. In contrast, Ia3reactivity was unaffected by a range of trypsin concentrationswhich greatly reduced Nd2 reactivity (Fig. 5B). Treatment ofthe mucin with formalin had little effect on the reactivity ofeither Ia3 or Nd2 (Table 4).

The sensitivity of Nd2 reactivity to denaturing conditionsdemonstrated by EIA was confirmed by gel electrophoresis (Fig.6). When SDS was present, no reactivity was observed afterpretreatment with mercaptoethanol (Fig. 6, lanes c and e),although some reactivity was detectable if SDS was omitted(Fig. 6, lanes g and /). Heating the sample (Fig. 6, lanes g andh) caused some reduction in reactivity (compare Fig. 6, lane gto lane i). In contrast, mercaptoethanol did not affect Ia3immunoreactivity but did result in a greater penetration of Ia3

376

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

MONOCLONAL ANTIBODIES REACTIVE WITH PANCREATIC CANCER MUCIN

1201

100

ftHU

80

60

4O

2O

2.0 3.0 4.O 5.0 6.0LOG OF NEURAMINIDASE CONCENTRATION (nU/ml)

O

OusgD

120

100

80

60

4O

20

2.0 3.0 4.0 5.0 6.0

LOG OF TRYPSIN CONCENTRATION (ng/ml)

Fig. 5. Effect of neuraminidase and trypsin on Ia3 and Nd2 ¡mmunoreactivity.Submaximal concentrations of SW1990 mucin were adsorbed overnight to 96-well polystyrene microtiter plates. A, Clostridium perfringens neuraminidase inPBS added to the wells and incubated for 2 h at room temperature. B. incubationwith bovine pancreatic trypsin for 5 h at room temperature. A, Ia3 reactivity withnative SW1990 mucin; •.Nd2 reactivity with native SVV1990 mucin; O. Nd2reactivity with partially deglycosylated SVV1990 mucin. No capture antibody wasused in these studies.

Ia3i—n-

Nd2+SDS -SDS

-+- ME

-205 K

-45 Ka b cd e f

Fig. 6. Ia3 and Nd2 immunoreactivity after polyacrylamide electrophoresis ofSW1990mucins. Lanes a-f, electrophoresis with SDS. Lane a. Ia3 reactivity afterheating mucin with d-mercaptoethanol (ME). Lane b, Ia3 reactivity when mucinwas heated without d-mercaptoethanol. Lane c. Nd2 reactivity when mucin washeated with J-mercaptoethanol. Lane d. same as lane c except without d-mercaptoethanol. Lane e, same as lane c except that mucin was not heated. Lane f, nomercaptoethanol and the mucin was not heated. Lanes g-j, SDS was not used.Otherwise the samples were the same as for lanes c-f. The molecular weightmarkers, myosin (M, 205,000) and ovalbumin (M, 45,000). were only applicableto the SDS system.

reactivity into the 7% gel (Fig. 6, lanes a and b).Nd2 retained reactivity even after partial deglycosylation of

SW1990 mucins by HF (Fig. 5). This suggested the possibilitythat Nd2 reacts, at least in part, with the peptide portion of themucin. Nd2 reactivity with partially deglycosylated mucin hada similar sensitivity to SDS, heat, and mercaptoethanol as

described above for native mucin (results not shown), and itwas also sensitive to trypsin digestion (Fig. 5B) but not toneuraminidase treatment (Fig. 5/1). Treatment of native mucinswith HF for 3 hr at room temperature completely abolishedNd2 reactivity.

Neither Ia3 nor Nd2 reactivity was affected by up to 50 mMof the monosaccharides present in native SW1990 mucins:fucose, galactose, jV-acetyl-galactosamine, or /V-acetyl-glucosa-mine (16). Finally, neither EIA nor immunostaining of thinlayer plates demonstrated any Ia3 or Nd2 reactivity with neutralor acidic glycolipids isolated from SW1990 membranes.

Comparison of Ia3 and Nd2 with Other Peptide and Carbohydrate Epitopes of Mucins. Four cancer-associated MoAbs,B72.3 (26), DU-PAN-2 (5), NS19-9 (3), and SPan-1 (7), whichreact with sialic acid-containing carbohydrate epitopes weretested for their cross-reactivity with Ia3 and Nd2 over a wide

range of concentrations of the purified antibodies (1 Mg/ml to1 mg/ml). No competition was observed. In addition, Ia3 didnot compete with Nd2 and vice versa, although Nd2 and Ia3exhibited self-inhibition.

