UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

In vitro phenotypic and functional characterization of epithelial progenitors present in the normal adult… Stingl, John 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2000-487229.pdf [ 7.81MB ]
Metadata
JSON: 831-1.0099500.json
JSON-LD: 831-1.0099500-ld.json
RDF/XML (Pretty): 831-1.0099500-rdf.xml
RDF/JSON: 831-1.0099500-rdf.json
Turtle: 831-1.0099500-turtle.txt
N-Triples: 831-1.0099500-rdf-ntriples.txt
Original Record: 831-1.0099500-source.json
Full Text
831-1.0099500-fulltext.txt
Citation
831-1.0099500.ris

Full Text

IN VITRO PHENOTYPIC AND FUNCTIONAL CHARACTERIZATION OF EPITHELIAL PROGENITORS PRESENT IN THE NORMAL ADULT HUMAN BREAST by John Stingl B.Sc, The University of British Columbia, 1988 M.Sc, The University of British Columbia, 1993 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in Department of Anatomy We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 2000 ©John Stingl In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Department of Date DE-6 (2/88) ABSTRACT The developmental relationships between the different epithelial cell lineages in the normal human mammary gland are not well defined. To characterize the progenitor activity of freshly isolated human breast epithelial cells (HBEC), 2- and 3-dimensional culture systems were developed that optimized their clonal growth, and fluorescence-activated cell sorting was used to characterize the progenitors of the different types of colonies obtained. These studies identified one type of progenitor cell that generates colonies of varying sizes with an alveolar-like morphology and generally contains cells expressing markers of luminal cells; i.e.: the apical glycoprotein MUC1 and epithelial cell adhesion molecule. They also express relatively high levels of erbB-2, retain rhodamine 123, express variable levels of the epidermal growth factor receptor and CD44v6 and express low levels of the ct6 integrin, the histo-blood group antigen type 2 and the common acute lymphoblastic leukemia antigen. A second type of progenitor identified is one that generates colonies composed of cells that express myoepithelial lineage markers and variable numbers of cells that express luminal cell markers. These progenitors generate colonies in collagen gels that have a branching duct-like morphology. These bipotent progenitors express the epidermal growth factor receptor, erbB-2, epithelial cell adhesion molecule, the oc6 integrin, the common acute lymphoblastic leukemia antigen, variable levels of MUC1, and retain low levels of rhodamine 123. Single cell cultures verified the clonal origin of both types of colonies. Epidermal growth factor at 10 ng/ml within the breast epithelial colony assay promoted the survival of cells enriched for luminal cell progenitors, as well as the migration of keratin 14 expressing cells in a manner that could be interpreted as ductal elongation. Serial passaging of highly enriched populations of bipotent progenitors indicated that some cells produced in the mixed and pure myoepithelial marker* ii colonies can generate colonies composed only of cells expressing myoepithelial characteristics, but not luminal characteristics. These latter results suggest that the culture conditions used compromise both the self-renewal of human breast epithelial stem cells and their differentiation into the luminal lineage. iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iv LIST OF FIGURES vii LIST OF TABLES xi LIST OF ABBREVIATIONS xii ACKNOWLEDGMENTS xv FORWARD xvi CHAPTER I: GENERAL INTRODUCTION 1 The Problem 1 The Human Mammary Gland 3 Characterization of the Cell Lineages in Vivo 4 Development of the Human Mammary Gland 5 Evidence for Stem Cells in the Nonmalignant Mammary Gland 8 Factors that Control Mammary Gland Development 11 Breast Cancer and Stem Cells 14 Normal Breast Epithelial Cell Culture 16 Malignant Breast Epithelial Cell Culture 19 General Rationale 20 CHAPTER II: GENERAL METHODS 22 Dissociation of Normal Human Mammary Tissue and Initiation of Primary Cultures 22 Fluorescence Activated Cell Sorting 23 Cell Culture 25 iv Collagen Gel Culture 26 Immunocytochemistry 26 MTT Assay 28 Statistical Analysis 29 CHAPTER IH: DEVELOPMENT OF AN IN VITRO HUMAN BREAST EPITHELIAL CELL COLONY ASSAY 30 Introduction 30 Experiment 1: Can Single HBECs Form Colonies in Vitro? 31 Methods 31 Results 32 Discussion 32 Experiment 2: Development of a 2-D HBEC Colony Assay 33 Methods 34 Results 37 Discussion 39 Experiment 3: Characterization Colonies Generated in the Presence of NIH 3T3 Feeders 40 Methods 41 Results 44 Discussion 48 Experiment 4: Development of a 3-D Collagen Gel HBEC Colony Assay 52 Methods 53 Results 54 Discussion 57 v CHAPTER IV: CHARACTERIZATION OF HUMAN BREAST EPITHELIAL CELL PROGENITORS 131 Introduction 131 Experiment 1: Characterization of HBEC Progenitors 131 Methods 132 Results 135 ' Discussion 143 Experiment 2: Influence of EGF and NDF on HBEC Precursors 147 Methods 148 Results 149 Discussion 150 Experiment 3: Do Mixed/Myoepithelial Colonies Contain Luminal Cell-Restricted Progenitors? 151 Methods 151 Results 152 Discussion 153 CHAPTER V: GENERAL DISCUSSION 176 General Discussion 176 Future Directions 181 REFERENCES 188 vi LIST OF FIGURES Page # Figure 1: Influence of HMF, M2-10B4 and BMS on HBEC colony formation. 60 Figure 2: Influence of FfMF, M2-10B4 and BMS feeders on average HBEC 62 colony size. Figure 3: HBEC colony phenotypes observed in culture in the presence of HMF feeders. 64 Figure 4: Distribution of colony phenotypes observed in HBEC cultures maintained in the presence of HMF feeders. 66 Figure 5: Influence of HMF and NIH 3T3 feeders on HBEC colony formation. 68 Figure 6: HBEC colony morphologies observed in single cell cloning experiments. 70 Figure 7: Range of HBEC colonies observed in culture in the presence of NTH 3T3 feeders. 72 Figure 8: Expression of K8/18, K19, MUC1 and Ep-CAM in colonies composed of closely arranged cells. 74 Figure 9: Expression of 5E11 and 4H1 epitopes in colonies composed of closely arranged cells. 76 Figure 10: Expression of CD44v6 in colonies composed of closely arranged cells. 78 Figure 11: Lack of expression of K14 in colonies composed of closely arranged cells. 80 Figure 12: Lack of expression of BGA in colonies composed of closely arranged cells. 82 V l l Figure 13: Expression of K14 in a colony composed of closely arranged cells. 84 Figure 14: Expression of K6 in colonies composed of closely arranged cells. 86 Figure 15: Expression of K8/18 in colonies of a mixed phenotype. 88 Figure 16: Expression of K19 and MUC1 in colonies of a mixed phenotype. 90 Figure 17: Expression of K14 in colonies of a mixed phenotype. 92 Figure 18: Expression of K8/18 in mixed and dispersed cell colonies. 94 Figure 19: Expression of BGA in mixed and dispersed cell colonies. 96 Figure 20: Expression of K6 in mixed and dispersed cell colonies. 98 Figure 21: Heterogeneous expression of Ep-CAM in mixed colonies. 100 Figure 22: Expression of the 5E11 and 4H1 epitopes in mixed colonies. 102 Figure 23: Expression of the Ki-67 cell proliferation antigen in mixed colonies. 104 Figure 24: Expression of MUC1 and K14 in a colony generated from a single cell. 106 Figure 25: Expression of BGA and K8/18 in colonies composed of retractile dispersed cells. 108 Figure 26: Expression of CD44v6 in HBEC colonies. 110 Figure 27: Expression of Ep-CAM and the 5E11 and 4H1 epitopes in normal human mammary tissue. 112 Figure 28: Inhibition of 5E11 antibody binding by 4H1 antibody. 114 Figure 29: Squamous metaplastic differentiation within mixed/myoepithelial 116 cell colonies. Figure 30: Influence of EGF on mixed/myoepithelial cell-restricted colony morphology. 118 viii Figure 31: Expression of K6 in HBEC colonies containing elevated masses of cells. 120 Figure 32: Expression of collagen IV in dispersed cell cultures. 122 Figure 33: Relationship between seeding density and colony formation in collagen gels. 124 Figure 34: Gross and ultrastructural characteristics of small spherical colonies in collagen gel culture. 126 Figure 35: Expression of K14 and CD44v6 in small spherical colonies in collagen gel culture. 128 Figure 36: Gross and ultrastructural characteristics of large branched colonies in collagen gel culture. 130 Figure 37: General protocol for characterizing HBEC progenitors. 155 Figure 38: Dot plots showing the pattern of expression of HMFG-2 and Ep-CAM in relation to CALL A in non-cultured and cultured mammary tissue. 157 Figure 39: Distribution of colony phenotypes among HBEC progenitors expressing varying levels of MUC1 and CALLA. 159 Figure 40: Colony phenotypes generated by progenitors expressing varying levels of MUC1 and CALLA. 161 Figure 41: Expression of Ep-CAM and retention of Rhl23 among luminal cell-restricted and mixed/myoepithelial cell-restricted progenitors. 163 Figure 42: Expression of erbB-2 and Ep-CAM among luminal cell-restricted and mixed/myoepithelial cell-restricted progenitors. 165 Figure 43: Expression of EGFR and Ep-CAM among luminal cell-restricted and mixed/myoepithelial cell-restricted progenitors. 167 Figure 44: Expression of BGA and Ep-CAM among luminal cell-restricted and mixed/myoepithelial cell-restricted progenitors. 169 Figure 45: Expression of a6 integrin and CALL A among luminal cell-restricted and mixed/myoepithelial cell-restricted progenitors. 171 Figure 46: Influence of NDF on luminal cell-restricted and mixed/myoepithelial cell-restricted progenitors. 173 Figure 47: Influence of EGF on luminal cell-restricted and mixed/myoepithelial cell-restricted progenitors. 175 Figure 48: Growth and differentiation potentials of HBEC progenitors. 185 Figure 49: Hypothetical relationship between different HBEC progenitors and differentiated progeny. 187 x LIST OF TABLES Page # Table 1: Expression of luminal and myoepithelial cell markers in closely arranged, mixed and dispersed cell colonies. 45 Table 2: Progenitor content of 19 different HBEC primary cultures. 142 Table 3: Frequency of progenitors types among first, second and third passage cultures. 152 xi LIST OF ABBREVIATIONS ct6, A6 - <x6 integrin A:M - acetone:methanol (1:1) ANOVA - analysis of variance APAAP - alkaline phosphatase anti-alkaline phosphatase BGA - histo-blood group antigen type 2 BMS - bone marrow stroma BSA - bovine serum albumin CALLA - common acute lymphoblastic leukemia antigen cAMP - cyclic 3', 5' adenosine monophosphate CDM - chemically defined medium CE - cloning efficiency CT - cholera toxin D - dimensional DNAse - deoxyribonuclease E2 - estrogen ECM - extracellular matrix EDTA - ethylenediaminetetraacetic acid EGF - epidermal growth factor EGFR - epidermal growth factor receptor EMA - epithelial membrane antigen Ep-CAM - epithelial cell adhesion molecule ER - estrogen receptor ESA - epithelial specific antigen xii FACS - fluorescence activated cell sorting; F12/DME/H - Hams F12/Dulbecco's modified Eagle's medium/Hepes FBS - fetal bovine serum FGF2 - fibroblast growth factor 2 / basic fibroblast growth factor FITC - fluorescein isothiocyanate h - hour HBEC - human breast epithelial cells HC - hydrocortisone HFN - Hanks balanced salt solution supplemented with 0.02% sodium azide and 2% FBS HGF - hepatocyte growth factor HMF - human mammary fibroblasts Ig - immunoglobulin IGF - insulin-like growth factor INS - insulin K6 - keratin 6 K8/18-keratins 8/18 K14 - keratin 14 K19 - keratin 19 L - luminal cell-restricted progenitor M - mixed/myoepithelial cell-restricted progenitor MMTV - mouse mammary tumor virus MTT - l-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide MUC1 -MUC1 glycoprotein NDF - neu differentiation factor xiii P - progesterone PBS - phosphate buffered saline PE - phycoerythrin PI - propidium iodide PR - progesterone receptor PRL - prolactin PTHrP - parathyroid hormone related protein rbc - red blood cell Rhl23 - rhodamine 123 SF - serum-free medium SLC - small light cell SMA - smooth muscle actin STV - saline-trypsin-versene TDLU - terminal ductal lobular unit TEB - terminal end bud TGF-a - transforming growth factor-a TGF-P - transforming growth factor-J} tris - tris hydroxymethyl amino methane hydrochloride w - weak xiv ACKNOWLEDGEMENTS I would like to thank my supervisor, Dr. Joanne Emerman and my co-supervisor Dr. Connie Eaves for their support and guidance that they have shown me over the years. I would also like to thank the members of my supervisory committee, Drs. Nelly Auersperg, Bruce Crawford and Cai Roskelley for their insightful advice. Particular thanks is extended to Darcy Wilkinson for putting up with me and for her conscientious help, to Dianne Reid for helping me out at the TFL, and to Drs. Wayne Vogl and Elaine Humphries for sharing their technical skills. I am also deeply indebted to Gayle Thornbury for performing the FACSorting and for her matchmaking skills. I would also like to extend my gratitude to Mr. Roman Babicki, The British Columbia Medical Services Foundation, British Columbia Foundation for Non-Animal Research and the Department of Anatomy for their financial support during this thesis. XV FORWARD Portions of this thesis have been previously published or submitted for publication as follows: Emerman J.T., Stingl J., Petersen A., Shpall E.J. and Eaves C.J. (1996) Selective growth of freshly isolated human breast epithelial cells cultured at low concentrations in the presence or absence of bone marrow cells. Breast Cancer Research and Treatment 41: 147-159. Stingl J., Eaves C.J., Kuusk U. and Emerman J.T. (1998) Phenotypic and functional characterization in vitro of a multipotent epithelial cell present in the normal adult human breast. Differentiation 63: 201 -213. Stingl J., Eaves C.J. and Emerman J.T. (2000) Characterization of normal human breast epithelial cell subpopulations isolated by fluorescence-activated cell sorting and their clonogenic growth in vitro. Journal of Mammary Gland Biology and Neoplasia (in press). Stingl J., Zandieh I., Eaves C.J. and Emerman J.T. (2000) Characterization of normal adult human breast epithelial progenitor cells. Development (submitted). xvi CHAPTER I: GENERAL INTRODUCTION The Problem The probability of a woman in North America being diagnosed with breast cancer at some time in her life is approximately 12 %, and the lifetime probability of dying of this disease is 3-4 % (Feur et al 1993). Although early detection procedures have increased survival of women with early disease, there has been relatively modest progress in the treatment of advanced disease. This lack of success is due to many problems. One of them is the incredible complexity of the disease. Many genes have been implicated either in the initiation or progression of breast cancer, and considering that only a fraction of the genes present in the genome have been identified, this number will likely increase further. Another problem related to our lack of success is the deficiency of a suitable in vitro model for studying human breast cancer. Aside from epidemiological studies to identify risk factors, the most valuable tools to date for studying human breast cancer have been the use of animal models and human and rodent cell lines in culture. Most of the current knowledge regarding mammary gland development and the factors that regulate it was generated using the mouse and rat as the model under study. Current genetic manipulation of these animals in which specific genes can be overexpressed in a tissue specific manner, or mutated to result in loss of function are very powerful techniques for identifying genes important in normal development and carcinogenesis. Although these studies have provided valuable information, the use of rodents always raises potential differences with the human species. Immortalized cell lines isolated from cultured normal mammary tissue and from malignant breast cells have been extremely valuable in analyzing intracellular pathways involved in breast cell responses to various factors. Unfortunately, these cell lines do not reflect the heterogeneity present within tumors or between tumors of different individuals. l Methodologies in which cells isolated from normal human mammary tissue can be propagated in vitro were initially developed in the late 1970's and early 1980's (Taylor-Papadimitriou et al 1977, Gaffney et al 1976, Papadimitriou et al 1977b, Stampfer et al 1980). Following this, numerous studies were done in which cultures were initiated using biopsy and mastectomy material in an attempt to study the behaviour of malignant human breast epithelial cells (HBEC; Smith et al 1981, Taylor-Papadimitriou et al 1989, Emerman and Wilkinson 1990). However, in 1993 Ethier reported it was the normal HBEC component of tumor samples that were the predominant cells in many cultures initiated with mixed populations of normal and malignant cells (Ethier et al 1993). This observation was supported by the paradoxical fact that pure populations of breast tumor cells (isolated from pleural effusions) are extremely difficult to propagate in vitro (reviewed in Ethier 1996). Typical success rates for these cultures is approximately 10 % (Cailleau et al 1974). Furthermore, the cell type that predominates in cultures initiated with biopsy and mastectomy specimens typically expressed markers characteristic of the myoepithelial lineage. There are two general lineages of cells present within human mammary epithelium, the luminal cells, of which a subset synthesizes milk proteins, and the underlying myoepithelial cells. The majority (-80 %) of mammary tumors exhibit characteristics of luminal epithelial cells (Rudland et al 1993, Bartek et al 1985a, Heatley et al 1995, Sloane and Ormerod 1981, Delahaye et al 1997, Daikee et al 1987). The emergence of cells expressing myoepithelial markers in human mammary cultures was particularly evident in experiments in which HBEC cultures were passaged. This promotes the loss of cells that express luminal markers and the selection of cells that express myoepithelial markers. Another difficulty in the human mammary cell culture field has been the inability to distinguish the type of HBEC being maintained in vitro. It was not until the identification of lineage-specific proteins and the development of antibodies that are specific for them that the heterogeneity of 2 cells within breast cell cultures was recognized (Bartek et al 1985a, Taylor-Papadimitriou et al 1989, Dairkee et al 1986, Karsten et al 1993). Even at this date, many authors use the over-simplistic categorization of HBEC as being luminal or myoepithelial. As will be explained in this thesis, heterogeneity exists even within these populations. The development of in vitro assays that monitor the growth and differentiation capabilities of primitive and mature hematopoietic progenitors has been key to the analysis of hematopoiesis. To address the lack of a suitable HBEC culture system, it was the objective of this thesis to develop a HBEC colony-forming cell assay that might draw from principles established in the hematopoietic field. Such an assay could then address some of the basic questions that remain unanswered human breast cell development: What are all the different subtypes of HBEC, and which ones have proliferative potential? When HBEC divide to form a colony, are the types of cells produced the same or are they composed of different cell types? And most importantly, what is the frequency, phenotypic features and lineage potential(s) of mammary stem cells? The Human Mammary Gland The human mammary epithelium is a compound tubuloalveolar gland. Milk is synthesized within the differentiated acini of the gland, and the secretory product is carried in a series of ramified ducts to the nipple (Osborne 1991). The smallest ducts, the intralobular ducts, together with clusters of alveoli, form the lobules or terminal ductal lobular units (TDLUs). The mammary parenchyma is composed of two general lineages of epithelial cells. Luminal epithelial cells line the lumen of the ducts and alveoli of the mammary tree and, underlying these, are the flattened smooth muscle-like myoepithelial cells. Another cell type present in the mammary gland is the basal clear (chief) cell. Unfortunately, the biological properties of these morphologically defined cells is not clear. They have been postulated to represent stem cells (Smith and Medina 1988, Smith et al 1990), although some consider them to represent myoepithelial cell precursors (Sapino et al 1990). In the ducts, the myoepithelial cells form a continuous sheath, whereas in the region of the alveoli, their distribution is much more sparse as the cytoplasmic processes of the cells embrace the luminal cells in a basket-like fashion (Emerman and Vogl 1986). In humans, the mammary parenchyma is typically embedded within a fibrous collagenous stroma rich in mammary fibroblasts. Also present within the stroma are mast cells, plasma cells and, more peripherally, adipocytes. A basement membrane separates the epithelial compartment from the stromal compartment. Characterization of the Cell Lineages in Vivo Although HBEC lineages are classified in the scientific literature as being luminal and myoepithelial, these terms are likely to be over-simplistic for describing the heterogeneous cell types present within the human mammary epithelium. To circumnavigate this, the terms luminal and myoepithelial will be reserved to describe cells in vivo. Cells in vitro will be described based on their expression of a variety of marker proteins since this permits a more precise description of the cells under study. Smooth muscle actin (SMA)(Skalli et al 1986, Gugliotta et al 1988), keratin 14 (K14)(Dairkee et al 1985, Taylor-Papadimitriou et al 1989), common acute lymphoblastic leukemia antigen (CALLA/CD10)(Gusterson et al 1986), ct6 integrin (Koukoulis et al 1991) and the intermediate filament vimentin (Guelstein et al 1988) are generally localized within the basal cell layer of the normal adult mammary epithelium. Exceptions to this include expression of K14 in luminal cells of the large ducts (Purkis et al 1990) and expression of vimentin in fibroblasts (Moll et al 1982). The most commonly used markers to identify cells of the luminal compartment include keratin 18 (K18)(Taylor Papadimitriou et al 1989), keratin 19 (K19) (Bartek et 1985 a & b, Taylor Papadimitriou et al 1989) and MUC1 (Burchell et al 1983, Ceriani et al 1977). Heterogeneous expression of K18 (Rudland and Hughes 1989) and K19 4 (Bartek et al 1985a&b, Bartek et al 1990) within the cells of the luminal compartment has been reported. These K19" luminal cells are positioned as single cells and small clusters within the small ducts and within the TDLUs of the gland (Bartek et al 1985, Bartek et al 1990). MUC1 is a highly glycosylated apical plasma membrane protein characterized by variable numbers of tandem repeats (reviewed in Patton et al 1995). Immunocytochemical studies have demonstrated considerable heterogeneity between MUC1 epitopes expressed on different luminal cells in vivo (Burchell et al 1983, Edwards and Brooks 1984, Edwards et al 1986). Another marker that is not routinely used to identify cell lineages, but was examined herein is the epithelial cell adhesion molecule (EpCAM). EpCAM, which is also known as epithelial specific antigen (ESA)(Litvinov et al 1994) or EGP-40 (Simon et al 1990), is a homophilic Ca2+-independent cell adhesion molecule specific for most epithelial cells (Momburg et al 1987, Latza et al 1990). When normal resting adult mammary tissue is stained for EpCAM, expression is found to be localized primarily to the basal and lateral cell