Muc-1 is a major mucin glycoprotein expressed on the surface of mammary epithelial cells. It has attracted considerable attention as it is expressed in an aberrant form on many breast tumor cells. Here we describe studies using a recently obtained cDNA probe of Muc-1 expression during lactogenic development in the mouse. Northern blot analysis demonstrated that Muc-1 is expressed at all stages of lactogenic development but its levels are increased very significantly during midpregnancy and into lactation. The basis of this was examined using CID-9 mammary epithelial cell cultures. It was found that in the presence of insulin Muc-1 mRNA levels were increased by both hydrocortisone and prolactin, with the combination of the three hormones supporting maximum expression. Muc-1 mRNA levels were also modulated by culturing cells on a basementmembrane-like extracellular matrix that promoted mRNA levels 5-to 10-fold above levels in cells cultured on plastic tissue culture dishes. Immunocytochemical studies using monoclonal antibodies to carbohydrate epitopes on Muc-1 demonstrated that while Muc-1 was found at all developmental stages, it became increasingly sialylated during the course of pregnancy and into lactation. Additionally, we found that while Muc-1 is tightly polarized to the apical surface of the epithelium of lactating and pregnant mice it exhibited a less-polarized distribution on a small proportion of ductal cells in virgin mice. We conclude that the expression of Muc-1 is regulated at several different levels and by a number of different factors. We speculate that this may reflect different functional roles for Muc-1 at different stages of mammary development.

MUC1 is a highly glycosylated mucin expressed on the surface of mammary epithelial cells (Mather, 1987). It has previously been referred to as the polymorphic epithelial mucin (PEM), and has been shown to be aberrantly expressed on most carcinomas (Arklie et al., 1981; Burchell et al., 1987, 1989; Girling et al., 1989). The structure of the human MUC1 molecule has been determined on the basis of gene and cDNA cloning, and the modifications in its structure associated with tumorigenesis have been characterized (Gendler et al., 1987,1988,1990; Ligtenberg et al., 1990; Siddiqui et al., 1988). Recently it has become apparent that there are strong similarities among the epithelial mucins expressed on different tissues and that the MUC1 core protein is the basic unit of not just the mammary mucin, but also the lung, pancreas, Fallopian tube, ovary, uterus, kidney, salivary gland and stomach (reviewed by Zotter et al., 1988).

While the importance of mucin molecules as tumor markers is clear and well accepted, very little information is available on the function of these molecules and on their regulation. Consequently, it has not been possible to assess the full significance of expression of mucins on tumor cells and to determine whether their expression contributes to the survival of the tumor. In order to develop a better understanding of the function of the MUC1 mucin, we have turned our attention to the mouse Muc-1 molecule, with the expectation of being able to modulate its activity in animal models and consequently gain an understanding of its function on the surface of epithelial cells. (The human mucin gene locus is designated MUC1, the mouse homologue Muc-1.) To this end we recently cloned a cDNA molecule coding for the mouse Muc-1 and determined the complete sequence of the protein core (Spicer et al., 1991). A comparison of the sequence of the mouse Muc-1 with the previously determined human sequence demonstrated that the cytoplasmic tail, the transmembrane domain and the Ó-glycosylation sites were well conserved but that the mouse Muc-1 did not exhibit the structural polymorphisms that characterize the human MUC1 (Spicer et al., 1991).

In this paper we describe studies using a cDNA probe for the mouse Muc-1 to monitor expression of Muc-1 during lactogenic development of the mouse mammary gland. We also present data on the hormonal regulation of Muc-1 mRNA in cultured mouse mammary epithelial cells. These molecular biology studies are complemented by an immunocytochemical analysis of Muc-1 distribution in the lactogenic mammary gland using monoclonal antibodies reactive with mouse Muc-1. Results of this study demonstrate that Muc-1 is regulated by all three lactogenic hormones, insulin, hydrocortisone and prolactin, and that while it is expressed at all stages of pregnancy, its oligosaccharide structure is developmentally regulated. Additionally we describe an unexpected finding that although Muc-1 is tightly polarized to the apical surface of the epithelium of lactating and pregnant mice, it exhibits a nonpolarized distribution on the surface of a few ductal epithelial cells in non-pregnant mice.

RNA isolation and Northern blot analysis

Total RNA was isolated using a single-step acid guanidinium isothiocyanate, phenol/chloroform extraction protocol as described by Chomczynski and Sacchi (1989). RNA (10 g) was loaded into each well of a formaldehyde-containing gel and resolved under denaturing conditions. The samples were then transferred to a nylon membrane (Hybond-N, Amer-sham) and hybridized first with probes for Muc-1 and then reprobed using a β-casein probe. High specific activity probes (1 ×109 to 2 ×109 cts min-1 μg_° DNA) were generated using random primers and gel-purified inserts from Muc-1 and β-casein plasmids. Hybridization was carried out at 42°C overnight in a solution containing 6 x SSC (SSC is 0.15 M NaCl, 0.015 M sodium citrate, pH 7.0), 5 × Denhardt’s (1 × Denhardt’s is 0.02% bovine serum albumin, 0.02% Ficoll, 0.02% polyvinylpyrrolidone), sodium dodecyl sulfate (0.1%), NaH2PO4 (25 mM), Na4P2O7 (1.5 mM), formamide (50%), and salmon sperm DNA (100 μg/ml). Blots were washed at 68°C and exposed to X-ray films at -70°C for 4 hrs (β-casein) and 12 hrs (Muc-1).

