During the development of the mammary gland, duct-lining epithelial cells progress through a program of expansive proliferation, followed by a terminal differentiation that allows for the biosynthesis and secretion of milk during lactation. The role of gap junction proteins, connexins, in the development and function of this secretory epithelium was investigated. Connexins, Cx26 and Cx32, were differentially expressed throughout pregnancy and lactation in alveolar cells. Cx26 poly-(A)(+) RNA and protein levels increased from early pregnancy, whereas Cx32 was detectable only during lactation. At this time, immunolocalization of connexins by confocal microscopy and immunogold labeling of high-pressure frozen freeze-substituted tissue showed that both connexins colocalized to the same junctional plaque. Analysis of gap junction hemichannels (connexons) isolated from lactating mammary gland plasma membranes by a rate-density centrifugation procedure, followed by immunoprecipitation and by size-exclusion chromatography, showed that Cx26 and Cx32 were organized as homomeric and heteromeric connexons. Structural diversity in the assembly of gap junction hemichannels demonstrated between pregnant and lactating mammary gland may account for differences in ionic and molecular signaling that may physiologically influence the onset and/or maintenance of the secretory phenotype of alveolar epithelial cells.

Immunolocalisation of type XIV collagen/undulin in the human mammary gland revealed greater deposition in the interlobular stroma than in the intralobular stroma. The interlobular stroma is located between the breast lobules and their associated intralobular stroma. Fibroblasts isolated from the interlobular stroma synthesised 3- to 5-fold more type XIV collagen/undulin than intralobular fibroblasts, but synthesised type I and type IV collagens in similar amounts. The differential expression of type XIV collagen/undulin was maintained with passage in culture. The results suggest a role for type XIV collagen/undulin in stabilising dense collagen fibrils. The maintenance of two types of structurally distinct stromas may be important during developmental processes in the mammary gland.

Inter- and intralobular mammary fibroblasts have been separated from normal human breast tissue and cultured to study the differential expression of ectoenzymes present within the stroma of the normal gland and associated with breast cancers. Specific ectoenzymes were identified by indirect immunofluorescence and quantified by flow cytometry and semi-quantitative PCR. A consistent difference was noted between the two fibroblast sub-populations at early passage in respect of dipeptidyl peptidase IV (DPP IV) and aminopeptidase N (APN) expression. Early passage intralobular fibroblasts were positive for APN but negative for DPP IV, as seen in the intact tissue. However, with continued sub-culture they gradually began to express DPP IV, until at later passages they became indistinguishable from the interlobular fibroblasts, which were APN and DPP IV-positive at all stages in culture, as they are in intact tissue. Neutral endopeptidase (NEP/CALLA/CD10) is not expressed by normal adult breast fibroblasts but is found in the stroma associated with over 60% of breast cancers. It was up-regulated in vitro on both inter- and intralobular fibroblasts, with final levels that were significantly (< 14 times) higher on the former in all pairs of preparations from individual donors analysed. This difference persisted with continued passage, and levels of the ectoenzyme and its messenger RNA were further up-regulated by hydrocortisone in both populations. These results demonstrate that phenotypically distinct cultures of human mammary fibroblast sub-populations can be used to study the regulation of these stromal ectoenzymes.(ABSTRACT TRUNCATED AT 250 WORDS)

The immunocytochemical distribution of three cell surface peptidases was investigated in samples of developing infant breast ranging in age from newborn to 9.5 months. We have previously demonstrated that in the adult breast these enzymes identify subpopulations of epithelial cells and fibrosblasts. We therefore wished to address two questions: (a) At what stage in breast development can fibroblast subpopulations be identified, and (b) Is the distribution of these peptidases related to cellular differentiation and morphogenesis? At the histological level there was a cuff of stromal cells closely associated with the developing ductular and lobular structures. At all stages of ductular and lobular development the fibroblasts in this layer were consistently negative for dipeptidyl peptidase IV (DPP IV) and clearly distinguished from the fibroblasts in the surrounding matrix, some of which expressed DPP IV in an age-dependent manner. Within the infant breast aminopeptidase N (APN) was localised to luminal epithelial cells and all fibroblasts, whilst neutral endopeptidase (NEP) was specifically localised to myoepithelial cells. These results are considered in relation to the role of stromal-epithelial interactions during morphogenesis and the proposed function of these enzymes.