membranes of luminal epithelial cells (Latza 1990). Stromal cells and mesothelial cells do not express this protein (Momburg et al 1987, Latza 1990). Development of the Human Mammary Gland Human mammary gland development begins as bilateral ingrowths of solid buds of epithelial cells from ventral surface ectoderm into the underlying mesenchyme during the fifth developmental week (Kellokumpu-Lehrinen et al 1987). The cells of these epithelial buds are uniformly keratin 19+ and do not express keratin 18 or the myoepithelial markers SMA, vimentin or keratin 14 (Anbazhagan 1998). As the epithelial bud penetrates the underlying mesenchyme, the mesenchymal cells begin to differentiate into mammary fibroblasts which in turn synthesize a collagenous stroma. In males, testosterone induces condensation of mammary mesenchyme around the stalk of the epithelial downgrowth which results in the destruction of the mammary 5 bud (Kratochwil and Schwartz 1976, Wasner et al 1983). In females, the penetrating downgrowths branch to form 15-25 solid chords of epithelial tissue. By 20 weeks, these epithelial downgrowths have canalized to form a rudimentary system of ducts composed of both luminal and basal cell layers (Tobon and Salazar 1974, Anbazhagan et al 1998). The most distal dilated tips of the extending ducts are composed of multiple layers of epithelial cells and are referred to as terminal end buds (TEBs)(Russo and Russo 1987). Initially these primitive ducts show little branching, but over time the TEBs cleft and divide to promote formation of a primitive mammary tree. As development continues, there is increased extension of the ductal tree and development of small lobules via double layered outpocketings referred to as alveolar buds (Russo and Russo 1987, Anbazhagan 1991). By the third trimester of pregnancy, the mammary epithelium begins to become secretory under the influence of maternal hormones (Russo and Russo 1987, Anbazhagan 1991). At the time of birth, the mammary gland is a heterogeneous mix of mammary ducts at various stages of development, with marked variability in ductal and lobuloalveolar development observed among tissues isolated from different specimens (Anbazhagan 1991). Ultrastructurally, the gland at this state is composed of K19+luminal cells and SMA + and vimentin+ basal cells (Anbazhagan 1998). Expression of keratin 14 in these basal cells is strongest in the large ducts and weakest in the distal portions of the gland, whereas a reciprocal pattern of expression is observed for vimentin. This differentiated state is maintained for approximately 1-2 months after birth, then the gland undergoes involution to regress back to its primitive ductal/lobuloalveolar state and remains so until puberty. At puberty with the onset of the menstrual cycle there is an increase in fatty and collagenous stroma, ductal elongation, lateral and dichotomous branching and lobule formation (Monaghan et 6 al 1990). Most lobules formed at this stage are composed of a terminal duct and approximately 10 alveolar buds (type I lobule) (Russo and Russo 1987). The lobule is embedded within a loose irregular connective tissue matrix, whereas the matrix in between different TDLU's is of a more dense type (Cardiff and Wellings 1999). Most of the mammary tree is composed of the typical luminal and myoepithelial cell layers, but regions of multilayering are observed in TEBS (Russo and Russo 1987) and at sites of lateral branching and lobule formation (Monaghan et al 1990). As the individual approaches the fourth decade of life, each ovulatory cycle promotes the formation of new budding structures in the present lobules (Vorherr 1974) such that the lobules become more complex (type II lobules) in that the alveolar buds have cleaved into small ductules resulting in more sac-like structures associated with each terminal duct (Russo and Russo 1987, Cardiff and Wellings 1999). During pregnancy, there is further ductal and lobuloalveolar development at the distal aspects of the mammary tree such that much of the stroma becomes filled with epithelium (Tobon and Salazar 1975). Most of this glandular proliferation occurs during the first trimester of pregnancy, whereas the latter stages are associated with the synthesis of secretory products. The main secretory products in human milk are caseins, a-lactalbumin, lactoferrin, immunoglobulins, lactose, fatty acids, vitamins and lipids (reviewed in Neville et al 1983). Following pregnancy, there is complete involution of the secretory apparatus via apoptosis (Strange et al 1992). Resting mammary glands in parous women have on average more type III lobules than nulliparous women (types I and II)(Russo et al 1992). However, some type I lobules are still present following pregnancy and have been hypothesized to represent a reserve of stem cells for further pregnancies (Dawson 1935). At menopause, there is involution of the mammary gland such that the lobules become atrophied and the ducts dilated (Cardiff and Wellings 1999). 7 Evidence for Stem Cells in the Nonmalignant Mammary Gland In this thesis, the term stem cell is defined as that cell, upon transplantation into an epithelium-free mammary gland, can renew an entire functional normal mammary epithelium (except for the connection to the nipple) as well as renew itself, with behaviour of these stem cells responsive to all the appropriate normal physiological control stimuli. The term multipotent progenitor is here defined as a cell that can generate more than one type of progeny, whereas bipotent progenitor is defined as a cell that can generate two types of progeny. Stem cells and progenitors can be differentiated from one another by the relatively restricted proliferation potential of the progenitors. The existence of stem cells in the human mammary gland has not been unequivocally demonstrated, although a pool of regenerative cells must be present to account for the cyclical epithelial cell proliferation (and subsequent apoptosis) observed during the menstrual cycle (Anderson et al 1982) and the massive alveologenesis (and subsequent involution) associated with pregnancy. Smith and colleagues, using retrovirally marked mouse mammary epithelial cells injected at limiting dilutions into cleared mouse mammary fat pads, have demonstrated the existence of alveolar-restricted and ductal-restricted progenitors as well as mammary stem cells (Smith 1996, Kordon and Smith 1988). Stem cells are known to be distributed throughout the resting mammary tree since transplantation of fragments from throughout the tree can generate the mammary epithelium (Hoshino et al 1964, Daniel et al 1968, Smith and Medina 1988) with the highest frequency of regenerative cells in the end buds and the lowest frequency in lactating alveoli (Smith and Medina 1988). The concept of ductal and alveolar-restricted progenitors was previously suggested by experiments in which fragments of murine mammary glands were serially transplanted into non-pregnant and pregnant hosts (Daniel et al 1968, Daniel et al 1971). In these studies it was observed that ductal development decreased with mitotic age of the transplant cells and not the chronological age of the host. However, if the 8 host injected with serial transplanted cells became pregnant, alveologenesis still occurred regardless of poor ductal development. This indicates that the cells responsible for alveolar development are distinct from those responsible for ductal development. Unfortunately, the phenotypic profile of these different progenitors is not known. Numerous studies have been performed to identify proliferating cells within both the rodent (Dulbecco et al 1982, Smith and Medina 1988, Sapino et al 1990, Chepko and Smith 1997, Zeps et al 1998) and human (Ferguson et al 1985, Ferguson 1988, Joshi et al 1986, Perusinghe et al 1991) mammary epithelial trees. A general consensus is that cell division is predominantly distributed within the luminal cell population, although cell division within the basal cell subpopulation is also observed. As well, cells exhibiting characteristics of differentiated myoepithelial cells have very low proliferation rates. The hierarchy of cell proliferation in the terminal structures of the human mammary tree is as follows: TEBs > type I lobule > type II lobule = terminal duct > type III lobule (Russo and Russo 1987). Morphological candidates that have been previously proposed to represent mammary stem cells include the cap cells that line the TEBs of the elongating ducts (Williams and Daniel 1983, Rudland 1991, reviewed in Rudland et al 1997), the small light cell (SLC) present in the rodent mammary gland (Chepko and Smith 1997), basal clear cells (Smith and Medina 1988, Smith et al 1990) and K19' luminal epithelial cells (Bartek et al 1990). The cap cells, which are only present in the developing ductal tree, are characterized as having a phenotype intermediate between luminal and myoepithelial cells in that they express both luminal and myoepithelial related proteins at low levels (e.g. vimentin, SMA and MUC1). Although these cells have been postulated by some investigators to represent the stem cells of the rodent (Williams and Daniel 1983, Ormerod and Rudland 1984) and human (Rudland 1991, Rudland et al 1997) mammary gland, others have suggested that this cell is merely a myoepithelial progenitor that paves the 9 way for ductal elongation during development (Sapino et al 1993). Chepko and Smith performed an exhaustive electron microscopic study of the resting mouse and rat mammary epithelium and identified three division-competent, structurally distinct cell populations (Chepko and Smith 1997). Based on identification of mitotic pairings and morphological intermediates, they have proposed a hierarchy of proliferation and differentiation from a stem cell (small light cell - SLC) to differentiated secretory and myoepithelial cells. Another cell speculated to represent a stem cell is the basal clear cell (Smith and Medina 1988), which corresponds to one of the immature intermediates of Chepko's hierarchy. The basal clear cell is distributed throughout the mammary epithelium at all stages of mammary development and differentiation, and is particularly predominant during the luteal phase of the menstrual cycle (Vogel et al 1981). In the mouse, these cells express K14 and as well as keratin 6 (K6), the latter being localized within the mammary epithelium only in proliferating cells (Smith et al 1990). However, the conclusion that these cells represent a stem cell population is controversial since an electron microscopy study of the murine mammary epithelium has concluded that these cells are a myoepithelial precursor (Smith et al 1984). Luminally positioned keratin 19' epithelial cells have also been postulated to represent a stem cell population in the resting adult human breast (Bartek 1990). They have been postulated to represent a stem cell compartment since they are relatively abundant in small ducts and in lobules (which are regions where cell proliferation would occur during pregnancy) and they do not secrete casein during pregnancy. Unfortunately morphological studies examining cap cells, SLC, basal clear cells and K19" luminal cells are limited by the fact that they do not link morphology with proliferative and differentiative capacity. Combining immunocytochemistry with autoradiography can provide information regarding which cells are proliferating at a given time, but it does not allow for elucidation of what type of progeny is generated or the ultimate fate of these cells. 10 Factors that Control Mammary Gland Development Mammary gland development in the embryonic and fetal stage, at least in rodents, appears to be independent of systemic hormones estrogen (E2) and progesterone (P) (reviewed in Imagawa et al 1994). Instead, this early development is dependent on close range interactions between the developing parenchyma and the underlying mesenchyme. One factor implicated in these interactions is parathyroid hormone-related protein (PTHrP). PTHrP is synthesized and secreted by the epithelial cells of the mammary buds, and the receptors for this ligand are located on the underlying mammary mesenchyme (Dunbar et al 1998, Wysolmerski et al 1998). In male mice, PTHrP secretion by the epithelium induces androgen receptor upregulation in the mammary mesenchyme (Dunbar and Wysolmerski 1999), which in turn causes destruction of the mammary bud. In females, PTHrP is essential for ductal bud ingrowth into the mesenchyme since PTHrP KO mice do not progress past the mammary bud stage (Wysolmerski et al 1998). The mechanism on how the stroma regulates epithelial morphogenesis in response to PTHrP is not known. Many growth factors and hormones have been implicated in the control of pubertal mammary gland development. The main players that will be briefly introduced here are E 2 , P, epidermal growth factor (EGF), insulin-like growth factors (IGFs), hepatocyte growth factor (HGF), neu differentiation factor/heregulin (NDF), prolactin (PRL), and transforming growth factor-P (TGF-P). Of all the hormones implicated in mammary gland development, none has received more attention than E2. This is due to the fact that peripubertal mammary development coincides with menarche and that there is a strong link between estrogen exposure and breast cancer (Kelsey and Berkowitz 1988). There are 2 receptors for E 2 , estrogen receptor-a (ERa) and estrogen receptor-P (ERJ3). ERP is not implicated in mammary gland development (at least not in the 1 1 mouse) since ERB KO mice display normal mammary gland development (Krege et al 1998). Cells expressing ERa are more frequently located within the lobules as compared to the interlobular ducts (Petersen et al 1987). More specifically, ERa has the highest frequency in type I lobules and lowest in type III lobules (Russo et al 1999). Exposure to E2 at puberty is associated with ductal elongation, although this is likely through an indirect mechanism mediated via the stroma (Daniel et al 1987, Haslam and Nummy 1992, Cunha et al 1997, Walden et al 1998). The current model on how E2 exerts its influence is that E2 binds to stromal cells adjacent to the TEB of the developing duct and promotes the release of a paracrine factor which then promotes proliferation and migration of the cells of the TEB. Possible candidates for this paracrine factor are EGF (Coleman et al 1988, Snedecker et al 1992, Wiesen et al 1998), transforming growth factor-a (TGF-a)(Snedecker et al 1992), IGF-1 (Walden et al 1998) and hepatocyte growth factor (HGF) (Yang et al 1995). The main factors regulating lobuloalveolar development appear to be P, PRL and NDF. Estrogen also has a role, but it may be limited in upregulating progesterone receptors (PR) in ER + luminal epithelial cells in sexually mature virgin mice (Haslam et al 1988) and by stimulating PRL release from the pituitary (Chen and Meites 1970, Lieberman et al 1978). The dual requirement of P and E2 for (maximal) HBEC proliferation is supported by the observation that cell proliferation in vivo is maximal during the luteal phase of the menstrual cycle (Masters et al 1977, Anderson et al 1982, Going et al 1988). In the human mammary gland, ER and PR colocalize within the same cells, but these ER+/PR+ cells are also non-proliferating cells (Clarke et al 1997, Russo et al 1999). Exposure of the developing mammary gland to P is associated with ductal branching and lobuloalveolar development since PR knock-out mice display long ducts with little branching and no alveologenesis (Lydon et al 1996, Brisken et al 1998). Conversely, mice overexpressing PR display extensive side branching (Shyamala et al 1998). The effect of 12 PR is also thought to be via paracrine interactions since the decreased lobuloalveolar development in PR KO mice can be rescued by mixing with PR+ cells (Brisken et al 1998). The nature of this paracrine factor is not known. Prolactin also plays an important role in mammary gland morphogenesis since this hormone is implicated in ductal side branching (via the stroma) and alveologenesis directly via the epithelium (Brisken et al 1999). In vitro studies using murine cells have demonstrated that prolactin, insulin, glucocorticoids, an intact basement membrane (laminin) and appropriate cell shape are the minimum requirements for the induction of milk protein synthesis (Emerman and Pitelka 1977, Streuli et al 1991, Roskelley et al 1994). Ductal branching, lobuloalveologenesis and induction of casein synthesis can also be induced by NDF (Yang et al 1995, Jones et al 1996). Transforming growth factor 0 is a negative regulator of epithelial cell proliferation (Valverius et al 1989). TGF-f3 (TGF-01 and TGF-p3) can be localized to both the luminal and myoepithelial cells of the mouse mammary epithelium and sequestered in the fibrous extracellular matrix surrounding growth quiescent mammary ducts (Silberstein et al 1990, Robinson et al 1991). This factor is involved in inhibition of ductal budding (Silberstein et al 1992), ductal elongation (Silberstein and Daniel 1987, Kordon et al 1995), lobuloalveologenesis (Gorska et al 1998) and is thought to be a key regulator in ensuring evenly spaced mammary ductal branching (reviewed in Daniel et al 1996). A mechanism by which this growth inhibition is regulated involves changes in the amount of extracellular matrix (ECM) surrounding growth-inhibited epithelial (Stampfer et al 1993, Daniel et al 1996). Virgin transgenic mice expressing mutant TGF-P type II receptors under control of the MMTV promoter exhibit inappropriate alveolar development and casein synthesis (Gorska et al 1998), a phenomenon similar to that observed in mice overexpressing the matrix metalloproteinase stromelysin-1 (Sympson et al 1994, Witty et al 1995). 13 The above list of hormones and growth factors is by no means complete. Many others are involved, including: fibroblast growth factors (Gomm et al 1997a, reviewed in Rudland et al 1997), activins/inhibins (Robinson and Hennighausen 1997), epimorphin (Hirai et al 1998), colony stimulating factor-1 (reviewed in Sapi and Kacinski 1999) and undoubtedly others yet to be discovered. Breast Cancer and Stem Cells Cancer theory suggests that it is the mammary stem cells, or their immediate progeny that are the cells in which the oncogenic process is initiated. Considering that multiple genetic mutations are required to transform a normal cell into a malignant epithelial cell with metastatic potential, the initial target cell must be one with a relatively high proliferative potential to permit the expansion of the cell pool to accommodate the low frequency of subsequent genetic mutations. With regards to human breast cancer, correlative data is available to date to support the concept of stem cells, or their proliferative competent progeny as the targets for malignancy. Although breast cancer is usually associated with adult women, scientific evidence supports the notion that events early in life influence predisposition to breast cancer. Trichopoulos has hypothesized that maternal hormones in pregnancy influence the probability of daughters having breast cancer (Trichopoulos 1990). Aside from familial history, the strongest link to breast cancer is exposure to E2 (Kelsey and Berkowitz 1988). Trichopoulos's hypothesis is supported by the observation women with breast cancer have atypical cerebral asymmetry significantly more often than women with no history of breast cancer (Sanderson et al 1992). Atypical cerebral asymmetry is thought to be linked to altered intrauterine hormone concentrations (Geschwind and Galaburda 1985). The variation in the degree of development of the human mammary epithelial tree at birth (Anbazhagan et al 1991) may be due to variations in the hormonal environment in utero (Anbazhagan and Gusterson 1992, Anbazhagan and Gusterson 1994). Unfortunately, no 14 correlative data is available linking degree of mammary gland development at birth with breast cancer risk in humans. Animal studies have demonstrated a direct link between degree of mammary gland development and susceptibility to mammary tumor formation after exposure to a carcinogen (Russo and Russo 1980). Mammary cancer risk is highest in animals where the presence of TEBs (proliferative cells) was maximal (e.g., in young virgin animals). Somewhat analogous to this are the studies examining the exposure of Japanese women to radiation during the atomic bombing of Nagasaki and Hiroshima (Tokunaga et al 1979, Tokunaga et al 1987). These studies have demonstrated that radiogenic breast cancer risk decreases with increasing age at exposure, with irradiation prior to menarche conferring the greatest risk. There are several mechanisms that may explain why early events could play a role in susceptibility to breast cancer risk. One is that risk for carcinogenesis is proportional to the number of stem cells in the pool and their rate of proliferation (Russo and Russo 1980, Albanes et al 1988). Another is that if there is an in utero mutation, exposure of the altered cell to agents that cause this cell to proliferate will expand the pool of mutated cells. X-chromosome inactivation analysis of normal human breast tissue has demonstrated that contiguous regions of a breast (e.g., a large duct and its associated lobules) contain the progeny of a single progenitor (Tsai et al 1996). Conceivably, one of these cells or one of its progeny could expand and provide a large pool of cells susceptible for the next step in multi-step breast tumor progression (Fujii et al 1996). In the adult human gland, most breast tumors become evident in the region of the TDLLTs (Wellings 1975). This indicates the cells that are the targets for later steps in tumor progression are located at this site. Breast cancer risk correlates positively with the number of menstrual cycles a woman will experience in her lifetime. Not surprisingly, the cyclical menstrual cycle associated HBEC proliferation occurs in the TDLU's of the mammary tree, with 15 maximal proliferation during the luteal (maximal E 2 + P) phase of the cycle (Anderson et al 1982, Longacre and Bartow 1986, Potten et al 1988). The concept that mammary stem cells phenotypically resemble those of the luminal lineage rather than a myoepithelial lineage is supported by the fact that the vast majority of human mammary tumors exhibit the luminal cell markers K18 (Rudland et al 1993), K19 (Bartek et al 1985a, Heatley et al 1995), MUC1 (Sloane and Ormerod 1981) and EpCAM (Delahaye et al 1997), but typically do not express the myoepithelial specific proteins SMA (Bocker et al 1992b) and K14 (Dairkee et al 1987, Heatley et al 1995). This concept also is supported by the study of the rat mammary tumor cell line Rama 25, a line that expresses luminal cell specific proteins. This line also expresses stem cell characteristics in that it can synthesize casein in response to differentiation agents (Rudland et al 1983) and generate sublines that exhibit a gradation of characteristics ranging from luminal to myoepithelial (Rudland et al 1986, Rudland and Hughes 1991, reviewed in Rudland 1997). Two similar SV-40 transformed human mammary cell lines that also express luminal characteristics (SVE3 and Huma 7) and exhibit stem cell behaviour have also been described (Rudland et al 1989). Unlike the Rama 25 cell line, however, no immunoreactive casein has been detected when cultured under lactogenic conditions. Normal Breast Epithelial Cell Culture Numerous in vitro experimental model systems have been developed to study normal mammary gland development and differentiation (reviewed in Ip and Darcy 1996 and in Ethier 1996). When normal human mammary epithelial cells isolated from reduction mammoplasties or from human milk are cultured on tissue culture plastic, good epithelial cell proliferation is observed (Stampfer et al 1980, Smith et al 1981, Petersen et al 1987, Emerman and Wilkinson 1990, Ethier et al 1993). However, the overall epithelial organization in these cultures obviously does 16 not resemble that observed in vivo (e.g., acini/lumen formation, organization of a myoepithelium between luminal epithelial cells and the basement membrane). Mammary epithelial cells, even those obtained from midpregnant mice, maintained in 2-dimensional (2-D) cultures do not polarize, deposit a basement membrane or synthesize milk proteins in response to lactogenic hormones (Emerman and Pitelka 1977, Tonelli and Sorof 1982, Streuli and Bissell 1990, Close et al 1995). It has become apparent that epithelial-stromal interactions and cellular shape/architecture are key regulators of gene function (reviewed in Bissell et al 1999). To promote a greater degree of differentiation, a variety of 3-dimensional (3-D) culture systems have been developed to mimic the physiological environment. These culture systems either rely on mammary epithelial cells maintained on (Emerman and Pitelka 1977) or in (Flynn et al 1982, Tonelli and Sorof 1982) a floating collagen gel, or in a reconstituted laminin matrix (Barcellos-Hoff 1989, Darcy et al 1991, Streuli et al 1991). Mouse mammary epithelial cells, obtained from either virgin (Tonelli and Sorof 1982, Darcy et al 1995) or mid-pregnant rodents (Emerman et al 1977), can be induced to synthesize milk proteins when maintained in 3-D collagen or laminin matrices in the presence of the lactogenic hormones PRL, INS and glucocorticoids (or lineolic acid instead of glucocorticoids - Levay Young et al 1987). To date, no milk protein synthesis has been demonstrated by HBEC cultured in a basement membrane or collagen matrix, although cultured luminal H B E C will deposit a basally located basement membrane, express sialomucins apically and form small (~8 cells) growth arrested acini (Petersen et al 1992). Other than the small acinar structures described by Petersen and colleagues (Petersen et al 1992), no good organotypic ductal growth from singly isolated HBEC has been demonstrated using reconstituted basement membranes or collagen gels. Even in experiments where the mammary epithelium is recapitulated, the starting material was epithelial organoids (Darcy et al 1991) or the structures generated represent the re-aggregation rather than the proliferation of H B E C (Gomm et al 1997b). 