HYBRIDOMA Volume 10, Number 2, 1991 Mary Ann Liebert, Inc., Publishers

Two Monoclonal Antibodies

Preferentially Expressed

Identify Antigens

Normal Human Breast Cells Versus Breast Cancer Cells on

G. PANCINO,1 M.H. MORTADA,1 C. CHARPIN,2 E. OSINAGA,1 P. DE CREMOUX,3 B. BETAILLE,' M.G. GOBERT,1 F. CALVO,3 and A. ROSETO1 '

(Division d'lmmunocytologie Appliquée) Universitéde Compiègne, BP. 649, 60206 Compiègne Cedex France d'Anatomopathologie Faculté" de Médecine, Boulevard Jean Moulin 13385 Marseille- France * Pharmacologie Expérimentale, INSERM U204, Institut d'Hématologie, Centre Hayem, Hôpital Saint Louis, 75475 Paris Cedex 10

DICA

2Laboratoire

ABSTRACT In order to obtain antibodies with specificity toward normal mammary epithelial antigenic determinants, we immunized BALB/c mice with normal milk cells and screened the hybridomas against an undifferentiated breast cancer cell line H466B, peripheral blood lymphocytes and normal fibroblasts. Two hybridomas were generated, which produced BA, (IgG..) and CA. (IgM) monoclonal antibodies (MAbs).

These MAbs did not react with 5 breast cancer cell lines. In cryostat sections of normal human breast tissue, BA. was reactive with 6/6 and CA. was reactive with 12/13 specimens both showing an apical staining of epithelial cells. Conversely staining of malignant cells in breast cancer biopsies was observed in 4/33 specimens with BA, and in 4/19 specimens with CA.. Computerized image analysis (SAMBA) of immunostained sections showed homogeneous distribution of staining, with a high percentage of stained cell surfaces in normal breast (mean percentages of positive surfaces : BA, : 75 % and CA. : 82 %) while, in malignant samples, staining was heterogeneous, with a mean percentage of positive surface of 25 % for BA, and 12 % for CA.. Both MAbs reacted strongly with human milk fat globule membranes (HMFGM) and skimmed milk. FPLC size exclusion chromatography of skimmed milk showed that CA. and BA, reactive materials eluted in distinct peaks in high molecular weight ranges. Eaectrophoretic separation of HMFGM followed by CA. staining detected a high molecular weight reactive band (Mr 380-600 kDa). CA. ana BA, reactivity was reduced by protease treatment of the antigen but was not affected by neuraminidase digestion, by methanol extraction or by Na-metaperiodate oxidation. After perchloric acid treatment of HMFGM, BA, activity was lost while the CA. activity was found in the soluble fraction. The results reported suggest that the two MAbs identify two distinct novel epitopes of normal breast cells.

INTRODUCTION The identification of molecules preferentially expressed either in normal human breast cells or cancer cells is an approach to the study of differentiation of the mammary gland and may provide new diagnostic and prognostic tools. A heterogeneous expression of antigenic determinants in normal epithelial cells of mammary gland was first revealed by polyclonal antibodies prepared against human milk fat globules (HMFG) (1) and later by monoclonal antibodies (2-4). Such heterogeneity was also demonstrated in human breast cancer cell lines (5,6). Surface antigen expression was shown to be more variable in breast cancer tissue than in normal tissues (7,8) and was observed in other epithelial tissues (review

9).

241

Using MAbs variations in intensity of expression or in cellular localization of antigens (10-14), as well as modifications in surface antigen glycosylation (15) or excessive sialylation and antigenic masking (16) in breast cancer cells ,

some

have been showed. In order to

identify new molecules specifically or predominantly expressed in normal breast, we produced monoclonal antibodies using a strategy of immunization with the differentiated normal milk cells, and screened versus a phenotypically undifferentiated breast cancer cell line, H466B (17). In this paper we report the generation of two MAbs which react with antigens preferentially expressed on normal breast epithelial cells. MATERIALS AND METHODS

Milk cells and HMFGM

preparation

Freshly recovered human milk from healthy lactating mothers was centrifuged for 15 min at 2000 g. After removing cream and skimmed milk, the cell pellet was washed three times in PBS with 1 % BSA. Cells were tested for viability by trypan blue exclusion ( 90 % living cells) and counted. May-Grenwald-Giemsa coloration was performed to define non-epithelial cell contamination. HMFGM isolation from milk cream was performed as previously described (19,20) with minor modifications. Briefly, the cream was washed three times in 0.9 % NaCl at 4°C and resuspended to 33 % (v/v) in the same medium. After one freeze-thaw cycle, membranes were isolated by centrifugation at 100,000 xg for 1h. The crude membrane pellet was resuspended in 10 mM Tris (pH 7.6) 1mM phenyl methylsulfonyl fluoride (PMSF), 0.5 % NP.Q. Skimmed milk was centrifuged at 100,000 xg for 1h before utilization. Cell lines The H466B cell line was derived from a poorly differentiated human breast adenocarcinoma and lacks estrogen receptors (17). The other human breast cancer cell lines used in this work were ZR-75-1, T47D, MDA-MB-231 (18). The cell line was derived from human fetal lung fibroblasts.