Immunocytochemistry procedures

Mammary tissue was obtained from 4-to 10-week-old virgin, 15-day pregnant and 2-to 10-day lactating mice. Tissue was fixed in formaldehyde (3.5%) in phosphate-buffered saline (PBS) at room temperature for 1 hour, rinsed in PBS containing ammonium chloride (50 mM) for a further hour and embedded in OCT (Bayor Diagnostics, Slough, UK). Frozen sections (10 μm) were cut at -30°C and stored at —20°C until processed for immunostaining. Antibodies used for immunostaining were mouse monoclonal antibodies that were generated against human MUC1. These have been described previously, as antibodies LBL1,2,3,4,5 and 6. For the experiments described here, a cocktail of the six antibodies was used. Preliminary experiments demonstrated that this mixture of antibodies exhibited cross-reactivity with mouse Muc-1. However, initial experiments also revealed that the use of mouse monoclonal antibodies to stain mouse sections led to high background staining due to the detection of endogenous mouse immunoglobulins by fluorescently tagged goat anti-mouse second antibodies. To overcome this problem we developed a blocking procedure whereby tissue sections were treated with monovalent goat anti-mouse Fab fragments prior to incubation with antibodies. This led to elimination of virtually all non-specific staining. Fab fragments were generated from goat anti-mouse IgG by papain digestion. For this purpose, IgG (5 mg/ml) was incubated with papain (50 μg/ml) in sodium acetate (0.1 M), pH 5.5, containing cysteine (50 mM), and EDTA (1 mM), at 37°C for 24 h. Digestion was stopped by addition of iodoacetamide (75 mM). The effectiveness of the digestion was monitored each time by SDS-PAGE analysis. This mixture of Fab and Fc fragments was used without futher purification as a blocking agent. Sections of fixed tissue were first incubated in PBS containing bovine serum albumin (0.2%) at room temperature for 30 min and subsequently in the Fab-containing blocking solution (100 μg/ml) for 30 min. Sections were fixed a second time in formaldehyde (3.5%) and glutaraldehyde (0.1%) in PBS for 30 min to crosslink the anti-mouse Fab fragments to the endogenous mouse IgG. After rinsing in PBS containing ammonium chloride (50 mM), the sections were then incubated at room temperature in antibodies to Muc-1 for 1 hour, rinsed 3 times each in PBS/BSA (0.2%), and then incubated in Texas red-conjugated goat anti-mouse immunoglobulins, for 1 hour at room temperature. Sections were viewed by fluorescence microscopy using a Nikon inverted microscope equipped with epifluorescence optics and photographed using Kodak Tri × film.

Neuraminidase digestions

Previous work has demonstrated that the epitopes recognised by the LBL series of antibodies can be partially masked by sialic acid (Moss et al., 1988). In some experiments tissue sections were digested with neuraminidase following formaldehyde fixation and ammonium chloride washing. Sections of tissue were digested with neuraminidase (typeV) in PBS at a concentration of 50 mU/ml at room temperature for 30 min.

Cell culture

COMMA-l-D cells (Danielson et al., 1984) and their derived variant CID-9 cells (Schmidhauser et al., 1990) were grown in DMEM/F12 (1:1) containing fetal calf serum (5%), gentamy-cin (50 μg/ml), and insulin (5 μg/ml). The medium was changed every other day. For expanding the culture, the cells were split using a ratio of 1:5. For hormonal and matrixdependent induction of differentiation the cells were plated either on plastic culture dishes or on a reconstituted basement membrane (EHS) derived from the EHS tumor at a concentration of 8 ×104 cells/cm2 (Kleinman et al., 1986). Fetal calf serum (2%) - to support attachment to the plastic - insulin (5 μg/ml), and/or hydrocortisone (1 μg/ml), and/or prolactin (ovine, 3 μg/ml), were administered to the experiments as indicated in the figure legends. At 24 hours after plating, the medium was changed to serum-free conditions, while retaining the hormones as necessary. The cells were then cultured in the defined conditions for a further 5 days with medium being changed daily.