The immunocytochemical distribution of the cell-surface enzyme dipeptidyl peptidase IV (DPP IV) has been studied in the human breast at the light and ultrastructural level. The presence of the enzyme was demonstrated on the cell membranes of interlobular fibroblasts, whilst intralobular fibroblasts were DPP-IV-negative. A fluorograph, after immunoprecipitation of S-35-methionine-labelled proteins of fibroblasts from primary breast cultures with an anti-serum to DPP IV, demonstrated a band at 135 kDa consistent with the presence of the enzyme. The clear delineation of 2 functionally distinct subpopulations of breast fibroblasts was maintained in benign fibro-adenomas and cystosarcoma phyllodes, both tumour types having growth characteristics of intralobular stroma. This observation has important implications for both normal breast biology and for breast carcinogenesis.

The ectoenzyme aminopeptidase N (APN) was localized in the normal human breast at both the light microscopic and the ultrastructural level. APN was expressed on intralobular and interlobular fibroblasts and on the apical surface of some luminal epithelial cells. This enzyme was not detected on either myoepithelial cells and their associated basement membrane or capillary endothelium. Furthermore, the staining pattern was maintained in benign and malignant breast disease. APN belongs to a family of enzymes that hydrolyze peptides in the extracellular space. As with other ectoenzymes present in the breast, APN expression is restricted to specific cell types. This pattern of expression may indicate a role for these enzymes in the biology of the normal breast.

A method of culturing human breast epithelium is described in which viable explants can be maintained in protein-free medium while retaining the capacity of responding to added hormones and growth factors for at least 7 days. Culture parameters were chosen to provide maximum sensitivity of detection of proliferative responses by autoradiography. Under basal conditions, the mean thymidine labeling index of the explants was 0.08%. After stimulation with insulin, hydrocortisone, and cholera toxin (I,H,CT), a combination known to stimulate proliferation in human breast epithelium in vitro, the mean labeling index was 15.7%. Stimulation of explants with epidermal growth factor (EGF) and transforming growth factor (TGF)-alpha resulted in mean labeling indices of 6.6 and 10.8%, respectively. Autoradiography at the ultrastructural level demonstrated that in I,H,CT-stimulated explants the majority of the labeled cells were luminal, with only 1.5% being basal cells. In contrast, after EGF and TGF-alpha basal cells accounted for 11.5 and 18.5% of the labeled population. These results indicate that this system provides an in vitro assay of proliferative activity in the normal human breast that enables comparisons to be made between both the luminal and the basal cells in the explants and their counterparts in monolayer culture prepared from flow sorted cells. Thus, growth responses dependent on cell-to-cell interactions or stromal modulation can be identified.

The progressive lowering of temperature (PLT) method of embedding for electron microscope immunolabelling has been examined with the objective of formulating a standardized protocol which can be applied to a wide variety of samples. The methods described cover fixation, processing of samples by the PLT method, embedding in Lowicryl HM20 and subsequent immunolabelling. Each of the steps in the fixation and embedding protocol has been assessed for its potential to retain both morphology and antigenicity. Comparison of samples embedded in Lowicryl K4M and HM20 at -25 degrees C indicate an increased membrane contrast in HM20 sections, and a further improvement in morphology when samples were embedded in HM20 at -50 degrees C rather than at -25 degrees C. The results of applying the methods described are demonstrated in a range of samples of both mammalian and botanical origin, which includes solid tissues, cells in suspension, and cells grown in vitro as a monolayer culture and embedded in situ. Samples processed by this method have been immunolabelled using a wide range of antibodies recognizing nuclear, cytoplasmic, cell membrane and extracellular matrix antigens.

Seventy-two samples of infant breasts, aged from newborn to 2 years, were collected at necropsy. Whole-mount preparations and histological sections were made. A system of classification was devised to study the extent of the structural development of the ductal system (morphological types I, II, and III) and the functional differentiation of the lining epithelium (functional stages I to V). There was no correlation between the age of the infant and the type of development of the ductal system. In contrast, the epithelial differentiation followed a chronological pattern, starting with secretory changes and apparently going through a period characterized by apocrine metaplasia before post-secretory involution. These epithelial changes were not associated with the morphological type of the ductal system. There were no distinguishing features between the breasts from the two sexes. Immunoperoxidase staining for actin and kappa-casein was carried out to study the myoepithelial cells and secretory cells, respectively. Myoepithelial cells were present at all stages and prominent staining for casein was observed up to 2 months of age. Embryonic-type adipose tissue was seen in 7 cases, in one of which it was associated closely with the developing ductal system. Extramedullary hematopoiesis was observed in the periductal connective tissue until 4 months of age. This paper describes the most extensive anatomical and histological study of the human infant breast to date and lays the foundation for a detailed study of the epithelial and stromal changes that take place during human breast development.