MRC5

MCF?,

Human

peripheral

blood

lymphocytes (PBL) preparation

Mononuclear cells

Lymphoprep (Eurobio) Immmunization and

were

separated by gradient density centrifugation

on

and washed in PBS with 1 % BSA.

hybridoma production

Four-week old BALB/c mice were immunized by three intraperitoneal injections x 10 human milk cells at four-week intervals. Four days after the last injection, spleens were removed for cell fusion. The fusion of immune splenocytes and SP2/0 myeloma cells was performed using the method of Köhler and Milstein (21) with some modifications (22). of 5

Screening methodology The hybridoma supernatants were assayed using ELISA on glutaraldehyde-fixed cells, from human milk, H466B, PBL and MRCj-. These assays allowed selection of the hybridomas that were specifically reactive with membrane antigens of milk cells.

To eliminate antibodies that would react to common cytoplasmic antigens, indirect immunofluorescence on acetone-methanol-fixed cells was subsequently performed. Hybridomas reactive only with milk cells were then cloned twice by limiting dilution and expanded as ascitic tumors in pristane-primed BALB/c mice. Isotyping of monoclonal antibodies was performed by Ouchterlony gel

242

immunodiffusion

isotypes.

using

rabbit antisera (MILES)

directed

against specific

murine

ELISA cancer cell lines and MRC- were plated in 96-well culture dishes at a density of 6 x 10 cells per well and allowed to grow for 24 h. Milk cells and PBL were plated at 10 cells per well in 96-well dishes pretreated with 50 ul of poly-L-lysine (SIGMA) at 50 ug/ml, according to manufacturer's instructions. Cells were then incubated overnight at 4°C. The plates with cell monolayers or poly-L-lysine bound cells were then slightly fixed with 0.25 % glutheraldehyde in PBS for 5 min at 4°C. Non-specific sites were blocked with 1 % gelatin in PBS for 1h at 37°C and ELISA was processed using the two-step

Breast

(Nunc)

avidin-biotin-peroxidase method (Vectastain previously described (13).

ABC

Kit,

Vector

Laboratories)

as

ELISA with HMFGM extract (from 60 ug/ml to 1 ug/ml), skimmed milk (75 ug/ml), lactalbumin and lactoferrin (Sigma) (from 10 to 0.1 ug/ml) and human purified IgG and IgA (10 ug/ml) were performed, drying 50 ug/well of protein in 0.1 M carbonate buffer pH 9.6 in microtiter plates (NUNC). They were subsequently processed for ELISA, as with plated cells. Peroxidase coupled rabbit anti human Ig (H+L) were used as positive controls for assays with human immunoglobulins. In each assay, irrelevant mouse MAbs (IgG- or IgM) were used as negative controls.

Enzymatic digestion

and chemical treatment of HMFGM and

cancer

cells

coated with HMFGM extracts (60 ug/ml) in 0.1M carbonate buffer H466B and T470 monolayers were exposed to enzymatic treatment with 100 ul/well of trypsin (Flow laboratories) (2.5 mg/ml), subtilisin (SIGMA Chemicals) (1 mg/ml), protease type VI (SIGMA) (2 mg/ml), in TRIS 10 mM pH 8,2 mM Vibrio cholera neuraminidase (Boehring) 50 U/ml in acetate buffer 50 mM CaClp, and 1mM CaCl.. pH Incubation was carried out at 37°C for 1h. Methanol lipid extraction was performed for 30 min at 4°C. Control plates were incubated with dilution buffers under the same conditions. After treatment, plates were washed and blocked with 1 % gelatin in PBS ; 100 ul of MAbs were added to the wells and incubated at room temperature for 1h. The remaining steps of the assay were the same as for ELISA. Na-metaperiodate oxidation (23) was performed on microtiter wells coated with skimmed milk. After rinsing with 50 mM sodium acetate buffer, pH 4.5, NalO. (Merck) was added at concentrations ranging from 0.1 to 20 mM in the same buffer and incubated for 1h in the dark. After washing with sodium acetate buffer and further incubation with 1% glycine in PBS for 30 min, reactivity with BA, and CA. MAbs was determined by ELISA. HMFGM lysates were subjected to perchloric acid (PCA) precipitation at 4°C (0.6M final concentration) PCA soluble fraction was neutralized with 1.2 M K0H, dialyzed against PBS and concentrated. The precipitated pellet was dissolved with 0.3 M NaOH. Reactivity of PCA soluble supernatant and of precipitated fraction with MAbs was tested by ELISA.

