(CANCER RESEARCH 49, 786-793. February 15, 1989]

Complexity of MAM-6, an Epithelial Sialomucin Associated with CarcinomasJohn Hilkens, Femke Buijs, and Marjolijn Ligtenberg1

Division of Tumor Biology, The Netherlands Cancer Institute (Antoni van Leeuwenhoek Huis), Plesmanlaan 121, 1066 CX Amsterdam, The Netherlands

ABSTRACT

The complexity of epithelial sialomucins was investigated by immu-noprecipitation and membrane immunofluorescence, using monoclonalantibodies (MAbs) against MAM-6 and other sialomucins. MAbs againstMAM-6 immunoprecipitated from a variety of sources either one or twosialylated glycoproteins with apparent molecular weights of over 400,000under reducing as well as nonreducing conditions. The electrophoreticmobility of each MAM-6 glycoprotein as isolated from serum, milk, andcell lines of different individuals showed considerable variation. Thedifferences in molecular weight of the MAM-6 glycoproteins were alsoreflected at the level of MAM-6 precursors which are less heavilyglycosylated. Therefore, large differences in apparent molecular weight(150,000 and over) are most likely due to a variable protein backbone.We used this molecular polymorphism to prove that 11 MAbs againstdifferent sialomucins, obtained from various investigators, precipitatedsialomucins generated from common precursor molecules.

The pattern of reactivity of the MAbs with carcinoma cell lines wascomplex. All but the two MAbs, directed against putative carbohydrateepitopes, immunoprecipitated the precursor molecule from each cell line.However, some of them were unable to immunoprecipitate the matureform of MAM-6 from these cell lines. These results indicate that thoseepitopes are masked, probably due to cell line- or possibly cell type-dependent variations in glycosylation of the epithelial sialomucin. Evenwithin a single cell line mature molecules with different epitopes wereobserved. The differential reactivity of the MAbs was confirmed bymembrane immunofluorescence. These results show that MAM-6 belongs to a family of epithelial sialomucins with a polymorphic proteinbackbone and extensive variation in glycosylation.

INTRODUCTION

Mucus glycoproteins (mucins) are present in the glycocalyxof epithelial cells of many secretory tissues, often show tumor-associated alterations (1-3), and constitute convenient markersfor certain types of tumors. Mucins contain large amounts of0-linked glycans of varying length and degree of branching.These glycoproteins are frequently used as tissue and serummarkers for cancer diagnosis (4-9), but little is known aboutthe regulation of their expression and their function in normaland tumor cells.

Many investigators have raised MAbs2 to mucins or mucin-

like molecules. The molecular weight of these mucins and theirtissue distribution vary considerably. For instance, DeKretseret al. (10) described M Ab OM-1 which reacts mainly withovarian carcinomas and rarely with breast or colon carcinomas.M Ab Cal reacts with a wide variety of carcinomas and a fewnoncarcinomas, but with only a very limited number of normaltissues (11). MAb SM-3 preferentially reacts with carcinomasand shows very little or no reaction with normal epithelium(12). In addition, other MAbs directed to mucins with a broad

Received 8/11/88; revised 11/3/88; accepted 11/11/88.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.

1Supported by the Netherlands Cancer Foundation (K. W. F.).2The abbreviations used are: MAb, monoclonal antibody; cDNA, complemen

tary DNA; RAM, rabbit antiserum to mouse immunoglobulins; NMS, normalmouse serum; FITC, fluorescein isothiocyanate; DMEM, Dulbecco's modifiedEagle's medium; PBS, phosphate-buffered saline; PCS, fetal calf serum; SDS,sodium dodecyl sulfate; PAGE, polyacry'lamide gel electrophoresis; HMFG,human milk fat globule; PUM, polymorphic urinary mucin.

reactivity on normal tissues and carcinomas have been described.

We have raised MAbs of the latter type which are directed toan epithelial sialomucin designated MAM-6. MAM-6 is definedby several MAbs (notably, 115D8,115F5, and 139H2) directedagainst different epitopes on the antigen (13, 14). Immunohis-tological studies have shown that MAM-6 is present at theapical side of most ductal and alveolar cells in normal tissuesand in the glycocalyx and the cytoplasm of almost all carcinomas.

Part of the MAM-6 cDNA has been cloned in a Xgtl 1 vector.3Because several of the MAbs directed to MAM-6 also reactedwith the MAM-6-/3-galactosidase fusion protein which wasobtained by infecting Escherichia coli with this recombinantphage, their epitopes are present on the protein backbone ofthe molecule.

We have recently shown that the antigen immunoprecipitatedby anti-MAM-6 MAbs from ZR-75-1 breast cancer cells consists of two high-molecular-weight glycoproteins of approximately 450,000 and 650,000 (15). The biosynthesis of eachglycoprotein proceeded very similarly through at least threedistinct precursors. The first precursors were detectable within1 to 2 min of biosynthetic labeling and rapidly processed byproteolytic cleavage to yield a second set of lower molecularweight precursors. Both sets of precursors possessed TV-linkedglycans. Maturation occurred by extensive O-linked glycosylation leading to M, 500,000 and 700,000 precursors after 30min and to the mature glycoproteins after 180 min. The latterstage of maturation probably involves mostly sialylation.