A number of MoAbs reported to react, at least in part, withmucin peptides were also tested (Table 5). These included DF3(10), HMFG-1 (8), HMFG-2 (8), SM-3 (12), 115D8 (9), and139H2 (9). None affected Ia3 or Nd2 binding. The polyclonalantibody against the repetitive peptide sequence of small intestinal/colonie mucin (13) also had no effect. We also comparedIa3 antigen levels in pancreatic cancer sera to levels of mucinousantigens assayed by three commercially available antibody assays, CAI5-3 (27), CAI9-9 (3), and DU-PAN-2 (5) (Fig. 7).Ia3 levels did not correlate with CAI5-3 (r = 0.090), ÇA19-9(r = 0.05), or DU-PAN-2 (r = 0.449). Moreover, the Ia3

determinant was elevated in 1 of 32, l of 38, and 6 of 35 caseswhich were not elevated in the CAI5-3, CAI9-9, and DU-PAN-2 determinants, respectively.

Ia3 antigen was compared to CAI9-9, 115D8, and DF3antigens in serum by SDS-PAGE and Western blotting (Fig.8). All three types of antigen migrated much more slowly thanthe M, 205,000 molecular weight marker (myosin). DF3 reactivity showed the polymorphic characteristic previously reported (1). The patterns of 115D8 and DF3, both componentsof the CAI5-3 sandwich assay (27), were similar (results notshown). While none of the sera had DF3 (or 115D8) reactivityremaining in the 4% stacking gel, both CAI9-9 and Ia3 hadmuch more slowly moving components. Thus, Ia3 antigenresembled CAI9-9 antigen more than DF3/115D8 antigens.However, like DF3/115D8 antigens, both Ia3 and CAI9-9reactivity patterns differed with the individual.

Table 5 Antibodies distinct from lai and Nd2

Reference

Monoclonal antibodies against pancreatic cancer mucinsB72.3Ia3/Nd2DU-PAN-2NS19-9SPan-1

26m

S37

Monoclonal antibodies against breast mucin peptideDF3115D8139H2HMFG-1HMFG-2SM-3

Polyclonal antibody against colonie mucin peptideanti-MRP

109988

II

13' Data in this report.

377

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

MONOCLONAL ANTIBODIES REACTIVE WITH PANCREATIC CANCER MUCIN

100000

— 10000

BM

n

1000

100

IO

lOOOOO

~ 10000

^

~ 1000

S too

.*••••

10 100

CA15-3 (uniti/mi)

1000

10

»•

10 100 1000 10000 100000

CA19-9 (units/ml)100000

—10000

I

1OO01

100

1010 100 1000 10000 lOOOOO

DU PAN 2 (unite/ml)Fig. 7. Comparison of Ia3. CAI5-3, CAI9-9, and DU-PAN-2 ancigen levels

in pancreatic cancer sera. Vertical lines, cul-off values recommended by themanufacturers for the commercial assays, CAI5-3. CAI9-9, and DU-PAN-2.Horizontal lines, cut-off value for the Ia3 assay (mean + 2 SD = 1070 units/ml).The correlation coefficients were as follows: CAI5-3, r = 0.090; ÇA19-9, r =0.05; and DU-PAN-2. r = 0.449.

Ia3 CA19-9 DF3

4%5%

-205 Ka bcde abed e abed e

Fig. 8. SDS-PAGE and Western blotting of sera. Antigens in human sera wereseparated by electrophoresis on a 4?c/5cc gel and transferred electrophoreticallyto nitrocellulose membranes, as described in "Materials and Methods." TheWestern blots were immunostained with Ia3, NS19-9. and DF3 monoclonalantibodies, rabbit secondary antibodies, and radioiodinated Protein A, as described in "Materials and Methods." Lanes a-d, pancreatic cancer sera; lane e.

normal serum.