Radiolabelling Muc-1 and immunoprecipitation

COMMA-l-D cultures were radiolabelled with [3H]glucosa-mine (10 μCi/ml) for 24 h in medium containing insulin (5 μg/ml) and fetal calf serum (5%). Thereafter, medium was removed and cells scraped into PBS. Cell and medium samples were then digested with neuraminidase for 30 min as described above and processed for immunoprecipitation using procedures that we have previously described (Parry et al., 1987). Briefly, cells were extracted with an immunoprecipitation buffer composed of NaCl (0.15 M), Tris-Cl (50 mM), pH 7.5, sodium deoxycholate (0.5%), NP40 (0.5%) and aprotinin (10 μg/ml) at 4°C for 1 hour. The extract was clarified by centrifugation and the Muc-1 in the supernatant was immunoprecipitated using the mouse monoclonal antibody cocktail described above. Samples were incubated with antibody for 2 hours at 4°C and then with a rabbit anti-mouse IgM antibody (40 μg/ml) at 4°C overnight. The complex was collected on Protein A-Sepharose beads, washed, dissolved in electrophoresis sample buffer and resolved by SDS-PAGE procedures. The dried gel was processed for fluorography and exposed to X-ray fim at —70°C for 7 days.

Developmental regulation of Muc-1 during pregnancy and lactation

mRNA

The expression of mRNA for Muc-1 in mammary glands of virgin, pregnant and lactating mice was probed by Northern blot analysis using a [α-32P]dCTP-labelled probe corresponding to the mouse repeat region. Muc-1 mRNA (2.2kb) was detectable in virgin glands but levels were very significantly increased by day 10 of pregnancy, reaching maximum levels by day 14 or 15 (Fig. 1). Levels were significantly reduced late in lactation (day 14).

Fig. 1.

Northern blot analysis of Muc-1 mRNA levels in mammary tissue from virgin (V), pregnant (10, 14, 15 and 20 days), and lactating (L, 14 days) animals. Equal quantities (10 μg) of RNA were loaded in each well, and blots were probed with [32P]dCTP-labelled probes. Autoradiograms were exposed for 12 hours.

Fig. 1.

Northern blot analysis of Muc-1 mRNA levels in mammary tissue from virgin (V), pregnant (10, 14, 15 and 20 days), and lactating (L, 14 days) animals. Equal quantities (10 μg) of RNA were loaded in each well, and blots were probed with [32P]dCTP-labelled probes. Autoradiograms were exposed for 12 hours.

Muc-1 glycoprotein

Monoclonal antibodies to Muc-1 were used to localize Muc-1 in frozen sections of tissue from virgin, pregnant and lactating mice. A cocktail of six antibodies reactive with different carbohydrate epitopes was used for all experiments. These antibodies were generated in mice that had been immunized with human milk fat globule membranes, and their specificity in reacting with human MUC1 has been well documented previously. The reactivity of these antibodies with mouse Muc-1 was demonstrated in cultures of COMMA-l-D cells by radiolabelling cells with [3H]glucosamine for 24 hours and then immunoprecipitating Muc-1 as described in the Materials and methods. Immunoprecipated mucin was analysed by SDS-PAGE procedures, followed by autoradiography of the dried gel. The principal species detected migrated as a single band with a molecular weight greater than 200 ×103 (Fig. 2). This contrasts with the doublet at this size that is usually seen in analyses of human MUC1 and reflects the absence of a polymorphism in the mouse Muc-1 gene, as we have previously reported (Spicer et al., 1991).

Fig. 2.

The reactivity of mouse monoclonal antibodies with mouse Muc-1 glycoprotein. Cultures of COMMA-l-D cells were radiolabelled with [3H]glucosamine (100 μCi/ml) for 24 hours and cell and medium samples were immunoprecipitated as described in Materials and methods. Immunoprecipitates were resolved by SDS-PAGE and the precipitated proteins detected by autoradiography. Lane M contains medium proteins and lane C cellular proteins. The antibody precipates a single major band from the cell but not from the medium. This migrates as a single band at the top of the 5% polyacrylamide gel (S, top of stacking gel).

Fig. 2.

The reactivity of mouse monoclonal antibodies with mouse Muc-1 glycoprotein. Cultures of COMMA-l-D cells were radiolabelled with [3H]glucosamine (100 μCi/ml) for 24 hours and cell and medium samples were immunoprecipitated as described in Materials and methods. Immunoprecipitates were resolved by SDS-PAGE and the precipitated proteins detected by autoradiography. Lane M contains medium proteins and lane C cellular proteins. The antibody precipates a single major band from the cell but not from the medium. This migrates as a single band at the top of the 5% polyacrylamide gel (S, top of stacking gel).

Sections were first treated with neuraminidase to expose maximally epitopes recognised by the set of monoclonal antibodies (Moss et al., 1988), and the distribution and relative amounts of Muc-1 were assessed (Fig. 3). Tissues from each of the developmental stages were stained under identical conditions and photographed for the same exposure times, thus permitting a comparison of the apparent quantities of Muc-1 present on the cell surfaces. It was found that the level of Muc-1 on the epithelium from the lactating gland was significantly greater than that on the virgin gland, while tissue from pregnant animals expressed similar levels of Muc-1 to those found in tissue from lactating animals. This immunofluoresence study then was consistent with the analysis of Muc-1 mRNA expression and demonstrated conclusively that within the epithelial cell population, Muc-1 expression was developmentally regulated. Additionally, of course, the proportion of epithelial cells composing the gland increases during the course of lactogenesis. Consequently, the very large increase in Muc-1 mRNA that is detected during glandular development (Fig. 1) is a reflection of both an increase in epithelial cell numbers in the gland and a significant increase in Muc-1 mRNA in each cell.