In view of the apparently discordant results obtained byvarious investigators using different MAbs directed againstthese clinically relevant mucins, we have investigated the relationship between the epithelial mucins detected by 11 MAbsraised by us and by others. For this purpose we analyzed thedifferences in molecular weight of the mucins and the biochemical basis for the differences in reactivity of these MAbs withvarious cell lines by immunoprecipitation and immunofluorescence techniques.

MATERIALS AND METHODS

Monoclonal Antibodies and Antisera. MAbs used in this study arelisted in Table 1. They are all directed to mucins. RAM was preparedby immunizing rabbits with affinity-purified mouse immunoglobulins.RAM used for 125Ilabeling was purified by Protein A-Sepharose affinity

chromatography. When RAM was used for the preparation of preformed complexes, it was partially purified by 50% ammonium sulfateprecipitation. NMS was prepared from nonimmunized BALB/c mice.FITC-labeled goat anti-mouse IgG was obtained from Nordic, Tilburg,The Netherlands.

Cells. The following human breast cancer cell lines were used (fordetailed description, see Ref. 25). ZR-75-1 cells were a gift from Dr.G. Malinckoff, Centocor, Malvern, PA. T47D and MDA MB 157 cellswere obtained from the American Type Culture Collection, Rockville,MD. BT-20 cells were kindly donated by Dr. J. Taylor-Papadimitriou,ICRF, London. MCF-7 cells were obtained from Dr. F. Prop, University of Amsterdam. H134 is a cell line derived from an ovarian card

3M. Ligtenberg, A. Gennissen, H. Vos, and J. Hilkens, manuscript in prepa

ration.

786

on July 7, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

COMPLEXITY OF MAM-6

Table 1 Monoclonal antibodies to epithelial sialomucins

MonoclonalantibodyHMFG-2CalDF3USDS115G2115F5139H2140C1OM-1E29SM-3AntigenMucinEpitectinMucinMAM-6MAM-6MAM-6MAM-6MAM-6SGAEMAMucinReportedM,>250,000350,000-390,000>400,000>400,000>400,000>400,000>400,000>400,000360,000>260,000>250,000Epitope"PCPUCP3P3P3P3P3PRef.12,

16,1711,18,1920-2313131314,1514102412

°P, protein epitope; C, epitope involving carbohydrates; U, unknown epitope.

noma obtained from Dr. M. Uitendaal, State University, Maastricht.Ovcar-3 and Capan-1 cells are derived from an ovarian and a pancreaticcarcinoma, respectively, and were obtained from the American TypeCulture Collection. The A549 cell line was derived from a bronchioal-veolar carcinoma and was obtained from the American Type CultureCollection. All cells were maintained in DMEM, supplemented with10% fetal calf serum.

Human Sera and Skim Milk. Serum samples containing high levelsof MAM-6 were collected from advanced breast cancer patients (7).Human skim milk was prepared by removing the fat from fresh humanmilk by centrifugation at 40,000 x g. Protease inhibitors were added,and the preparation was stored at -20°C.

Metabolic Labeling. Subconfluent cells, grown in 25-cm2 Falcon

tissue culture flasks, were washed once in PBS and subsequently cultured in DMEM, supplemented with 5% PCS, containing 0.45 mg ofglucose per ml (i.e., 10% of the usual concentration) when ['Hjgluco-

samine labeling was performed and no threonine for labeling with[3H]threonine. After l h the cells were biosynthetically labeled for the

appropriate time by adding 100 ¿tCiof labeled compounds (Amersham;specific activity of D-[6-3H]glucosamine, 33 Ci/mmol, and of L-[3-3H]threonine, 17 Ci/mmol).

Antibody lodination. Affinity-purified antibodies were labeled with'"I using the iodogen method as described by Fraker et al. (26).

Immunoprecipitation Procedures. MAM-6 was immunoprecipitatedwith preformed complexes prepared of MAbs and RAM as describedby Hilkens and Buys (15). For sequential precipitation experiments,metabolically labeled cell lysates were precleared with preformed complexes prepared with a specific MAb. The supernatant of the specificprecipitation was incubated a second time with preformed complexesprepared from the same antibodies to remove any residual antigen.Subsequently, the supernatant depleted of molecules reactive with thefirst MAb was used for immunoprecipitation with a second MAb. Thismaterial was analyzed by SDS-PAGE.

SDS-Polyacrylamide Gel Electrophoresis and Autoradiography. Immunoprecipitated samples were dissolved in 50 ¿ilof 0.1 M Tris/HClbuffer, pH 6.8, containing 2% SDS, 6 M urea, 7% /3-mercaptoethanol,10% glycerol, and 0.15% bromophenol blue. The sample buffer contained 10 HIMiodoacetamide instead of /3-mercaptoethanol when thesamples were examined under nonreducing conditions. After boilingfor 2 min the samples were cleared by centrifugation (10,000 x g, 5min) and analyzed on 4 to 10% SDS-polyacrylamide gradient gels witha 3% stacking gel according to Laemmli (27). High-molecular-weightmarkers (BioRad) were used as reference. After electrophoresis the gelswere fixed and stained with Coomassie brilliant blue and prepared forfluorography.

