Research in VeterinaryScience 1995,58, 272-276

Establishment of a cell line (MCM-B2) from a benign mixed tumour of canine mammary gland
B. P. PRIOSOERYANTO, S. TATEYAMA, R. YAMAGUCHI, K. UCHIDA, Department of Veterinary Pathology,

Faculty of Agrieulture, Miyazaki University, Miyazala" 889-21, Japan

SUMMARY
A cell line was established from a benign mixed tumour of the canine mammary gland. Light microscopy of the cells cultured on plastic dishes revealed monolayer colonies. Cells that grew within the collagen gel matrix formed large three-dimensional colonies with a branching pattern. Immunohistochemically,these cells reacted intensely with anti-vimentin antiserum and mildly with anti-desmin antiserum. Ultrastruc~tral examination revealed a large nucleus, intraeytoplasmic organelles and intermediate filaments, which varied among cells. The cells possessed an abnormal chromosome number, an average of 80 per cell. Histologically, the xenografted tumour of cultured cells was similar to anaplastic carcinoma and reacted strongly with antivimentin antiserum, mildly with anti-desmin antiserum, and weakly with anti-keratin antiserum. The average chromosome number of cells form the xenografted tumour was the same as that of the original cultured cells. These findings suggest that the cell line might be derived from stem cells or atypical cells, and that it should be useful as a model for the study of cell differentiation and proliferation in canine mammary mmours. MAMMARY tumours are common in dogs, ranking second in prevalence after skin tumours (Brodey et al 1983, Madewell and Theilen 1987, Moulton 1990). About 50 per cent of these tumours are benign, and of these, the mixed tumour is the most common type. They generally have a complex histological pattern and have been the subject of many reports (Fowler et al 1974, Bostock 1975, Tateyama and Cotchin 1977, Gilbertson et al 1983, Destexhe et al 1993, Hellmen et al 1993). The complex cell components of canine mammary tissue may be responsible for these histological features. It has been well documented that the histological origin and behaviour of canine mammary gland tumours and such features as their steroid hormone receptor status may affect the risk of mammary carcinoma in a manner similar to that in human breast cancer (Martin et al 1984, Moulton 1990). as a result dog tumours may provide some advantages over rodent tumours for increasing the understanding of the aetiology, pathogenesis and therapy of mammary'tumours (Madewell and Theilen 1987). Some comparative studies have already been reported (Sasadaira 1976, Chrisp and Spangler 1980, Raynaud et al 1981, Martin et al 1984). Since the mammary gland is a complex organ consisting of a mixed population of cells, including stem cells (Daniel and Silberstein 1987), in vitro studies have also expressed this complexity (Dunnington et al 1983, Schmid et al 1983, Tateyama et al 1990), and a number of permanent cell lines have been established in vitro from canine mammary tumours, most of which were epithelial cells (Wolfe et al 1986, Van der Burg et al 1989). Owen et al (1977) established two cell lines, which were identified as fibroblasts and epithelial cells. However, there has been only one report of the establishment of a canine mammary tumour cell line with an intermediate-type morphology (cell line 229) suggesting a stem cell (Hellmen 1992). The identification and establishment of stem cell is of interest because of their relevance to the behaviour of mammary cancer and their participation in the development of the mammary gland. The purpose of this study was to establish a cell line derived from the stem cells of a canine mammary gland tumour for further study of the differentiation and proliferation of cell components in canine mixed tumours.

M A T E R I A L S AND M E T H O D S

Isolation of cell line

A tumour mass 3 cm x 5 cm was obtained surgically from the mammary gland of a 10-year-old female pointer dog. The tumour mass had appeared two years previously and an X-ray examination confirmed that there was a metastasis in the lung. Histologically, the tumour was composed of a mixed population of cells with chondromucoid tissue. The major cellular components were elongated spindle-shaped and polygonal cells (Fig 1). A double-layered architecture was common in the epithelial and myoepithelial cells of the acini. Mitotic figures were occasionally visible. From these findings, the case was diagnosed as a benign mixed tumour of the mammary gland (Moulton 1990). The tissue was minced and digested for eight hours at 37C in a humidified atmosphere of 5 per cent carbon dioxide in air with 4 mg m1-1 collagenase (Wako; 232 U m g -1) in Dulbecco's Modified Eagle's medium (DME) and Ham's Nutrient Mixture F-12 (Sigma) containing 10 per cent fetal calf serum (FCS), 100 iu m1-1 penicillin and 100 gg m1-1 streptomycin. The digested tissue was filtered through nylon mesh cloth (80 gm), centrifuged at 184 g for 10 minutes, and cultured as described by Tateyama et al (1990). Briefly, the isolated cells were cultured in 90 m m diameter plastic dishes (Sumilon, Sumitomo Bakelite Medical) in DME/F-12 medium containing 10 per cent FCS, 100 iu m1-1 penicillin, 100 gg m1-1 streptomycin, 10 gg m1-1 insulin, and 1 ~g m1-1 hydrocortisone together with 18 x 18 m m coverslips for further immunohistochemical and electron microscopical studies. The culture dishes were maintained in 5 per cent carbon dioxide in air at 37C and observed

