We have used immunological reagents for characterizing epithelial cell types in the adult rat mammary gland. Some were sera against purified proteins; others were monoclonal antibodies to mammary cells. On the basis of marker distribution we have identified 10 different cell types. We have established the developmental connection between these types by adopting the principle that cells displaying the same markers are directly related to each other. The results concur with those previously obtained by studying the growth response of mammary cells to the hormonal changes of the estrus cycle [Dulbecco, R., Henahan, M.

(1); they also, however, generate cells offusiform shape (fibroblast like) that do not form domes. It has been suggested that RAMA 25 cells are derived from a mammarv stem cell and that the fusiform cells may be the equivalent of the myoepithelial cells of the normal mammary gland (1). If so, the formation of fusiform cells in the cultures might be the in vitro equivalent ofa normal differentiation step and thus represent an interesting model system for studying differentiation in vitro.The present work was carried out to determine whether the cuboidal-to-fusiform transition observed in RAMA cultures indeed reproduces a differentiation step occurring in the animal. In this context, we asked two questions: (i) Is this transition a general, or at least a frequent, property of rat mammary cell lines and (ii) how does the distribution of available markers in the cuboidal and the fusiform cells in vitro compare with those of mammary epithelial and myoepithelial cells in vivo?The results show that the cuboidal-to-fusiform transition can be regarded as a general property of mammary cells. The transition, however, does not correspond to that of a stem cell to a myoepithelial cell. Rather, the data suggest that the fusiform cells are a new cell type. MATERIALS AND METHODSCell Lines. The RAMA 25 (cuboidal) and RAMA 4 and 29 (fusiform) lines were obtained from D. Bennett; the LA7 line (cuboidal) was a derivative of RAMA 25 (2). The F2408 line (rat fibroblast) was obtained from W. Eckhart.Tumor Induction. Tumors were induced in 50-dav-old virgin female Wistar-Furth rats by single intravenous injection of Nnitrosomethylurea (K & K) at 5 mg/100 g of body weight (3). The day after injection, the animals were put on a high fat diet (20% lard/36.5% cornstarch/7.5% dextrose/25% casein/5% alphacel/4% salt mixture/2% vitamin fortification). Seven ofthe 18 rats injected developed mammary tumors within a year.Isolation of RANI Cell Lines. Cells were isolated from a slowgrowing adenocarcinoma of the fifth right inguinal mammary gland, which appeared 7 months after injection. Histologically, this carcinoma conformed to a papillary adenocarcinoma. A fairly soft part of the tumor was cut into small fragments and digested with collagenase and hyaluronidase (4). The cell suspension was plated on a feeder layer of RAMA 29 (1) cells previously exposed overnight to mitomycin C (0.5 jig/ml). The medium was Ham's F-12 that had been conditioned for 24 hr by RAMA 29 cells and then centrifuged and filtered through a Nalgene filter unit (0.45-pL pore size) to remove cells. It was supplemented with 20% fresh Dulbecco's modified Eagle's medium/5% fetal calf serum/hydrocortisone (50 ng/ml)/insulin (50 ng/ml). In subsequent transfers, medium conditioned by LA7 cells (2) was sometimes used, without discernible difference in the results.Cloning. When the cultures became confluent, they were transferred bv the use of trypsin (250 mg/liter)/EDTA (90 mg/ liter) in Tris-buffered saline (pH 7.5). After three such transfers, the cells were cloned. A culture w...

We describe reversible changes of intermediate filaments of fibroblastic cells associated with changes in the functional state of the cells. The changes are revealed by comparing the immunofluorescence patterns given by a monoclonal antibody and a polyclonal serum, both recognizing vimentin. The state of the filaments depends on culture density; this effect cannot be attributed to the nutritional state of the cells, their growth rate, or substances released into the medium. It seems to depend mainly on the aggregation of filaments during strong cell movements. The possible significance of these findings for the functional role of intermediate filaments is discussed.Intermediate filaments, 8-10 mm in diameter, are present in most cells. They are made up of different proteins that share some common epitope (1). In fibroblastic cells, the main constituent of the filaments is vimentin (2, 3), with a molecular weight of 58,000; smaller related peptides (molecular weight, >40,000) are also observed (4, 5) and are thought to be cleavage products of vimentin by a specific protease (6). Vimentin molecules associate to form protofilaments: various models have been proposed for this association (7,8). Protofilaments are associated in bundles; in muscle cells and nucleated erythrocytes, they are crosslinked by another protein, synemin (9, 10); protofilaments are aggregated by cationic proteins (11). Vimentin undergoes phosphorylation, which may have a regulatory significance (12,13). Under the action of colchicine, intermediate filaments collapse into a coiled mass near the nucleus (14); collapse is also caused by intracellular injection of a monoclonal antibody to a common antigen of the intermediate filaments (15) and, in lymphocytes, by capping of surface molecules (16). This and other evidence shows that intermediate filaments are connected to the plasma membrane (10, 17). They also form a cage around the nucleus (18) and appear to be connected to the nuclear membrane (19 for 2 hr; they were then fixed in neutral formalin for 20 min, washed in phosphate-buffered saline, and dried. They were overlaid with photographic emulsion, developed after 1 wk, and counterstained with hematoxylin.Characterization of Material Stained by Hybridoma TllA9e. The material stained by Tl1A9e was characterized by the immunological blotting technique. Cytoplasmic extracts of sparse or confluent NIH 3T3 cultures were prepared (23) by lysing monolayers in 1% (wt/vol) deoxycholate/1% (vol/vol) Nonidet P-40/1 mM phenylmethylsulfonyl fluoride/7 mM NaCI/1 mM MgCl2/100 mM NaF/7 mM Tris-HCI, pH 7.4. The lysates were centrifuged at 4°C for 10 min at 3,000 X g to sediment nuclei. The supernatants were run on 10% NaDodSO4 gels and blotted to nitrocellulose paper (Schleicher & Schuell; BA85, 0.45 ,tm) (24). The nitrocellulose paper was incubated in undiluted culture supernatant of T1lA9e hybridoma or a 1:30 dilution of Tl1A9e mouse serum. Bound TllA9e was located by using either l"I-labeled goat anti-mouse Ig or rabbit anti-mouse IgM followed...