17 A seemingly endless combination of growth factors and hormones have been tested over the years to determine the optimal combination for maximal normal mammary epithelial cell growth and differentiation. With regards to differential growth effects on cells expressing characterisitics of luminal and myoepithelial cells, few studies can be interpreted since little distinction was made between the different mammary epithelial cell subtypes. More recent research that recognizes mammary epithelial cell heterogeneity is revealing that HBEC with varying morphologies have distinct growth requirements in vitro (Petersen and van Deurs 1988, Ethier et al 1993, Gomm et al 1997b, Kao et al 1997). Commonly used media have compositions that range from being chemically defined and relatively simple (Emerman and Wilkinson 1990), chemically defined but complex (Petersen and van Deurs 1987, Gomm et al 1997b, Pechoux et al 1999), chemically defined except for the presence of bovine pituitary extract (Hammond et al 1984), to serum containing media (O'Hare et al 1991, Ethier et al 1993). With the exception of the very complex chemically defined medium (CDM)5 (Gomm et al 1997b) and the related CDM6 (Pechoux et al 1999), serum-free media appear to be most suitable for the selection and proliferation of cells that express myoepithelial markers (Ethier et al 1993, Gomm et al 1997b, Pechoux et al 1999). In fact, the only essential additive required for the proliferation of these cells (at non-clonal densities) is insulin (Petersen and van Deurs 1988, Gomm et al 1997b). This relative lack of exogenous growth factor dependence may be explained by the observation that myoepithelial cells secrete TGF-a, a known mitogen for these cells (Smith et al 1989). As well, expression of smooth muscle actin by these dispersed cells is inversely correlated with their growth state (Petersen and van Deurs 1988). Survival and expansion of cells expressing luminal cell characteristics is most dependent on the presence of fetal bovine serum (FBS) (Ethier et al 1993, Gomm et al 1997), although good proliferation of these cells is also observed when maintained in CDM5 (Gomm et al 1997b), CDM6 (Pechoux et al 1999) or in a low calcium 1 8 medium (Berthon et al 1992). The incorporation of EGF (Gomm et al 1997b) and HGF (Pechoux et al 1999) are key ingredients in promoting survival and proliferation of cells expressing characteristics of luminal HBEC. Fetal bovine serum is not used on a routine basis because it promotes the proliferation of mammary fibroblasts that will eventually dominate the cultures (Emerman and Wilkinson 1990). Other additives commonly included in media to promote the proliferation of normal mammary epithelial cells include cholera toxin, hydrocortisone, insulin, transferrin and bovine serum albumin (BSA). Cholera toxin activates adenylate cyclase (Lad et al 1980) and promotes the proliferation of mammary epithelial cells (Taylor-Papadimitriou et al 1980, Yang et al 1980). Hydrocortisone (Cortisol) is the prototypic glucocorticoid involved directly and indirectly with most metabolic processes, particularly in influencing the responsiveness of cells to other hormones (Tepperman and Tepperman 1987). Insulin is required for general anabolic processes within the cell, however the concentrations commonly used in vitro (1-5 ug/ml) are also mitogenic via interaction with the IGF I receptors (King et al 1980). Iron is essential for a wide variety of metabolic processes, but is insoluble in aqueous solutions (reviewed in Ponka et al 1998). Transferrin is an iron binding protein that is commonly included in culture media Bovine serum albumin (BSA) is included in the culture media because it can act as a carrier protein for insoluble agents (e.g., glucocorticoids and fatty acids)(Tepperman and Tepperman 1987). Malignant breast epithelial cell culture Malignant HBEC are notoriously difficult to grow and to distinguish from contaminating nonmalignant HBEC in vitro (reviewed in Ethier 1996). At first, this may appear paradoxical, since malignant cells are thought to be proliferating continuously and in a deregulated manner in vivo. However, one has to consider that most cell culture media have been designed to maximize normal mammary epithelial cell proliferation. As well, primary human mammary tumors in vivo 19 have tumor doubling times that range from 44 to 1869 days (mean = 212 days)(Fournier et al 1980). Assuming an exponential growth rate, this equates to approximately an average of 20 years to grow from an single cell to a tumor 2 cm 3 (Fournier et al 1980). Analysis of normal and malignant cell growth rates in vitro also support the concept that tumor cells (as identified by reduced neotetrazolium reductase activity - see below) have relatively slow doubling times (>120 h) compared to normal cells expressing luminal markers (= 48 h) and myoepithelial markers (= 24 h) maintained under the same conditions (Petersen and van Deurs 1987). Several novel systems have been reported to promote the selective growth of malignant H B E C . These include the use of a reconstituted basement membrane (Petersen et al 1992), culture conditions that simulate the microenvironment of breast tumors (Bergstraesser and Weitzman 1993, Dairkee et al 1995) and optimization of other culture parameters (Pandis et al 1992). General Rationale The characteristics of mammary stem cells, the developmental relationship between the different H B E C lineages and the heterogeneity o f cell types that may exist within the lineages is not well known. A s well , the identification and propagation of malignant H B E C in vitro, particularly in the presence o f normal H B E C , is difficult to perform on a routine basis. A s a first step in solving these problems, three specific objectives are proposed: 1. Development of an in vitro HBEC colony assay. The objective o f these studies was to identify culture conditions that promote the growth of single isolated normal H B E C progenitors. Variables examined included culture substrates (2D plastic vs 3-D collagen gel) and the influence of different feeder layers. 20 2. Characterization of HBEC colonies A next objective was to characterize any colonies generated in the HBEC colony with regards to gross colony morphology and expression by the cells produced a panel of epithelial and HBEC lineage-specific proteins. Such experiments would determine if HBEC clonogens are lineage-restricted or not and the types of structures that might be generated in vitro (e.g., ductal vs alveolar). 3. Characterization of HBEC progenitors A final objective was to begin to characterize the different types of HBEC clonogens that might be detected in vitro in terms of the lineage markers and growth factor receptors they express, and their ability to retain metabolic dyes and how these features might correlate with the types of colonies they produce. Developmental inter-relationships between the different progenitors and the influence of the growth factors EGF and NDF would then be examined. Besides characterizing progenitors, correlation of expression of cell markers with colony-forming properties provides a basis for developing strategies to purify specific HBEC progenitor populations for future studies. These specific objectives would not necessarily identify stem cells in the human breast since this would require an in vivo model. However, the experimental approaches pursued in this thesis were considered an important first step to understand the numbers and types of HBEC progenitors that exist and to develop methodologies that would allow their isolation at high purities. Once these objectives are obtained, candidate stem cell populations could then be assayed in a yet to be tested in vivo system for true stem cell activity. 21 CHAPTER II: GENERAL METHODS This section describes the methods used in most of the experiments described. Specific additions or changes to this general methodology are described in the Methods section of each study. Dissociation of Normal Human Mammary Tissue and Initiation of Primary Cultures Normal tissue from reduction mammoplasties were transported from the operating room on ice in Ham's F12/Dulbecco's modified Eagle's medium [F12/DME; 1:1 (v:v); StemCell Technologies, Vancouver, B.C., Canada] supplemented with 10 mM Hepes (H; Sigma Chemical Co., St. Louis, Mo., U.S.A.) and 5% calf serum (Gibco Laboratories, Grand Island, N.Y., U.S.A.). Upon delivery, the tissue was aseptically transferred to sterile petri dishes, minced with scalpels and dissociated in F12/DME/H supplemented with 2% bovine serum albumin (BSA, Fraction V; Gibco Laboratories, Grand Island, N.Y.), 5 ug/ml insulin, 0.5 ug/ml hydrocortisone (HC), 10 ng/ml cholera toxin (CT), 300 U/ml collagenase and 100 U/ml hyaluronidase (all from Sigma) at 37 °C for 18 h in a Bellco dissociation flask (Bellco Glass, Inc., Vineland, New Jersey, U.S.A.). After the dissociation, the epithelial cell rich-pellet was collected by centrifuging the cell suspension at 80 g for 4 min followed by one wash with F12/DME/H. The supernatant from the first centrifugation was saved to serve as a source of human mammary fibroblasts (HMF) to use as feeder layers (see "Preparation of Feeders"). Typically, a large red blood cell (rbc) pellet cosedimented with the epithelial cell pellet of the first centrifugation. Unless the pellet was frozen in liquid nitrogen (in which the rbc do not survive) for later use, the rbc were selectively lysed to decrease their frequency in HBEC primary cultures. Rbc were lysed by a 15 min incubation at 37°C of the rbc/epithelial cell pellet in 2 ml of primary culture medium (F12/DME/H supplemented with 1 mg/ml BSA, 1 ug/ml INS, 0.5 ug/ml HC and 10 ng/ml CT) supplemented with 5% FBS in which 9 ml of 0.155 M NH4CI and 1 ml of 0.170 M tris hydroxymethyl amino methane hydrochloride (tris; Fisher Scientific, Vancouver, B.C.) at pH 22 7.65 was added. The suspension was then centrifuged at 100 g for 5 min and the supernatant discarded. The rbc-depleted cell epithelial pellet was then resuspended in primary culture medium supplemented with 5% FBS and aliquoted into 25 cm2 or 75 cm2 tissue culture flasks at a subconfluent density (approximately 25%). After 24 h, the media was changed to primary culture medium and the cultures maintained for a further 2 days in vitro before harvesting the cells for analysis and/or sorting using fluorescence-activated cell sorting (FACS). This initial 3 day culture period allows non-adherent blood cells and debris to be washed away, while allowing the organoids to flatten out onto the tissue culture plastic and to become easier to harvest into a single cell suspension. Exposure to a minimal growth medium (no EGF or serum; serum is present only in the first 24 h to promote cell attachment) during this period reduces proliferation and HBEC lineage selection while maintaining cell viability. To passage HBEC, subconfluent cultures were treated with Ca 2 + and Mg2+-free saline solution containing 0.05% trypsin (Gibco) and ethylenediamine tetraacetic acid (EDTA; 0.025%; Sigma) (STV) for several minutes until all the cells have detached. An equal volume of primary culture medium supplemented with 5% FBS was then added and the suspension was centrifuged at 100 g for 5 min. The resulting pellet was then diluted in primary culture medium supplemented with 5% FBS and 10 ng/ml EGF and the suspension transferred to new culture vessels at a ratio of 1:3. After a 24 hour cell attachment period, the growth medium was changed to primary culture medium supplemented with 10 ng/ml EGF (now called serum free (SF) medium). Fluorescence Activated Cell Sorting. Two to six day old primary cultures of mammary tissue were incubated with prewarmed STV solution and gently agitated until all cells were released from the plastic substratum. An equal volume of Hank's balanced salt solution supplemented with 0.02% sodium azide (w/v; Sigma) and 2% FBS (HFN) was added to the STV solution and the cell suspension was centrifuged at 23 100 g for 5 min. After disposal of the supernatant, 1 ml of HFN, 100 ul of dispase (Collaborative Biomedical Products, Bedford, MD) and 50 ul of 1 mg/ml DNAse I (Sigma) were added to the epithelial cell pellet and the suspension warmed to 37 °C in a water bath for 5 min to promote disaggregation of any cell clumps. Following gentle pipetting to promote disaggregation of any cell clumps, the suspension was diluted with 10 ml of cold HFN and filtered through a 20 um mesh (BioDesign Inc., New York, NY). The filtrate was centrifuged at 100 g for 5 min and the resulting pellet incubated with 1 ml of 10% human serum (Red Cross) in HFN for 30 min to reduce non-specific antibody binding. The pellet was then aliquoted into (typically) four 12 x 75 mm polystyrene tubes. The first tube was stained with control antibodies, the second and third with single label antibodies (conjugated directly or indirectly to either R-phycoerythrin (PE) or fluorescein isothiocyanate (FITC)) and the fourth was double labelled with the selected sort markers. The single label cells were used for setting single color parameter compensation. Thirty minute incubations were used for all antibodies followed by two 3 ml HFN washes. On the last wash, the cells were diluted with 1 ug/ml of PI in 10 ml of HFN and the suspension filtered a second time through a 20 um mesh. PI was added to the last wash to allow identification and exclusion of PI+ (dead) cells from subsequent analysis or sorting procedures. The filtrate was then centrifuged and the resultant pellet resuspended with 300-600 ul of HFN supplemented with 0.1 mg/ml DNAse I. Viable (PI) cells were sorted by a FACStarPLUS (Becton Dickinson, Mountain View, CA) into 1.5 ml microfuge tubes containing SF medium supplemented with 10 % FBS. During all staining and sorting procedures, the cells were kept in the dark and maintained on ice. An important consideration for separating cells by flow cytometry is achievement of a single cell suspension since the cells must pass through a small orifice (70 um on the FACSTAR P L U S ) when 24 being sucked into the flow cytometer. As well, the purity of the sort is limited by the ability to disaggregate dissimilar cells. The protocols described herein relied on a short pre-sort culture period, treatment with a variety of enzymes (trypsin, dispase and DNAse), inclusion of the metabolic inhibitor sodium azide, two filtration steps through a 20 um mesh (one at the beginning of cell labelling and one immediately prior to cell sorting) and maintenance of the cells on ice in an attempt to generate single cell suspensions. As well, very large events (potential cell clumps) measured on the forward light scatter channel (a measure of cell size) on the flow cytometer were gated out of any further analysis. Cell Culture All the cell culture experiments described in this thesis, unless otherwise stated, involved seeding unsorted or FACS-sorted FfBEC isolated from primary cultures at clonal (1 cell/culture vessel) or low cell densities (< 800 cells/cm2). In experiments in which cells were seeded on glass coverslips (for immunostaining in which acetone was required as the fixative) or in 24- and 96-well plates, the substratum was first collagen-coated prior to seeding of the HBEC in an attempt to increase the cloning efficiency of the seeded population. The growth medium used during the initial attachment phase of all culture experiments, unless otherwise noted, was SF medium supplemented with 5% FBS. After 24-48 h, the medium was changed to SF medium and the cultures maintained for a further 7-12 days. Cells seeded in the smaller culture vessels received 1-3 media changes during the week, whereas those seeded in the larger 60 mm dishes did not receive any media changes. Cultures were terminated by fixation in -20°C acetone:methanol (A:M)(1:1, v/v). Following air drying, the cultures were rehydrated with H2O and stained with Wright-Giemsa to increase the contrast of HBEC colonies on the culture vessel. In experiments in which the cultured cells were to be analyzed by immunocytochemical staining, they were processed as described in the "Immunocytochemistry" section. 25 Collagen Gel Cultures For seeding cells in 3-D collagen gel cultures, unsorted or sorted HBEC from primary cultures were mixed at the desired density (typically 5 x 10^  -10 4 cells/ml of gel) with a solution of rat tail collagen dissolved in 0.1% acetic acid, 10X F12/DME/H (Sigma), NaOH and SF medium, growth factors and hormones (Richards et al 1983). Aliquots of 0.5 ml of this solution were then pipetted into individual wells of 24-well plates which had 0.3 ml of cell-free collagen gel. In some experiments, a feeder layer was seeded in the culture vessel prior to being overlaid with the collagen gel. Following gelling of the collagen, 0.5 ml of growth medium was added to the wells. The growth media used in the collagen gel experiments, unless otherwise indicated, was SF medium. The gels were also rimmed with a pipette tip to float the gels. Failure to do so resulted in the colonies having a spiked rather than a rounded appearance and the colonies growing as concentric swirls within the gel, presumably along tension lines within the matrix. All 3-D cultures were then incubated for 10 -14 days. Colonies were scored by first removing the gel from the well, gently detaching and discarding the cell-free collagen gel underlay and placing the HBEC-containing gel on a plastic grid. The number of colonies (>3 cells) present was then enumerated directly using a dissecting microscope. Immunocytochemistry Expression of epitopes by cultured HBEC were detected by a variety of immunocytochemical methods. Cells cultured on collagen-coated glass coverslips were fixed in acetone at -20 °C for 1 min and air dried prior to detection of epitope expression by immunostaining using the alkaline phosphatase anti-alkaline phosphatase (APAAP) technique (Johnston et al., 1992). Briefly, the samples were rehydrated with tris buffer (0.1 M tris in saline, pH 7.6), preblocked for 30 min with 5% human serum and 5% normal rabbit serum in dilution buffer (tris buffer supplemented 26 with 1% BSA (w/v) and 0.1% sodium azide) and incubated with the primary mouse antibody for 30 min. The preparations were then washed, incubated with the secondary antibody (a rabbit anti-mouse IgG (DAKO, Mississauga Ont, Canada) diluted at 1:200 with 40% human serum) for 30 min and washed again in tris buffer. After another wash, the APAAP complex (1:50, calf intestinal alkaline phosphatase and mouse monoclonal anti-alkaline phosphatase; DAKO) was added for 1 h, the samples were then washed again with tris buffer, followed by the addition of an alkaline phosphatase substrate solution containing New Fuchsin for 12-15 min and a final wash performed before mounting. Cells cultured on tissue culture plates were immunostained and visualized by one of two methods. The cells were either fixed in acetone:methanol (1:1) at -20 °C for 1 min and air dried prior to visualization using the APAAP technique as described above. Alternatively, fluorescence microscopy was used as the visualization method if epitope expression was difficult to detect by the alkaline phosphatase technique or double labelling was required. Briefly, the cells were fixed in -20 °C methanol for 5 min, rinsed with phosphate buffered saline (PBS) and stored in PBS at 4°C until staining day. Cells were preblocked with PBS dilution buffer (PBS supplemented with 10% human serum and 1% BSA) for 30 min and then incubated for a further 30 min with the primary mouse antibody of interest diluted in PBS dilution buffer. Following three washes with PBS, the cells were incubated with a goat anti-mouse antibody conjugated to PE. In some staining experiments, the cells were double labelled with a FITC-conjugated mouse monoclonal antibody (EMA-FITC) that recognizes the MUC1 glycoprotein. Cultures were not coverslipped, but were covered with PBS solution. Fluorescence was visualized using the LaserSharp imaging program (Bio-Rad Laboratories, Mississauga, Ontario, Canada) and a inverted Zeiss Axiovert fluorescence microscope (Carl Zeiss Canada, North York, Ontario, Canada). A 488 nm argon ion laser (Bio-Rad) was used as the excitation light source. To control 27 bleeding of the FITC emission on the PE detector in double labelling experiments, single labelled controls were first analyzed and the photodetectors adjusted accordingly. Collagen gels containing HBEC colonies were either snap frozen in liquid nitrogen-cooled isopentane and serial sectioned, or the gels were fixed overnight in 1% paraformaldehyde in PBS and then embedded in paraffin wax and serial sectioned. Some slides were selected for staining with hematoxylin and eosin, and the remainder were used for immunocytochemistry. Since only one epitope (MUC1) that was examined generated a strong signal in the wax sections, the wax sections were studied mainly for general morphology, whereas the cryosections were used primarily to detect expression of specific epitopes using the APAAP technique. Positive controls for all immunocytochemistry included sections of normal human mammary tissue, normal human skin, cultured breast epithelial cells or cytospins of breast epithelial cell lines or HBEC primary cultures, always fixed in the same manner as the test sample. Negative controls consisted of cells that had the primary antibody replaced with a 10% dilution of normal serum corresponding to the species in which the primary antibody was derived or an isotype matched control antibody at the same or higher concentration. Negative controls for all experiments presented in this thesis were always unstained. All photographic prints generated were scanned into digital format and the images manipulated in Photoshop 4 (Adobe Systems Incorporated, San Jose, CA) without altering the integrity of the data. MTT Assay Growth of HBEC cultured in 96-well plates was monitored using the colorimetric MTT assay (Mossman 1983). Powdered MTT (l-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) (Sigma) was dissolved at 1 mg/ml in warmed serum-free and phenol red-free 28 F12/DME/H (Sigma) at a concentration of 1 mg/ml. After filter sterilization through a 0.2 um filter, 100 ul of MTT solution were added to each well of a 96-well plate in which all the growth medium has been removed. Following a 5 h incubation period at 37°C, 100 ul of 20% formol in saline were added to each well. Following 30 min at room temperature, the MTT-formol solution was removed and 100 ul of anhydrous isopropanol (Fisher Scientific) were added to each well to dissolve any formazan crystals. After 30 min, the plates were gently agitated and the absorbance values of each well was determined at 540 nm using a 96-well microplate reader (Biotek Instruments, Winooski, Vermont). Statistical Analysis An overall analysis of variance (ANOVA) for the influence of different feeders and growth factors on HBEC was performed. Post-hoc analysis using Tukey's test was used to identify significant arms of the analysis. 