DISCUSSION

This report describes the characterization of two new MoAbs,Ia3 and Nd2, produced against purified pancreatic cancer mu-cins. The Ia3 determinant contains sialic acid but is not presenton glycolipids isolated from the same tissue source as the

mucins and differs from the epitopes identified by such sialicacid-requiring MoAbs as B72.3 (26), CAI9-9 (3), DU-PAN-2(5), and SPan-1 (4). Moreover, levels of Ia3 antigen in seravaried independently of those identified by the other mucinmarkers CAI5-3, CAI9-9, and DU-PAN-2. SDS-PAGEanalysis of sera showed that Ia3 reacted with antigens withdifferent mobilities from those recognized by DF3 or 115D8,components of the breast cancer mucin assay, CAI5-3 (27).While DF3 or 115D8 reacted with polymorphic components inthe molecular weight range of 300,000 to 400,000 (1,9), mostof the Ia3 components migrated more slowly. In this respectIa3 also differed from the M, 300,000 pancreatic cancer-associated antigen MUSE 11 (28) but resembled the CAI9-9 antigens, which also migrated more slowly. While Ia3 and DF3/115D8 antigens differ from each other in mobility, both havethe common feature that each pattern is characteristic of theindividual. The variations in the molecular size observed withthe breast cancer mucin MoAbs are now thought to be due tolength polymorphism of the gene encoding for the apomucin(15). The differences in mobilities seen with Ia3 antigens maybe due to similar length polymorphisms of the apomucin ormay be due to differences in degree or type of glycosylation.

Serum levels of Ia3 antigen were elevated in a significantnumber of patients with GI malignancies (50.4%). However,Ia3 was not elevated in sera of most patients with non-GItumors (5.4%), although Ia3 was present in a high percentageof malignant non-GI epithelial tissues such as breast (66.7%)and bladder (100%). The positive rate of Ia3 for pancreaticcancer was at least as high as that of carcinoembryonic antigen(CEA) or tissue polypeptide antigen (TPA) (29), although notas high as that of CA19-9 (29) or SPan-1 (4). Since the Ia3determinant is distinct from those of CAI9-9 and DU-PAN-2,it may be used in combination with those assays for increasedsensitivity or in cases where it is not possible to apply theothers. For example, Ia3 antigen was elevated in 17% of thepancreatic cancer cases not elevated in DU-PAN-2. The Ia3assay could also be used for individuals of the Lewis a~/b~

phenotype, who may lack the CAI9-9 epitope (3). Finally, ourdata show that, in addition to being elevated in pancreaticcancer serum, Ia3 antigen was elevated in a significant percentage of gastric cancer patients (63%). Thus, Ia3 may also beapplied to gastric cancer diagnosis.

While Ia3 reactivity was destroyed by neuraminidase andperiodate, Nd2 reactivity could only be reduced by a maximumof 40% by either of those agents. The loss of Nd2 reactivityafter protease digestion and protein-denaturing treatments suchas sulfhydryl bond reduction indicates that Nd2 immunoreac-tivity may involve both carbohydrates and peptide, with theintegrity of the peptide backbone being a major determiningfactor. This may explain the lack of reactivity of Nd2 withcompletely deglycosylated SW1990 mucins. Colon cancer mucins deglycosylated in a similar fashion (3 h, room temperature)underwent extensive peptide cleavage, while partially deglycosylated mucin (1 h, 0°C)retained the integrity of the mucin

core peptide (17). Thus, it is possible that damage to relativelyrare Nd2-reactive peptide regions caused by the more strenuousconditions could account for the lack of Nd2 reactivity withcompletely deglycosylated mucins. Similar examinations of therelative sensitivities to neuraminidase, periodate, and proteasehave been used for the initial identification of MoAbs that reactwith peptides (28, 30), including DF3. Although the reactivityof DF3 with breast mucin was greatly reduced by neuraminidase(30), it was still useful in identifying mucin associated cDNAs(10). However, Nd2 did not react with the breast mucin-asso-

378

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

MONOCLONAL ANTIBODIES REACTIVE WITH PANCREATIC CANCER MUCIN

ciated peptide sequences recognized by MoAbs such as 115D8(9), 139H2 (9), DF3 (10), HMFG-1 (8), HMFG-2 (8), and SM-3 (11). It also did not recognize the tandem repeat segment ofintestinal apomucin recognized by anti-MRP (13). Furtherstudies utilizing new sequences identified by Nd2 in cDNAcloning experiments may provide more precise information onits epitope.