Immunoblotting. Human skim milk or serum proteins were separatedby SDS-PAGE (4 to 10% gradient gels) as described above. The proteinswere blotted onto nitrocellulose filters (Schleicher and Schuell) according to Towbin et al. (28). After blotting, the nitrocellulose filters werepretreated for 30 min with 0.05% Tween 20 in PBS, covered withWhatman 3CHR filter paper, soaked in 5 ml of PBS containing 5 xIO6cpm/ml of '"¡-labeled MAb 139H2 (specific activity, 5 to 8 x IO4

cpm/ng) and 2% NMS, and incubated for 30 min at room temperature.The filters were washed 4 times in PBS containing 0.05% Tween 20,air dried, and subjected to autoradiography using Kodak XS-1 films.

Immunofluorescence. Cells were grown to confluency, detached with

PBS containing 5.10 4M d¡sodiumEDTA, and washed once in DMEMplus 10% FCS. Subsequently, cells were incubated (30 min, 0°C)in

DMEM and appropriate dilutions of the different MAbs followed byincubation with FITC-labeled goat anti-mouse IgG antibodies for 30min, 0°C.The fluorescence intensity was measured by the FACSCAN

(Becton and Dickinson), and the fluorescence index was calculated asFl - Bl—¿�——,where Fl is fluorescence intensity of a specific MAb, and Bl

Blis fluorescence intensity of a nonreactive MAb.

RESULTS

Characterization of MAM-6 from ZR-75-1 Cells. We haveimmunoprecipitated MAM-6 with MAb 139H2 from lysates ofZR-75-1 cells, labeled for 4 h with [3H]threonine. The immu-noprecipitates were analyzed by SDS-PAGE. The fluorogramrevealed two bands of approximately 450,000 and 650,000under reducing conditions (Fig. 1). Both under reducing andnonreducing conditions the same bands were observed, indicating that MAM-6 does not consist of homo- or heterodimerslinked by disulfide bonds.

As we will show below, the M, 450,000 as well as the M,650,00 mature glycoproteins were also detected in inuminoblotting experiments using MAb 139H2 and obviously carrythe same epitope, confirming that both glycoproteins werespecifically and independently precipitated.

B

kDA

-200

-ÃŽ16

- 92

- 68

- 45

Fig. 1. Fluorogram of [3H]threonine-labeled MAM-6 from ZR-75-1 cellsimmunoprecipitated by MAb 139H2 and analyzed by SDS-PAGE. 4 to 10% gel,under reducing conditions (Lane A) and nonreducing conditions (Lane B).

787

on July 7, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

COMPLEXITY OF MAM-6

Characterization of the MAM-6 Antigen from Human Carcinoma Cell Lines, Milk, and Serum. To investigate the nature ofthe two glycoproteins in human cell lines, various carcinomacell lines were labeled for 4 h with [3H]glucosamine. The cellswere lysed, and MAM-6 antigens were immunoprecipitated byM Ab 139H2 and analyzed by SDS-PAGE under reducing conditions. Fluorograms revealed a pattern of either one or twobroad bands with different mobilities for each cell line (Fig. 2).The variations in electrophoretic mobility of MAM-6 glycoproteins from breast carcinoma cell lines (Fig. 2, Lanes A to F)were similar to those from an ovarian (Lane G) and lung (LaneII) tumor cell line.

This result raised the question whether the differences inapparent molecular weight were due to cell line-dependent,posttranslational modifications in tissue culture or were alsopresent in vivo. To answer this question, human skim milk fromdifferent individuals and serum from advanced breast cancerpatients were analyzed on immunoblots. Fig. 3, A and H, showsthat also in serum and milk either one or two glycoproteinswith different electrophoretic mobilities were detected by MAb139H2, in the same molecular weight range and with similarvariations in electrophoretic mobility as those observed withMAM-6 from cultured cell lines (Fig. 2), indicating that thedifferences in apparent molecular weight were due to variationsamong individuals.

Many MAbs Are Directed against the Same Family of Sialo-mucins Which Are Generated from a Common Precursor Molecule. We have generated 5 different MAbs against the MAM-6antigen with slightly different reactivities on tissue sections(13). Cross-blocking experiments showed that these MAbs weredirected to different epitopes, although some epitopes wereoverlapping (e.g., 139H2 and 115F5) (13, 14). We have investigated whether these MAbs were directed against the samemolecules. For this purpose, lysates of ZR-75-1 and T47D cells,labeled with ['Hjglucosamine for 4 h, were immunoprecipitated

with each of the five MAbs. The immunoprecipitates were

ABCDEFGH

kDa

t l-

.200

.116Fig. 2. Fluorogram of a 4 to 10% SDS-polyacrylamide gradient gel showing

MAM-6 immunoprecipitated with MAb 139H2 from lysates of breast, lung, andovarian carcinoma cell lines labeled for 4 h with [3H]glucosamine. Breast cancercell lines: Lane A, MCF-7; Lane B, ZR-7S-1; Lane C, T47D; Lane D, MDAMB 157; Lane E, BT-20; Lane F, CAMA-1. Lane G, HI 34 ovarian carcinoma cellline; Lane H, AS49 lung carcinoma cell line.

ABCDEF - ABCDE

kDa

III »I •¿�

200r

•¿�•i.

Fig. 3. Autoradiogram of MAM-6 blotted on nitrocellulose sheets after SDS-PAGE on a 4 to 10% gradient gel. Antigen was detected with '"I-labeled MAb

139H2. Autoradiogram /4, human skim milks. Autoradiogram li. Lane A, humanskim milk; Lanes B to /;, serum from breast cancer patients.

analyzed by SDS-PAGE under reducing conditions (Fig. 4). Allfive MAbs precipitated the same set of two glycoproteins withthe same variation in molecular weight in both cell lines as hasbeen shown above for MAb 139H2. Because all antigens showedthe same variable pattern in each cell line, it is very likely thatall five MAbs reacted with the same molecules.