272

Canine cell line from benign mixed tumour

273

FIG 1 : Histological appearance of the original canine tumour showing spindle cell proliferation surrounding the alveoil (A) and cartilage formation (B). Haematoxylin and eosin x 100

daily with a phase-contrast microscope (Nikon, DIAPHOT- Electron microscopy

TMD).
The cells were subcultured by washing them with 0.2 mM ethylenediamine tetra-acetic acid in phosphatebuffered saline (EDTA-PBS)and digested with 0.05 per cent trypsin in EDTA-PBS, at 37C for seven minutes. A DME/F12 medium containing 10 per cent FCS was added to stop the trypsinisation. Viable cells were diluted to 5.0 x 105 cells m1-1, and plated on a 50 mm dish. The cells were stocked in a medium consisting of DME/F-12 containing 10 per cent FCS and 10 per cent dimethylsulphoxide (DMSO) and stored at -80C. Agar cloning Freshly trypsinisised single cells (1.0 x 104) were layered in 2 ml of a 10 per cent FCS medium containing 0.375 per cent (w/v) agar in a 50 mm dish. This layer was placed on top of a 2 ml baselayer of the same medium containing 0-5 per cent agar. The cells were incubated at 37C in a humidified atmosphere containing 5 per cent carbon dioxide in air; after 14 days colonies larger than 10 cells were scored (Van der Burg et al 1989). Growth in collagen Acid-soluble type I collagen solution from porcine tendon (Cellmatrix IA; 3 mg m1-1, pH 3.0) was used and reconstituted according to the manufacturer's recommendations. For fixed gel-embedded cultures, the cell suspension was centrifuged at 184 g for five minutes, and 1.5 ml of collagen solution was added to the cell precipitate. The cell suspension with collagen gel was spread on to a 35 mm dish, incubated for 15 minutes, and overlaid with 2 ml DME/F-12 supplemented with 10 per cent FCS, and observed daily by phase-contrast microscopy. For a floating gel-embedded culture, the cell culture was prepared as for the fixed gel-embedded culture before it was overlaid with DME/F-12, and the gel was then released from the dish wall.

The cells were grown on coverslips, and the gel with embedded cells was fixed with 1.5 per cent glutaraldehyde in 0.1 M cacodylate buffer, pH 7.4, for one hour and then with 1"5 per cent osmium tetroxide. After dehydration in graded ethanols, the Cells were embedded in Spurr's resin. Ultrathin sections were stained with uranyl acetate and lead citrate and examined with an electron microscope (Hitachi H-800 MU).

Immunohistochemistry The following primary antibodies were used: mouse monoclonal antibodies (mAb) against human a-smooth muscle actin, human vimentin (Dakopatts), human oestrogen receptor (Novocastra Laboratories) and rabbit antisera against human keratin and human a-desmin (Dako Corporation). A coverslip with attached cells was washed in PBS, fixed in cold acetone (-20C) and stored in a freezer for 30 minutes. The cells were reacted with each of the primary antibodies overnight at 4C. The attached antibodies were visualised by an avidin-biotin-peroxidase complex (ABC) system (Vectastain, Vector Laboratories).

Growth curve and chromosome studies The cells were plated in triplicate at a concentration of 5-0 x 102 cells m1-1 in a 35 mm dish. The cells were counted every day for five days. Doubling times were determined during the exponential growth phase (Campling et al 1992). For chromosome studies, the cells were treated with 0.1 ~tg m1-1 demecolcine (Sigma) for three hours, trypsinised, washed and treated with a hypotonic solution of 0.075 M potassium chloride and then fixed with methanol-acetic acid. Preparations of cells were air dried and stained with Giemsa. Chromosomes spread from 150 metaphases were counted and the results averaged (Gibson et al 1991).