Microplates (pH 9.6) or with

5Í5,

Molecular

weight determination

of CA. and BA, reactive

antigens

Skimmed milk was subjected to FPLC gel filtration chromatography using a Superóse 6 HR column (Pharmacia). Elution was performed with 0.05 M phosphate buffer pH 7, at a flow rate of 0.4 ml/min. The reactivity of the eluted fractions with CA. and BA, was determined by ELISA. Analysis of HMFGM extracts was performed by direct immunoreaction on gel after SDS-PAGE separation on 3-15 % polyacrylamide gels as described (24). Briefly after fixation in 10 % acetic acid 25 % isopropyl alcohol, gels were neutralized in 25 mM Tris-HCl pH 7.2, 0.13 M NaCl (TBS) and incubated overnight with 5 % BSA in TBS. Gels were then incubated with MAbs CA. or BA, for 4h, and washed

243

extensively with (Biosys) reaction

TBS. After 2h incubation with peroxidase, the goat anti-mouse was revealed with 4-chloro-1-naphtol.

Immunofluorescence

staining

Indirect immunofluorescence was done on live or fixed cells. For live cell assays, 10 cells in 100 ul PBS, 1 % BSA and 0.1 % sodium azide were incubated with 100 ul culture supernatants for 60 min at 4°C, washed and resuspended in 200 ul of FITC-labelled sheep anti-mouse immunoglobulin (Biosys). After 60 min. incubation at 4°C, cells were washed and mounted in glycerol buffer on glass slides and examined under a Zeiss epifluorescence microscope. Alternatively, cells grown on slides or cytocentrifuged cells were fixed in methanol acetone (2:1) for 10 min. at 4°C. Fixed slides were processed for 11F as above. Negative controls (irrelevant mouse monoclonal antibodies) were run simultaneously with each experiment.

Immunoperoxidase studies Five urn sections from frozen blocks of normal or pathological breast tissues stained using an avidin-biotin-peroxidase kits (Vectastain ABC Kit, Vector Laboratories) as previously described (25). Immunoperoxidase staining was revealed with 3-amino-9-ethylcarbazole (SIGMA). Slides were counter stained with Mayer's hemalun and mounted in Glycergel (DAK0). Multiparametric analysis of immunostained surfaces was assessed with a computerized system of image analysis referred to as SAMBA (T1TN, Grenoble, France) (26). Application programs have been specifically developed for tissue sections (27). were

RESULTS

Generation and

screening

of

hybridomas.

Splenocytes of immunized BALB/c mice were fused with SP2/0 myeloma cells and plated on 96-well microplates at approximately 4 x 10 cells per well. Hybridoma growth was observed in 90 % wells. Seven out of the 200 hybridomas supernatants

reactive with human milk cells using ELISA did not bind either to H466B cells or PBL and MRC¡. fibroblasts. I1F assays were then performed with methanol-acetone-fixed cells, and three hybridomas whose supernatants were reactive exclusively with milk epithelial cells were cloned. The secreted MAbs were BA, of the IgG- isotype, IgM CA. and DC.. to

Immunocytochemical reaction

with breast

cancer

cell lines.

Culture supernatants of the three hybridomas were tested for reactivity with the breast cancer cell lines MCF_, T47D and ZR-75-1, which have steroid receptors and MDA-MB-231, which lacks estrogen and progesterone receptors. ELISA and 1IF on methanol-acetone fixed cells failed to show binding of BA, and CA., while DC. was T47D and ZR-75-1 cell lines. BA, ana CA. antibodies weakly reactive with were thus further studied and characterized.

MCF?,

Histochemical sections :

reactivity

The distribution of

of MAbs with normal and

positive

pathological

frozen breast tissue

immune reactions is shown in Tables 1 and 2.