Subsequently, we investigated whether six additional MAbs,listed in Table 1, raised by various other investigators to mucinor mucin-like glycoproteins, were also directed to the sameepithelial sialomucin, MAM-6. For this purpose we used theseMAbs to immunoprecipitate the mucins from cell lysates ofZR-75-1 and T47D cells labeled for 90 min and 3 h, respectively, with [3H]threonine. These labeling periods should alsoreveal the early MAM-6 precursor molecules (15). All MAbs,except Cal, detected the same precursor molecules in ZR-75-1as well as T47D cells, but not all MAbs immunoprecipitatedmature molecules from these cell lines (Figs. 5 and 6). In bothcell lines MAbs 139H2 and DF3 immunoprecipitated the mature and precursor form of MAM-6. MAbs SM-3 and HMFG2precipitated the mature and precursor form of MAM-6 fromT47D cells, but only the precursor form of the molecule fromZR-75-1 cells. In contrast, MAbs E29 and OM-1 precipitatedthe precursor and mature form of MAM-6 from ZR-75-1 cellsbut only the precursor molecules from T47D cells. Even afterlonger exposure times no mature molecules could be observedin T47D cells with these MAbs.

The differences in immunoprecipitation results with the various MAbs could be explained by higher and lower affinities ofthe MAbs. However, all the MAbs were able to precipitate theprecursor forms, with the exception of MAb Cal. ConsequentlyMAbs E29, HMFG2, SM-3, and OM-1 have only a reducedaffinity for the mature glycoproteins of either cell line. Sincematuration of the MAM-6 precursor mainly occurs by glyco-sylation, these results strongly suggest that the reduced affinityis due to different terminal glycosylation reactions in both celllines. In conclusion, all MAbs, except Cal, detect moleculeswith the same protein backbone, but the mature sialomucinmolecules are differentially glycosylated in the two cell linesinvestigated, by a process that affects the antigenic profile ofMAM-6.

MAb Cal was unable to immunoprecipitate any materialfrom either cell line, but was able to precipitate the matureform of MAM-6 from BT-20 cells (Fig. 7). No precursor

788

on July 7, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

COMPLEXITY OF MAM-6

Fig. 4. Fluorogram of a 4 to 10% SDS-polyacrylamide gradient gel showing MAM-6immunoprecipitated from cell lysates of ZR-75-1 (fluorogram A) and T47D cells (fluoro-gram B) labeled 4 h with ['HJglucosamine.Lane A, NMS control; Lane B, MAM-6 precipitated with M Ab 115D8; Lane C, MAb115F5; Lane D, MAb 115G2; Lane E, MAb139H2; Lane F, MAb 140C1.

I

AB C D E FA B C D E F

k Dak Da

200- 200-

ne

<v

// CÕff /kDa

kDa

. 200

. 116

200 .

116

Fig. 5. Immunoprecipitation of the epithelial sialomucin with various MAbsfrom ZR-75-1 breast cancer cells. Cells were labeled with [3H]threonine for 90min, and MAM-6 was immunoprecipitated with various MAbs and analyzed on from T47D breast cancer cells. Cells were labeled with [3H)threonine for 3 h, and4 to 10% SDS-polyacrylamide gradient gels. *, mature forms; arrowheads, MAM-6 was immunoprecipitated with various MAbs and analyzed on 4 to 10%

Fig. 6. Immunoprecipitation of the epithelial sialomucin with various MAbs

precursor forms. SDS-polyacrylamide gradient gels. *, mature forms; arrowheads, precursor forms.

molecules could be observed with this MAb in any of the cell from each cell line a different subset of the mature moleculeslines tested. This result indicates that the Cal epitope is (a)only present in a restricted number of cell lines, and (b) mostcritically dependent on terminal glycosylation.

MAM-6 Mucins from ZR-75-1 Cells Are Composed of Subclasses Expressing Different Epitopes. By comparing the relative intensities of the precursor and mature forms of the mucinprecipitated by the various MAbs it appeared that within eachcell line the ratios between them were different, suggesting that

was precipitated. This result suggests that each epitope couldbe masked to a different degree depending on the cell line andpossibly its histological origin.

To investigate further whether certain MAbs were directedto different populations of molecules within one cell line,MAM-6 was immunoprecipitated with MAb 139H2 from celllysates of ZR-75-1 cells, labeled overnight with [3H]threonine.

Subsequently, the lysates were precleared twice with MAb789

on July 7, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

COMPLEXITY OF MAM-6

/// sA B C D E F G

kDa

.200

Fig. 7. Immunoprecipitation of the epithelial sialomucins from the BI -20breast carcinoma cell line. Cells were labeled with [3H]threonine for 3 h, and theantigen was precipitated with various MAbs. *. mature forms; arrowheads,precursor forms.

139H2 and used for immunoprecipitation with MAbs 11SD8,115F5,115G2,140C1, and in addition MAb 139H2 as control.The third precipitation with MAb 139H2 was analyzed by SDS-PAGE and revealed that the Mr 450,000 and 650,000 matureMAM-6 molecules, reactive with MAb 139H2, were clearedcompletely from the lysate. None of the other MAbs was ableto precipitate additional MAM-6 molecules from the lysatescleared with MAb 139H2 (not shown). Also MAb 115D8 and140C1 cleared all MAM-6 molecules detected with this set ofMAbs, indicating that MAbs 139H2, 115D8, and 140C1 reactwith the same population of molecules in ZR-75-1 cells.