274

B. P. Priosoeryanto, S. Tateyama, R. Yamaguchi, K. Uchida

FIG 2: Cell culture at mid-passage (45th) showing a subconfluent monolayer with long cytoplasmic bridges (A) and duct-like structural appearance on the collagen gel culture (B). Phase-contrast; A x 316, B x 164

Growth in athymic nude mice

A suspension of 5 x 107 cells was inoculated subcutaneously into six nude mice (KSN nude SLC strain, Nippon SLC Corporation). Seven weeks after inoculation, all the mice were killed, and the nodules that had developed at the injection site were processed for routine histological and immunohistochemical examinations, and the determination of chromosome number.

multilayer appearance. The cells within the collagen gel formed three-dimensional colonies, organoid structures with a branching pattern and duct-like structures (Fig 2b). Ultrastructural studies revealed large nuclei, numerous free ribosomes, endoplasmic reticulum, mitochondria and intermediate filaments (Fig 3). These morphological and ultrastructural characteristics of the cultured cells were maintained during extended periods of growth in culture for more than 70 passages over a year.

RESULTS

Immunohistochemistry
The cultured cells were strongly stained with antivimentin mAb (Fig 4) and mildly with anti-desmin antiserum, but were negative for the antibodies against keratin and oestrogen receptor. Two types of staining patterns for vimentin-positive cells were encountered, one running parallel to the cell margin and another forming a network throughout the whole Cytoplasm. The same reactivities were also expressed in the cells that grew on the fixed and floating gel cultures.

Light and electron microscopy

Light microscopy of the cells grown on plastic dishes revealed that they were elongated to polygonal in shape, had large nuclei and often two or more prominent nucleoli. They formed monolayer colonies (Fig 2a) but did not form any organoid structures. A common feature of these cells was their interconnection by long cytoplasmic bridges. The cells that grew on the gel surface showed a monolayer or

FIG 3: Ultrastructure of the cultured cells, showing discontinuous border between cell surface and collagen gel (CG). x 6000

FIG 4: Immunohistochemical examination of the cultured cells showing positive reaction to vimentin antibody. Avidin biotin peroxidase complex method x 275

Canine cell line from benign mixed tumour

275

Growth in athymic nude mice

All the mice inoculated with the cultured cells had developed a tumour mass around 0.3 cm in diameter within two weeks of inoculation. The tumours grew rapidly to be 1 cm x 2 cm seven weeks after inoculation, when the mice were killed. Histologically, the mmour was composed of oval to elongated small and large pleomorphic cells with bizarre nuclei and numerous threads of chromatin (Fig 5). The cytoplasm was eosinophilic and mitotic figures were fore"to seven per high power field. The immunohistochemical examination of xenografted tissues showed that they were strongly reactive with anti-vimentin mAb, mildly with antidesmin antiserum, and weakly with anti-keratin antiserum. In vitro immunohistochemical and chromosome studies of cultured cells obtained from the xenografted mmour revealed the same results as in the original cultured cells.
Growth and chromosome studies

serum was required for its growth, with the best concentration being 10 per cent. The cell line was able to grow and form colonies in soft agar culture with a cloning efficiency that ranged from 4.1 to 20.2 per cent. The analysis of the chromosome numbers in 150 metaphase spreads of the cells from the early (6th), middle (38th) and late (67th) passages and also from the xenografted tumour, revealed a modal number of 80, a slight increase from the normal number (2N=78).