29 CHAPTER IH: DEVELOPMENT OF AN IN VITRO HUMAN BREAST EPITHELIAL CELL COLONY ASSAY Introduction The development of a HBEC colony assay would be a useful tool in studying normal mammary gland biology since FACS analysis of phenotypically distinct HBEC subpopulations might then identify subtypes with proliferative potential. Immunocytochemical staining of daughter cells to detect HBEC lineage specific proteins could also identify the fate of such clonogens. Some of the desired features of an in vitro HBEC colony assay would be as follows: 1. Permits detection of all the different progenitors present within the input population. 2. Promotes the appropriate differentiation of milk producing alveolar cells, ductal luminal cells and myoepithelial cells. 3. Supports the survival of mammary epithelial stem cells. 4. Does not require the inclusion of serum or feeder cell layers, or other undefined agents within the culture environment. 5. Allows the identification of the different cell types produced based on their morphology without the additional need for immunostaining. 6. Is technically easy to perform. The features listed above represent the ideal situation. In reality, attainment of some or all of these features may not be, in the short term, realistic. The methodologies of the in vitro HBEC colony assay and the classification of the resultant colonies during the course of this thesis underwent continuous evolution and refinement. Every attempt has been made to present results in a logical manner, but there is some discontinuity between the methodologies used in different experiments. 3 0 Five general experiments were performed in the development of an in vitro HBEC colony assay: 1. Demonstration that single HBECs can form colonies in vitro. 2. Development of a 2-D HBEC-colony assay. 3. Characterization of the colonies generated in 2-D cultures. 4. Development of a 3-D HBEC colony assay. 5. Characterization of the colonies generated in 3-D (collagen gel) cultures. Experiment 1: Can Single HBECs Form Colonies in Vitro? In these experiments, cells isolated from first passage mammary cultures as well as freshly isolated HBEC were utilized. First passage HBEC were used initially because these offered a source of enriched HBEC suspensions (HMF do not passage well in SF medium) which made use of FBS in the colony assay cultures possible. Methods The cloning attachment of the FACStar^EUS w a s u s e c j to seed single viable HBEC isolated from HBEC cultures into individual collagen-coated flat bottomed wells of 96-well microtiter plates, each of which was preloaded with 100 ul of F12/DME/H supplemented with 5% FBS, 1 mg/ml BSA, 1 ug/ml INS, 0:5 ng/ml HC, 10 ng/ml CT and 10 ng/ml EGF. In 3 experiments used to assess cloning efficiency, first passage HBEC were cloned and maintained in SF medium supplemented with 5% FBS. For 10 cloning experiments, HBEC from primary cultures were cloned and maintained in serum-containing medium only for the first 48 h. After this initial attachment phase, the medium was switched to SF medium. In all experiments, growth medium was changed twice weekly. After 9-14 days in vitro, the plates were fixed in A:M and colonies (clusters of 3 or more cells) were scored directly under a microscope. In the cloning experiment using first passage material and serum containing medium, selected colonies were stained using 3 1 the pan keratin IgG antibody mixture AEI and AE3 (Woodcock-Mitchell et al. 1982)( ICN Biomedicals, Inc., Costa Mesa, CA) at a 1:400 dilution (2.5 ug/ml) in conjunction with the APAAP procedure to detect keratin containing cells. Results When HBEC isolated from primary cultures were seeded at 1 cell/well and cultured in SF medium, the mean frequency of colony formation was 2.4 ± 0.5% (range = 0.1 - 4.4%; 360 -960 cells cloned/sample; 10 samples tested). Although the frequency of wells containing more than 1 cell/well at the time of seeding was not recorded during these experiments, later cloning experiments (see below) revealed that the frequency of wells containing more than one cell 12 h post seeding was approximately 5.3 ± 1.6% (range = 2.5 - 9.4%; 180 - 240 wells seeded/sample; 4 samples analyzed). The mean size of these colonies was 18 ± 6 cells/colony (range = 2 - 200 cells; 149 colonies scored; 10 samples analyzed). Unfortunately, the low cloning efficiency observed with HBEC isolated from primary cultures is not statistically significant from this background rate. When HBEC isolated from first passage cultures were seeded at 1 cell/well and cultured in SF medium supplemented with 5% FBS, the mean frequency of colony formation increased to 16.8 ±3.2% (range - 10.4 - 20.8%; 144 - 192 cells cloned/sample; 3 samples analyzed). The mean colony size of these colonies was 6.3 ± 2.6 cells/colony (range = 2-50 cells; 90 colonies scored; 3 samples analyzed). Sixteen out of 27 (59%) selected colonies expressed cytokeratins as determined by immunocytochemistry. Discussion The results presented here demonstrate that HBEC seeded at clonal densities do not proliferate well, despite being maintained in a growth medium that supports good HBEC proliferation at 3 2 non-clonal densities (Emerman and Wilkinson 1990). Cloning of HBEC isolated from first passage normal breast cell cultures suggests that HBEC can form colonies in vitro. However, the single cell origin of these keratin expressing colonies remained to be unequivocally proven. It is not discernible if the higher cloning efficiency observed with the passaged HBEC is due to passaged starting material or to the presence of serum within the culture medium. Addition of serum to cultures initiated with freshly isolated HBEC is routinely avoided since it promotes the proliferation of fibroblasts that contaminate the initial seeding population. These contaminating fibroblasts will rapidly dominate the cultures making the morphological identification of HBEC difficult. Serum also causes epithelial cells to lose their typical morphology and become very "spikey" and somewhat fibroblast-like which further compromises their morphology. Unfortunately, this effect is also not reversible by switching the cultures back to a serum-free medium. Experiment 2: Development of a 2-D HBEC Colony Assay In an attempt to increase the cloning efficiency of single isolated HBEC, a variety of stromal derived feeder layers were examined for their ability to influence the clonogenic growth of HBEC seeded at low densities in vitro. The feeders layers that were tested were irradiated (growth arrested but viable) HMF, irradiated human fibroblast-containing adherent layers subcultured from primary cultures of normal human bone marrow (bone marrow stroma - BMS), the mouse bone marrow stromal cell line M2-104 (Lemoine et al., 1988) and the mouse embryonic fibroblast cell line NTH 3T3. Normal mammary fibroblasts have been demonstrated to influence mammary epithelial cell behaviour in vitro (Smith et al 1981, Haslam 1986). Bone marrow fibroblasts were an obvious cell type to evaluate considering the propensity of malignant HBEC to metastasize to the bone marrow. The M2-10B4 cells were examined because they promote the survival of hematopoietic progenitors (Lemoine et al 1988, Sutherland et al 1991). 33 Methods The HMF, BMS, M2-10B4 and NTH 3T3 feeders were prepared as follows: Passaged HMF: Freshly dissociated HMF in the supernatant from the first centrifugation of the freshly dissociated reduction mammoplasty specimens were pelleted by centrifugation at 100 g for 10 min. The pellet was resuspended in F12/DME/H supplemented with 5% FBS and 5 ug/ ml INS (HMF media) and seeded in 75 cm2 flasks. Human mammary fibroblast cultures were passaged at a ratio of 1:5 by treatment with STV. Harvested HMF in suspension at 106 cells/ml in HMF medium were irradiated at 2.5 x 103 to 5 x 103 cgy and were seeded at 5 x 103 to 104 cells/cm2 into the wells of 24 - and 96 - well plates or into 60 mm dishes. Prior to initiation of the experiment, the feeder was rinsed twice with prewarmed basic medium. M2-10B4 murine bone marrow stromal cell line: M2-10B4 cells were routinely passaged in RPMI + 10% FBS. Harvested M2-10B4 cells were irradiated at 8 x 103 cgy and seeded at 104 cells/cm2 into 24 - well plates in RPMI + 10% FBS. Prior to initiation of the experiment, the feeder was rinsed twice with prewarmed basic medium. Passaged human bone marrow stroma (BMS): Human fibroblast-containing adherent layers subcultured from a primary culture of normal human bone marrow cells (Eaves et al., 1991) were harvested and irradiated at 1.5 x 103 cgy and seeded at 104 cells/cm2 in human long-term culture medium (enriched a medium supplemented with 12.5% FBS, 12.5% horse serum, 10"4 M mercaptoethanol and IO"4 Solucortef; StemCell). Prior to initiation of the experiment, the feeder was rinsed twice with prewarmed basic media. 34 NIH 3T3 mouse embryonic fibroblasts: NTH 3T3 cells (ATCC) were maintained and passaged as per HMF. Harvested NTH 3T3 cells at 106 cells/ml HMF media were irradiated at 5000 cgy and were seeded at 5 x 103 cells/cm2 into 60 mm dishes in either HMF media. The first set of feeder experiments examined the influence of HMF, BMS and M2-10B4 cells on HBEC cloning efficiency. In these experiments, freshly dissociated cells isolated from 3 reduction mammoplasties were incubated with pre-warmed STV solution supplemented with 5 U/ml dispase for 4 min at room temperature to disaggregate cell clumps. Following washing and resuspension with HMF medium, the cell suspension was filtered through a 20 um mesh. The cells in the filtrate were then counted by trypan blue dye exclusion on a hemocytometer and seeded at 7.5 x 102 and 1.5 x 103 cells/cm2 in SF medium supplemented with 5% FBS in the absence or presence of irradiated HMF, BMS or M2-10B4 cells. After 24 h, the medium was replaced with fresh SF medium and the culture maintained for a further 4 days. At the end of the culture period, the cultures were fixed with A:M, stained with Wright-Giemsa and the colonies scored based on their morphology. In an attempt to determine if single HBEC could form colonies that contained more than one HBEC lineage in vitro, progenitor-enriched populations ( C A L L A ^ p C A M * - see Chapter IV) were isolated from 6 primary cultures and deposited using the FACS single cell deposition unit at 1 cell/well into 96-well plates preseeded with HMF feeders in SF medium supplemented with 5% FBS. The growth medium for the first 48 h was SF medium supplemented with 5% FBS, and after this initial attachment phase, the medium was switched to SF medium. Thereafter, the medium was changed 3 times per week. After a further 8 days, the colonies were fixed and scored under a microscope. Since the HMF feeders were cultured from reduction mammoplasty preparations and the frequency of keratin (AE1/AE3+) expressing cells in feeder suspensions is 35 1:500 to 1:1000, it is not inconceivable that some HBEC progenitors could be present within the feeder when seeded. To take into account any "background" colony formation, 6 x 102 wells were seeded with feeders alone and maintained and scored as per the cloning plates. A second set of feeder experiments examined the influence of HMF and NTH 3T3 cells on HBEC cloning efficiency. In these experiments unsorted and sorted cells isolated from 6 primary cultures of normal breast tissue were seeded at 100 cells/cm in SF medium supplemented with 5% FBS into 60 mm (21 cm2) dishes in the absence or presence of irradiated HMF or NTH 3T3 cells. Forty-eight hours later, the medium was changed to SF medium and the cultures maintained for a further 6 to 9 days with no media changes. Following colony formation, the cultures were fixed in A:M and stained with Wright-Giemsa and the colonies scored based on their morphology. In the final 4 cloning experiments, (EpCAM1"; see Chapter IV), HBEC were isolated from primary cultures and deposited using the FACS at 1 cell/well in 96-well plates in SF medium supplemented with 5% FBS. Half of these plates were preseeded with irradiated NTH 3T3 cells. Nine hours later, those plates that were not preseeded with NTH 3T3 feeders were scored under a microscope to ensure that only 1 cell was deposited in every well. Wells containing more than one cell were first identified and excluded from further analysis. Irradiated NTH 3T3 feeders were then added in SF medium supplemented with 5% FBS. Twenty-four hours later, the medium was changed to SF medium and the culture maintained for 7 days. In all experiments, colonies (clusters of 3 or more cells) were scored directly under a microscope after 9-14 days. 36 Results Colony yields from freshly dissociated HBEC on the HMF, BMS and M2-10B4 feeders tended to be higher than in the controls but, except for HMF, these differences were not sufficiently large or reproducible to be statistically significant. In the case of HMF, a 3.4 ± 0.8 -fold stimulation of HBEC colony formation was observed by comparison to cultures without feeders (Figure 1). However, as shown in Figure 2, the growth-promoting effects of all 3 types of fibroblast-containing feeders was evident from the size of the colonies attained by 5 days. Colonies were typically composed of 20 - 30 cells in the presence of the feeders after 5 days, whereas those in the absence of feeders contained <10 cells. When the 5-day old cultures maintained on HMF feeders were examined on the basis of the colony morphologies seen, 3 general types of colonies were observed. Colonies had morphologies ranging from being composed of entirely closely arranged cells, entirely of loosely highly refractile cells or mixtures of both (Figures 3 A and B). Note that the dispersed cells are located peripheral to the more closely arranged cells (Figure 3B). The distribution of the colony morphologies in the 3 cultures maintained on HMF is shown in Figure 4. Based on the strong and consistent growth promoting effects of HMF on HBEC proliferation, HMF were later utilized as a feeder in HBEC cloning experiments with particular emphasis on generating colonies exhibiting mixed characteristics. Cloning of enriched HBEC progenitor populations on HMF resulted in 17 + 6.1% of the seeded cells forming a colony (range = 9.0 -47.2%; 120 - 360 cells cloned/sample; 6 samples analyzed). No background HBEC colony formation was observed using the HMF feeder. Colony sizes were typically 50-300 cells/colony. With the exception of 3 colonies that consisted of closely arranged cells with indistinct cell borders, all colonies had morphologies consistent with myoepithelial-like (e.g., highly refractile and teardrop shaped) cells after 10 days in culture. 37 High frequencies of colony formation by HBEC could be obtained in the presence of HMF. However, as demonstrated by the large colonies composed of teardrop shaped cells generated in the single cell seeding experiments, the closely arranged cells of the mixed colonies of HBEC were not maintained under these conditions. To identify feeders that promote the production of the closely arranged indistinct cells in the mixed colonies, a comparison of HMF and NTH 3T3 feeders was performed. Colony yields from HBEC isolated from primary cultures on HMF and NTH 3T3 feeders were 46 ± 14 and 55 + 13 -fold higher than those maintained in the absence of feeders (4 primary cultures analyzed). A 3.7-fold increase in the frequency of colonies composed of only closely arranged cells was observed in the presence of NTH 3T3 feeders when compared to HMF feeders (6 primary cultures analyzed). Although the identification of cells being dispersed and teardrop shaped vs closely arranged with indistinct cell borders in the mixed colonies was based only on their morphology and arrangement, it is readily apparent that the frequency of colonies exhibiting mixed characteristics was much higher when the progenitors were cultured on NTH 3T3 feeders than on HMF feeders or in the absence of any feeders (Figure 5). When HBEC were seeded at 1 cell/well onto NTH 3T3 feeders, mean frequencies of colony formation of 10.9 + 2.3% (range = 5.18 - 16.9%; 420 - 540 cells cloned/sample; 4 samples analyzed) were observed. The morphology of the colonies generated in the presence of the NTH 3T3 feeders ranged from cells being closely arranged with one another (Figure 6A) to cells being loosely arranged (Figure 6B), as well as morphologies intermediate between these two extremes (Figure C). This latter type of colony is composed of dispersed loosely arranged cells and more non-refractile closely arranged cells (arrowheads). Colony sizes generated in the presence of feeders typically contain 20 to > 100 cells/colony. 38 Discussion All feeders examined influence HBEC progenitor activity either by increasing their cloning efficiency and/or by increasing the number of daughter cells each generates. Initial experiments comparing the clonogenic growth of HBEC maintained on HMF, BMS or M2-10B4 cells suggested that HMF would be the most suitable to use as feeders. Three types of colonies could be identified by colony morphology when maintained in the presence of HMF feeders. Some colonies were composed of cells with morphologies suggestive of luminal cells (i.e., closely arranged cells with indistinct cell borders); others generated colonies composed of cells with morphologies suggestive of myoepithelial cells (i.e., highly refractile, dispersed teardrop-shaped cells) (See Experiment 3 below for full characterization of the different colonies). Interestingly, clusters of cells exhibiting characteristics of both lineages were also observed at a low (1 - 7%) frequency in these experiments. The presence of such mixed clusters suggests the existence of bipotent progenitors, although the possibility that such clusters might represent coincident colonies of pure closely arranged and dispersed cells cannot be ruled out by these experiments. Unfortunately, single cell cloning onto irradiated HMF failed to resolve this question since all colonies obtained from such efforts were composed exclusively of either cells exhibiting a closely arranged morphology or a dispersed cellular morphology. When normal HBEC isolated from reduction mammoplasties are serial passaged in serum-free medium, there is a gradual loss of the closely arranged cells and a concurrent increase in the dispersed cell phenotype (Ethier et al 1993, and our own laboratory experience). This suggests that (a subset of) the closely arranged cells are generating the dispersed cells with concurrent exhaustion of the closely arranged cell phenotype. This concept is supported by two observations. The first is that the dispersed teardrop-shaped cells surround the closely arranged 39 cells in the mixed colonies (as shown in Figure 3B). This suggests that the former are derived from the latter. Secondly, mixed clusters of cells which are observed after 5 days in vitro (feeder experiments) are no longer seen after 10 days (cloning experiments). It is apparent that although the presence of HMF increases HBEC cloning efficiency, it also appears to accelerate the exhaustion of closely arranged HBEC and the formation of dispersed teardrop-shaped cells. This prompted the examination of NTH 3T3 cells as a potential feeder instead of HMF. As it turns out, NTH 3T3 cells promote the generation of the dispersed cells as well as causing a 3.7-fold increase in the cloning efficiency of progenitors that generate colonies containing only closely arranged cells. Furthermore, single cell cloning of HBEC on irradiated NTH 3T3 cells demonstrated the existence of progenitors that generate colonies containing only closely arranged cells as well as progenitors that generated colonies that, on a gross morphological level, appeared to be composed of mixed cell types (See Chapter IV for full characterization of these colonies). Evidence of a third type of progenitor, the dispersed cell-restricted progenitor was also obtained. This latter progenitor is suspected of being a more mature progenitor that is a descendant of bipotent progenitors that have lost their ability to generate closely arranged daughter cells. Experiment 3: Characterization of the Colonies Generated in the Presence of NTH 3T3 Feeders Colonies generated in vitro from HBEC progenitors were also examined for the presence of cells expressing a variety of luminal and myoepithelial cell-specific markers. This permitted a more rigorous analysis of how specific colony morphologies might correlate with the generation of cells of specific lineages. Aside from expression of the luminal (K8/18, K19, MUC1, EpCAM) and myoepithelial (K14 and SMA) markers briefly described earlier, colonies were characterized on the basis of expression of histo-blood group antigen type 2 (BGA), a splice variant of the hyaluronate receptor (CD44v6) and keratin 6 (K6). Histo-blood group antigen type 2 has been 40 previously identified as a reliable marker for basally derived cells in vitro (Karsten et al 1993), whereas expression of CD44v6 has been localized to the myoepithelial cell layer in vivo (Fox et al 1994). Expression of K6 was examined since this marker, when coexpressed with K14, is a marker for basal clear cells. Keratin 6 is also expressed by a proportion of luminally situated epithelial cells in the mouse mammary epithelium (Smith et al 1990) and by body cells of end buds (Smith et al 1990, Sapino et al 1993). Methods For these studies, cultures initiated with single or low densities of FTBEC isolated from primary cultures of normal human mammary tissue were stained with Wright-Giemsa as previously described or were examined for the expression of a variety of epitopes by immunocytochemical methods. Cultures were maintained at 750 cells/cm2 in the absence of any feeder and at 100 cells/cm2 in the presence of a NTH 3T3 feeder, unless otherwise noted. The following epitopes were examined by immunocytochemistry using fluorescence and/or the APAAP procedure in 3-8 different cultures: Luminal cell specific epitopes Mammary gland mucin MUC1: Three different mouse monoclonal antibodies recognizing this epitope were examined. They included antibody IgGl clone HMFG-2 (a gift from Dr. J. Taylor-Papadimitriou, Imperial Cancer Research Fund, UK) used as an undiluted supernatant, IgGl clone 214D4 (a gift from Dr. J. Hilkens, The Netherlands Cancer Institute) used at a 1:100 dilution of supernatant and IgG2 clone E29 (referred to as EMA or epithelial membrane antigen in this thesis) directly conjugated to FITC (DAKO) used at a 1:10 dilution. Epithelial cell adhesion molecule: Antibody IgGl clone VU-1D9 (Novocastra, Newcastle upon Tyne, UK) was diluted 1:100 (0.4 ug/ml). 