When the ZR-75-1 cell lysate was cleared twice with MAb115F5, subsequent immunoprecipitation of additional MAM-6 molecules was observed with MAbs 115D8, 139H2, and140C1, but not with 115G2 (Fig. 8).

The same sequential immunoprecipitation experiments wererepeated with lysates of ZR-75-1 cells, pulse-labeled with[3H]threonine for 90 min, using the same MAbs, except 115G2.The 90-min pulse labeling was intended to show the M, 200,000and 350,000 early and the M, 500,000 and 700,000 late precursors. The same results were obtained with the M, 500,000and 700,000 precursors as with the mature MAM-6 moleculesin the previous experiment. However, the early precursor molecules were completely removed by the clearance with MAb

tKd

-330

-200

-116

-92

Fig. 8. Sequential precipitation of MAM-6 with various MAbs. Cell lysates of[}H]threonine-labeled ZR-7S-1 cells, overnight labeling, were precleared with

11SFS and subsequently immunoprecipitated with MAbs 11SF5 (Lane B), 115D8(LaneC), 115G2 (Lane D), 139H2 (Lane E), 140C1 (Lane F), and NMS (LaneG). Lane A shows MAM-6 precleared with MAb 115F5.

115F5, and no further early precursors could be precipitatedwith MAb 139H2 (Fig. 9) and MAbs 115D8 and 140C1 (notshown). These experiments show that all MAM-6 early precursor molecules contained the 115F5 epitope, which was notdetectable on a subpopulation of mature MAM-6 molecules.The 115F5 epitope is present on the protein backbone of MAM-6 because the MAb binds to the MAM-6-ß-galactosidase fusionprotein expressed in E. coli.3 The M, 500,000 and 700,000 late

precursors and the mature molecules are extensively glycosy-lated in contrast to the early precursors (15). Therefore, it islikely that the 115F5 epitope is masked by carbohydrates on asubset of the M, 500,000 and 700,000 precursors and themature molecules. An alternative explanation could be thatMAb 115F5 precipitated MAM-6 incompletely due to a loweraffinity. This seems unlikely since MAb 115F5 cleared theprecursor completely and has approximately the same affinityas 139H2.4 These results support the notion that there is a

family of sialomucins which have a differentially glycosylatedprotein backbone.

4S. Zotter, P. C. Hageman, A. Lossnitzer, J. van den Tweel, J. Hilkens, W.

J. Mooi, and J. Hilgers. Monoclonal antibodies to epithelial sialomucins recognize epitopes at different cellular sites in adenolymphomas of the parotid gland,manuscript submitted for publication.

790

on July 7, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

COMPLEXITY OF MAM-6

kDa

-200

B DFig. 9. Sequential precipitation of pulse-labeled MAM-6 by 115F5 and 139H2.

Cell lysates of [3H]threonine-labeled ZR-75-1 cells, 90-min labeling, were pre-cleared with MAh 115F5. Lane A shows MAM-6 precleared with 115F5. Subsequently, MAM-6 was immunoprecipitated from the precleared lysates with MAb115F5 (Lane B) and 139H2 (Lane C). Lane D shows a longer exposure of LaneC to confirm that all precursor molecules were cleared by M Ah I151-5.

Table 2 Reactivity ofMAbs against epithelial sialomucins with various cell linesexpressed as relative fluorescence index

CelllinesMAbs139H2

DF-3HMFG-2E29OM-1CalSM-3ZR-75-174

3213199

1110T-

47D23

199

130.669BT-2025

1511103

1512H1348

44151.8

73Ov-car-39

82

60.47

2Capan-

129

229

230.8

119

Detection of Cell Surface-associated Sialomucins by Immu-nofluorescence-activated Cell Sorter Analysis. The presence ofcell surface-associated sialomucins was also determined by indirect immunofluorescence and analyzed by fluorescence-activated cell sorting. The results are expressed as fluorescenceindex and are presented in Table 2. The MAbs precipitatingthe mature molecules strongly reacted in the immunofluorescence assay. The MAbs which showed a low reactivity (fluorescence index, -10 or less) did not precipitate the mature molecule, confirming the immunoprecipitation results describedabove. Because the expression of the epithelial sialomucin genemay be variable in each cell line, only the reactivity of the

various MAbs within one cell line can be compared. For example, the reactivity of the MAbs with the epithelial sialomucinin Ovcar-3 cells was very low. This was not due to the maskingeffect but merely to a low expression of the MAM-6 gene, sinceonly low amounts of the precursor glycoprotein could be precipitated (not shown).

DISCUSSION

There are several reports concerning MAbs directed againsthigh-molecular-weight mucins present on carcinoma cells (Table 1). It is very difficult to compare these mucins by theirmolecular weight as determined in SDS-PAGE, since mostauthors used different cell lines or tumors and electrophoreticsystems to characterize their antigens. Therefore, we analyzedthe antigens from a few well-defined cell lines in a side-by-sidecomparison by immunoprecipitation with MAbs raised by usand by other investigators.