Fig. 3.

Immunofluorescence analysis of Muc-1 expression in (A) 7-week-old virgin, (B) 14-day pregnant, and (C) lactating tissue sections. The upper and lower micrographs represent paired fluorescence and phase-contrast micrographs, respectively. All sections were stained under identical conditions and fluorescence micrographs were each photographed for 8 seconds to permit a direct comparison of the intensity of staining at each developmental stage. Clearly, while Muc-1 is expressed at all stages of lactogenic development, its levels are significantly increased during the early stages of pregnancy and are maintained at those levels into lactation. (Note: the exposure times chosen for photography are clearly not optimal for all of the stages of development but are shown in this way to permit a direct comparison between Muc-1 staining at each developmental stage.) ×1000.

Fig. 3.

Immunofluorescence analysis of Muc-1 expression in (A) 7-week-old virgin, (B) 14-day pregnant, and (C) lactating tissue sections. The upper and lower micrographs represent paired fluorescence and phase-contrast micrographs, respectively. All sections were stained under identical conditions and fluorescence micrographs were each photographed for 8 seconds to permit a direct comparison of the intensity of staining at each developmental stage. Clearly, while Muc-1 is expressed at all stages of lactogenic development, its levels are significantly increased during the early stages of pregnancy and are maintained at those levels into lactation. (Note: the exposure times chosen for photography are clearly not optimal for all of the stages of development but are shown in this way to permit a direct comparison between Muc-1 staining at each developmental stage.) ×1000.

Regulation of Muc-1 mRNA in culture: hormone and extracellular matrix influences

It is possible that the significant increase in mRNA observed in early pregnancy could be a consequence of regulation of Muc-1 by the lactogenic hormones insulin, hydrocortisone and prolactin, and that hormonally stimulated transcription of the Muc-1 gene leads to increased mRNA levels. This, of course, has been well established for the set of skim milk proteins that are synthesized by mammary epithelial cells. The matrix- and hormone-responsive mammary epithelial cells, CID-9, derived from the COMMA-l-D mouse mammary epithelial cell strain were used to examine hormonal effects. Cultures were established on tissue culture plastic and on an extracellular matrix gel isolated from the EHS tumor, and maintained in serum-free medium in the presence of selected lactogenic hormones for 5 days following cell attachment. Four hormone regimes were examined: insulin alone, insulin plus hydrocortisone, insulin plus prolactin, and insulin plus hydrocortisone and prolactin. The response of the cells was monitored using cDNA probes specific for β-casein and Muc-1. Under all hormonal conditions tested, expression of Muc-1 mRNA was significantly higher in cells cultured on EHS matrix than in cells on plastic although it was always detectable in cells cultured on plastic (Fig. 4). Moreover, the cells cultured on EHS matrix were hormonally responsive, while cells on plastic were not. In cells cultured on EHS matrix Muc-1 mRNA was present in significant quantities in the presence of insulin alone, but was increased 4-to 6fold by the combination of either hydrocortisone or prolactin with insulin. Maximum expression of the Muc-1 mRNA was detected in the presence of all three hormones, insulin, hydrocortisone and prolactin. This contrasts with the hormonal regulation of β-casein seen on reprobing the blot with a β-casein cDNA, which demonstrated that casein mRNA was expressed only in the presence of insulin, hydrocortisone and prolactin or insulin and prolactin alone. No β-casein mRNA could be found in CID-9 cells cultured in insulin alone, in insulin plus hydrocortisone, or in cells cultured on plastic under any conditions. This response is consistent with previously published data on /Lease in mRNA regulation and demonstrates that the CID-9 cells represent a good model system for determining the hormonal basis for Muc-1 expression. The less-stringent hormonal requirements for Muc-1 mRNA expression are consistent with its expression at all developmental stages of the gland, including the virgin gland. The elevated levels of Muc-1 mRNA seen in pregnant and lactating tissue are a consequence then of modulation of this basal level by prolactin and hydrocortisone.

Fig. 4.

(A) Northern blot analysis of Muc-1 expression in cultured CID-9 mouse mammary epithelial cells. Cultures were established on plastic (2,4,6 and 8) or on EHS-derived matrix (1,3,5,7) and maintained in the presence of either insulin alone (7,8), insulin + hydrocortisone (5,6), insulin+prolactin (3,4), or insulin+hydrocortisone-l-prolactin (1,2). The concentrations of each of the hormones are described in the Materials and methods. (B) The same samples probed with a β-casein cDNA probe. Equal quantities (10 μg) of RNA were loaded on the gel.

Fig. 4.