BA, : Positive staining was observed mainly in normal breast (6/6) and non-malignant disorder (7/9) : staining involved the apical part of the cytoplasmic membrane (Fig 1a, b). Only a few malignant tumours were BA, positive (4/34) (Fig 1c, d). The immunoprecipitate was diffusely spread throughout the -

244

Table 1 DISTRIBUTION OF POSITIVE IMMUNOREACTIONS WITH BA. AND CA. IN NORMAL BREAST NON-MALIGNANT AND MALIGNANT BREAST DISORDERS -

Normal breast Adenosis Fibroadenoma

Phyllod

tumor

(Gr.lT

Severe epitheliosis Intraductal carcinoma Invasive adenocarcinoma Axillary lymph node métastases Cerebral métastases

BA,

CA,

"676" T7T 17T IJT T7T "072"

12/13

WT T7T

~27T

"57ÏT 374"

T/T

"Ü7T

3/26

2/15 NT

CELLULAR DISTRIBUTION OF POSITIVE IMMUNOREACTIONS

TABLE 2 -

BA, Cellular distribution

Intraglandular deposit

Cell apex

Cytoplasmic Cytoplasm

benign

optical density =

mean

r

=

range

malignant

T

membrane

SAMBA analysis % of positive cell surface Histogram of integrated

benign

malignant

focal diffuse

m

CA,

+ ++

m=75%

m=25%

(r=50-94%)

(r=7-50%)

m=82%

m=12%

| (r=65-89%) (r=5-34%) Homegeneous Heterogeneous Homogeneous Heterogeneous (monomodal)

(multimodal)|(monomodal) (multimodal) I

cytoplasm, with some enhancement of the cell membrane. Two of the three positive primitive cancers were of grade 3 with vascular and lymphatic invasions. Most of the samples of normal breast were positive (12/13). The normal cells in lobules, ductules and ducts were stained on the apex of the cell membrane and adjacent cytoplasm. Positive staining was also observed within the lumen of normal ductules (Fig 2a). In contrast, only 10 of 20 benign lesions were positive (Fig 2b) and most of the malignant primary adenocarcinomas were negative (14/16). However, 2 out 3 lymph node métastases were stained. In positive tumours staining focally or totally involved the cell cytopasm (Fig 2c,d). CA.

:

epithelial

Computerized image analysis (SAMBA)

of

immunostaining distribution (Table 2)

SAMBA analysis showed that the mean percentages of stained surfaces by BA, and CA. were higher in normal and non-malignant disorder s than in carcinomas (75 % ver us 25 % for BA, and 82 % versus 12 % for CA.) (Table 2). Moreover, the staining intensity with BA, and CA. was heterogeneous in tumors as compared to normal tissues, where staining was homogeneous as reflected by the optical density

histograms (Fig 3).

245

'

I"

«3» * te

&

Figure

;*s>~ ,-Ä

?C-^

1 MAb BA, and avidin-biotin peroxidase + AEC : a) normal breast and b) focally in epitheliosis ; positive staining is observed within the ductal lumens and in the apex of cell membrane (x750, x580), whereas c), d) in very agressive inflammatory invasive comedo carcinoma (with vascular invasion), positive staining may be observed diffusely spread in the cytoplasm with focal intracytoplasmic lumps (c : intraductal component, d : invasive component (x 120, x 750)).

HMFGM and milk

protein reactivity

Using ELISA, both BA, and CA. showed strong binding to HMFGM and skimmed The two MAbs did not bind to alpha-lactalbumin, lactoferrin, human IgG and Enzymatic digestion

and chemical treatment of HMFGM and

The two reactive antigens were similarly treatments. Methanol and neuraminidase did not modify

cancer

cells

affected by the different reactivity with CA. and BA,. Trypsin reduced binding of the MAbs by about 80 % and 70 %, respectively. Protease VI and subtilisin treatment virtually abolished MAbs reactivity (Fig 4). Periodate mild oxidation of the antigens did not affect CA. and BA, reactivity (Fig 5). To detect whether these epitopes expressed on normal cefls could be masked by abnormal sialylation on cancer cells, H466B and T47D cells were treated by neuraminidase before ELISA. No reactivity could be observed. Evidence of

246

M

Figure

"

2 MAb CA. and avidin-biotin peroxidase -i- AEC : a) in normal breast, positive staining consists in dark intraluminal precipitate and appears in cells as a thick enhancement of the cell membrane (x220) b) in papillomas, the staining is confined to the apical part of the cell membrane (x300), whereas c) in carcinomas, some tumour cells display only diffuse or partial cytoplasmic staining (x750), d) with some focal enhancement along the cell membrane (x1200).

effectiveness of neuraminidase digestion was shown by the diminution of DF~ to H466B cells after treatment. After PCA fractionation of HMFGM, CA. reactivity was found in acid-soluble fraction using ELISA, while the alkaline-redissolved precipitate did not show CA. binding. Conversely to CA., BA, activity was lost after PCA treatment ; no binding was revealed by ELISA to either acid-soluble fraction or precipitated proteins.

binding

Molecular weight determination of CA. and BA, reactive

antigens.