DISCUSSION The cell line established was named Miyazaki University canine mammary gland mixed turnout, and designated MCM-B2. The isolation and maintenance of this turnout cell line was achieved by adapting dissociative techniques to obtain significant improvements in cell cultivation. The methods used have been extended to other cell types with the same results (data not shown). The MCM-B2 cells showed intense immunoreactivity with a mAb for vimentin, which is known to be an intermediate filament of mesenchymal cells in vivo (Virtanen et al 1981), and mild immunoreactivity with anti-desmin antiserum, but they were negative for anti-keratin antiserum. The staining pattern for vimentin was in agreement with that of an earlier report (Den[: et al 1985). The immunohistochemical and morphological examinations confirmed that the morphological characteristics of the MCM-B2 cells were in almost complete conformity with those of fibroblasts. In contrast, when the cells were transplanted into nude mice, they showed a similarity to the anaplastic carcinoma described by Hampe and Misdorp (1984), and also expressed keratin intermediate filament, known as a hallmark of epithelial cells (Henderson and Webber 1981), whereas the cultured cells obtained from xenografted tissue did not express keratin. The authors' experiments and others (Rohol et al 1991, Hellmen 1992) have shown that cells in culture can exhibit immunohistochemical phenomena not observed in the original tissue or transplant cells, both in normal and malignant cells, for example, the expression of keratin. The expression of different types of intermediate filament in cell culture, might indicate phenotypic diversity (Lichtner et al 1987) and different stages of development from stem cells, as shown in the rat mammary cell line RAMA 37 (Dunnington et al 1983) and human breast carcinoma cell line PMC 42 (Whitehead et al 1983a). There are pluripotent stems cells located in the mammary end bud, defined as cap cells which can differentiate into both mammary ductal epithelial and myoepithelial cells (Williams and Daniel 1983). There are also unidentified cell types in their associated tumours, consisting of cells in a state of differentiation or of different phenotypes (Hellmen 1992). Attempts to identify mammary stem cells in clonal culture have been reviewed by Rudland et al (1980b) and Whitehead et al (1983a, b). Transplantation of the Rama 25 cell line, whose basic morphology is cuboidal, into nude mice gives rise to tumours that may contain regions of both spindle-shaped and epithelial cells, again suggesting that these cells are able to differentiate (Rudland et al 1980a). It is possible that the MCM-B2 cells differentiated into epitheliaMike cells when transplanted into nude mice, and adopted a lower degree of differentiation during cell culture. This idea was supported by the different expression of intermediate filaments with the same chromosome number in the original and xenografled cultured

A growth curve was plotted for cells growing in media supplemented with 10 per cent Fcs at the 69th passage (Fig 6). The average time for population doubling was eight hours, and a confluent monolayer was attained after 48 hours. Observations of this cell line showed that fetal calf

FIG 5: Histological appearance of xenografted tumour showing large round to spindle-shaped cells. The large round cells were weakly positive for keratin whereas the spindle-shaped cells were negative. Haematoxylin and eosin x 143

10 7 10 6 lO s
Z

10 4
r~ 10 3 10 2

4 DAYS

FIG 6: Growth curve of MCM-B2 cells growing in DME/F-12 supplemented with 10 per cent FCS at the 69th passage. Each point indicates the mean of three determinations

276

B. P. Priosoeryanto, S. Tateyama, R. Yamaguchi, K. Uchida