41 Keratin 8/18: Mouse monoclonal antibody IgGl clone 5D3 (Novocastra) diluted 1:40 (2 ug/ml). Keratin 19: Mouse monoclonal antibody IgGl clone RCK 108 (DAKO) diluted 1:80 (0.5 ug/ml). Myoepithelial cell markers Keratin 14: Mouse monoclonal antibody IgG3 clone LL002 (Novocastra) diluted 1:20 (5 ug/ml). Smooth muscle actin: Mouse monoclonal IgG2 clone 1A4 (DAKO) used at 1:30 dilution (3.6 ug/ml). Histo-blood group antigen H type 2: Mouse monoclonal IgM clone 92FR-A2 (DAKO) diluted 1:50 (total protein concentration = 0.53 mg/ml). All cultures stained by immunocytochemistry to detect this epitope were fixed in 1% paraformaldehyde in PBS. CD44v6 variant of the hyaluronate receptor: Mouse monoclonal IgGl clone 2F10 (R&D, Minneapolis, MN) diluted 1:1000 (1 ug/ml). Miscellaneous epitopes Keratin 6: Mouse monoclonal antibody IgG2 clone NCL-CK6 (Novocastra) used at 1:40 dilution (1 ug/ml). Collagen type IV: Mouse monoclonal IgGl clone CIV 22 (DAKO) used at a 1:50 dilution (1.6 ug/ml) Cell proliferation associated nuclear protein Ki-67: Mouse monoclonal IgGl clone Ki-67 (DAKO) used at 1:50 dilution (9.8 ug/ml). Uncharacterizedantibodies 5E1I and 4H1: These two antibodies were raised from mice injected with a human mammary carcinoma cell line (T-47D). Mouse monoclonal IgGl clones 5E11 and 4H1 (both gifts from Dr. Peter Lansdorp, Terry Fox Laboratories, Vancouver, B.C.) were used at 1:100 dilution and undiluted, respectively. 42 Negative control and secondary antibodies An IgGl mouse monoclonal recognizing dextran (also a gift from Dr. Peter Lansdorp) diluted to 5 ug/ml was used as a negative control for all staining experiments in which the primary test antibody was a non-FITC conjugated mouse monoclonal IgG subtype antibody. An exception to this were the staining experiments using Ki-67. In these cases, the anti-dextran IgG concentration was increased to 10 ug/ml. An IgGl mouse monoclonal negative control antibody (DAKO) diluted 1:10 was used as the negative control for EMA-FITC staining. A mouse IgM negative control antibody (DAKO) diluted 1:10 (10 ug/ml). A goat anti-mouse IgG (H+L) directly conjugated to PE (Jackson ImmunoResearch Laboratories, Inc., West Grove, P.A., U.S.A.) used at 1:125 dilution in the fluorescence microscopy studies. In two pilot experiments to induce collagen IV and SMA expression in dispersed teardrop-shaped cells in culture, second and third passage cultures enriched for these were grown to approximately 20% confluence in SF medium. Following a change to EGF- and insulin-free SF medium, the cultures were maintained for a further 14 days. Cultures were then terminated and fixed with methanol and stained using monoclonal antibodies to detect SMA and collagen IV. An antibody blocking experiment was performed to determine if the antibodies 4H1 and 5E11 recognize the EpCAM epitope. To do this, serial dilutions (up to 1:1000 dilution of culture supernatant) of either the 5E11 antibody or the 4H1 antibody were tested to see if they could block the binding of the EpCAM specific antibody Ber-Ep4 (directly conjugated to FITC) to T47-D mammary tumor cells. As well, the ability of 4H1 to block 5E11-FITC antibody binding to T47D cells was also examined. Antibody binding to cells was analyzed on a flow cytometer. 43 Results When HBEC were seeded at low densities and then cultured for 9 to 12 days, a spectrum of colony phenotypes could be observed within individual cultures (Figures 7A-G). Colonies ranged from being composed entirely of closely arranged cells with indistinct cell borders (Figure 7A-C) to colonies composed exclusively of cells exhibiting a dispersed cell phenotype (Figure 7G), with others containing a mixture of these cell types (Figures 7D-F). In the smaller of the colonies with the closely arranged cell phenotype, the cells formed small rounded alveoli-like structures (see example in Figure 7A). In the larger of these colonies, the cells located at the periphery were more closely arranged than the more centrally located cells which suggested the formation of a central lumen (Figures 7B and C). This effect was particularly evident when the cells were cultured in the presence of NIH 3T3 cells, but not in the presence of HMF. Cell cloning experiments verified that colonies such as these could be derived from a single cell (e.g., see Figure 6A). The cells in these compact colonies uniformly expressed the typical luminal epitopes K8/18 (Figure 8A), K19 (Figure 8B), MUC1 (Figure 8C) and EpCAM (Figure 8F). Cells of these colonies were also uniformly positive for expression of the uncharacterized 5E11 (Figure 9B) and 4H1 (Figure 9D) epitopes. These colonies also expressed the CD44v6 hyaluronate receptor (Figure 10). Cells of these colonies did not express the myoepithelial cell-specific K14 (Figures 1 IB and E) and BGA (Figure 12); however, occasional colonies composed predominantly of cells expressing luminal cell markers have some peripheral cells that are K14+ (Figure 13). Cells in colonies composed of closely arranged cells typically did not express K6 (Figure 14A), however there were occasional colonies in which some cells do expressed K6 (Figure 14B). The expression of luminal and myoepithelial cell-specific markers by the cells of the different colony types is summarized in Table 1. 44 TABLE 1: Summary of the expression of luminal and myoepithelial markers by HBEC in the different colony types. Luminal cell-specific markers Myoep specii ithelial cell -ic markers Misc. markers CELL TYPE K8/1 8 K19 MUC 1 EpCA M 5E1 1 4H1 CD44 v6 K14 BG A Ki-67 K6 Cells of compact colonies + + + + + ± + n.d. -/+ Central cells of mixed colonies + + + + + ± + -I± + Dispersed cells of mixed colonies -/+ -/± + + + ±/+ ±/+ Cells of dispersed colonies -l± + + + n.d. n.d. = not done Mixed colonies typically contained a central core of cells that were either piled up on one another or appeared flattened, but generally had poorly defined cell borders and were relatively non-refractile. The cells expressed the typical luminal epitopes K8/18 (Figures 15A and D), K19 (Figure 16A), MUC1 (Figure 16C; using the 214D4 antibody clone), but were negative for the myoepithelial-specific K14 (arrows in Figures 17B and E). Emerging from the periphery of these colonies were cells expressing keratin 14 (Figures 17A and E). These emerging cells also expressed keratins 8/18 (arrows in Figure 15A, also see Figure 18B) and BGA (Figure 19B) in a heterogeneous fashion. As shown in Figures 19A - D, expression of BGA appeared to be strongest in those cells exhibiting the strongest "myoepithelial cell" morphology (e.g., dispersed cell arrangement and highly refractile). As shown in Figures 20A and B, centrally located cells as well as some of the peripherally located dispersed cells (arrows) expressed K6. The 45 distribution of EpCAM (Figure 21), 5E11 and 4H1 (Figure 22) epitopes among the mixed colonies tended to be heterogeneous. The observation of centrally located non-reffactile cells that were highly positive for these epitopes and highly retractile dispersed cells that expressed lower levels of these epitopes (arrows in Figures 21 and 22) was a consistent pattern. Immunofluorescent staining with the 5E11 monoclonal antibody typically gave a very strong reaction, whereas staining with the 4H1 clone tended to be much weaker. To determine if the peripherally located dispersed cells represented an end stage cell, or in fact were still proliferating, mixed colonies were stained with antibodies to the cell cycle associated nuclear protein, Ki-67. Expression of Ki-67 in these mixed colonies was highest among the peripherally located cells (Figure 23). Mitotic figures were also observed within the peripherally located dispersed cells (arrows in Figure 30B). Figure 24 shows a 9 day-old colony generated from a single bipotent progenitor. Double immunofluorescence staining of this colony to detect expression of MUC1 (green; using the EMA antibody clone) and keratin 14 (red) demonstrated that cells expressing a luminal marker and cells expressing a myoepithelial marker can be generated from a single progenitor and this does not represent a culture artefact. Some of the centrally located cells expressed low levels of both MUC1 and keratin 14 (asterisk in Figure 24). As well, expression of both these proteins could be co-localized within the same cell (arrow in Figure 24). Homogeneous colonies composed entirely of highly retractile dispersed cells expressed keratin 14 (Figure 1 IE), BGA (Figure 25B) and CD44v6 (arrow in Figure 26). Most of these colonies did not express K8/18 or K6 (Figure 20D), but a heterogeneous expression of these markers was observed (e.g., K8/18+ cells in the colony on left side of Figure 25E and K6 + cells in Figure 20C). No expression of MUC1 (using all 3 antibody clones) or K19 was observed in the cells of these colonies. 46 In normal human mammary gland in vivo, the antibodies 5E11 and 4H1 recognize epitopes that are localized to the basal and lateral cell membranes of cells within the luminal cell compartment (Figures 27C and E), in a manner similar to EpCAM (Figure 27A). Likewise, the 5E11 and 4H1 epitopes have a similar distribution to EpCAM among HBEC in vitro. Analysis of HBEC primary cultures with regards to 5E11, 4H1 and EpCAM expression revealed a similar pattern of distribution. Considering the similarities in the staining patterns with each of these antibodies, antibody blocking experiments were initiated to determine if any antibodies recognize the same epitopes. Results demonstrated that 5E11 and 4H1 recognize the same epitope, but do not block binding of the Ber-EP4 antibody to EpCAM (Figure 28). In some cultures, colonies containing K14+ cells exhibited features of squamous metaplasia (e.g., cells become very flattened and grew on top of other cell types)(Figure 29). The presence of these cells appeared to be dependent on the patient from whom the sample was obtained, since they were prevalent in cultures from some samples, but were not detected at all in others. Many of the centrally located cells of these colonies also contained somewhat flattened cells (Figure 3 OB). The presence of these was found to be highly dependent on EGF (compare Figure 30A to Figure 30B). Many colonies containing K14+ cells also contained a central core of cells that formed elevated masses of concentrically arranged cells expressing very high levels of K6 (Figure 31 A). It is suspected that cells in these elevated masses were undergoing squamous metaplasia. It is interesting to note that the flattened squamous metaplastic cells and the elevated masses of cells were predominant in first passage cultures, but not later. Preliminary experiments in which CT was deleted from the culture medium suggest that the formation of the large flattened squamous cells is CT-dependent. However, the elevated masses of cells still develop in a CT-depleted culture medium. 47 When HBEC were cultured at low densities (750 cells/cm2) in SF medium and in the absence of a NTH 3T3 feeder, no expression of smooth muscle actin was observed in any HBEC colonies regardless of their morphology or the culture conditions used. Expression of collagen rv could be detected in cultures maintained in a growth factor reduced (EGF- and insulin-free) environment (Figure 32A). The expression of smooth muscle actin and collagen rv was not examined for HBEC colonies generated on NTH 3T3 feeders. Discussion These studies have shown that two distinct categories of HBEC progenitors can be identified by analysis of the colonies they produce in vitro. The first, termed luminal-restricted progenitor, generates colonies of cells that express typical luminal cell proteins. The second, termed the bipotent or mixed progenitor, is unique in that it generates colonies composed of some cells expressing luminal cell-specific proteins and other cells expressing myoepithelial cell specific proteins. Cells expressing low or high levels of both luminal (MUC1) and myoepithelial (K14) are also observed within the latter type of colonies. A third type of progenitor, termed the myoepithelial cell-restricted progenitor, generates progeny expressing predominantly myoepithelial lineage markers. These progenitor types correlate in some respects with previously described HBEC that proliferate in vitro (Dairkee at al 1986, Edwards et al 1986, Taylor-Papadimitriou et al 1989, Mork et al 1990, O'Hare et al 1991, Karsten et al 1993, Kao et al 1995, Pechoux et al 1999) and the general observation of closely arranged HBEC with indistinct cell borders expressing typical luminal markers (e.g., K8 A 8 , K19 and MUC1) and dispersed refractile cells expressing basal cell markers (e.g., K14 and BGA) appears to be a general one. It should be noted that the latter 48 cells do not express SMA, and thus cannot be considered fully differentiated myoepithelial cells. Expression of SMA in these cells, at least in vitro, is inversely correlated with cell proliferation (Petersen and van Deurs 1988). However, no SMA could be detected in the cultures examined here, even when the levels of growth factors were reduced (by omission of EGF and insulin). This contrasts with previous suggestions that differences in culture media and antibody clones tested might explain the observed expression of SMA by Go teardrop shaped cells in some studies. The presence of K14+/K187K197SMA" cells within the luminal cell compartment in vivo has also been previously described (Bocker et al 1992a) and may be related to a similar vimentin+/K18+ cell within the luminal cell compartment (Pechoux et al 1999). These unique cells might represent newly formed myoepithelial cells that have not become basally positioned and fully differentiated. Electron microscopic studies of human mammary epithelium demonstrate that differentiated myoepithelial cells have low proliferation frequencies (Ferguson et al 1985, Ferguson 1988, Joshi et al 1986). Expression ofthe cell proliferation associated epitope Ki-67 by newly formed K14+/SMA' cells suggests that these cells are not differentiated end stage cells. Some of the peripherally located cells that have a dispersed phenotype express keratin 6. Since a dispersed phenotype is also associated with expression of K14, and co-expression of K6 with K14 is a marker of basal clear cells (in the mouse mammary epithelium), it is speculated that the K6+/K14+ cells observed in vitro in this study correspond to the basal clear cells described in the mouse mammary gland by Smith et al (1990). However, considering that results here and elsewhere (Taylor-Papadimitriou et al 1987) have failed to detect K6 in normal resting adult human mammary epithelium, it appears that the dual expression of K6 and K14 is not be suitable for identifying basal cells in the human mammary epithelium in vivo. The present study has also provided several unique observations concerning the proliferative and differentiation potential of individual HBEC progenitors. The most significant is that a single 49 progenitor can generate both mammary lineages in vitro. The cellular arrangement of the mixed colonies obtained consists of cells expressing luminal specific proteins surrounded by cells expressing predominantly myoepithelial specific proteins. This supports the concept that a subset of H B E C with luminal characteristics generate cells with myoepithelial characteristics. A similar conclusion was reached in recent reports in which bulk cultures of purified K18 + cells were found to generate K14 + progeny, but not vice versa (Kao et al 1995, Pechoux et al 1999). Another unique observation here was the heterogeneity found to be displayed by luminal HBEC progenitors in vitro. Although the closely arranged cells of both the colonies generated by luminal-restricted and bipotent progenitors expressed the same luminal-specific epitopes (e.g., K8/18, K19, MUC1, EpCAM), those in the former colonies tended to "ball up", whereas this was a less obvious feature of the mixed colonies. The closely arranged cells in the mixed colonies could be further differentiated from those in the colonies composed of only cells expressing luminal-specific proteins by the more prevalent expression of K6 by the former. Finally, the cells expressing luminal proteins of the mixed colonies to give rise to K14 + cells, whereas those of the colonies composed of cells expressing only luminal-specific proteins do not. It is interesting to note that balling up of HBEC of the colonies containing cells expressing only luminal-specific proteins was not as evident when the HBEC were cultured in the absence of or in the presence of HMF. Since H B E C colonies tend to grow in the spaces between the NTH 3T3 feeder cells, it suggests that paracrine effects are responsible for this phenomenon. Previous studies analyzing the distribution of lineage markers among unseparated (Dairkee at al 1986, Taylor-Papadimitriou et al 1989, Mork et al 1990, Karsten et al 1993) and FACS-sorted H B E C (O'Hare et al 1991) have generally described most HBEC as having either luminal or myoepithelial epitopes. However in each of these studies, minor subpopulations of cells expressing "inapropropriate" lineage markers have also been noted. For example, subpopulations 50 of H B E C with a classical "luminal" morphology but expressing the myoepithelial associated proteins vimentin (Mork et al 1990), keratin 14 (Taylor-Papadimitriou et al 1989, O'Hare et al 1991, Karsten et al 1993) and CALLA/CD10 (Karsten et al 1993) have been described. The expression of the luminal cell associated protein K8/18 in HBEC having a "myoepithelial" teardrop-shaped morphology has also been reported (O'Hare et al 1991, Karsten et al 1993); results which are in agreement writh those presented here. In vivo, human mammary epithelial cells also express "inappropriate" lineage epitopes. As previously mentioned, a subset of luminal HBEC express K14 but not SMA (Bocker et al 1992a). Similarly, Pechoux and colleagues reported the presence of luminally-positioned cells expressing both vimentin and K18 (Pechoux et al 1999). Such incongruencies can be accounted for if the luminal and myoepithelial compartments are considered as being composed of a hierarchy of H B E C at varying stages of differentiation towards terminally differentiated progeny. Assuming that a subset of luminal HBEC are the source of myoepithelial cells, the presence of cells with luminal morphologies expressing varying degrees of myoepithelial proteins might identify cells that are just beginning a transition into myoepithelial cells. Likewise, cells with myoepithelial morphologies expressing varying numbers of luminal epitopes (e.g., K8/18 and EpCAM) might suggest the presence of newly forming myoepithelial cells still expressing residual luminal cell-specific proteins. This concept of a hierarchy of cells is also supported by the large range of colony morphologies observed within single cultures. Expression of the CD44v6 form of the hyaluronate receptor is localized to the basal cell compartment in vivo (Fox et al 1994). However, when HBEC are placed into culture, all cells express this protein. This is in agreement with a previous study that reported upregulation of this receptor in cultures seeded with cells FACS-sorted on the basis of expression of luminal and 51 myoepithelial-specific proteins (Cooper et al 1998). Consequently it is concluded that this upregulation likely represents culture artefact. The presence of cells undergoing squamous metaplasia in primary cultures of human mammary tissue has been previously described (Edwards et al 1984, O'Hare et al 1991). Cells undergoing this process are characterized as large flat cells that are usually located superficial to other cell types and thus are difficult to visualize. This process is restricted to the central regions of K14+ cell-containing colonies and is not observed in colonies composed solely of cells expressing luminal cell markers (O'Hare et al 1991). In the present studies, many K14+ cell-containing colonies also contain cells that have a flattened morphology, but not to the same degree as the cells exhibiting squamous metaplasia. These flattened cells are typically located in the center of the colony and are surrounded by tear-drop shaped migrating K14+ cells. The presence of these flattened cells is highly dependent on the presence of EGF. Considering that the process of squamous metaplasia is cell proliferation-dependent (Schaefer et al 1983), it is speculated that these flattened cells are in the early stages of this differentiation pathway. Experiment 4: Development of a 3-D Collagen Gel HBEC Colony Assay Numerous reports have illustrated the advantages of 3-D culture systems over 2-D systems in promoting a greater degree of breast epithelial cell differentiation. Therefore, the use of a 3-D culture system was examined to see if a greater degree of morphological differentiation of HBEC within colonies would further facilitate their characterization. Both collagen gels (Emerman et al 1987, Durban et al 1985, Levay-Young 1987) and reconstituted basement membrane (also known by its' commercial name Matrigel) (Barcellos-Hoff 1989, Darcy et al 1991, Streuli et al 1991) permit casein synthesis by murine epithelial cells when cultured in the presence of lactogenic hormones. However, HBEC seeded in Matrigel become growth-arrested (Petersen et 52 al 1992, Dr. J. Gomm, personal communication). Since the primary use of a HBEC colony assay is to detect HBEC with proliferative potential, collagen gels were selected because it has been reported that they permit good HBEC proliferation and a partial degree of differentiation (e.g., some deposition of a basement membrane, partial retention of tissue architecture) (Foster et al 1983, Smith et al 1987). It should be noted that these experiments were performed before NTH 3T3 cells were recognized to be the preferred feeder for HBEC colony growth. As a result, all collagen gel experiments utilized HMF feeders. Methods The preparation of rat tail collagen type I was performed as previously described (Richards et al 1983). Flow sorted MUC1 + and CALLA + HBEC (see Chapter IV) as well as non-sorted HBEC isolated from primary cultures of normal mammary tissue were mixed at the desired density (typically 5 x IO2 -10 4 cells/ml of gel) with a solution of rat tail collagen dissolved in 0.1% acetic acid, 10X F12/DME/H (Sigma), NaOH and SF medium, growth factors and hormones. Aliquots of 0.5 ml of this solution were then pipetted into individual wells of 24-well plates in which an irradiated HMF feeder layer had been previously established and overlaid with 0.3 ml of cell-free collagen gel. The human mammary fibroblasts were irradiated at 2,500 cgy and seeded at 104 cells/cm2 in F12/DME/H supplemented with 5% FBS and 5 ug/ml INS. Prior to addition of the first layer of cell-free collagen gel, the feeder layers were rinsed twice with F12/DME/H. Once the collagen layer containing the suspended HBEC had gelled, an additional 0.5 ml of SF medium was added on top. To plate single HBEC in collagen gels (4 experiments), the cloning attachment of the FACS was used to deposit single HBEC expressing either MUC1 (detected using the HMFG-2 antibody clone) or CALLA (see Chapter IV) into the individual wells of a 96-well plate each of which already contained an irradiated mammary stromal cell feeder layer and 35 ul of SF medium. Once the HBEC were deposited, 100 ul of collagen gel 53 were added to the well and after the gel had set, 100 ul of SF medium or SF medium supplemented with 5% FBS were also added to each well. All 3-D cultures were then incubated for 14 days at 37 °C in a humidified atmosphere consisting of 5% CO2 and 95% air. Colonies (except for those in the cloning experiments) were scored by first removing the gel from the well, gently detaching and discarding the cell-free collagen gel underlay and placing the HBEC-containing gel on a plastic grid. The number of colonies present was then enumerated directly using a dissecting microscope. In the case of the single cell culture experiments, the colonies were scored directly in the gels in the wells of the 96-well plates using a dissecting microscope. Results Cells isolated from primary cultures of normal breast tissue and cultured at 5 x 10^  to 5 x 10^  cells/ml in collagen gels on top of irradiated HMF produced colonies at frequencies of 0.7 - 7.8% (mean = 3.1 ± 0.8%; 10 primary cultures analyzed). Colony yields under these culture conditions were proportional to the input cell seeding density, although the relationship was non-linear (Figure 33). In an attempt to confirm that the colonies observed in 3-D cultures had been generated by the clonal proliferation of individual HBEC (rather than representing the reaggregation of dissociated HBEC), single HBEC expressing either MUC1 + or CALLA + were seeded into collagen gels. Under these conditions, a total of 9 colonies were obtained from 1440 cells isolated from 4 different primary cultures. All colonies, except for one (Figure 33B), were small and barely scorable. These results show that clone formation in 3-D matrices is possible but is probably suboptimal and leaves somewhat unresolved the question of whether the colonies seen at higher plating densities are all of single cell origin. 