After size exclusion chromatography of skimmed milk, the antigenic activity of CA. was found in the void volume of the columns (separation range for globular proteins : 5X10 ) and in a second peak with maximal reactivity at 330 kDa. BA, reactive,substances eluted in a single broad peak with maximal reactivity at about 1.2 x 10 Da (Fig 6). After SDS-PAGE separation of HMFGM extracts, immunostaining of the gels with CA. revealed a reactive band of apparent Mr 380-600 kDa, absent in a MCF- membrane

247

Figure

3 SAMBA analysis of BA, (a,b) CA. (c,d) immunostaining ; a) c) in normal breast tissue, the histograms of integrated optical densities reflect the homogeneity of BA, and CA. staining whereas b) and d) in carcinomas, positivity is heterogeneous.

O.D Without treatment Methanol Neuraminidase

Trypsin Protease VI

Figure 4 Chara cterization of CA. and BA¿ antigens on HMFGM ËTTSA

Subti 1isin

performed as described in "Materials and Methods". a) Binding of CA. to HMFGM without treatment (1), after methanol extraction (2) , after digestion with Vibrio Ch olera neuraminidase (3), trypsin ( 4), protease VI (5) and subtilisin (6). b) Binding of BA, to HMFGM without treatment (1), after methanol extraction (2) , after enzymatic digestion with neuraminidase (3), trypsin (4), p rotease VI (5) and subtilisin (6) Ordinate : O.D at 405 nm. was

Without treatment

248

Figure

5

Na-metaperiodate treatment skimmed milk. Binding of CA. and BA, after antigen treatment

was

of

tested by ELISA.

1BE-|2i recognizing carbohydrate epitope present in milk -(28), was used to assess the effectiveness of periodate MAb

treatment.

5

10

Na-METAPERIODATE (mM)

lysate same

run

method

concomitantly (Fig 7), or by immunoblotting.

BA

,

staining

could not be obtained

using

the

DISCUSSION In order to obtain MAbs which react specifically with a normal cell we followed a strategy of immunization using terminally differentiated normal cells and screening of hybridomas with a poorly differentiated cancer cell line. Three hybridoma supernatants BA,, CA. and DC-, were shown to bind strongly to milk cells, but to be unreactive with H466B cells. Further analysis of their reactivity with other breast cancer cell lines of various stages of differentiation revealed weak DC- binding with cancer cells. MAbs BA, and CA., respectively IgG- and IgM, were selected for further characterization. In vitro ZR-75-1, T47D and MDA-MB-231). they did not bind to four cancer cell lines Immunohistochemical studies on frozen sections of normal, non-malignant, and tumoral breast revealed that BA, and CA. are preferential markers of normal

population,

(MCF?,

II

Fraction

Figure

Number

6 Size exclusion FPLC of skimmmed milk using a superóse HR/6 columnFlow rate 0.4 ml/mn. Elution profile was monitored at 280 nm. CA. and BA, reactivity with eluted fraction was assessed by ELISA (absorbance : 405 nm). Calibration standard : A : Thyroglobuline (MW: 669.000), B: Ferritin (MW: 440.000), C: Catalase (MW: 232.000), D: Aldolase (MW: 158.000).

249

200 Figure

.92.4

7 MAb CA. immunostaining of gel after 3-15 % SDS-PAGE of HMFGM and membranes NP-40 extracts. (MCF7 membranes were prepared as described in 14)). a) MCF7 membranes : no reactivity is found, b) HMFGM: a high molecular weight reactive bound is

MCF?

visualized.

tissues. Indeed, BA, strongly stained 100 % normal and 71 % of benign pathological breast sections versus 11 % of malignant breast lesions. Similarly, CA. snowed strong reactivity with 92 % of normal breast, 50 % of benign lesions and only 12 % of primitive carcinomas. Both BA, and CA. staining of breast cancers involved a minority of cells and was heterogeneously distributed, as shown by SAMBA analysis of immunostained surface. Conversely, in benign specimens, a mean of 75 % and 82 % of the cell surface was stained by BA, and CA., respectively, with a very homogeneous pattern. This staining was polarized, apical in normal cells and diffuse in the cytoplasm of cancer cells. The finding that CA. stained 2 of 3 breast cancer métastases requires further studies on a large panel of samples. The two MAbs differed in the percentage of reactive non-malignant breast samples, and particularly in lymph node metastasis staining BA, did not stain the two métastases reactive with CA.. FPLC filtration chromatography of milk revealed that the native antigens recognized by CA. and BA, are very high-molecular-weight species eluting at different volumes. The CA.-reactive antigen was further identified as a 380-600 kDa protein by polyacrylamide gel electrophoresis. Failure of detecting BA, reactive bands after SDS-PAGE and direct staining of the gel or immunoblotting may be due to loss of reactivity after antigen denaturation. This hypothesis is consistent with BA, reactivity loss after boiling the antigen in Laemli buffer using ELISA or dot Dlot (results not shown). Antigen purification is currently in progress and will permit further characterization of the reactive molecules. We studied the effects of antigen enzymatic digestion and chemical treatment on the reactivity of BA, and CA. to further characterize the antigens. While protease digestion reduced or abolished binding, Vibrio cholera neuraminidase did not affect MAb reactivity, nor did methanol lipid extraction. In order to determine whether