During the development of the rat mammary gland, ducts are formed from end-buds, which contain the stem cells. In this process a lumen is formed in the semisolid mass of the end-bud, and the cells acquire polarity. We have studied this process by following the localization of three inframembranous proteins present in the cells'of both end-buds and ducts; microvillin, the microvillar protein p80, and the desmosomal plaque protein p205. We find that the development of ducts is accompanied by a redistribution of these proteins, which in immature parts of the end-buds are found together in the cell. Microvillin and p80 go together to the apical pole of the cells, in contact with the lumen, whereas p205 goes to the basal surface, in contact with cells of the myoepithelial lineage. The acquisition of polarity occurs at the same time as a lumen begins to form by local'gaps between clls. It seems likely that the redistribution of the inframembraneous proteins is the consequence of the localization of surface glycoproteins that affect in opposite ways the adhesion between the cells. characterized. In these cells microvillin is an inframembranous protein present, in association with other proteins, in microvilli. We show here that the distribution of microvillin in cells of end-buds and ducts of the rat mammary gland undergoes changes correlated with the formation of the lumen. The changes concern not only the cell types in which the protein is expressed but also the part of the cell in which it is localized. 13y using the monoclonal antibody it is possiThe publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 5763 ble to follow the redistribution of components of the outer layer of the cytoplasm that accompanies the development of cell polarity and lumen formation.In this study we have also used antibodies to two other proteins with inframembranous localization: a protein of the brush border, which is also present in the microvilli of mammary cells (p80) (4), and a component of demosomal plaques (p205) (5). We show that they also change localization within cells as they differentiate. The pattern of redistribution of p80 is exactly the same as that of microvillin, whereas that of p205 differs, taking a complementary path.MATERIALS AND METHODS Animals. Sprague-Dawley rats, either 3 or 7 weeks old, were used. Cryostat sections were cut at 5 ,tm thickness from whole fourth and fifth glands frozen at -30'C.Reagents. Rabbit antisera to the microvillar 80-kDa protein was generously provided by A. Bretscher (4), to the demosomal plaque 205-kDa protein by D. R. Garrod (6), and to collagen IV by L. A. Liotta. Monoclonal antibody 9B16-12 specific for microvillin has been described (3). Immunofluorescence. Immunofluorescence was carried out by the sandwich technique as described (1). RESULTSIn normal mammary glands of 3-or 7-week-old rats microvillin is restricted to the epi...

We have isolated two monoclonal antibodies to cytokeratins and determined their cell specificities. They display interesting localization within the rat mammary gland. One (lA10) shows specificity for myoepithelial cells; the other (24B42) is specific for lumenal cells at various stages of development. These two monoclonal antibodies and three others to cytokeratin previously isolated were used in conjunction with antibodies to myosin and collagen IV to confirm and extend our previous findings on epithelial cell types and development within the mammary gland.Antibodies, especially monoclonal antibodies, are very useful for identifying cell types using immunocytological methods. This approach has been used in our laboratory for studying the pathways of cell differentiation in the mammary gland (1). Antibodies to cytokeratins are especially useful in this context because there are many different kinds of cytokeratins, which display considerable cell specificity.The cytokeratins are expressed by a multigene family that is developmentally regulated (see refs. 2 and 3 for reviews). They are expressed mostly in epithelia but also in mesothelial cells (4). Every tissue has a complex pattern of these peptides, which can be distinguished by their isoelectric point and electrophoretic mobility in gels. Many tissues share the same peptides (5, 6). The distribution of the peptides, ascertained by either extraction or by immunohistochemical methods using polyclonal antisera, differs in different organs or in different parts of the same organ-for instance in different layers of the epidermis. Similar peptides are found in different mammalian species. Using monoclonal antibodies (7), the existence of both peptide-specific and common epitopes was observed (8). Monoclonal antibodies with considerable specificity (8-10), as well as some with broad crossreactivity (11), were identified by immunohistochemistry. Using polyclonal antisera to partially purified peptides it was shown by immunohistochemistry that cytokeratins are not uniformly distributed in the mammary gland (11-13). Quantitative differences were observed between myoepithelial and epithelial cells, and in the latter cells, between ductal and secretory alveolar cells (12)(13)(14).Cytokeratins show considerable crossreactivity among species. For instance, Lane (15) has shown that monoclonal antibodies to cytokeratins from PtK1 cells recognize cells in the rat mammary gland. We have found that one of these antibodies (Le6l) is highly selective for certain cell types (1). By taking advantage of this crossreactivity, we have prepared a set of monoclonal antibodies from mice immunized with total cow muzzle keratin. We report here on the properties of two antibodies that have considerable cell specificities. These antibodies are very useful for studying the cell types present in the rat mammary gland. They allow a further refinement of the pathway of cell differentiation in this organ. We also compare these monoclonal antibodies with three others from mice immunized w...