The immunoprecipitation and immunoblotting experimentsreported here show that MAM-6 is composed of either one ortwo high-molecular-weight glycoproteins with different mobilities on SDS-polyacrylamide gels. Usually, two glycoproteinswere precipitated, and both carried common epitopes, were verysimilarly processed, and were not linked by disulfide bonds,suggesting that they are distinct yet closely related molecules.The differences in apparent molecular weight of MAM-6 molecules within one cell line or among cell lines and specimensfrom different individuals may be due to differences in carbohydrate composition or differences in the protein backbone. Wehave shown that the early precursors of MAM-6 (without N-linked glycans) had a relatively similar difference in apparentmolecular weights as the mature MAM-6 molecules (15). Although O-linked glycans may already be present on the earlyprecursors of certain molecules (29-31), only a few, if any, arepresent on the early precursors of these sialomucins (32). Therefore, early O-linked glycosylation is unlikely to account for suchlarge differences in molecular weight (150,000 and over). Thedifference in mobility between both molecules must thereforebe due to variations in the protein backbone rather than in thecarbohydrates.

Here we have shown that the protein backbone of MAM-6and HMFG mucins are identical. Swallow et al. (33) demonstrated with a cDNA probe (34) that the HMFG gene isidentical to the polymorphic PUM gene (35). Therefore, it islikely that the variations in apparent molecular weight of thesialomucins, as observed here, are due to genetic polymorphism.The two bands, observed in specimens from most individuals,represent two codominantly expressed alíelesas has been shownby Karlsson et al. (35) and Hayes et al. (23). Additional variations in apparent molecular weight of the mature molecule dueto variable glycosylation cannot be excluded.

It is plausible to assume that the differences in mobility ofthe sialomucins in cell lines are also caused by genetic polymorphism. We have taken advantage of this polymorphism toinvestigate the relationship between the antigens precipitatedby various MAbs against mucins. All MAbs listed in Table 1,with the exception of MAbs Cal and 115G2 (the results with115G2 are not shown), immunoprecipitated the same allelicforms of the precursor molecules in ZR-75-1 and T47D cells.Since all MAbs (except Cal and 115G2) are directed to proteinepitopes (Table 1) differences in reactivity with the maturemucin must be due to posttranslational modifications leadingto masking of epitopes or conformational changes in the proteinbackbone. These modifications of the epitopes could occur by

791

on July 7, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

COMPLEXITY OF MAM-6

the use of different attachment sites of the O-linked glycans orby alterations in length or branching of the glycans. Our resultsindicate that these MAbs are reactive with a family of sialo-mucins derived from the same precursor molecules but withdifferent fully processed molecules in each cell line.

The sequential precipitation experiments with 115F5, andthe different ratios of mature and precursor molecules precipitated with several of the other MAb studies, show that alsowithin one cell line there are differentially glycosylated sialo-mucins which have a common precursor.

A Cal-reactive glycoprotein was only precipitated from BT-20 cells. Although this glycoprotein had the same electropho-retic mobility as the mature sialomucin in this cell line, this isnot conclusive evidence that this glycoprotein belongs to thesame family of sialomucins. However, Griffiths et al. (36)proved that the Cal antigen and MAM-6 are identical in urineby comparing the antigen profiles from different individuals bySDS-PAGE. Cal did not react with the precursor of the epi

thelial sialomucins, indicating that it is also directed to anepitope depending on terminal glycosylation. This is in accordance with studies by Ashall et al. (18) who have shown that theCal epitope was sensitive to neuraminidase. Another MAb,115G2, did not react with the precursor molecules (not shown),indicating that also in this epitope carbohydrates are involved.

Differences in reactivity on tissue sections were observed withvarious MAbs directed to epithelial sialomucins. For example,MAb 115F5 reacts only with 10% of the colon carcinomas,while 115D8 reacts with almost all colon carcinomas. De-Krester et al. (10) reported that OM-1 reacted preferentiallywith ovarian carcinomas. The mature form of the OM-1 mneinwas indeed only weakly detectable in one of the breast carcinoma cell lines. However, MAb OM-1 reactive precursor molecules but no mature sialomucins were immunoprecipitatedfrom two ovarian carcinoma cell lines, Ovcar-3 and H134 (datanot shown). Therefore, we could not confirm the tissue preference of this MAb in vitro. Nevertheless, differential glycosylation of MAM-6 may occur in vivo in a tissue-specific mannerand may explain the difference in tissue preference of thevarious MAbs directed to sialomucins. This has been moststrikingly shown by Zotter et al.4 using, in part, the same group

of MAbs. They showed that these MAbs could be rankedaccording to at least two different types of reactivities in aden-olymphomas of the parotid gland.

Tissue-specific determinants on mucins have been reportedearlier by Gold et al. (37). They showed that pancreatic mucinspossess specific determinants that are unique to the pancreas.Also blood group determinants which often occur on mucinsare expressed in a tissue-specific manner (2).

In view of the differential glycosylation of the sialomucins itis not surprising that there are also MAbs with a strong preference for carcinomas relative to the normal epithelium as hasbeen reported by Burchell et al. (12) for MAb SM-3. It isconceivable that the tumor specificity of the SM-3 epitope isgenerated in the same way as the epitopes with tissue preferencediscussed above.