carbohydrates

particpate

as

antigenic

determinants,

antigens

were

submitted to mild periodate oxidation, but no reactivity modifications were observed. These results, taken together, suggest the protein nature of the antigens. As the MAb LlCR-L0N-M-fi (15,29) binds to an epitope of luminal plasma membranes of normal mammary epithelium which is masked by sialylation in most breast carcinomas, we submitted H466B and T47D cell lines to neuraminidase treatment. This treatment did not allow any binding of BA, and CA.. Different behaviors of BA, and CA. were found after PCA fractionation of HMFGM. CA. antigen was found in the acid-soluble fraction, contrary to BA, antigen, which lost its reactivity after acid tratment.

250

The tissue distribution and high molecular weight of the CA. and BA, reactive antigens differed from those of some of the main milk proteins : ^ and K caseins, proteins secreted by the fully differentiated breast gland, have been detected only in lactating human breast tissue using specific monoclonal antibodies, but not in normal resting breast tissue, as recently reported (30). Moreover, lactalbumin, and lactoferrin did not show reactivity with BA, and CA. using ELISA. The immunocytochemical and histochemical patterns of CA. and BA, reactivity suggest that epitopes recognized are also different from the HMFGM determinants described thus far. For instance, HMFGM gp 70 and gp 155 (31,32) were detected by immunoperoxidase studies in most breast cancer tested. Similarly, most monoclonal and polyclonal antibodies (4,7,11,33-35) raised against HMFGM were reactive with a high percentage of breast cancer tissue specimens and were also reactive with such breast cancer cell lines as MCF?. However studies using a double determinant raioimmunoassay on human milk or breast cancer patients sera showed a cross-reactivity of MAb CA. but not of MAb BA, with the breast epithelial mucin defined by MAbs DF- (12) and ED„ (36) suggesting a co-expression of CA., DF_ and on the same molecufe (unpublished results). EDfi epitopes The immunocytochemical and biochemical data reported here indicate that BA, and CA. identify two new distinct epitopes of the normal resting and lactating human breast gland. They could be useful markers in the study of breast epithelial cell differentiation, in vivo analyzing their expression in epithelial cells in the the different sections of the mammary gland, in vitro using normal breast epithelial cells subjected to different culture conditions. Moreover the modulation of CA. and BA, epitopes expression is actually studied using in vitro models of oncogene transformation of normal breast cells. Preliminary results indicate that BA, reactivity with normal cells disappears in SV40 transormed cells (Ph. Berthon, personal communication). Similar studies have shown a negative correlation of a Mr 155 kDa glycoprotein, termed NEA 155, with the malignant transformation of human mammary epithelia (37). The relationship between BA, antigen and NEA 155 needs to be investigated.

ACKNOWLEDGMENT We wisk to thank mothers donors of milk and maternity ward staff of St Joseph Hospital, Compiègne and of Lariboisière Hospital, Paris for their helpful collaboration. We thank Or Nora Berois for her precious help, Dr M.N. Lavaut for skilled technical assistance, Mr B. Boursin for photographical work and Mrs C. Guilbert for typographical assistance. This study was financially supported by grants from MRT, Biosys SA and ARC. BIBLIOGRAPHY

1) Peterson, J.A., Bartholomew, J.C., Stampfer, M. and Ceriani, R.L. (1981)

Analysis of expression of human mammary epithelial antigens in normal and malignant breast cells at the single cell level by flow cytofluorimetry. Exp. Cell Biol., 49, 1-14. 2) Arklie, J., Taylor-Papadimitriou, J., Bodmer, W., Egan, M. and Millis, R. (1981) Differentiation antigens expressed by epithelial cells in the lactating breast

are

also detectable in breast

Int. J. Cancer, 28, 23-29. (1983) Modulation of phenotype in

cancers.