Nd2 antigen was not elevated in the sera of patients withmalignancies, although it was present in a high percentage oftumor tissues. There may be several explanations for this. Onepossibility is that, although the various Nd2 assays in this studycorrelated well with purified antigen levels, they were, nevertheless, not sensitive enough to detect circulating Nd2 antigen.Alternatively, the Ia3 and Nd2 epitopes may be carried bydifferent populations of mucins, only one of which is releasedinto the bloodstream. At least three different types of mucinsexist, which differ in tissue distribution (10, 13, 31, 32). Onetype was originally characterized in breast tissue and is encodedby the mucin gene MUC 1 (33), which is located on chromosome 1 (33). The second type is associated with small intestinemucins (13) and is coded by the mucin gene MUC 2, locatedon chromosome 11; the third type, MUC 3, located on chromosome 7, was also originally identified in small intestine (32).All three mucin genes are polymorphic (13, 14, 32). Recently,the products of both the MUC 1 and the MUC 2 genes weredetected in pancreatic cancer tissue (31, 34). However, it is notknown at this time how many different types of mucins arepresent in each normal and cancerous tissue type or what factorsinfluence the production and release of specific mucin types.

The Nd2 epitope was not detected in normal GI tissues,although it was present in a high percentage of malignanttissues. This cannot be explained simply by a loss of immuno-reactive antigen during formalin fixation of tissue since Nd2reactivity is unaffected by formalin treatment of purified mucins. As has been suggested for SM-3, a MoAb directed againstdeglycosylated breast cancer mucin which was also detectedonly in malignant tissues (12), Nd2 may recognize parts of themucin protein core which are exposed in malignant but notnormal tissues. Thus, the ability of Nd2 to detect such malignancy-associated alterations can be utilized to study how bio-synthetic pathways of mucin in normal and malignant cellsdiffer from each other. Moreover, the high specificity exhibitedby Nd2 for tumor tissues and the absence of the Nd2 epitopefrom sera suggest that it may be used as the recognition agentfor immunotargeting of therapeutic drugs or tracers.

In conclusion, the principal findings of this study are thattwo new MoAbs directed against purified pancreatic cancermucins, Ia3 and Nd2, identify malignancy-associated carbohydrate and peptide structures in mucins which appear not tohave been previously described. Thus, Ia3 and Nd2 may have awide range of application including: (a) diagnostic screening ofsera and tissues; (b) immunotargeting of drugs or tracers tomalignant tissue; (c) elucidation of the biosynthesis and processing pathways of different types of mucins in normal andcancerous cells; and (d) characterization of the pancreatic cancer mucin polypeptide(s).

ACKNOWLEDGMENTS

The authors would like to thank Whitney Ryan for her technicalassistance, Dr. Bader Siddiki for helpful suggestions, Dr. Y. S. Chungfor providing the CAI9-9 and DU-PAN-2 kits, and Rita Burns for herhelp in the preparation of this manuscript.

REFERENCES

1. Hayes, D. F., Sekine, H., Ohno, T., Abe, M., Keefe, K.. and Kufe. D. W.Use of a murine monoclonal antibody for detection of circulating plasmaDF3 antigen levels in breast cancer patients. J. Clin. Invest.. 75: 1671-1678,1985.

2. Fisken, J.. Leonard. R. C. F., and Roulston, J. E. Immunoassay of CAI 25in ovarian cancer: a comparison of three assays for use in diagnosis andmonitoring. Disease Markers. 7:61-67, 1989.

3. Magnani, J. L., Steplewski, Z., 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.

4. Chung. Y. S., Ho, J. J. L.. Kim. Y. S.. Tanaka, B. Nakata. B., Hiura. A.,Motoyoshi, H., Satake, K., and Umeyama, K. The detection of humanpancreatic cancer-associated antigen in the serum of cancer patients. Cancer(Phila.), 60: 1636-1643, 1987.

5. Metzgar. R. S., Rodriguez, N., Finn, O. J., Lan, M. S., Daasch, V. N.,Fernsten, P. D., Meyers, W. C., Sindelar, W. F., Sandler, R. S. and Seigier,H. F. Detection of a pancreatic cancer-associated antigen (DU-PAN-2 antigen) in serum and ascites of patients with adenocarcinoma. Proc. Nati. Acad.Sci. USA. 81: 5242-5246. 1984.