(A) Northern blot analysis of Muc-1 expression in cultured CID-9 mouse mammary epithelial cells. Cultures were established on plastic (2,4,6 and 8) or on EHS-derived matrix (1,3,5,7) and maintained in the presence of either insulin alone (7,8), insulin + hydrocortisone (5,6), insulin+prolactin (3,4), or insulin+hydrocortisone-l-prolactin (1,2). The concentrations of each of the hormones are described in the Materials and methods. (B) The same samples probed with a β-casein cDNA probe. Equal quantities (10 μg) of RNA were loaded on the gel.

Immunocytochemical analysis of Muc-1 sialylation in the lactogenic gland

While reactivity of antibodies to neuraminidase-treated sections of tissue from lactating gland was very strong (Fig. 3), it was consistently found that binding of these antibodies to untreated sections of tissue from lactating animals was weak and patchy (Fig. 5B). Similar results were obtained for sections of tissue from pregnant animals although the untreated sections stained more strongly and more uniformly than did those from lactating tissue (Fig. 6B). However, an enhancement of staining was again observed by prior neuraminidase digestion of the sections (Figs 5A, 6A). In tissue from virgin mice, the antibodies reacted exclusively with the ductal epithelium but, in contrast to what was observed in the tissues from pregnant and lactating glands, the antibodies bound without the need to treat sections with neuraminidase (Fig. 7B). Moreover, no change in staining was detectable upon predigestion with neuraminidase (Fig. 7A).

Fig. 5.

Frozen sections of tissue from lactating mammary glands stained with antibodies to Muc-1 protein followed by Texas red-conjugated goat anti-mouse immunoglobulins. (A,B) Paired phase-contrast and fluorescence micrographs of sections stained with (A) and without (B) neuraminidase treatment prior to antibody binding. ×1000.

Fig. 5.

Frozen sections of tissue from lactating mammary glands stained with antibodies to Muc-1 protein followed by Texas red-conjugated goat anti-mouse immunoglobulins. (A,B) Paired phase-contrast and fluorescence micrographs of sections stained with (A) and without (B) neuraminidase treatment prior to antibody binding. ×1000.

Fig. 6.

Frozen sections of tissue from mid-pregnant animals stained with antibodies to Muc-1 as described for Fig. 4. (A) With neuraminidase; (B) without neuraminidase. ×1000.

Fig. 6.

Frozen sections of tissue from mid-pregnant animals stained with antibodies to Muc-1 as described for Fig. 4. (A) With neuraminidase; (B) without neuraminidase. ×1000.

Fig. 7.

Frozen sections of tissue from virgin mice stained with anti Muc-1, as described in the legend to Fig. 4. (A) With neuraminidase treatment; (B) without neuraminidase. No enhancement of staining was detected following neuraminidase treatment. (C) Examples of occasional ducts that did not exhibit a fully polarized Muc-1 distribution. ×1000.

Fig. 7.

Frozen sections of tissue from virgin mice stained with anti Muc-1, as described in the legend to Fig. 4. (A) With neuraminidase treatment; (B) without neuraminidase. No enhancement of staining was detected following neuraminidase treatment. (C) Examples of occasional ducts that did not exhibit a fully polarized Muc-1 distribution. ×1000.

The requirement for neuraminidase digestion of tissues to expose epitopes recognized by the monoclonal antibodies is indicative of extensive sialylation. We have previously demonstrated that the antibodies used in this study react with oligosaccharide epitopes of Muc-1 (which probably accounts for their cross-species reactivity), and have shown that antibody binding is indeed reduced by sialylation. The need to digest sections from lactating and pregnant animals to expose reactive epitopes, while the same epitopes in tissues from virgin animals are already exposed, demonstrates a differential degree of sialylation of Muc-1 in virgin mice compared with pregnant and lactating animals. These studies reveal a developmentally regulated modulation of sialylation of Muc-1 during the course of pregnancy and lactation.

In sections from pregnant and lactating tissue, antibodies localized exclusively to the apical surface of the epithelium with no detectable staining of basal or lateral membranes. In the virgin gland, however, a small proportion of ducts exhibited distinct staining of the lateral and basal surfaces of the cells in addition to the apical surface (Fig. 7C). Staining at the apical surface was always much stronger than at the basolat-eral surface but the non-polarized distribution was reproducibly observed in about 5% of ducts. In this respect then there is developmental regulation of membrane polarity in a small proportion of mammary epithelial cells.

In this paper we demonstrate that the epithelial mucin, Muc-1, is expressed at all stages of development of the adult mammary gland, but the levels of expression, the composition of its oligosaccharide chains, and its polarized distribution on the cell surface, are each developmentally regulated. Additionally we establish, for the first time, the hormonal basis for regulation of Muc-1 mRNA expression.