3) Chang, S.E. and Taylor-Papadimitriou, J.

cultures of human milk epithelial cells and its relation to the expression of a membrane antigen. Cell. Diff. 12, 143. 4) Edwards, P.A.W. and Brooks, I.M. (1984) Antigenic subset of human breast epithelial cells distinguished by monoclonal antibodies. J. Histochem. Cytochem. 32, 531. 5) Hand, P.H., Nuti, M., Colcher, D. and Schlom, J. (1983) Definition of antigenic heterogeneity and modulation among human mammary carcinoma cell populations using monoclonal antibodies to tumor associated antigens. Cancer Res. 43, 728-735. 6) Peterson, J.A., Ceriani, R.L., Blank, E.W. and Osvaldo, L. (1983) Comparison of rates of phenotypic variability in surface antigen expression in normal and cancerous breast epithelial cells. Cancer Res. 43, 4291-4296.

251

7) Burchell J., Durbin, H. and Taylor-Papadimitriou, J.

expression

(1983) Complexity of antigenic determinants, recognized by monoclonal antibodies HMFG-2, in normal and malignant human mammary epithelial cells. of

HMFG-1 and J. Immunol. 131, 508-513.

8) Ceriani, R.L., Peterson, J.A. and Blank, E.W. (1984) Variability in surface

antigen expression

of

human

breast

epithelial

cells

cultured

from

normal

breast, normal tissue peripheral to breast carcinomas and breast carcinomas. Cancer Res. 44, 3033-3039. 9) Edwards, P.A.W. (1985) Heterogeneous expression of cell-surface antigens in normal epithelia and their tumo urs, revealed by monoclonal antibodies. Br. J. Cancer 51, 149-160. 10) Sloane, J.P. and Ormerod, M.G. (1981) Distribution of epithelial membrane

antigen in normal and neoplastic tissue and its value in diagnostic tumour pathology. Cancer, 47, 1786-1795. 11) Hilkens, J., Buijs, F., Hilgers, J., Hageman, Ph., Calafat, J., Sonnenberg, A. and Van der Valk, M. (1984) Monoclonal antibodies against human milk-fat globule membranes detecting differentiation antigens of the mammary gland and its tumour Int. J. Cancer, 34, 197-206. 12) Kufe D., Inghirani G., Abe M., Hayes D., Justi-Wheeler H. and Schlom J. (1984) Differential reactivity of a novel monoclonal antibody (DF.J with human malignant versus benign breast tumours. Hybridoma 3, 223-232. 13) Pancino G., Charpin C, Calvo F., Guillemin M-C. and Roseto A. (1987) A novel

(7Bin)

14)

15)

16) 17)

18)

19) 20) 21) 22)

23) 24) 25)

26)

with differential reactivity between human mammary monoclonal antibody carcinoma and normal breast. Cancer Res. 47, 4444-4452. Pancino G.F., Le Doussal V., Mortada M.H., Berthon P., Osinaga E., Calvo F. and Roseto A. (1989) Characterization and distribution in normal and tumoral human tissues of breast cancer-associated antigen defined by monoclonal Cancer Res. 49, 7078-7085. antibody Burchell J., Gendler S., Taylor-Papadimitriou I., Girling A., Lewis A., Millis R. and Lamport D. (1987) Development and characterization of breast cancer reactive monoclonal antibodies directed to the core protein of the human milk mucin. Cancer Res. 47, 5476-5482. Foster C.S. and Neville A.H. (1987) Expression of breast epithelial differentiation antigens in human primary breast cancer. INCI, 79, 613-622. Calvo F., Goubin G., Gauville C, Jabrane N., Magdelenat H., de Cremoux P. and Marty M. (1986) A new human breast carcinoma cell line with dedifferentiation characteristics in vitro and a transforming c-Ki-ras gene. Proc. Am. Soc. Clin. Oncol. 5, 94. Engel L.W. and Young (1978) Human breast carcinoma cells in continuous culture : a review. Cancer Res. 38, 4327-4339. Keenan T.W., Morré J.D., Olson D.E., Yunghans W.W. and Patton S. (1970) Biochemical and morphological comparison of plasma membrane and milk fat globule membrane from bovine mammary gland. J. Cell Biol. 44, 80-93. Kobylka D. and Carraway K.L. (1972) Proteins and glycoproteins of the milk fat globule membrane. Biochim. Biophys. Acta 288, 282-295. Köhler G. and Mil stein C. (1975) Continuous culture of fused cells secreting antibody of predefined specificity. Nature 256, 494-497. Roseto A., Scherrer R., Cohen J., Guillemin M-C, Charpilienne A., Feynerol C. and Peries G. (1983) Isolation and characterization of anti-rotavirus immunoglobulins secreted by cloned hybridoma cell lines. J. Gen. Vivol. 64, 237-240. Woodword M.P., Young W.W. and Bloodgood (1985) Detection of monoclonal antibodies specific for carbohydrate epitopes using periodate oxidation. J. Immunol. Methods 78 : 143-153. Hoff S.D., Matsushita Y., Ota D.M., Cleary K.R., Yamori T^, Hakomori S. and Trimura T. (1989) Increased expression of sialyl-dimeric Le antigen in liver métastases of human colorectal carcinoma. Cancer Res. 49, 6883-6888. Charpin C, Lissitzky J.C., Jacquemier J., Lavaut M.N., Kopp F., Pourreau-Schneider N., Martin P.M., Toga M. (1986) Immunohistochemical detection of laminin in 98 human breast carcinomas : a light and electron microscopic study. Hum. Pathol. 17, 355-365. Brugal G. (1984) Image analysis of microscopic preparations. Methods and