Luminal and myoepithelial cells have been separated from normal adult human breast epithelium using fluorescence activated cell sorting. Their isolation was based on the exclusive expression of two surface antigens, epithelial membrane antigen (EMA) and the common acute lymphoblastic leukaemia antigen (CALLA/CD10/neutral endopeptidase 24.11). Sorted luminal and myoepithelial cells displayed distinctively different morphologies when maintained in monolayer culture, differences which were enhanced by the addition of hydrocortisone, insulin and cholera toxin to the culture medium. The EMA-positive cells formed an attenuated monolayer with indistinct cell boundaries while CALLA-positive cells, by contrast, formed tightly packed arrays of refractile cells. The distribution of the cell type-specific markers cytokeratin 18 (luminal cells) and smooth muscle alpha-actin (myoepithelial cells) indicated that the sorted populations were approximately 98% pure. However, a significant minority (approximately 15%) of sorted luminal cells consistently expressed the basal-cell marker cytokeratin 14 in culture. A marked difference was noted in the proliferative behaviour of the two types of sorted cells, with myoepithelial cells dividing rapidly in response to the humoural additives, in contrast to the luminal cells which proliferated slowly. Both types of sorted cells could be cloned in the presence of feeder layers of mouse fibroblasts. Clones of luminal and myoepithelial cells were also distinctive; all "spread" luminal clones were similar in appearance to each other, although some cellular heterogeneity, including squamous metaplasia, was observed in "compact" myoepithelial clones. Both types were shown to have retained their original surface markers and to exhibit different cytoskeletal antigenic phenotypes when they were re-analysed after a 3-week growth period. Both spread and compact phenotypes were obtained when separately isolated ducts and alveoli were cloned. This detailed characterization of cells isolated from the human breast epithelium by flow cytometry provides the basis for further studies of luminalmyoepithelial interactions and growth responses of purified cell types in vitro.

The microanatomical and histological appearance of the human breast has been studied during puberty. The macroscopic architecture of the mammary epithelial tree was identified and correlated with the histological appearance of material excised from defined regions of the breast preparations. Between ages 13 yrs and 15 yrs the human breast shows evidence of ductal elongation and branching, with lobules formed by lateral and dichotomous branching. The majority of ducts are lined by a two-layered epithelium consisting of recognisable myoepithelial and luminal cells. Less-well-defined multilayered regions were also observed in some areas, apparently at the site of lateral branching or early lobular development.

A new cell line Rat mammary (Rama) 900 was isolated from the ascitic version of the SMT-2A metastasising rat mammary tumour by stepwise adaptation of the tumour cells to tissue culture. The cells grew mainly as loosely-adherent aggregates, and were dependent during the first 18 passages in vitro on a feeder layer of mesothelial-like cells (Rama 950) obtained from the same tumour. Subcutaneous injection of Rama 900 cells in fat pads of syngeneic Wistar Furth rats yielded anaplastic primary tumours and extensive, gross metastases including those in lungs, lymph nodes, liver and bones, similar to the parental transplantable tumour. The extent of metastatic spread from subcutaneous fat pads was increased by passage 17 in vitro for the Rama 900 cells. A similar extent of metastatic spread was achieved at earlier times by injecting the original cells with the non-tumorigenic Rama 950 cells in vivo. Subcutaneous injection of Rama 900 into thymectomised rats or MF1 nu/nu mice yielded fewer tumours, most of which regressed. No metastases occurred in the thymectomised rats and fewer metastases, mainly in lungs but not in lymph nodes, livers or bones, were seen in the nude mice. The ascitic tumours formed by intraperitoneal injection of nude mice contained both anaplastic rat cells similar to Rama 900 and mouse mesothelial-like cells similar to Rama 950. Although these anaplastic ascites cells failed to yield any tumours in syngeneic or thymectomised rats, they still produced tumours and metastases, including those in lymph nodes, in nude mice.