Both immunocytochemical and Northern blot analyses demonstrate the presence of Muc-1 in the virgin gland, consistent with many previous studies using human mammary tissue. These, levels, however, are clearly modulated during the course of lactogenesis as a consequence of hormonal action. We were able to estabfish the basis for this using cells cultured on EHS-derived matrix, which has previously been shown to promote optimal expression of other milk protein genes (Li et al., 1987; Barcellos-Hoff et al., 1989). Significantly, we found that Muc-1 was expressed in the presence of insulin alone, and that either hydrocorti-sone or prolactin could increase Muc-1 mRNA levels beyond this basal level. The combination of all three hormones resulted in the maximum expression of Muc-1. While this maximal response is obtained under the same hormonal conditions previously found for some of the skim milk proteins, including j3-casein, it is clear that the hormonal basis for Muc-1 expression is distinctive in that it is not totally dependent upon prolactin. There is then a constitutive level of Muc-1 expression that is observed in the presence of insulin alone and this can be increased by either hydrocortisone or prolactin, accounting for the increased expression of Muc-1 in pregnancy and lactation. The finding of a constitutive level of expression is consistent with the presence of the mucin in several non-mammary epithelial cells, including pancreas, lung and stomach (Zotter et al., 1988). It is likely that this constitutive expression is under the control of tissue-specific promoters. To date an analysis of genomic sequences 5’ to the mouse Muc-1 gene has not revealed identifiable hormone response elements that can account for the observed hormonal regulation of Muc-1 (Spicer et al., 1991). Futher analysis of gene regulatory sequences in the light of the data presented here will be most interesting.

Two aspects of the effects of culture substratum on Muc-1 expression are notable. Firstly, the EHS matrix promotes expression of Muc-1 above and beyond expression on plastic in all combinations of hormones. This is exactly what has been demonstrated for other milk proteins and is what is confirmed here for β-casein mRNA (Li et al., 1987; Barcellos-Hoff et al., 1989). It is significant, however, that while β-casein mRNA is not expressed on plastic, mRNA for Muc-1 is easily detectable. This is consistent with the idea presented above in relation to the hormone experiments, that there is a constitutive level of Muc-1 expression that can be induced by either matrix components or hormones. A second aspect of the substratum effect is that culture on EHS matrix promotes hormone responsiveness. This is seen for Muc-1 as well as for /kcasein. While the detailed mechanisms by which hormone action and culture substrata are related remain to be elucidated, selection of an appropriate matrix substratum as used here is clearly important in determining the hormonal response of Muc-1. Previous studies have demonstrated that permitting hormones to interact with the basal surface of the cells is also crucial for secretory differentiation of mammary epithelial cells (Parry et al., 1987).

The immunocytochemical analysis of Muc-1 using mouse monoclonal antibodies was possible because of the blocking procedure that we developed to eliminate completely the background staining associated with using mouse monoclonal antibodies on mouse tissue sections. This permitted a careful analysis of the polarity of Muc-1 on the cell surface at all stages of glandular development. Two aspects of the polarized distribution of MUC-1 were significant. Firstly, the exclusive localization of Muc-1 to the apical surface of epithelia in pregnant and lactating mice is of interest in terms of the role of tight junctions in the maintenance of polarized membrane domains (Parry et al., 1990). Previous electron microscopy studies have established that tight-junction strands are organized into occluding junctional complexes only in lactating tissue and that epithelia from virgin and pregnant animals have short disorganized tight-junction strands (Pitelka, 1978). It is clear then, from the data presented here that restriction of Muc-1 to the apical domain occurs even in the absence of complete tight junctions. Previous work from our laboratory on the polarization of Muc-1 has led us to propose that Muc-1 is restricted in its distribution by interactions with a submembraneous actin-containing cytoskeleton and can polarize its distribution in cultured cells that are prevented from forming tight junctions (Parry et al., 1990). The in vivo observations presented here are consistent with this model and provide a clear demonstration of apical membrane polarization in vivo in the absence of complete tight junctions. The second aspect of polarization of Muc-1 that is of interest is the existence of a small fraction of ductal epithelial cells in the virgin gland that express some Muc-1 on their basolateral surfaces. We have attempted to determine if the cells expressing basolateral Muc-1 constitute a particular population of ductal cells, but have been unable to distinguish them from other ductal epithelia that exhibit clear polarization of Muc-1. It is possible that Muc-1 expression on the basolateral membrane performs a distinct physiological function in this restricted set of cells. It is, however, also possible that this set of epithelia represent a population of immature cells that are in the course of differentiating into fully polarized epithelia and have yet to develop the machinery needed to polarize completely their plasma membranes. In this latter case, basolateral localization of Muc-1 would simply represent an intermediate step in the formation of the mature epithelium and would not necessarily have a major biological function.

A final aspect of Muc-1 regulation that emerged from this study was the developmental regulation of terminal sialylation of the oligosaccharides. This was clearly established on the basis of a requirement for neuraminidase digestion of Muc-1 for antibody binding. This observed progressive increase in sialylation during pregnancy and lactation is presumably indicative of hormonal induction of a sialyltransferase during the course of pregnancy. The altered form of Muc-1 oligosaccharides in the virgin and lactating states is interesting in that it may reflect a different functional role for Muc-1 in the virgin and lactating gland. Muc-1 in the virgin gland is clearly required for some epithelial cell functions, but in the lactating gland the bulk of the synthesized Muc-1 is secreted in association with the milk fat globule membrane and presumably performs a physiological function related to neonatal growth or survival. It is also possible that it is required for membrane secretion and formation of polarized membrane domains. In future studies considering these potential functions of the mucin it will clearly be important to evaluate whether the composition of the oligosaccharide chains can influence the functional characteristics of the molecule.