7B.Q.

252

27)

28)

29) 30) 31) 32)

33)

34)

35) 36) 37)

Achievements in Experimen -al Pathology. Ed. G. Jasmin, L. Proschek, Montreal, Pub : S. Karger, Basel, 11, 1-33. Charpin C, Martin P.M., Devictor B., Lavaut M.N., Habib M.C., Andrac L., Toga M. (1988) Multiparametric study (SAMBA 200) of estrogen receptor immunocytochemical assay in 400 human breast carcinomas : analysis of estrogen receptor distribution heterogeneity in tissues and correlations with dextran coated charcoal assays and morphological data. Cancer Res. 48, 1578-1586. Pancino G., Charpin C, Osinaga E., Bétaille B., Le Roy M., Calvo F. and Roseto A. (1990) Characterization and distribution in human tissues of a glycoproteic antigen defined by monoclonal antibody 1 BE-? raised against T47D human breast cancer cell line. Cancer Res. In Press. Foster C.S., Neville A.M. (1984) Monoclonal antibody L1CR-L0N-M18 identifies impaired expression and excess sialylation of the I(Ma) cell surface antigen by primary breast carcinoma cells. Hum. Pathol. 15, 502-513. Earl H.M., Mcllhimey R.A.J., Wilson P., Güster son B.A. and Coombes R.C. (1989) Immunohistochemical study of B- and K-casein in the human breast and breast carcinomas, using monoclonal antibodies. Cancer Res. 49, 6070-6076. Iman A., Taylor C.R. and Tokés Z.A. (1984) Immunohistochemical study of the expression of human milk fat globule membrane glycoprotein 70. Cancer Res. 44, 2016-2022. Iman A., Drushella M.M., Taylor C.R. and Tokés Z.A. (1986) Preferential expression of a Mr 155.000 milk-fat-globule membrane glycoprotein on luminal epithelium of lobules in human breast. Cancer Res. 46, 6374-6379. Ceriani R.L., Peterson J.A., Lee J.P., Moneada R. and Blank E.W. (1983) Characterization of cell surface antigens of human mammary epithelial cells with monoclonal antibodies prepared against human milk fat globule. Som. Cell. Genet. 9, 415-427. Cordell J., Richardson T.C., Pulford K.A.F., Ghosh A.K., Gatter K.C., Hegdermann E., Mason D.Y. (1985) Production of monoclonal antibodies against human epithelial membrane antigen for use in diagnostic immunocytochemistry. Br. J. Cancer, 52, 347-354. Price M.R., Edwards S., Pawell M. and Baldwin R.W. (1986) Epitope analysis of monoclonal antibody NCRC-11 defined antigen isolated from human ovarian and breast carcinomas. Br. J. Cancer 54, 393-400. Mortada M., Pancino G., Charpin C, Schlageter M-H., Calvo F. and Roseto A. detects a circulating antigen in (1990) A novel monoclonal antibody breast cancer patient sera. C.R. Acad. Sei. Paris, 311, 63-68. Iman A. (1988) Identification of a normal cell-surface glycoprotein whose expression is inhibited in mammary eptihelial cells transformed with oncogenes. Abstracts of the second International Conference of Anticancer Research. Saronis, Grèce, October 11-15. Anticancer Res. 8, 1141.

(EDg)

Address reprint requests to : Dr. Gianfranco PANCINO Division d'Immunocytologie Appliquée UTC- BP. 649 60206 COMPIEGNE Cedex France -

Received for publication July 20, 1990 Accepted November 19, 1990

253

Two monoclonal antibodies identify antigens preferentially expressed on normal human breast cells versus breast cancer cells.

In order to obtain antibodies with specificity toward normal mammary epithelial antigenic determinants, we immunized BALB/c mice with normal milk cell...
6MB Sizes 0 Downloads 0 Views