cells. The question whether cultured mammary cells become less differentiated in vitro has not yet been answered (Hellmen 1992). According to the chromosome number of the MCM-B2 cell, which was hyperploid, variations in the chromosome number have usually occurred in canine mammary mmours (Hellmen et al 1988) and also in established cell lines (Owen et al 1977, Norval et al 1984, Wolfe et al 1986). These variations may be associated with the instability of chromosomes that occurs during mmour progression (Hellmen 1992). The identification of stem cells is important because, in rats, the incidence of tumours is correlated directly with the density of the terminal end buds. The susceptibility of the mammary gland to carcinogens is due not only to the presence of terminal end buds as undifferentiated structures but also to their biological properties such as their high mitotic activity and high DNA labelling index. The incidence of tumours decreases markedly when the mammary gland undergoes total differentiation, and this lower proliferative activity seems to be one of the properties that account for its lower susceptibility to carcinogens (Russo et al 1979). The potential to differentiate in various directions depends on the intrinsic characteristics of the tumour in combination with environmental stimuli. It appears that MCM-B2, when transplanted into nude mice, shows an epithelial-like cell differentiation, that may be influenced by the suitable in vivo environment. The histogenesis of MCM-B2 is, therefore, not clear. The available information points to a multipotential cell type, which is not yet known, arising during tumorigenesis in canine mammary mixed tumours. A knowledge of the morphological and tumorigenic behaviour of MCM-B2 in nude mice should help to clarify this question and such studies are now in progress. REFERENCES
BOSTOCK, D. E. (1975) The prognosis following the surgical excision of canine mammary neoplasms. European Journal of Cancer 11,389-396 BRODEY, R. S., GOLDSCHMIDT, M. H. & ROSZEL, J. R. (1983) Canine mammary gland neoplasms. Journal of the American Animal Hospital Association 19, 61-90 CAMPLING, B. G., HARWORTH, A. C., BAKER, H. M., GREER, D. L., HOLDEN, J. J. A., BRADLEY, W. E. C., PYM, J. & DEXTER, D. F. (1992) Establishment and characterization of a panel of human lung cancer cell lines. Cancer 69, 2064-2074 CHRISP, S. E. & SPANGLER, W. L. (1980) The malignant canine tumor as a model for the study of human breast cancer. In The Canine as a Biomedical Research Model. Washington DC, US Department of Energy, Technical Information Center. pp 331-349 DANIEL, C. W. & SILBERSTEIN, G. B. (1987) Postnatal development of the rodent mammary gland. In The Mammary Gland. Development, Regulation, and Function. Eds M. C. Neville and C. W. Daniel. New York, Plenum Press. pp 3-36 DENK, H., WEYBORA, W., RATSCHEK, M., SOHAR, R. & FRANKE, W. W. (1985) Distribution of vimentin, eytokeratins, and desmosomal-plaque proteins in human nephroblastoma as revealed by specific antibodies; co-existence of ceil groups of different degrees of epitheiial differentiation. Differentiation 29, 88-97 DESTEXHE, E., LESPAGNARD, L., DEGEYTER, M., HEYMANN, R. & COIGNOUL, F. (1993) Immtmohistochemical identification of myoepithelial, epithelial, and connective tissue ceils in canine mammary Wanors. Veterinary Pathology 30, 146-154 DUNN1NGTON, D. J., HUGHES, C. M., MONAGHAN, P. & RUDLAND, P. S. (1983) Phenotypic instability of rat mammary tumour epithelial cells. Journal of the National Cancer Institute 7, 1227-1240 FOWLER, E. H., WILSON, G. P. & KOESTNER, A. (i974) Biologic behaviour of canine mammary neoplasms based on a histogenetic classification. Veterinary Pathology 11,212-229 GIBSON, C. A., VEGA, J. R., BAUMRUCKER, C. R., OAKLEY, C. S. & WELSCH, C. W. (1991) Establishment and characterization of bovine mammary epithelial cell lines. In Vitro Cellular and Developmental Biology 27A, 585-594 GILBERTSON, S. R., KURZMAN, I. D., ZACHRAN, R. E., HURVITZ, A. I. & BLACK, M. M. (1983) Canine mammary epithelial neoplasms: Biologic implications of morphologic characteristic assessed in 232 dogs. Veterinary Pathology 20, 127-142