We have previously encountered the issue of Muc-1 sialylation in studies of the apparent heterogeneous expression of Muc-1 on tumor cell surfaces (Moss et al., 1988). Working with cultured 734B cells we found that cell to cell heterogeneity in expression of MUC1 could be partially accounted for by heterogeneous sialylation of the mucin that resulted in blocking of epitopes and consequential failure of monoclonal antibodies to bind to MUCl-expressing cells. On the basis of the data reported here it would appear that the heterogeneous expression of sialylation has a developmental basis and that the tumor cell heterogeneity reflects the tendency of the 734B cells to adopt phenotypic characteristics of several developmental states of the mammary gland.

It is also possible that the stage-specific sialylation observed for the mouse Muc-1 can account for a discrepancy between our findings in the mouse and earlier work on the guinea pig mammary mucin, which was found by monoclonal antibody staining to be expressed only in the lactating gland (Johnson and Mather, 1985). While species variation certainly cannot be ruled out, the possibility that in the guinea pig studies the monoclonal antibodies used recognized only the fully sialylated form of Muc-1 also needs to be considered.

We thank Leslie Blackie and Rabib Talhouk for tissue samples, and J. Rosen for the β-casein probe. This research was supported by the Department of Energy, Office of Health Effects Research under Contract DEAC03-76SF 00098.