Differences in reactivity of some of the MAbs as studied herewith various cell lines have also been reported by Griffith et al.(38) and by Abe and Kufe (39). The latter authors suggestedthat MAbs DF3, 115D8, and Cal belong to a single family ofglycoproteins based on competition studies. Here we have extended these studies and show that: (a) several more MAbsdefine new members of this family; (b) each member is differentially expressed on different cell lines; (c) this is due todifferential glycosylation of the same protein backbone; and (d)

differences in molecular weight of the mucins in various celllines are due to genetic polymorphism. We suggest referring tothis complex family of epithelial sialomucins as episialins.

ACKNOWLEDGMENTS

We thank Dr. J. Minke, Dr. R. Michalides, and Dr. H. Ploegh fortheir advice and critical reading of the manuscript; Dr. J. Taylor-Papadimitriou and Dr. J. Burchell for providing us with MAbs HMFG-2 and SM-3; Dr. M. Bramwell for MAb Cal; and Dr. T. Dekretser forMAb OM-1.

REFERENCES

1. Hakomori, S. Tumor associated carbohydrate antigens. Annu. Rev. Intmu-nol., 2: 103-126, 1984.

2. Feizi, F. Demonstration by monoclonal antibodies that carbohydrate structures of glycoproteins and glycolipids are oncodevelopmental antigens. Nature (Lond.), 314: 53-57, 1985.

3. Zotter, S., Lossnitzer, A., Hageman, Ph., Delemarre, J. F. M., Hilkens, J.,and Hilgers, J. Immunohistochemical localization of the epithelial markerMAM-6 in invasive malignancies and highly displastic adenomas of the largeintestine. Lab. Invest., 57: 193-199, 1987.

4. Heyderman, E., Strudley, I., Powell, G., Richardson, T. C., Cordell, J. L.,and Mason, D. Y. A new monoclonal antibody to epithelial membraneantigen (EMA)-E29. A comparison of its immunocytochemical reactivitywith polyclonal anti-EMA antibodies and with another monoclonal antibody.HMFG-2. Br. J. Cancer, 52: 355-361, 1985.

5. Hageman, Ph. C., Peterse, J. L., Hilkens, J., and Hilgers, J. Monoclonalantibodies in breast tumor pathology. In: D. Ruyter, S. Warnaar, and G.Fleuren (eds.), Application of Monoclonal Antibodies in Tumor Pathology,pp. 167-190. Dordrecht: Martinus Nijhoff, 1987.

6. 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.

7. Hilkens, J., Kroezen, V., Bonfrer, J. M. G., De Jong-Bakker, M., andBruning, P. F. MAM-6, a new serum marker for breast cancer monitoring.Cancer Res., 46: 2582-2587, 1986.

8. Hilkens, J., Bonfrer, J. M. G., Kroezen, V., Van Eykeren, M., Nooyen, W.,De Jong-Bakker, M., and Bruning, P. F. Comparisons of circulating MAM-6 and CEA levels and correlation with the estrogen receptor in patients withbreast cancer. Int. J. Cancer, 39:431-435, 1987.

9. Hayes, D. F., Zurawski, V. R., and Kufe, D. W. Comparison of circulatingCal5.3 and CEA levels in patients with breast cancer. J. Clin. Oncol., 4:1542-1550, 1986.

10. DeKretser, T. A., Thorne, H. J., Jacobs, D. J., and Jose, D. G. The sebaceousgland antigen defined by the OM-1 monoclonal antibody is expressed at highdensity on the surface of ovarian carcinoma cells. Eur. J. Cancer Clin. Oncol.,21:1019-1035, 1985.

11. McGee, J. O. D., Woods, J. C, Ashall, F., and Bramwell, M. E. A newmarker for human cancer cells. 2. Immunohistochemical detection of the Caantigen in human tissues with the Cal antibody. Lancet, 2: 7-10, 1982.

12. Burchell, J., Gendler, S., Taylor-Papadimitriou, J., Girling, A., Lewis, A.,Millis, R., and Lamport, D. Development and characterization of breastcancer reactive monoclonal antibodies directed to the core protein of thehuman milk mucin. Cancer Res., 47: 5476-5482,1987.

13. Hilkens, J., Buijs, F., Hilgers, J., Hageman, Ph., Calafat, J., Sonnenberg, A.,and Van der Valk, M. Monoclonal antibodies against human milk-fat globulemembranes detecting differentiation antigens of the mammary gland and itstumors. Int. J. Cancer, 34: 197-206, 1984.

14. Hilkens, J., Kroezen, V., Buijs, F., Hilgers, J., Van der Vliet, M., De Voogd,W., Bonfrer, J., and Bruning, P. F. MAM-6, a carcinoma associated marker:preliminary characterization and detection in sera of breast cancer patients.In: R. L. Ceriani (ed.). Monoclonal Antibodies and Breast Cancer, pp. 28-42. Boston, Martinus Nijhoff, 1985.

15. Hilkens, J., and Buijs, F. Biosynthesis of MAM-6, an epithelial sialomucin;evidence for a rare proteolytic cleavage step in the endoplasmic reticulum. J.Biol. Chem, 263:4215-4222, 1988.

16. Taylor-Papadimitriou, J., Peterson, J. A., Arklie, J., Burchell, J., Ceriani, R.L., and Bodmer, W. F. Monoclonal antibodies to epithelium specific components of the human milkfat globule membrane: production and reactionwith cells in culture. Int. J. Cancer, 28: 17-21, 1981.