54 Colonies generated within collagen gels conformed to one of two morphologies. The first type of colony was typically small and spherical (Figure 34A) and generally consisted of a homogeneous simple cuboidal epithelium arranged around a centrally located lumen (Figure 34B). The nuclei of cells of these small spherical colonies tended to be basally located, suggesting that some polarity had been established. Immunocytochemistry failed to detect any basement membrane specific collagen type IV. These latter two observations suggested that the extent of polarization was less than that exhibited by luminal cells in vivo. The cells in these small spherical colonies expressed the luminal keratins 19 (Figure 34C) and 8/18 (Figure 34D). However, the expression of both these in many cases was abnormal in that they were localized within the colony adjacent to the extracellular matrix. This staining is unlikely to be non-specific since this pattern was observed only with these particular antibodies (and about 5% of the time with the K14 specific antibodies - see below) and not with the control antibodies or with any of the other antibodies tested. Cells of these colonies also express the MUC-1 glycoprotein, but it was distributed throughout the cytoplasm (Figure 34F), rather than being apically located as seen in vivo. About 5% of these colonies also showed low expression of K14 in a basal location within the epithelium (Figure 3 5 A) whereas all colonies showed a low expression of CD44v6 in a basal location (Figure 35B), but were consistently negative for smooth muscle actin. The second type of colony generated in 3-D collagen gel cultures had a gross morphology reminiscent of branched ducts (Figure 36A). Figure 36B shows a hematoxylin and eosin-stained cross section of one of these colonies fixed after 14 days in culture. This colony contains a heterogeneous population of cells, the centrally located cells being larger and paler staining than those located at the periphery. Occasionally, a lumen was observed in such colonies, typically in the larger ones. Immunohistochemical analysis revealed the expression of MUC-1 in these 55 colonies to be highly variable from one sample source to another, with the majority of colonies not containing any MUC1 positive cells. Expression of CD44v6 in the large branched colonies was localized to the peripherally located cells, but not the centrally located cells (Figure 36E). Cells of the large colonies were also negative for EpCAM. Cells undergoing squamous metaplasia were also observed in these colonies (e.g., Figure 36B). When MUC-1 + cells were present, they were usually among the more centrally located cells, especially in the larger colonies (e.g., Figure 36C, which is the same colony as Figure 36B). Most colonies generated from these progenitors also contained cells expressing the luminal cell specific keratins 8/18 (Figure 36F). However, cells expressing keratin 19 were seen only rarely. The distribution of all of these keratin-expressing cells appeared random. All large branched colonies examined contained cells that expressed keratin 14 with a gradation in staining intensity from the cells at the periphery to a weaker level or no staining in the centrally located cells (Figure 36G). The cells in these were consistently negative for smooth muscle actin and the basement membrane protein collagen type IV. In an attempt to increase the degree of luminal and myoepithelial cell differentiation observed in collagen gel cultures, a series of pilot experiments (1-5 samples per experiment) examining the effects of 10 ng/ml HGF, 10"8 M E 2 , 10 ng/ml P, 20 ng/ml NDF, 20 ng/ml keratinocyte growth factor, 3 ng/ml TGF-P, 0.01 - 1.0 ug/ml retinoic acid, 2.5 - 10% FBS and low growth factor medium (e.g., switching to insulin- and EGF-free SF medium once colonies had formed) on HBEC within collagen gels were initiated. The endpoints of these experiments included altered cloning efficiency, altered colony morphology, cell polarity and collagen IV deposition. These pilot experiments demonstrated a decreased cloning efficiency in the presence of TGF-P and at 56 higher (>0.1 ug/ml) concentrations of retinoic acid (data not shown). Examination of sections of colonies demonstrated no consistent influence of any of these factors on colony architecture. Discussion Two types of colonies were seen when HBEC were cultured in collagen gels. The first were small spherical colonies reminiscent of alveoli. The second were colonies containing large branching structures, which on a gross morphological level, is reminiscent of branching ducts observed in vivo. Considering that the existence of alveolar and ductal stem cells has been demonstrated in mouse mammary glands (Smith 1996), the generation of such colonies leads to the hypothesis that alveolar and ductal progenitors are being detected. Interestingly, the ultrastructure of the colonies generated within the collagen gel, particularly the large branched ones, do not resemble normal mammary epithelium seen in vivo. Occasional colonies are observed where a central lumen has apically localized MUC1, however this is usually the exception. Most large branched colonies contain few cells expressing luminal-specific epitopes, and colonies are composed of SMA7K14+/CD44v6+ cells. In hindsight, this may not be surprising considering that HMF feeders were used, which in previous experiments, promoted proliferation of dispersed refractile cells and poor survival of HBEC that typically express luminal-specific proteins. Cholera toxin activates adenylate cyclase and results in increased levels of cyclic 3',5' adenosine monophosphate (cAMP, Lad et al 1980, Holmgren 1981), which in turn is associated with the development of squamous metaplasia (e.g., concentric arrays of cells with excessive keratinization) in mammary gland organ culture (Schaefer et al 1983). The growth medium used in the experiments presented here contained CT and did not contain retinoic acid, an inhibitor of squamous metaplasia (Huang et al 1986). As a result, the presence of regions of squamous 57 metaplasia in these large colonies is not surprising. What is interesting to note is that this metaplasia was seen only in the large branched colonies. Similarly in the 2-D cultures only the mixed/myoepithelial cell-colonies exhibited these properties. Organ culture studies with mouse mammary glands has demonstrated that it is a relatively rare cell within the mammary epithelium that can undergo squamous metaplastic differentiation since exposure of glands to cAMP and prostaglandins reveals that only discreet small foci of single or few cells within the epithelium undergo this process (Schaefer et al 1983). The same studies demonstrated that induction of lobulo-alveolar development and differentiation and induction of squamous metaplasia are reciprocal processes. A question that then arises is whether the lobuloalveolar precursors and epidermoid (squamous metaplasia) precursors are common or separate pools of precursor cells. Later studies utilizing serial passaged (homogeneous) mouse mammary cell lines exhibiting stem cell characteristics demonstrate that these cells are susceptible to squamous metaplasia. Since these cell lines were passaged to select for ability to regenerate the mouse mammary epithelium, it would be expected that if lobuloalveolar and epidermoid cells are separate populations the epidermoid precursors would be lost during passaging. However, the persistence of epidermoid precursors among these stem cell lines suggested that the two types of precursors may actually be identical or closely related (Schaefer et al 1984). The fact that squamous metaplasia was also seen in the colonies obtained in vitro supports the concept that the progenitors of these colonies are cells that have regenerative activity in vivo. Similarly a previous study utilizing human mammary epithelial cell lines revealed that only those cell lines exhibiting stem cell characteristics (i.e., could generate daughter cells that could express either luminal- or myoepithelial-specific proteins) were able to undergo squamous metaplasia when cultured within a collagen gel (Rudland et al 1991). 58 Figure 1: Influence of HMF, M2-10B4 and BMS on HBEC colony formation after 5 days in vitro. Freshly dissociated uncultured cells from 3 separate breast reduction samples were seeded in the absence or presence of pre-established viable irradiated HMF, M2-19B4 and BMS feeders at 7.5 x 102 to 1.5 x 103 cells/cm2 in SF medium supplemented with 5% FBS. After 24 hours, the medium was changed to SF medium and the culture maintained for a further 4 days. At the end of the culture period, the cultures were fixed with A:M, stained with Wright-Giemsa and the total number of colonies scored based on morphology. Results are presented as fold-increase in cloning efficiency over that obtained in the absence of feeders. Asterisks indicate significant differences from controls (no feeders, p<0.05). 59 60 Figure 2: Influence of HMF, M2-10B4 and BMS feeders on average H B E C colony size after 5 days in vitro. Freshly dissociated cells from 3 different breast reduction samples were seeded in the absence or presence of pre-established viable irradiated HMF, M2-19B4 and BMS feeders at 7.5 x 102 to 1.5 x 103 cells/cm2 in SF medium supplemented with 5% FBS. After 24 hours, the medium was changed to SF medium and the culture maintained for a further 4 days. At the end of the culture period, the cultures were fixed with A : M , stained with Wright-Giemsa and the mean number of cells in 10 randomly selected colonies from each sample in each condition was counted. Asterisks indicate significant differences from controls (no feeders, p<0.05). 61 62L Figure 3: Colony phenotypes observed in the presence of FfMF after 5 days in vitro. Colonies are composed of either dispersed or closely arranged cells (A), or are composed of both cell types (B). Note in this latter type of colony, the dispersed cells appear to bud from the central core of tightly arranged cells. Bar = 250 um. 63 Figure 4: Distribution of compact, dispersed and mixed colonies generated after 5 days in vitro in the presence of a HMF feeder and SF medium. Results from 3 independent experiments are shown. 65 Figure 5: Influence of HMF and NTH 3T3 feeders on HBEC colony formation. Unsorted HBEC isolated from a 3-day primary culture were seeded at 500 cells/cm2 in the absence of any feeder (A), in the presence of an irradiated HMF feeder (B) or in the presence of irradiated NTH 3T3 feeders (C), maintained for 11 days in SF medium and then stained with Wright's Giemsa. Note that the majority of the colonies in the presence of the HMF feeder are composed only of dispersed teardrop-shaped cells and that only a few colonies containing only closely arranged cells are observed (arrows). In the presence of the NTH 3T3 feeders, abundant closely arranged cell colonies are observed in addition to the dispersed cell colonies (arrowheads). Bar = 103 um. ) 67 18 Figure 6: Morphologies of the 3 general types of colonies observed in the single cell culture experiments. Colonies are composed of closely arranged cells (A), loosely arranged cells (B) or exhibit characteristics of both (C). Colonies composed of the closely arranged cells typically have smooth borders, and the cells at the periphery of the colony tend to pile upon one another. The colony illustrated in C is composed of a central core of non-refractile closely arranged cells (arrowheads) surrounded by dispersed cells. Bar in panels A and B = 100 um. Bar in panel C = 250 um. 69 Figure 7: Range of FfBEC colony morphologies observed after 11 days in culture in the presence of NTH 3T3 feeders. All colonies were generated from the same original primary FfBEC culture. Colony phenotypes include those composed of only closely arranged cells (A-C), mixed phenotypes (D-F) and those composed predominantly of loosely arranged cells (G). Bar = 250 um. 71 It Figure 8: Expression of K8/18 (A), K19 (B), MUC1 (using the 214D4 antibody clone, C) and EpCAM (F) in colonies composed of closely arranged cells. Colonies shown in panels A - D were grown on glass coverslips, and thus are not necessarily as "rounded" as those maintained on tissue culture plastic (panels E and G). IgG control staining for panels A - C is shown in panel D; IgG control staining for panel F is shown in panel H. Colonies in panels A - C were counterstained with Wrights Giemsa, whereas those in panel D were not. Bar in panels A, D and G - 250 um. Bar in panels B, C and E = 100 um. 73 7+ Figure 9: Expression of the uncharacterized epitopes recognized by the 5E11 (B) and 4H1 (D) antibodies by cells of colonies composed of closely arranged cells. Relatively strong uniform fluorescence of the 5E11 epitope is observed, whereas fluorescence by the 4H1 epitope is much weaker. IgG negative control staining is shown in panel F. Bar in panels A, C and E = 100 um. 75 Figure 10: Expression of CD44v6 in colonies composed of closely arranged cells (B). All cells express this epitope. IgG negative control staining is shown in panel D. Bar in panels A and C = 250 um. 77 78 Figure 11: Lack of expression of K14 in colonies composed of closely arranged cells (denoted by the arrows in Figures A, B and E). IgG negative control staining for K14 is shown in panels D and F. Bar in panels A, C, E and F = 250 um. 79 B c D *', T T T lakh- ? • _ go Figure 12: Lack of expression of BGA in a colony composed of closely arranged cells (B). IgM negative control staining is shown in panel D. Bar in panel A = 100 urn. Bar in panel C = 250 um. 8 1 Figure 13: Expression of K14 in a colony composed of non-refractile closely arranged cells. Only a few cells at the periphery of this colony express this epitope (arrows). IgG negative control staining is shown in panel D. Bars in panels A and C = 100 um. 83 B D Figure 14: Expression of K6 in colonies composed of closely arranged cells. Most colonies do not express this keratin (A), however occasional colonies containing cells expressing this protein are observed (B). IgG control antibody staining is shown in panel C. Bar = 100 um. 85 Figure 15: Expression of K8/18 (A and D) in colonies of mixed phenotype. Note that in panel A, a few of the peripheral migrating cells express intermediate levels of these proteins (arrowheads), whereas in panels C and D the highly retractile peripheral cells do not express these proteins (arrows). IgG negative control staining for panel A in shown in panel B; the negative control staining for panel D is shown in panel F. Cells in panel A were counterstained with Wright's Giemsa, whereas those in panel B were not. Bar in panel A = 100 um. Bar in panels B, C and E = 250 um. 87 D LL 88 Figure 16: Expression of K19 (A) and MUC1 (using the 214D4 antibody clone, C) in mixed colonies. Note in panel C the MUC1" cells (arrows) surrounding the MUC1 + centrally located cells. IgG negative control staining for panel A is shown in panel B; the negative control staining for panel C is shown in panel D. Bar in A - D = 100 um. 89 A * * ft * 10 Figure 17: Expression of K14 in colonies of mixed phenotype. K14 is restricted to the more peripherally located cells of these colonies (panels B and E), whereas the centrally located cells do not express this protein (arrows in panels A and E). IgG negative control staining for panel B is shown in panel D; the negative control staining for panel E is shown in panel F. Cells of panel E are counterstained with Wright's Giemsa. Bar in panels A, C, E and F = 250 um. 91 Figure 18: Expression of K8/18 in mixed colonies and in colonies composed of more loosely arranged refractile cells. Note that the colony on the left in panel A expresses K8/18 in a uniform fashion, whereas the one on the right does not express this keratin. Dispersed cells adjacent to the colony on the left also express this keratin (arrows). IgG negative control staining is shown in panel D. Bar = 250 um. 93 Figure 19: Expression of BGA in mixed and dispersed cell colonies. Note in panels A - D that BGA has a heterogeneous pattern of expression. Expression appears to be higher in cells that are less closely arranged and are more light refractile; however, the differences between those that do not express BGA (arrows) are subtle. IgM negative staining control in shown in panel F. Bar = 250 um. 95 Figure 20: Expression of K6 in mixed (panels A and B) and in dispersed (C and D) cell colonies. Note that in the mixed colonies, the centrally located closely arranged cells and some of the more peripherally located dispersed cells express K6 (arrowheads denote K6 + dispersed cells). Cells of colonies composed only of loosely arranged cells express K6 in a heterogeneous fashion. Some colonies contain cells expressing this protein (arrowheads in panel C), whereas others do not (panel D). IgG antibody control staining is shown in panel E. Bar =100 um. 97 Figure 21: Heterogeneous expression of EpCAM (panels B and F) in mixed colonies. Expression of EpCAM in mixed colonies is highest in those cells that are closely arranged (asterisk in panels A) and lowest in highly refractile dispersed cells (arrows in panels A, B, E and F). Colonies composed of closely arranged cells express EpCAM in a uniform manner (arrowhead in panel A). IgG negative control staining for panel B is shown in panel D; negative control staining for panel E is shown in panel H. Bar = 250 um. 99 100 Figure 22: Expression of the 5E11 (A) and the 4H1 (B) epitopes in mixed colonies. Immunofluorescence staining with the 5E11 monoclonal antibody yields a relatively strong signal. The epitope is localized to primarily to most cell types, however some dispersed cells express relatively low levels of this epitope (arrow in panel A). Immunofluorescence staining with the 4H1 monoclonal antibody yields a relatively weak signal. Localization of the epitope is particularly weak in dispersed cells (arrows in panels C and D). IgG negative control staining is shown in panel F. Bar = 250 um. 101 Figure 23: Expression of the cell proliferation associated antigen Ki-67 in colonies exhibiting mixed phenotypes (panels B and D). Immunofluorescence localization of Ki-67 is relatively weak, but can be localized to many peripherally located cells of HBEC colonies (arrows in panels A and C). IgG negative control staining is shown in panel F. Bar =100 um.. 103 B • * • Q D * • I * % * LL Figure 24: Expression of MUC1 (green; using the EMA antibody clone) and keratin 14 (red) in a 9 day-old colony generated from a single cell. Note that some of the centrally located cells express low levels of both of these proteins (asterisk), whereas expression of both these proteins at intermediate levels within a single cell is also observed (arrow). Immunostaining with negative control FITC and PE antibodies is shown in panel D. Bar = 250 um. 105 - B IT *• D * l Ok Figure 25: Expression of BGA (panel B) and K8/18 (panel E) among cells of colonies composed of dispersed HBEC. All the cells of these colonies express BGA, however heterogeneous expression of K8/18 is observed (see colony on the left in panel E). IgM negative control staining for panel B is shown in panel D; IgG negative control staining for panel E is observed in panel F. Cells in panel E have been counterstained with Wright's Giemsa, whereas those in panel F have not. Bar in panels A and C = 250 um. Bar in panels E and F = 100 um. 107 Figure 26: Expression of CD44v6 (panel B) in colonies composed of closely arranged cells and in colonies composed of loosely arranged cells (arrow in panel A). Cells of both types of colonies express CD44v6. IgG negative control staining is shown in panel D. Bar = 250 urn. 109 B 4 / 1 D Figure 27: Expression of EpCAM (using VU-1D9 antibody clone; panel A), and the 5E11 (panel C) and 4H1 (panel E) epitopes in normal adult human breast tissue. Expression of these molecules is localized to the cell membranes of luminal cells. Basally located cells are unstained (arrowheads), as is the stroma. IgG negative control antibody binding is shown is panels B, D and F. Bar in A-F = 50 um. i l l Figure 28: Histograms demonstrating the ability of the monoclonal antibody 4H1 to block binding of 5E11 antibody to T-47D cells. When T-47D cells pre-exposed to FITC conjugated control antibodies are analyzed on a flow cytometer, the dot plot in (A) is observed. When these cells are incubated with FITC-conjugated antibodies recognizing EpCAM (B) or 5E11 (E) in the presence of IgG anti-dextran antibodies, an increase in fluorescence intensity is observed. Note that the increased fluorescence due to EpCAM antibody binding (Ber-EP4 clone) is not blocked by pre-incubation with either 4H1 (C) or 5E11 (D). antibodies. However, pre-incubation of T-47D cells with 4H1 antibodies could block binding of 5E11 antibodies (F). 