Arkite
,
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 cancers
.
Int. J. Cancer
28
,
23
29
.
Barcellos-Hoff
,
M. H.
,
Aggeler
,
J.
,
Ram
,
T. G.
and
Bissell
,
M. J.
(
1989
).
Functional differentiation and alveolar morphogenesis of primary mammary cultures on reconstituted basement membrane
.
Development
105
,
223
235
.
Burchell
,
J.
,
Gendler
,
S.
,
Taylor-Papadimitriou
,
J.
,
Girling
,
A.
,
Lewis
,
A.
and
Millis
,
R.
(
1987
).
Development and characterization of breast cancer reactive monoclonal antibodies directed to the core protein of the human milk mucin
.
Cancer Res
.
47
,
54765482
.
Burchell
,
J.
,
Taylor-Papadimitriou
,
J.
,
Boshell
,
M.
,
Gendler
,
S.
and
Duhig
,
T.
(
1989
).
A short sequence, within the amino acid tandem repeat of a cancer-associated mucin, contains immunodominant epitopes
.
Int. J. Cancer
44
,
691
696
.
Chomczynski
,
P.
and
Sacchl
,
N.
(
1987
).
Single step RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction
.
Analyt. Biochem
.
162
,
156
159
.
Danielson
,
K. G.
,
Obom
,
C. J.
,
Durban
,
E. M.
,
Butel
,
J. S.
and
Medina
,
D.
(
1984
).
Epithelial mouse mammary cell line exhibiting normal morphogenesis in vivo and functional differentiation in vitro
.
Proc. Nat. Acad. Sci. USA
81
,
3756
3760
.
Gendler
,
S. J.
,
Burchell
,
J. M.
,
Duhig
,
T.
,
Lamport
,
D.
,
White
,
R.
,
Parker
,
M.
and
Taylor-Papadlmltriou
,
J.
(
1987
).
Cloning of partial cDNA encoding differentiation and tumor-associated mucin glycoproteins expressed by human mammary epithelium
.
Proc. Nat. Acad. Sci. USA
84
,
6060
6064
.
Gendler
,
S. J.
,
Lancaster
,
C. A.
,
Taylor-Papadimitriou
,
J.
,
Duhig
,
T.
,
Peat
,
N.
,
Burchell
,
J.
,
Pemberton
,
L.
,
Lalani
,
E. N.
and
Wilson
,
D.
(
1990
).
Molecular cloning and expression of a human tumor-associated polymorphic epithelial mucin
.
J. Biol. Chem
.
265
,
15286
15293
.
Gendler
,
S.
,
Taylor-Papadimitriou
,
J.
,
Duhig
,
T.
,
Rothbard
,
J.
and
Burchell
,
J.
(
1988
).
A highly immunogenic region of a human polymorphic epithelial mucin expressed by carcinomas is made up of tandem repeats
.
J. Biol. Chem
.
263
,
12820
12823
.
Girling
,
A.
,
Bartkova
,
J.
,
Burchell
,
J.
,
Gendler
,
S.
,
Gillett
,
C.
and
Taylor-Papadimitriou
,
J.
(
1989
).
A core protein epitope of the polymorphic epithelial mucin detected by the monoclonal antibody SM-3 is selectively exposed in a range of primary carcinomas
.
Int. J. Cancer
43
,
1072
1076
.
Johnson
,
V. G.
and
Mather
,
I. H.
(
1985
).
Monoclonal antibodies prepared against PAS-I butyrophilin and GP-55 from guinea-pig milk-fat-globule membrane bind specifically to the apical pole of secretory-epithelial cells in lactating mammary tissue
.
Exp. Cell Res
.
156
,
144
158
.
Kleinman
,
H. K.
,
McGarvey
,
M. L.
,
Hassell
,
J. R.
,
Star
,
V. L.
,
Cannon
,
F. B.
,
Laurie
,
G. W.
and
Martin
,
G. R.
(
1986
).
Basement membrane complexes with biological activity
.
Biochemistry
25
,
312
318
.
Lancaster
,
C. A.
,
Peat
,
N.
,
Duhig
,
T.
,
Wilson
,
D.
,
Taylor-Papadlmitriou
,
J.
and
Gendler
,
S. J.
(
1990
).
Structure and expression of the human polymorphic epithelial mucin gene: an expressed VNTR unit
.
Biochem. Biophysics Res. Comm
.
173
,
1019
1029
.
Li
,
M.-L.
,
Aggeler
,
J.
,
Farson
,
D. A.
,
Hatier
,
C.
,
Hassell
,
J.
and
Bissell
,
M. J.
(
1987
).
Influence of a reconstituted basement membrane and its components on casein gene expression and secretion in mouse mammary epithelial cells
.
Proc. Nat. Acad. Sci. USA
84
,
136
140
.
Ligtenberg
,
M. J. L.
,
Vos
,
H. L.
,
Gennissen
,
A. M. C.
and
Hilkens
,
J.
(
1990
).
Episialin, a carcinoma-associated mucin, is generated by a polymorphic gene encoding splice variants with alternative amino termini
.
J. Biol. Chem
.
265
,
5573
5578
.
Mather
,
I. H.
(
1987
).
Proteins of the milk-fat-globule membrane as markers of mammary epithelial cells and apical plasma membrane
.
In The Mammary Gland
(ed.
M. C.
Neville
and
C. W.
Daniel
), pp.
217
268
,
New York
,
Plenum Press
.
Moss
,
L.
,
Greenwait
,
D.
,
Cullen
,
B.
,
Dinh
,
N.
,
Ranken
,
R.
and
Parry
,
G.
(
1988
).
Cell-to-cell heterogeneity in the expression of carbohydrate-based epitopes of a mucin-type glycoprotein on the surface of human mammary carcinoma cells
.
J. Cell. Physiol
.
137
,
310
320
.
Parry
,
G.
,
Beck
,
J. C.
,
Moss
,
L.
,
Bartley
,
J.
and
OJakian
,
G. K.
(
1990
).
Determination of apical membrane polarity in mammary epithelial cell cultures: the role of cell-cell, cell-substratum, and membrane-cytoskeleton interactions
.
Exp. Cell Res
.
188
,
302
311
.
Parry
,
G.
,
Cullen
,
B.
,
Kaetzel
,
C. S.
,
Kramer
,
R.
and
Moss
,
L.
(
1987
).
Regulation of differentiation and polarized secretion in mammary epithelial cells in culture: Extracellular matrix and membrane polarity influences
.
J. Cell Biol
.
105
,
2043
2051
.
Pitelka
,
D.
(
1978
).
Cell contacts in the mammary gland
In Lactation A Comprehensive Treatise
, vol.
IV
(ed.
Larson
B.L.
), pp.
41
66
,
New York
,
Academic Press
.
Schmidhauser
,
C.
,
Bissell
,
M. J.
,
Myers
,
C. A.
and
Casperson
,
G. F.
(
1990
).
Extracellular matrix and hormones transcriptionally regulate bovine /5-casein 5’ sequences in stably transfected mouse mammary cells
.
Proc. Nat. Acad. Sci. USA
87
,
118
9122
.
Siddiqui
,
J.
,
Abe
,
M.
,
Hayes
,
D.
,
Shanl
,
E.
,
Yunis
,
E.
and
Kufe
,
D.
(
1988
).
Isolation and sequencing of a cDNA coding for the human DF3 breast carcinoma-associated antigen
.
Proc. Nat. Acad. Sci. USA
85
,
2320
2323
.
Spicer
,
A.
,
Parry
,
G.
,
Patton
,
S.
and
Gendler
,
S. J.
(
1991
).
Molecular cloning and analysis of the mouse homologue of the tumor-associated mucin, MUC1, reveals conservation of potential O-glycosylation sites, transmembrane, and cytoplasmic domains and a loss of minisatellite-like polymorphism
.
J. Biol. Chem
.
266
,
1509915109
.
Zotter
,
S.
,
Hageman
,
P. C.
,
Lossnitzer
,
A.
,
Moot
,
W. J.
and
Hilgers
,
J.
(
1988
).
Tissue and tumor distribution of human polymorphic epithelial mucin
.
Cancer Rev
.
11-12
,
55
100
.