Human embryonal carcinoma cells sometimes display the developmental potential of early embryonic stem cells. While available data do not clearly identify a counterpart of these tumor cells in normal development, previous comparisons of human embryonal carcinoma and yolk sac carcinomas indicated that these cell types are closely related, and suggested that embryonal carcinoma cells might resemble the progenitors of extraembryonic endoderm. To analyse further cell-differentiation lineage in these tumors, we produced monoclonal antibodies to cytostructurally associated antigens of human embryonal carcinoma cells. Spleen cells from mice immunized with a detergent-insoluble extract of cultured human embryonal carcinoma cells were fused to NS-1 myeloma cells, and hybridoma supernatants were screened by indirect immunofluorescence on the immunizing cell line, then on a panel of cell lines derived from human embryonal carcinomas, yolk sac carcinomas, and a range of neoplastic and normal tissues. Monoclonal antibody GCTM-1 stained the nuclei of all human cells tested and served as a positive control; this antibody immunoprecipitated proteins of 85 and 66 k Da from human embryonal carcinoma cells. GCTM-2 recognized an epitope on a 200-k Da extracellular protein present on the surface of embryonal carcinoma cells, and stained the surface of visceral yolk sac-type carcinoma and colorectal carcinoma cells as well. Enzymatic analysis of carbohydrate residues on the GCTM-2 antigen revealed that it was a keratan sulphate proteoglycan, and suggested that the epitope recognized by the antibody lies on the core protein. In immunoblots, antibody GCTM-3 bound to a 57-k Da cytoskeletal protein expressed in human embryonal carcinoma. This antibody decorated filamentous arrays in cell lines from human embryonal carcinoma, visceral yolk sac carcinoma, parietal yolk sac carcinoma (endodermal sinus tumour), and adenocarcinoma and large cell carcinoma of the lung. Antibody GCTM-4 recognized a determinant present on a 69-k Da polypeptide, associated with a component of the lysosomal compartment, which was expressed in embryonal carcinoma cells, but no other cell type tested. The results with this antibody panel thus allow distinction between human embryonal carcinoma and yolk sac carcinoma, but provide further evidence of a close relationship between these cell types.

Fragments of ductal and lobular epithelium ('organoids') produced by collagenase digestion of reduction-mammoplasty specimens were injected into athymic 'nude' mice. These heterospecific tissues were accepted without rejection, and the presence of xenografts was confirmed by cytology, immunocytochemistry and chromatin staining. Lactation, as confirmed by immunocytochemical and ultrastructural criteria, was observed in the grafted human epithelium during murine pregnancy at both intra- and extra-mammary sites.

Fragments of human breast epithelium, devoid of all stromal and basal lamina components, which maintain their in vivo topological organisation can be cultured for up to 28 days within a reconstituted rat-tail-derived collagen matrix. These organoids initially undergo a loss of structural and 3-dimensional organisation, typified by loss of lumina formed by epithelial cells, and myosin from myoepithelial cells. Their subsequent reorganisation is dependent on the presence of serum, insulin, hydrocortisone, and cholera toxin in tissue culture medium. After this preliminary phase, a reduction in the concentration of serum, insulin, hydrocortisone, and cholera toxin is necessary to allow the structural differentiation of epithelial and myoepithelial cells. The myoepithelial cells also regain their ability to produce the basal lamina component laminin. The use of bovine-dermal collagen as the matrix, rather than rat-tail-derived collagen is shown to result in more stable organisation and differentiation of the organoids. The successful use of single-cell pellets (derived by trypsinisation of the organoids) in place of organoids in such cultures illustrates that there is no requirement for pre-existing cell/ cell contact or topological organisation of cells prior to embedding within the collagen matrix.

The myoepithelial-type cell line, Rama 712, derived from a normal rat mammary gland, deposits an extracellular matrix containing type-IV collagen and other basement membrane proteins round its cellular periphery. After transformation with a temperature-sensitive mutant of Rous sarcoma virus (tsRSV) the cells fail to deposit an extracellular matrix at the permissive temperature (35 degrees C), but retain the capacity to do so at the non-permissive temperature (41 degrees C). The synthesis of type-IV collagen is not affected by the temperature shift. Rama 712 cells fail to form tumours in syngeneic rats. However, Rama 712-tsRSV cells form tumours that are locally invasive but fail to metastasize. In histological sections, the tumour cells stain with an antibody to type-IV collagen, but do not deposit any extracellular type-IV collagen. Cells isolated from the tumours (Rama 712T) remain temperature-sensitive for the extracellular deposition of type-IV collagen when grown in vitro. Rama 712, Rama 712-tsRSV and Rama 712T fail to produce any detectable type-I or type-IV collagenase at either 35 degrees C or 41 degrees C. These results show that in this system extracellular deposits of basement membrane proteins are lost from invasive tumours produced by myoepithelial-type cells by mechanisms other than those due to the production of collagenolytic enzymes.