HAMPE, J. R. & MISDORP, W. (1974) Tumours and dysplasias of the mammary gland. Bulletin of the WorldHealth Organization 50, 111-133 HELLMEN, E. (1992) Characterization of four in vitro established canine mammary carcinoma and one atypical benign mixed tumor cell lines. In Vitro Cellular and Developmental Biology 28A, 309-319 HELLMEN, E., BERGSTROM, R., HOLMBERG, L., SPANGBERG, I.-B., HANSSON, K. & LINDGREN, A. (1993) Prognostic factors in canine mammary tumors: A multivariate study of 202 consecutive cases. Veterinary Pathology 30, 20-27 HELLMEN, E., LINDGREN, A., LINELL, F., MATSSON, P. & NILSSON, A. (1988) Comparison of histology and clinical variables to DNA pIoidy in canine mammary tumors. Veterinary Pathology 25, 219-226 HENDERSON, D. & WEBBER, K. (1981) Immunoelectron microscopical identifications of two types of intermediate filaments in established epithelial cells. Experimental Cell Research 132, 297-311 LICHTNER, R. B., MOSKWA, P. S. & NICOLSON, G. L. (1987) Heterogenous expression of cytokeratins in metastatic mammary adenocarcinoma in vitro and in vivo. Invasion Metastasis 7, 367-383 MADEWELL, B. R. & THEILEN, G. H. (1987) Tumors of the mammary gland. In Veterinary Cancer Med ~ he. 2nd edn. Eds G. H. Theilen and B. R. Madewell. Philadelphia, Lea and Febiger. pp 327-344 MARTIN, P. M., COTARD, M., MIALOT, J. P., ANDRE, F. & RAYNAUD, J. P. (1984) Animal models for hormone-dependent human breast cancer. Relationships between steroid receptor profiles in canine and feline mammary turnours and survival rate. Cancer Chemotherapy and Pharmacology 12, 13-17 MOULTON, 1. E. (1990) Tumors of the mammary gland. In Tumors in Domestic Animals. 3rd edn. Ed J. E. Moulton. Berkeley, University of California Press. pp 518-552 NORVAL, M., MA1NGAY, J, & ELSE, R. W. (1984) Studies of three canine mammary carcinoma cell lines-I. In vitro properties. European Journal of Cancer and Clinical Oncology 20, I489-1500 OWEN, L. N., MORGAN, D. R., BOSTOCK, D. E. & FLEMANS, R. J. (1977) Tissue culture and transplantation studies of canine mammary carcinoma. European Journal of Cancer 13, 1445-1449 RAYNAUD, J. P., COTARD, M., ANDRE, F., MIALOT, J. P., ROLLAND, P. H. & MARTIN, P. M. (1981) Spontaneous canine mammary mmours. A model for human endocrine therapy? Journal of Steroid Biochemistry 15, 201-207 ROHOL, P. J. M., PRINSEN, I., RADEMARKERS, L. P. H. M., HSU, S.-M. & VAN UNNIK, J. A. M. (1991) Two cell lines with epithelial ceil-like characteristics established from malignant fibrous histioeytomas. Cancer 68, 1963-t 972 RUDLAND, P. S., BENNET, D. C. & WARBURTON, M. J. (1980a) Growth and differentiation of cultured rat mammary epithelial cells. In Hormones and Cancer. Eds S. Iacobelli, R. J. B. King, H. R. Lindner and M. E. Lippman. New York, Raven Press. pp 225-269 RUDLAND, P. S., ORMEROD, E. J. & PATERSON, F. C. (1980b) Stem ceils in rat mammary development and cancer: A review. Journal of the Royal Society of Medicine 73, 437-442 RUSSO, J., WILGUS, G. & RUSSO, I. H. (1979) Susceptibility of the mammary gland to carcinogenesis. I. Differentiation of mammary gland as a determinant of tumor incidence and type of lesion. American Journal of Pathology 96, 721-735 SASADAIRA, H. (1976) Comparative pathological studies on canine mammary gland tumors and human breast and salivary gland mixed tumors. Morphology of the neoplastic myoepithelial cells in the tumors. Bulletin of Azabu Veterinary College 1, 1-14 (In Japanese) SCHMID, E., FRANKE, W. W., GRUND, C., SCHILLER, D. L., KOLB, H. & PAWELETZ, N. (1983) An epithelial cell line with elongated mydid morphology derived from bovine mammary gland. Experimental Cell Research 146, 309-328 TATEYAMA, S. & COTCHIN, E. (1977) Alkaline phnsphatase reaction of canine mammary mixed tumours: a light and electron microscopic study. Research in Veterinary Science 23, 356-364 TATEYAMA, S., FURUK~WA, H., YAMAGUCHI, R. & NOSAKA, D. (1990) Plate and collagen-gel cultures of normal canine mammary epithelial cells. Research in Veterinary Science 49, 14-19 VAN DER BURG, B., VAN SELM-MILTENBURG, J. P., VAN MAURIK, P., RUTTEMAN, G. R., MISDORP, W., DE LAAT, S. W. & VAN ZOELEN, E. J. J. (1989) Isolation of autonomously growing dog mammary tumor ceil lines cultured in medium supplemented with serum treated to inactivate growth factors. Journal of the National CaneerInstitute 81, 1545-1551 VIRTANEN, I., LEHTO, V. P., LEHTONEN, E., VARTIO, T., STENMAN, S., KURKI, P., WAGER, O., SMALL, J. V., DAHL, D. & BADLEY, R. A. (1981) Expression of intermediate filaments in cultured cells. Journal of Cell Science 50, 45-63 WHITEHEAD, R. H., BERTONCELLO, I. WEBBER, L. M. & PEDERSEN, J. S. (1983a) A new human breast carcinoma cell line (PMC42) with stem ceil characteristics. I. MorphoIogic characterization.Journal of the National CancerInstitute 70, 649-661 WHITEHEAD, R. H., MONAGHAN, P., WEBBER, L. M. & VITALI, A. A. (1983b) A new human breast carcinoma cell line (PCM42) with stem ceU characteristics. II. Characterization of cells growing as organoids. Journal of the National Cancer Institute 71, 1193-1203 WILLIAMS, J. M. & DANIEL, C. W. (1983) Mammary ductal elongation: Differentiation of myoepithelium and basal lamina during branching morphogenesis. Developmental Biology 97, 274-290 WOLFE; L. G., SMITH, B. B., TOIVIO-KINNUCAN, M. A., SARTIN, E. A., KWAPIEN, R. P., HENDERSON, R. A. & BARNES, S. (1986) Biologic properties of cell lines derived from canine mammary carcinomas. Journal of the National Cancer Institute 77, 783 -792 Received Apri114, 1994 Accepted January 11, 1995