6. Rittenhouse, H. G., Manderino, G. L., and Hass, G. M. Mucin-type glyco-proteins as tumor markers. Lab. Med., 16: 556-560. 1985.

7. Ho, J. J. L., Chung, Y. S., Fujimoto, Y., Bi, N., Ryan, W., Yuan, S-Z., Byrd,J. C., and Kim, Y. S. Mucin-like antigens in a human pancreatic cancer cellline identified by murine monoclonal antibodies SPan-1 and YPan-1. CancerRes., 48: 3924-3931, 1988.

8. Gendler, S. J., Burche»,J. M.. Duhig, T., Lamport, D.. White, R., Parker,M., and Taylor-Papadimitriou, J. Cloning of partial cDNA encoding differentiation and tumor-associated mucin glycoproteins expressed by humanmammary epithelium. Proc. Nati. Acad. Sci. USA, 84: 6060-6064, 1987.

9. Hilkens, J., Buijs, F.. and Ligtenberg. M. Complexity of MAM-6. an epithelial sialomucin associated with carcinomas. Cancer Res., 49: 786-793, 1989.

10. Siddiqui, J., Abe. M., Hayes, D., Shani. E., Yunis, E., and Kufe, D. Isolationand sequencing of a cDNA coding for the human DF3 breast carcinoma-associated antigen. Proc. Nati. Acad. Sci. USA. 85: 2320-2323. 1988.

11. Burchell. J., Taylor-Papadimitriou, J., Bosheit, M., Gendler, S., and Duhig,T. A short sequence, within the amino acid tandem repeat of a cancer-associated mucin. contains immunodominant epitopes. Int. J. Cancer. 44:691-696, 1989.

12. Girling, A., Bartkova, J., Burchell, J., Gendler, S., Gillett, C., and Taylor-Papadimitriou, J. A core protein epitope of the polymorphic epithelial mucindetected by the monoclonal antibody SM-3 is selectively exposed in a rangeof primary carcinomas. Int. J. Cancer, 43: 1072-1076, 1989.

13. Gum, J. R., Byrd, J. C., Hicks, J. W., Toribara, N. W., Lamport, D. T. A.,and Kim, Y. S. Molecular cloning of human intestinal mucin cDNAs.Sequence analysis and evidence for genetic polymorphism. J. Biol. Chem.,264: 6480-6487. 1989.

14. Griffiths, B., Gordon, A., Burchell, J., Bramwell, M. E., Griffiths, A.. Price,M., Taylor-Papadimitriou, J., Zanin, D., and Swallow, D. M. The breasttumour-associated epithelial mucins and the peanut lectin binding urinarymucins are coded by a single highly polymorphic gene locus 'PUIW. DiseaseMarkers, 6: 185-194, 1988.

15. Ligtenberg, M. J. L., Vos, H. L.. Gennissen, A. M. C.. and Hilkens, J.Episialin, a carcinoma-associated mucin, is generated by a polymorphic geneencoding splice variants with alternative amino termini. J. Biol. Chem., 265:5573-5578, 1990.

16. Nardelli, J., Byrd, J. C., Ho. J., Fearney, F. J., Tasman-Jones, C., and Kim,Y. S. Pancreatic cancer mucin from xenografts of SW1990 cells: isolation,characterization, and comparison to colon cancer mucin. Pancreas. 3: 631-641, 1988.

17. Byrd, J. C., Lamport, D. T. A., Siddiqui, B.. Kuan, S-F., Erickson. R.,Itzkowitz, S. H.. and Kim. Y. S. Deglycosylation of mucin from LS174Tcolon cancer cells by hydrogen fluoride treatment. Biochem. J., 261: 617-625, 1989.

18. Yuan, S-Z., Ho. J. J. L., Yuan, M., and Kim, Y. S. Human pancreatic cancer-associated antigens detected by murine monoclonal antibodies. Cancer Res.,45:6179-6187, 1985.

19. Ho, J. J. L., Chung. Y-S., Ryan, W., Henslee, J. G., and Kim, Y. S. Evaluationof factors affecting the performance of enzymatic SPan-1 immunoassays forpancreatic cancer. J. Immunol. Methods, 125: 97-104, 1989.

20. Carlstedt, I., Lindgren, H., and Sheehan, J. K. The macromolecular structureof human cervical-mucus glycoproteins. Studies on fragments obtained afterreduction of disulphide bridges and after subsequent trypsin digestion.Biochem. J., 213: 427-435. 1983.