1 1 3 Figure 29: Squamous metaplastic differentiation within mixed/myoepithelial cell colonies. The lower panel is a higher magnification of the boxed region shown in the upper left panel. Note the presence of the large flat cells (arrows) that are positioned superficial to other cells. Bar = 250 um. 115 Figure 30: Phase contrast photographs of Wright-Giemsa stained colonies grown in the presence and absence of 10 ng/ml EGF. Note that in the presence of EGF, the centrally located cells become very squamous and the peripherally located cells become migratory. Many of the peripherally located cells are undergoing mitosis (arrows). Bars = 250 um. 117 118 Figure 31: K6 expression in a mixed/myoepithelial colony in which the centrally located cells have become very "balled up" (panel A). These "balled up" cells express K6 (arrow) whereas the cells that are migrating out from the center of the colony express low or no levels of this protein. IgG negative control staining is shown in panel B. Bar = 100 um. 119 Figure 32: Expression of collagen IV in dispersed cell cultures. Cultures enriched for cells with a dispersed cell morphology were grown to approximately 20% confluence in SF medium. Following a change to EGF- and insulin-free SF medium, the cultures were maintained for a further 14 days and then stained to detect collagen IV. As indicated by the arrows in panel A, collagen IV is localized to the cells and to the extracellular spaces. IgG negative control antibody staining is shown in panel B. Bar = 250 um. 121 111 Figure 33: (A) Typical relationship between seeding density and colony formation by FfBEC cultured within collagen gels in the presence of SF medium and FfMF. Unsorted FfBEC from a 3 day-old primary culture were seeded at varying densities and cultured for 14 days, then the gels were removed, placed onto a grid and the number of colonies present counted under a dissecting microscope. (B) Photograph of a colony generated from a CALLA + precursor after 10 days in collagen gel culture in SF medium supplemented with 5% FBS. The large black circles are air bubbles within the collagen gel. Bar = 50 um. 123 Figure 34: Phase contrast photograph of colonies generated from MUC1+/CALLA" precursors after 14 days of growth in a collagen gel in SF medium and in the presence of irradiated HMF feeders (panel A). A hematoxylin and eosin-stained cross section of a typical colony is shown in panel B. The cells form a somewhat disorganized simple cuboidal epithelium surrounding a central lumen. These cells express keratin 19 (panel C) and keratin 8/18 (panel D), but, as illustrated in the latter, the localization of these keratins is not necessarily evenly distributed throughout all the epithelial cells. As shown in panel F, MUC1 (detected using the HMFG-2 antibody clone) is not localized to the apical cell surface, but is instead distributed throughout the cytoplasm. IgG negative control antibody staining for panels C and D is shown in panel E; IgG negative control antibody staining for panel F is shown in panel G. Bar in panel A = 100 um. Bars in panels B-F = 20 um. Bar in panel G = 50 um. 125 D l i f e Figure 35: Cross sections of colonies generated from M U C 1 + / C A L L A ' precursors after 14 days in a collagen gel in SF medium and in the presence of a HMF feeder layer. Most cross sections do not contain any cells expressing K14, however occasional cells do express this protein (panel A). Expression of this protein is weak and localized in epithelial cells adjacent to the extracellular matrix. All colonies have cells that express low levels of CD44v6 in a pattern similar to K14 (panel B). IgG negative control antibody staining is shown in panel C. Bar = 50 um. 127 Figure 36: Phase contrast photograph of a colony generated from a MUC17CALLA + precursor after 14 days in a collagen gel (panel A). A hematoxylin and eosin-stained cross section of such a colony reveals a heterogeneous population of cells that exhibit early characteristics of squamous metaplasia (panel B). Note that the centrally located cells are large and pale staining, whereas the surrounding cells are more squamous. Immunocytochemical staining of the same colony to detect the HMFG-2 (MUC1) epitope reveals that the large pale cells express this mucin, whereas the surrounding cells do not (panel C). Peripherally located cells in these colonies express CD44v6 (panel E) and K14 (panel G). As shown in panel F, cells expressing K8/18 are randomly distributed in these colonies. IgG negative control antibody staining for panel C is shown in panel D; negative control antibody staining for panels E-G is shown in panel H. Bar in panel A = 200 um. Bar in panels C, D, E and G = 50 um. Bar in panels F and H = 20 um. 129 ISO CHAPTER IV: CHARACTERIZATION OF HUMAN BREAST EPITHELIAL CELL PROGENITORS Introduction Since it was clearly evident that different types of colonies containing cells expressing luminal and/or myoepithelial markers could be reproducibly and quantitatively obtained from single isolated precursors, it became feasible to use FACS-sorting to determine whether different colony types were produced by phenotypically distinct subpopulations and what features they might possess, (see general design illustrated in Figure 37). This strategy allows not only for a phenotypic characterization of the clonogenic cells, but also for the identification of parameters that permit highly purified progenitor cell populations to be obtained. Experiment 1: Characterization of HBEC Progenitors The cell sort markers analyzed were the epidermal growth factor receptor (EGFR, also known as (erbB-1), erbB-2, a6 integrin, CALLA, EpCAM, MUC1 and retention of the fluorescent dye Rhodamine 123 (Rhl23). The distribution of EGFR was examined among HBEC progenitors because its ligand is associated with ductal development in mice (Coleman et al 1988, Snedecker et al 1992, Haslam et al 1993). The distribution of ErbB-2 was examined because this receptor is activated by NDF (indirectly via erbB-3 or erbB-4), a ligand that stimulates lobulo-alveogenesis in the mouse (Yang et al 1995, Jones et al 1996). As well, this protein is commonly overexpressed in 25-30 % of human mammary tumors and is associated with a poor prognosis (Slamon et al 1987, Slamon et al 1989). Alpha 6 integrin and CALLA have previously been identified as basal cell markers (Koukoulis et al 1991, Karsten et al 1993, Gusterson et 1986), whereas EpCAM (also known as epithelial specific antigen) and MUC1 are luminal cell markers (Burchell et al 1983, Latza et al 1990, Simon et al 1990). Retention ofthe fluorescent dye Rhl23 was also studied since this has been demonstrated to be useful in identifying subsets of mouse 131 multipotent hematopoietic stem cells (Rebel et al 1996, Spangrude and Johnson 1990, Li and Johnson 1992). Poor retention of this dye is correlated to expression of the multidrug efflux pump P-glycoprotein (Chaudhary and Roninson 1991, Lee et al 1994). Methods Tissue source Normal mammary tissue isolated from women undergoing reduction mammoplasties were used for all experiments. Because of the greater difficulty in obtaining sufficient numbers of single cells from freshly dissociated breast tissue (required for FACS analysis and sorting), most of the sorting studies were undertaken with suspensions that had been cultured for an initial 3 -6 days. However, in 5 experiments, breast cells were sorted by FACS without any prior cell culture. In these, normal mammary tissue was enzymatically dissociated, treated with NH4CI and frozen as described earlier. To generate a single cell suspension, the thawed breast epithelial cell organoid suspension was incubated with prewarmed STV solution supplemented with 50 ug/ml DNAse I and 5 caseinolytic units/ml of dispase for 15 - 20 min with occasional gentle pipetting. Following filtration of the cell suspension through a 20 um mesh (BioDesign Inc., New York, NY), the filtrate was centrifuged and the pellet preblocked with human serum and stained for FACS as described previously. During FACS sorting only those experiments in which all HBEC subpopulations were sorted were used to quantitate the number of progenitors present within the different HBEC subpopulations and to infer the total progenitor content of the initial starting population. Sort markers Monoclonal antibodies recognizing EGFR (IgG2 clone 528 from Santa Cruz Biotechnology Inc., Santa Cruz, CA., U.S.A.; used at 2 ug/ml), erbB-2 (IgGl clone 9G6 from Santa Cruz used at 2 132 ug/ml), <x6 integrin (rat monoclonal clone GoH3 from PharMingen Canada, Mississauga, Ont., Canada, used at 5 ug/ml), BGA (mouse monoclonal IgM clone 92FR-A2 from DAKO diluted 1:50), CALLA (IgGl clone SS2/36 directly conjugated to FITC or PE from DAKO, used at 1:10 dilution), EpCAM (IgGl clone VU-1D9 from Novocastra used 0.4 ug/ml and IgGl clone Ber-EP4 directly conjugated to FITC from DAKO used at 1:10 dilution), MUC1 (IgG clones HMFG-2, 214D4 and E29 (also known as EMA) were used as previously described) and retention the fluorescent dye Rhl23 (Molecular Probes Inc., Eugene, Ore., U.S.A.) were used as the basis for cell sorting. For Rhl23 staining, 3-day HBEC cultures were first incubated for 45 min with 0.5 ug/ml Rhl23 dissolved in primary culture medium and then for 30 min with Rhl23-free primary culture medium prior to cell harvesting for FACS analysis (Ronot et al 1990). Retention of Rhl23 was measured on the FL1 (green fluorescence) channel. Secondary antibodies used included goat anti-mouse IgG (H+L) directly conjugated to PE (Jackson ImmunoResearch) used at 1:125 dilution and goat anti-rat IgG (H+L) directly conjugated to FITC (Jackson ImmunoResearch) used at 1:100 dilution. IgGl control antibodies directly conjugated to either FITC or PE (DAKO) used at a 1:10 dilution were used as negative controls for the CALLA and EpCAM primary antibodies that are directly conjuagated to FITC or PE. An IgGl mouse monoclonal antibody recognizing dextran diluted to 5 ug/ml was used as a negative control form the EGFR, erbB2 and EpCAM staining. Normal rat serum diluted to 10% in HFN served as the negative control for GoH3 antibody binding. A mouse IgM negative control antibody (DAKO) diluted 1:10(10 ug/ml) was used as a negative control for anti-BGA antibody binding. 133 Limiting dilution analysis To detect the frequency of MUC17CALLA+ HBEC that can generate MUC1 + progeny, limiting dilutions (102 to 8 x 103) of MUC17CALLA+ HBEC FACS sorted from 5 primary cultures were seeded in 6 replicate wells in 96-well plates in SF medium supplemented with 5% FBS. After 24 h, the medium was changed to SF medium and the culture maintained for a further 10 days with tri-weekly media changes. Cultures were terminated by fixation with acetone:methanol and all the wells stained using APAAP to detect the HMFG-2 (MUC1) epitope. The frequency of the progenitors was then inferred by determining the seeding density which resulted in 37% of seeded wells were negative for the presence of MUC1 + cells (Sharrock et al 1990). HBEC colony forming assay With the exception of some of the initial HBEC FACS sorting experiments using MUC and CALLA staining as the sort parameters (i.e., those in which sorted cells were analyzed for colony-forming ability in low density culture in the absence of any feeder or in collagen gel culture in the presence of a HMF feeder), all FACS-sorted HBEC subpopulations were analyzed for progenitor content using the 2-D HBEC colony assay (Experiment 2, Chapter ni). The initial analysis for detection of progenitor content in MUC1 and CALLA FACS-sorted populations relied on culturing HBEC at low density (800 cells/cm2 in 24-well plates) in SF medium for 10 -14 days. The HBEC colony assay used for all other analyses consisted of seeding HBEC (sorted or non-sorted) at 2x103 cells/60 mm dish (approximately 100 cells/cm2) in SF medium supplemented with 5% FBS and in the presence of viable but irradiated NTH 3T3 cells seeded at 6 x 103 cells/cm2. A minimum of 3 plates were seeded for each HBEC sample. Forty-eight hours later, the medium was changed to SF medium and the culture maintained for a further 7 to 10 days with no further medium changes. Plates were fixed in acetone: methanol and stained with Wright-Giemsa. The number of colonies containing cells that express only luminal cell-134 morphology and the number of colonies containing dispersed teardrop-shaped cells were then scored under a dissecting microscope. The progenitors that generated these morphologies were then defined as being luminal-restricted and mixed/bipotent, respectively. All 60 mm plates were number coded and scored blind. Results The staining pattern of the enriched HBEC population obtained from enzymatically dissociated non-cultured normal human mammary tissue after dual labelling of the cells with anti-CALLA and anti-MUCl (using the HMFG-2 clone) antibodies is illustrated in Figure 3 8 A. The predominant population was consistently negative for both of these surface markers (54% in the representative sample shown in Figure 38A). The next most common phenotype was a population of MUC17CALLA + cells (36% in Figure 3 8 A). A small population of MUC1 + /CALLA" cells (7% in Figure 3 8 A) could also be routinely identified, but a significant population of M U C 1 + / C A L L A + cells was not observed. After 72 h of culture of the same tissue sample, a small, weakly double-positive population became apparent and the proportion of double-negative cells declined slightly (Figure 38B). Preliminary analysis of the progenitor content in HBEC subpopulations sorted on the basis of MUC1 (using the HMFG-2 clone) and CALLA revealed that the double-negative and MUC1 + /CALLA" fractions were highly enriched for progenitors that generate colonies composed only of cells expressing luminal cell markers in 2-D cultures and small spherical colonies in collagen gel cultures. Conversely, the MUC17CALLA + subpopulation was highly enriched for progenitors of mixed/myoepithelial colonies when assayed in 2-D cultures and of large branched colonies in collagen gel cultures. Immunocytochemical staining to detect MUC1 135 and K14 in colonies generated from CALLA + precursors revealed the presence of bipotent progenitors within CALLA + sort fractions. Limiting dilution analysis revealed that the frequency of cells within the CALLA + sort population that could generate MUC1 + progeny (as detected by using the HMFG-2 antibody clone) in 2-D cultures (in the absence of any feeders and in the presence of SF medium) after 11 days in vitro was 1 cell in every 1800 ± 400 CALLA + cells (range = 1 in 500 - 3000; 5 samples analyzed). The next series of experiments examined the distribution of MUC1 and CALLA among FfBEC in primary culture, but using the 214D4 antibody clone instead of the HMFG-2 clone to detect MUC1. The rationale for this was that the 214D4 clone recognizes a greater proportion of the HBEC population than the HMFG-2 clone (Dr. J Hilkens, personal communication). A comparison of the expression of the HMFG-2 and 214D4 MUC1 epitopes in relation to CALLA by HBEC from primary cultures is shown in Figures 38C and D. In this example, only 10% of the total cell population stained with HMFG-2, whereas 42% of the cells stained positively with 214D4. The average proportion of primary cultured HBEC that expressed the HMFG-2 and 214D4 epitopes was 9 ± 3% and 40 ± 7% respectively (6 cultures analyzed). In the 214D4/CALLA dot plot, three general populations can be identified: MUC17CALLA +, M U C l ^ C A L L A 1 and MUC1+/CALLA". The distribution of colony phenotypes generated by seeding progenitors enriched within these three subpopulations is shown in Figure 39. In these experiments, colonies were classified (by gross morphology alone) as having either a luminal cell-restricted morphology (compact cell arrangement), a myoepithelial cell-restricted morphology (dispersed cell arrangement) or elements of both (mixed). Luminal cell-restricted progenitors were enriched the fraction of MUCl + /C ALL A" cells, whereas mixed/myoepithelial cell-restricted progenitors were enriched in the fractions of MUCl'Vc ALL A* and MUCl' /CALLA + cells. It was noted that colonies in the fraction of MUCl + /C ALL A" cells tended to be 136 much more "rounded up" than those generated from the MUCl^VCALLA* cells (e.g., compare Figures 40A and B). An unfortunate drawback of using CALLA and MUC1 (using the 214D4 antibody clone) to identify HBEC in vitro is that mammary fibroblasts upregulate expression of these epitopes when HBEC are maintained in culture, thus accounting for a large portion of the MUC-\-/CALLA± population that emerges in the 3 day primary cultures (Figure 38D). To obtain more definitive evidence of the phenotype of the different progenitors, HBEC from 3-day old primary cultures of normal mammary tissue were sorted on the basis of their expression of EpCAM (using the Ber-EP4 antibody clone). When seeded at low density, sorted EpCAM" cells consisted mainly of stromal cells and teardrop-shaped dispersed cells, but occasional daughter cells expressing either MUC1 or EpCAM were seen. The EpCAM + HBEC isolated from primary cultures could be subdivided according to their expression of CALLA (Figure 38E). CALLA"/EpCAM + progenitors generate progeny identical to the MUC1 + /CALLA" progenitors, whereas CALLA^EpCAM* cells generated progeny identical to those obtained from MUCl^CALLA* and MUC17CALLA + cells. A typical dot plot generated on the basis of expression of EpCAM (detected using the VU-1D9 antibody clone) and retention of Rhl23 is shown in Figure 41 A. The distribution of stromal and the different HBEC progenitors among the different sort fractions is shown in Figure 4IB. Approximately 90 ± 7% of the total number of progenitors were observed in the EpCAM1" fraction (4 primary cultures analyzed). Cells expressing EpCAM could be roughly divided into those that are Rhl23du11 and Rhl23b r i g h t (the typical sort gates for these 2 fractions is shown in Figure 41). Highly purified (>99 %) populations of mixed/myoepithelial progenitors could be 137 enriched within the RM23 d u l l/EpCAM + fraction (Figure 4IB; 5 primary cultures analyzed). Approximately 98 ± 1% of all luminal cell-restricted progenitors were contained within the gj l223b n8h t/EpCAM+ fraction, however a significant proportion (31 ± 6%) of mixed/myoepithelial progenitors were co-purified in this subpopulation. Stromal cells were eluted in the Rh 123bright/EpCAM'fraction. A typical dot plot generated when cells harvested from a short term primary culture were labelled with fluorescently tagged antibodies recognizing EpCAM (detected using the Ber-EP4 antibody clone) and erbB-2 is shown in Figure 42A. Contrary to what was observed in the Rhl23/EpCAM sort experiments, only 17 ± 8% of the total number of progenitors were found to be within the EpCAM* population. This variability was found to be attributable to the different clones of antibodies used in the Rhl23 sorting experiments (clone VU-1D9 from Novocastra) and in the erbB-2/EGFR sorting experiments (clone Ber-EP4 from DAKO). Double labelling of cells isolated from 3 different short term primary cultures with these two clones of antibodies, each conjugated to a different fluorochrome, revealed that the Ber-EP4 monoclonal only binds to 17 ± 4 % of the cells recognized by the VU-1D9 clone (3 primary cultures analyzed). The erbB-27EpCAM" subpopulation was enriched for few progenitors, and was made up stromal cells. Expression of erbB-2 is relatively epithelial progenitor-specific since 98 + 1% of all FfBEC progenitors detected expressed this receptor (Figure 44; 5 primary cultures unalloyed). When the erbB-2+ population was divided on the basis of expression of EpCAM, 78 ± 10% of the total number of progenitors present in the original FfBEC population were found to be in the erbB-2+/EpCAM" population and to be of the mixed/myoepithelial type. Only 3 ± 2% of all mixed/myoepithelial progenitors were obtained in the erbB-2+/EpCAM+ population. Expression 138 of erbB-2 was found to be several-fold higher in the erbB-2+/EpCAM+ group than in the erbB-2+/EpCAM" population (note the log scale used on both the x and >y axes in Figure 42A). A typical dot plot showing the distribution of EGFR in relation to EpCAM (detected using the Ber-EP4 clone) among HBEC is shown in Figure 43 A. The average frequencies of the EGFR" /EpCAM", EGFR+/EpCAM", EGFR7EpCAM+ and EGFR +/EpCAM + cells in the original HBEC population 36 ± 6, 34 ± 6, 16 + 5 and 14 + 4% respectively (n = 6 samples analyzed). Approximately 88 ± 5% of all mixed/myoepithelial restricted progenitors expressed EGFR (Figure 43B; 6 primary cultures analyzed) and expression of this receptor did not have any correlation to enrich for luminal cell-restricted progenitors. Sorted stromal cells were isolated from the EGFR+/EpCAM" fraction. No gross morphological differences were seen in the colonies containing cells expressing only luminal markers obtained from EGFRTEpCAM1" or EGFR+ZEpCAM* progenitors. A typical dot plot showing the distribution of BGA in relation to EpCAM is shown in Figure 44A. The frequency of BGA + cells present within HBEC primary cultures was highly variable since 4 to 61% of the cells express this epitope (mean ± s.e.m. = 20 + 7%, 7 primary cultures analyzed). Although mixed/myoepithelial progenitors were not particularly enriched in any one fraction (Figure 44B), approximately 80 ± 6% of all luminal cell-restricted progenitors were present at an average purity of 75 ± 6% within the BGATEpCAM1" subpopulation. A typical dot plot observed when cells from a short term primary culture were labelled with antibodies recognizing the basal cell markers a6 integrin and CALLA is shown in Figure 45A (the typical sort fractions analyzed are indicated within the boxed areas). The distribution of 1 3 9 stromal and the different HBEC progenitors among the different sort fractions is shown in Figure 45B. Stromal cells were concentrated in the a67CALLA+ population, whereas the total HBEC subpopulation was distributed along the diagonal of the dot plot (6 primary cultures analyzed). Luminal progenitors expressed weak (w) levels of these proteins, whereas mixed and myoepithelial cell-restricted progenitors expressed both ct6 and CALLA. The low cloning efficiency of the cells in the a67CALLA" cell fraction was probably due to their overgrowth by co-purified HMF that do not express CALLA. The frequency of EpCAM* HBEC within the a6" /CALLA" fraction is highly variable. Of 3 samples analyzed, the mean was 34 ± 30% (range = 3 -96%). Analysis of short term primary cultures initiated with reduction mammoplasty specimens isolated from 19 different patients indicated that the HBEC colony assay described in this study allowed detection of progenitors (in the initial primary culture) at a frequency of 5.8 ± 0.8% (range = 0.003 - 12.3%), although cloning efficiencies of up to 33% for some HBEC sorted subpopulations could be observed. This mean cloning efficiency was approximately 5-fold higher than that observed for freshly dissociated non-cultured HBEC. This discrepancy is likely due to contamination of the freshly dissociated breast epithelial cell suspension with non-epithelial cells as well as loss of cells with low progenitor activity within the primary cultures. Of all the progenitors assayed using short term cultured material, 84 ± 5% were of the mixed/myoepithelial phenotype. However, considerable heterogeneity was seen within this progenitor population (range = 22 - 100%, 19 primary cultures analyzed). Some progenitors generated true mixed colonies composed of cells expressing luminal markers and cells expressing myoepithelial markers, whereas others generated closely arranged colonies of 140 dispersed cells that expressed variable levels of EpCAM, and others generated colonies composed of dispersed clusters of EpCAM' cells. Assaying the EpCAM* and EpCAM" cells isolated from freshly dissociated noncultured mammary tissue revealed that virtually all HBEC progenitors (>99%) were EpCAM* fraction since the frequency of progenitors in the EpCAM' and EpCAM* subpopulations were 0.001 ± 0.001 and 0.4 ± 0.2% respectively (5 breast tissue samples analyzed). Approximately 27% of freshly dissociated breast cells expressed EpCAM (range = 3 - 44%, 5 samples analyzed). The patient age and the cloning efficiencies (e.g., the frequency of the different progenitor types in the HBEC primary culture) of the mixed/myoepithelial cell- and the luminal cell-restricted progenitors of the 19 different tissue specimens is shown in Table 1. No obvious correlation between patient age or the days spent in culture and the presence of the different HBEC progenitors is observed. Unfortunately the reproductive histories of the patients from whom these specimens were isolated were not available. 