17. Burchell, J., Durbin, H., and Taylor-Papadimitriou, J. Complexity of expression of antigenic determinants, recognized by monoclonal antibodies HMFG-1 and HMFG-2 in normal and malignant human mammary epithelial cells.J. Immunol., 131: 508-513, 1983.

18. Ashall, F., Bramwell, M. E., and Harris, H. A new marker for human cancercells. I. The Ca antigen and the CAI antibody. Lancet, 2: 1-6, 1982.

19. Bramwell, M. E., Bhavanandan, V. P., Wiseman, G., and Harris, H. Structureand function of the CA antigen. Br. J. Cancer, 48: 177-183, 1983.

20. Kufe, D., Inghirami, G., Abe, M., Hayes, D., Justi-Wheeler, H., and Schlom,

792

on July 7, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

COMPLEXITY OF MAM-6

J. Differential reactivity of a novel monoclonal antibody (DF3) with humanmalignant versus benign breast tumors. Hybridoma, 3: 223, 1984.

21. Sekine, H.. Ohno, T., and Kufe, D. W. Purification and characterization ofhigh molecular weight glycoproteins detectable in human milk and breastcarcinomas. J. Immunol., 135: 3610-3615, 1985.

22. Siddiqui, J., Abe, M Hayes, 1)., Shani, £.,Yunis, £.,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.

23. Hayes, D. F., Sekine, H., Marcus, D., Alper, C. A., and Kufe, D. W.Genetically determined polymorphism of the circulating human breast cancerassociated DF3 antigen. Blood, 71:436-440, 1988.

24. Cordell, J., Richardson, T. C, Pulford, K. A. F., Ghosh, A. K., Gatter, K.C., Heyderman, E„and Mason, D. Y. Production of monoclonal antibodiesagainst human epithelial membrane antigen for use in diagnostic immuno-cytochemistry. Br. J. Cancer, 52: 347-354, 1985.

25. Engel, L. W., and Young, N. A. Human breast carcinoma cells in continuousculture: a review. Cancer Res., 38:4327-4339, 1978.

26. Fraker, P. J., and Speck, J. C. Protein cell membrane iodinations withsparingly soluble chloroamide, l,3,4,6-tetrachloro-3</,6a-diphenyl-glycoluril.Biochem. Biophys. Res. Commun., 80: 849-857, 1978.

27. Lacrimili, U. K. Cleavage of structural proteins during the assembly of thehead of bacteriophage T4. Nature (Lond.), 227: 680-685, 1970.

28. Towbin, H., Staehelin, T., and Gordon, J. Electrophoretic transfer of proteinsfrom polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Nati. Acad. Sci. USA, 76:4350-4354, 1979.

29. Smms. G. J. A. M. Initial glycosylation of proteins with acetylgalactosami-nylserine linkages. Proc. Nati. Acad. Sci. USA, 76:2694-2698, 1979.

30. Jokinen, M., Anderson, L. C., and Gahmberg, C. G. Biosynthesis of themajor human red cell sialoglycoprotein, glycophorin A. J. Biol. Chem., 260:11314-11321,1985.

31. Patzelt, C., and Weber, B. Early O-glycosidic glycosylation of proglucagonin pancreatic islets; an unusual type of prohormonal modification. EMBOJ., 5:2103-2108, 1986.

32. Linsley, P. S., Kallestad, J. C., and Horn, D. Biosynthesis of high molecularweight breast carcinoma associated mucin glycoproteins. J. Biol. Chem., 26.?:8390-8397,1988.

33. Swallow, D., Gendler, S., Griffiths, B., Corney, G., Taylor-Papadimitriou,J., and Bramwell, M. The tumor-associated epithelial mucins are coded byan expressed hypervariable gene locus PUM. Nature (Lond.), 328: 82-84,1987.

34. Gendler, S. J., Burchell, 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.

35. Karlsson, S., Swallow, D. M., Griffiths, B., Corney, G., Hopkinson, D. A.,Dawnay, A., and Cartron, J. P. A genetic polymorphism of a human urinarymucin. Ann. Hum. Genet., 47:263-269, 1983.

36. Griffiths, B. A., Gordon, A., Burchell, J., Brauwell, M., Griffiths, A., Price,M., Taylor-Papadimitriou, J., Zanin, I)., and Swallow, D. The tumor-associated epithelial mucins and the peanut lectin binding urinary mucinsappear to be coded by a single highly polymorphic gene locus "TIM." Dis.Markers, 6:185-194, 1988.

37. Gold, D. V., Hollingsworth, P., Kremer, T., and Nelson, D. Identification ofa human pancreatic duct tissue-specific antigen. Cancer Res., 43: 235-238,1983.

38. Griffiths, A. B., Burchell, J. Gendler, S., Lewis, A., Blight, K., Tilly, R., andTaylor-Papadimitriou, J. Immunologia) analysis of mucin molecules expressed by normal and malignant mammary epithelial cells. Int. J. Cancer,Â¥0:319-327,1987.

39. Abe, M., and Kufe, D. W. Identification of a family of high molecular weighttumor-associated glycoproteins. J. Immunol., 139: 257-261, 1987.

793

on July 7, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1989;49:786-793. Cancer Res   John Hilkens, Femke Buijs and Marjolijn Ligtenberg  CarcinomasComplexity of MAM-6, an Epithelial Sialomucin Associated with

  Updated version

  http://cancerres.aacrjournals.org/content/49/4/786

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/49/4/786To request permission to re-use all or part of this article, use this link

on July 7, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from