21. 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.

22. Siddiqui, B., and Kim, Y. S. Effects of sodium butyrate, dimethyl sulfoxide.and retinoic acid on glycolipids of human rectal adenocarcinoma cells. CancerRes., 44: 1648-1652. 1984.

23. Sun, Q.. Siddiqui, B., Nudelman. E., Hakomori, S.-L, Ho, J. J. L., and Kim,Y. S. New murine monoclonal antibodies to a human colonie cancer-associated glycolipid, extended difucosylated Leyglycolipid. Cancer J., /: 213-220,1987.

24. Kannagi, R., Stroup, R., Cochran, N. A., Urdal, D. L., Young, W. W., Jr.,and Hakomori. S-I. Factors affecting expression of glycolipid tumor antigens:

379

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

MONOCLONAL ANTIBODIES REACTIVE WITH PANCREATIC CANCER MUCIN

influence of ceramide composition and coexisting glycolipid on the antige-nicity of gangliotriaosylceramide in murine lymphoma cells. Cancer Res..«.•4997-5005.1983.

25. Buehlcr. J.. and Macher. B. A. Glycosphingolipid immunostaining: detectionof antibody binding with an avidin-biotin enzyme system. Anal. Biochem.,158: 283-287. 1986.

26. Gold. D. V.. and Mattes. M. J. Monoclonal antibody B72.3 reacts with acore region structure of O-linked carbohydrates. Tumor Biol.. 9: 137-144.1988.

27. Hayes. D. F., Zurawski, V. R., Jr.. and Kufe, D. \\ . Comparison of circulatingCAI5-3 and carcinoembryonic antigen levels in patients with breast cancer.J. Clin. Oncol., 4: 1542-1550. 1986.

28. Ban. T.. Imai. K.. and Vachi, A. Imtnunohistological and immunochemicalcharacterization of a novel pancreatic cancer-associated antigen MUSEI 1.Cancer Res.. 49: 7141-7146. 1989.

29. Benini. L.. Cavallini. G.. Zordan. D.. Rizzotti. P.. Rigo. L.. Brocco. G..Perobelli. L.. Zanchetta, M.. Pederzoli. P.. and Scuro. L. A. A clinical

evaluation of monoclonal (CA19-9. CASO, CA12-5) and polyclonal (CEA,TPA) antibody-defined antigens for the diagnosis of pancreatic cancer. Pancreas, i: 61 -66. 1988.

30. Sekinc, H.. Ohno, T., and Kufe, D. Purification and characterization of ahigh molecular weight glycoprotein detectable in human milk and breastcarcinomas. J. Immunol., 135: 3610-3615, 1985.

31. Lan. M. S.. Hollingsworth. M. A., and Metzgar. R. S. Polypeptide core of ahuman pancreatic tumor mucin antigen. Cancer Res., 50: 2997-3001. 1990.

32. Gum. J. R.. Hicks. J. W., Swallow, D. M.. Lagace. R. L.. Byrd. J. C,Lamport, D. T. A.. Siddiki. B.. and Kim. Y. S. Molecular cloning of cDNAsderived from a novel human intestinal mucin gene. Biochem. Biophys. Res.Commun., 171: 407-415, 1990.

33. Griffiths. B., Gum. J., West, L. F., Povey, S.. Swallow, D. M., and Kim. V.S. Mapping of the gene coding for intestinal mucin to chromosome 11p 15.Cytogenet. Cell Genet., SI: 1008. 1989.

34. Yonezawa, S., Byrd. J. C., Ho.. J.. Dahiya, R.. Gum. J., and Kim, Y. S.Pancreatic and colonie cancer cells express different mucin genes. Gastro-enterology. 98: A240. 1990.

380

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from

1991;51:372-380. Cancer Res   Jenny J. L. Ho, Ning Bi, Pei-sha Yan, et al.   Monoclonal Antibodies Directed against Purified MucinCharacterization of New Pancreatic Cancer-Reactive

  Updated version

  http://cancerres.aacrjournals.org/content/51/1/372

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/51/1/372To request permission to re-use all or part of this article, use this link

Research. on January 23, 2020. © 1991 American Association for Cancercancerres.aacrjournals.org Downloaded from