141 T A B L E 2: Patient age, days in culture prior to HBEC assay, and cloning efficiencies (CE) of mixed/myoepithelial cell- (M) and luminal cell-restricted (L) progenitors. The ratio of mixed/myoepithelial cell to luminal cell-restricted progenitors for every 100 random progenitors is also given. itientAge Days MCE (%) LCE (%) Total CE(%) Ratio M: 20 3 5.6 1.0 6.6 85 : 15 21 4 6.7 0.7 7.4 91 : 9 26 3 11.9 0.4 12.3 97: 3 28 3 7.45 0.05 7.5 99 : 1 28 3 3.8 0.2 4.0 96 : 4 29 3 5.1 0.3 5.4 95 : 5 30 5 5.0 0.9 5.8 85 : 15 32 3 1.7 6.0 7.7 22 : 78 33 3 8.3 0.4 8.7 95 : 5 34 3 3.2 0.2 3.4 95 : 5 35 6 1.0 1.9 2.9 34 : 66 36 3 7.7 1.8 9.5 81 : 19 40 2 3.7 0.6 4.3 86 : 14 40 3 11.6 0.5 12.1 96 : 4 41 2 4.8 1.7 6.5 74 : 26 41 6 0.003 0 0.003 100 : 0 44 3 2.3 0.2 2.5 92 : 8 45 3 1.6 1.5 0.1 91 : 9 47 3 3.6 0.4 4.0 91 : 9 142 Discussion My results of assessing the clonogenic potential of immunophenotypically distinct HBEC demonstrates that there exist two separate progenitor populations with different biological properties. The first type of progenitor generates pure colonies composed of cells that express luminal cell-specific proteins. This progenitor expresses the luminal markers MUCl and EpCAM. The second type of progenitor generates colonies composed of cells expressing proteins specific for the luminal HBEC lineage as well as cells expressing proteins specific for the myoepithelial cell lineage. Since essentially all (>99%) HBEC progenitors from freshly dissociated breast tissue are localized within the EpCAM* fraction (as detected by the VU-1D9 antibody clone), it is concluded that these bipotent progenitors also exhibit luminal cell characteristics. This conclusion is further supported by the observation that the mixed colonies were composed of EpCAM7K14+/K19" cells surrounding the EpCAM*/K147K19* cells. Such a cellular arrangement suggests that it is the cells with a luminal phenotype that generates daughters expressing myoepithelial lineage markers, and not vice versa. Although both the luminal cell-restricted and bipotent progenitors colocalize with the EpCAM* fractions, the two can be phenotypically distinguished from one another by an increased expression of ct6 and CALLA, and a decreased retention of Rhl23 by the bipotent progenitors. The conclusion that cells expressing luminal markers are the source of cells expressing myoepithelial markers in vitro is also consistent with observations reported by others (Rudland et al 1991, Kao et al 1995, Kang et al 1997, Pechoux et al 1999), and are in concordance with the observations that mitotic cells within the human mammary epithelium in vivo are located exclusively within the luminal cell and basal clear cell compartment (Joshi et al, 1986; Ferguson, 1985; Ferguson, 1988). MUCl and CALLA were originally described to be markers of HBEC by O'Hare and colleagues (O'Hare et al 1991). These authors reported that fluorescently-labelled antibodies that recognize 143 MUC1 identify luminal cell progenitors, whereas those specific for CALLA identify myoepithelial cell progenitors. The results outlined in this thesis differ from these earlier findings in that the mixed colonies observed here were not recognized. As well, the presence of the very small balled up colonies composed of cells expressing luminal markers were not described. This may be due to the long culture period used before colonies were scored (28 days) in O'Hare's study. My results would suggest that these luminal marker+ cells seen here in mixed colonies might well have disappeared by 28 days, in which case such colonies would appear to contain only myoepithelial cells. The colony-forming abilities of flow sorted luminal and myoepithelial cells isolated from adult rat mammary glands has also been previously examined (Dundas et al 1991). The results obtained in these studies mirror the results described in this thesis in that pure luminal marker"1" cell colonies, pure myoepithelial marker+ cell colonies and mixed (K14+/K18+/K19+) colonies were all identified and the luminal cell-restricted and bipotent progenitors were found to be enriched in the luminal (monoclonal antibody 25.5+ - has a similar distribution in vivo as MUC1) but not the CALLA + sort fractions. Recently, a similar FACS sorting study was performed using freshly isolated mouse mammary tissue as the starting material (Smalley et al 1998). Colonies composed only of cells expressing myoepithelial proteins as well as mixed colonies are described, but surprisingly no homogenous luminal marker+ colonies are observed. The EGF receptor and the erbB-2 receptors are homologous transmembrane receptor tyrosine kinases. Autoradiography studies of peripubertal mice using 1 2 5I-EGF reveals this growth factor binds to the cap cells lining the TEB, the myoepithelial cells and the stromal cells surrounding the elongating duct. Minimal binding was observed within the luminal cell compartment (Coleman et al., 1988; Coleman and Daniel, 1990). A similar distribution of this receptor among 1 4 4 stromal and myoepithelial cells in the human mammary gland has been previously reported (O'Hare et al., 1989, Gompel et al., 1996). In these studies, detectable levels of receptor for EGF were present on stromal, mixed/myoepithelial cell-restricted progenitors and on approximately half of the luminal cell-restricted progenitors, although the absence of detectable levels of this receptor on luminal cell-restricted progenitors could not be correlated with any differences of the morphologies colonies they produced. The results obtained here showing the expression of erbB-2 on epithelial but not stromal cells is in agreement with a previous report that examined the distribution of this receptor in sections of human mammary tissue (Gompel et al, 1996). The observation that erbB-2 is expressed by all HBEC progenitors is of significant relevance to the identification of malignant breast epithelial cells in bone marrow aspirates. Aside from the keratins, it has been difficult to identify a cell marker that is expressed uniformly by all HBEC. The advantage of erbB-2 as a HBEC marker over the cytokeratins is that it is expressed on the cell surface, and thus could be used in bone marrow purging protocols in the treatment of patients with metastatic breast cancer. Considerable heterogeneity was observed when staining patterns with different clones of monoclonal antibodies reactive with identical molecules were compared. For example, the MUCl epitope recognized by the 214D4 antibody was found to be expressed more widely on HBEC than when the HMFG-2 clone was used. Such variability is consistent with the heterogeneous pattern of MUCl expression seen within luminal cells in vivo (Foster et al 1982, Edwards and Brooks 1984). With regards to HBEC enrichment strategies, a variety of combinations of antibodies are useful. For routine selection of cultured HBEC subpopulations, the marker combination of a6 integrin and CALLA has the advantage that the stromal, luminal, and bipotent plus myoepithelial subpopulations can be separately isolated using a single 145 combination of antibodies. Although some non-CALLA expressing stromal cells are eluted in the fraction of a67CALLA" cells, epithelial cells in this fraction could be purified by sorting via EpCAM. However, it remains undetermined if these a67CALLA" cells actually have different proliferative potentials from those eluted in the a 6 w e a k / C A L L A w e a k fractions. The antibody combination of a 6 integrin and CALLA would not be suitable for freshly dissociated mammary tissue since stromal cells will not have upregulated expression of CALLA (Dundas et al 1991, Clarke et al 1994). An untested, but potential strategy is to sort breast cells on the basis of expression of EpCAM (using the VU-1D9 clone) and a 6 integrin. Theoretically, these markers should allow for discrimination of luminal (EpCAM+/a6), bipotent (EpCAM+/a6+), myoepithelial (EpCAM7a6+) and non-epithelial (EpCAM/a6') subpopulations. Should sorted cells from freshly dissociated non-cultured tissue be required, EpCAM (using the VU-1D9 antibody clone) or erbB-2 used in combination with CALLA, a 6 integrin, or Rh 123, would be logical markers to identify subpopulations enriched for HBEC progenitors. A major aspect in the classification and isolation of different HBEC populations by flow cytometry is the procedure used to set the gates that distinguish different subpopulations. Analysis of HBEC on a flow cytometer typically does not always generate two obvious subpopulations on a dot plot, but instead a gradation of cells expressing varying levels of a cell marker (e.g. see CALLA vs a 6 dot plot in Figure 45 A) and a second parameter of interest (e.g. colony growth) that may not correlate perfectly with the expression of a given antigen. Progenitor fractions approaching 100% purity are possible; however, there is then usually a trade-off between sort gate stringency and cell yield. Cell yields of up to >2 x 105 sorted cells could be attained when provided with large enough samples, but on a more routine basis, yields of 5 x 104 to 105 sorted cells for each subpopulation were more the norm. 146 The. majority of mammary tumors have been found to express keratin 19, MUCl and ESA/EpCAM (Bartek et al 1985a, Latza et al 1990, Bailey et al 1996). Human breast epithelial cell selection strategies that utilize luminal cell-specific markers might therefore give the best enrichment of malignant HBEC. However, the data reviewed here show that nonmalignant luminal cells and their progenitors would be copurified in such an approach. EpCAM would likely be the most suitable marker for this since this epitope is specific for most epithelial cells, is not expressed by neural, muscular or connective tissue (Latza et al 1990), and appears to distinguish adenocarcinoma cells in serous effusions (Bailey et al 1996, Delahaye et al 1997). Alternatively, a variety of antibodies have been developed that recognize novel epitopes present on MUCl that are unmasked via aberrant glycosylation during tumor progression (Burchell et al 1987, Xing et al 1992, Croce et al 1997, Fiorentini et al 1997). As a result, these antibodies show preferential binding to malignant HBEC over normal HBEC. There was no correlation between the age of the tissue donors and the frequencies of the two types of colony-forming cells analyzed. This is in agreement writh previous published experiments that demonstrate success of primary culture is independent of the age of the tissue donor, but is instead dependent on the stage of mammary gland development at the time of specimen removal (Russo et al 1989). Experiment 2: Influence of EGF and NDF on HBEC Precursors In 1995 it was demonstrated that sequential administration of HGF and NDF (-0) promoted ductal elongation and lobuloalveologenesis, respectively, in mouse mammary glands maintained in organ culture (Yang et al 1995). Shortly thereafter, it was shown that implantation of NDF-a and -P-containing pellets into peripubertal mice stimulates ductal branching and lobuloalveologenesis (Jones et al 1996). Implantation of EGF containing Evlax pellets into 147 ovariectomized peripubertal mice has implicated a role of this growth factor as a promoter of ductal elongation (Coleman et al., 1988; Snedeker et al., 1992; Haslam et al., 1993). The existence of ductal and lobular precursors in the mouse mammary gland has been proven (Smith et al 1996), and it is the hypothesis of this thesis that the bipotent and luminal cell-restricted progenitors represent ductal and alveolar precursors, respectively. Since the effects of HGF on C A L L A + and MUC1 + sorted precursors has previously been described (Niranjan et al 1995), the effect of EGF and NDF-P on sorted HBEC was examined. Methods The effect of NDF-a was examined at two different time points during the course of this thesis. Initial experiments examined the effects of 2 - 50 ng/ml NDF on EpCAMVCALLA" cells in bulk culture (9 x 103 cells/cm2) and maintained in SF-medium supplemented with 2.5% FBS for 8 days. No feeders were used in this experiment. The influence of NDF on mixed/myoepithelial cell restricted progenitors was not tested at this time since methods to isolate these progenitors at sufficient numbers and free of stromal cells were not developed. Relative cell numbers were measured using the MTT assay. In the second set of experiments performed much later, the effect of 20 ng/ml NDF on HBEC progenitor activity in the HBEC colony assay (e.g., SF-medium and a NTH 3T3 feeder) was assessed. In experiments to assess the effect of EGF on HBEC progenitor activity, HBEC isolated from 7 primary cultures were sorted on the basis of EGFR and EpCAM and were seeded into the HBEC colony forming assay in the presence or absence of 10 ng/ml EGF, and the number of colonies generated in each condition scored. Colonies grown in the absence of EGF were not scored differentially with regards to having a luminal or mixed/myoepithelial morphology since their distinctive morphologies were lost under such conditions. In 3 of these experiments, some of the 148 resultant colonies maintained in the presence and absence of EGF were stained by immunocytochemistry to detect expression of keratin 14. The effect of EGF on CALLA + progenitors cultured within a collagen gel was also examined. FACS sorted CALLA + FfBEC from 3 primary cultures were seeded at 5 x 102 - 104 cells/ml gel in the presence of irradiated HMF as described in the General Methods chapter. The cells were maintained in F12/DME/H supplemented with 1 mg/ml BSA, 1 ug/ml INS, 0.5 ug/ml HC, 10 Tjg/ml CT, 15% FBS and with or without 10 ng/ml EGF. Medium was changed twice weekly, and after 10-14 days in vitro, the colonies were scored under a dissecting microscope and photographed. Results In the initial experiments examining the effect of NDF on luminal cell-restricted (EpCAM^/CALLA") progenitors, NDF is mildly growth inhibitory at concentrations of 2 ng/ml and above (p< 0.05, Figure 46A; 4 cultures analyzed). However, inclusion of 20 ng/ml NDF in the HBEC colony assay in the later experiments did not affect on the cloning efficiencies of either type of progenitor or the gross morphology of the resultant colonies (Figure 46B; 4 cultures analyzed). As illustrated in Figure 47A, deletion of 10 ng/ml EGF from the HBEC colony assay growth medium significantly altered the cloning efficiency of EGFRTEpCAM* progenitors (p < 0.05). Sorted cells enriched for luminal cell progenitors (>80% purity) were highly dependent on EGF in the culture medium, since deletion of this growth factor resulted in a >5-fold decrease in cloning efficiency of these progenitors (p< 0.05, 3 cultures analyzed). Although no gross morphological change was observed in colonies generated from luminal cell-restricted 149 progenitors other than a reduction in their size, mixed/myoepithelial cell colonies showed a greatly altered morphology (compare Figures 30A and B). In the absence of EGF, the latter type of colonies remained as clusters of tightly arranged cells. In the presence of EGF, peripherally positioned K14+ cells become migratory. Seeding CALLA + progenitors within 3-dimensionsal collagen gels in the presence and absence of 10 ng/ml EGF also revealed a role of EGF in HBEC migration (compare Figures 47 B and C). Discussion An influence of EGF on ductal elongation of the peripubertal mouse mammary gland has been unequivocally demonstrated (Coleman et al 1988, Snedeker et al 1992, Haslam et al. 1993, Wiesen et al 1999). Ductal elongation occurs via the cell division and migratory activities of the TEB. The cap cells that line the distal aspect of the TEB are one of the leading candidates for a stem cell (Rudland 1993, Rudland et al 1997, Williams and Daniel 1983), although others have argued that this cell is merely a myoepithelial progenitor that paves the way for ductal elongation (Sapino et al 1993). Data presented in these studies demonstrate that the migration of K14+/SMA" cells appear to be involved in what might be interpreted as ductal elongation of the mammary tree. The concept that basally derived cells are responsible for ductal morphogenesis is supported by the morphogenic effects of HGF on CALLA + but not MUC1 + HBEC cultured within collagen gels (Niranjan et al 1995). Branching morphogenesis of mammary epithelial cells in response to agents such as HGF and EGF is dependent on expression of epimorphin, a molecule restricted to the mammary stroma and the basally located cells of the mouse mammary gland (Hirai et al 1998). Although EGF did not influence the colony morphology of luminal cell-restricted colonies, the clonogenicity of luminal cell-restricted progenitors (particularly those that are EGFR") were 150 highly dependent on the presence of this growth factor. This somewhat paradoxical observation that EGFR"EpC AM* progenitors are dependent on EGF for colony formation suggests that these cells either express EGFR levels below the level that can be detected by antibody binding and FACS, or EGF exerts an effect via the NTH 3T3 stromal feeder layer. This latter explanation is supported by the observation that EGF regulated ductal elongation in vivo is regulated via EGFR in the stromal compartment (Wei sen et al 1999). Experiment 3: Do Mixed/Myoepithelial-Cell Colonies Contain Luminal Cell-Restricted Progenitors? To determine if cells present within mixed/myoepithelial-cell colonies can form secondary colonies that are composed of only cells that express luminal markers, or if they are ultimately restricted to produce K14+progeny, a serial passaging study was performed. Methods Eight day-old colonies generated from 5 different enriched populations (>99% homogeneity) of mixed/myoepithelial-cell progenitors were harvested with the 0.025% trypsin solution, treated with dispase and DNAse, filtered through a 20 um mesh to attain single cells, and the cells replated in a new (secondary) HBEC colony assay. Colonies were then maintained for a further 8 days, harvested and subcultured again and the procedure repeated once more. Prior to each split, triplicate plates were stained with Wright-Giemsa and the colonies scored. Colonies were categorized (by morphology alone) as having a close cellular arrangement, as having a dispersed cellular arrangement (or, if the cultures were senescing, a flattened vacuolated "fried egg" appearance), or containing mixed elements. 151 Results Assessment of replated cells showed that neither mixed nor pure colonies composed of dispersed cells could generate colonies of closely arranged cells in secondary or subsequent cultures. As shown in Table 2, the proportion of compact colonies with closely arranged cells to mixed/myoepithelial cell and pure dispersed cell colonies declined dramatically over 3 passages and, by the third passage, only pure dispersed cell colonies were obtained, many of which were vacuolated. In our experience, such cells are associated with cellular senescence. No evidence to support the concept that cells of the mixed/myoepithelial colonies could general luminal cell-restricted progenitors was obtained. The average colony size also declined in the first, second and third passages from 185 ± 22, to 69 + 10 and 6 ± 1 cells/colony (5 samples analyzed, 20 - 25 colonies scored per passage for each sample). Assuming all of the cells are divided, this equates to an average of 16.2 ± 0.4 cell doublings (range = 15.0 - 17.1; 5 cultures analyzed) of the initial of cells. Since these cultures were maintained for a total of 26 days, this would equate to 0.62 cell doublings/day or a cell cycle time of 35.8 hours. Table 3: The ratios of different types of colonies obtained with serially passaged cells. Results are based on assays of 5 different samples of enriched mixed/myoepithelial cells and are expressed as a percentage of the total number of colonies observed during that passage. Frequency of colony type (%) Passage number Closely arranged Mixed Dispersed 1 0.3+0.1 81.1 ±4 .4 18.7 ±4 .4 2 0 10.8 ±2 .0 89.2 ± 2.0 3 0 2.6 ±0.4 97.4 ± 0.4 152 Discussion These results indicate that, under the culture conditions used, the mixed/myoepithelial cell-restricted progenitors rapidly exhaust their capacity to produce progeny that express luminal markers but continue to produce cells that can differentiate into cells that express myoepithelial lineage markers. It may be that the culture conditions used in these experiments lacked the appropriate signals to permit the survival of luminal cell-restricted progenitors. Agents that might provide such signals could include serum (Ethier et al 1993), HGF (Pechoux et al 1999) and low calcium culture conditions (Berthon et al 1992) since these have been shown by others to promote the survival of cells expressing a luminal phenotype. The progenitors isolated from the HBEC primary cultures underwent at least 16 cell doublings before senescing. Assuming an extra 2 cell doublings in the 3 day pre-assay culture period, this brings the total number of doublings to 18, which is below the value (40 - 50 cell doublings) reported for mouse mammary stem cells (Kordon and Smith 1998). This suggests that the initial inoculum of cells does not contain true mammary stem cells and/or that the in vitro conditions used here are still suboptimal for mammary stem cells. 153 Figure 37: General protocol for studying the clonal growth of immunophenotypically distinct FfBEC isolated from primary cultures of HBEC. 154 Normal human mammary tissue Enzymatic dissociation Three day culture Cell harvesting and labelling for F A C S Seeding of immunophenotypically distinct cells in an in vitro H B E C colony forming assay 9-12 days Analysis of colonies Figure 38: Dot plots generated from FACS analyses of cells in 3-day primary FfBEC cultures. The dot plots in panels A and B were generated from a tissue source before (A) and after (B) 3 days of primary culture. The dot plots in panels C and D were generated from a separate 3 day primary culture. Cells were stained with FfMFG-2/goat anti-mouse-PE and CALLA-FITC (panels A-C), 214D4/goat anti-mouse-PE and CALLA-FITC (D) or with CALLA-PE and ESA-FITC (E). The crossed lines are adjusted such that cells stained with isotype control antibodies are all situated within the lower left quadrant. 156 C A L L A C A L L A C A L L A C A L L A E p - C A M Figure 39: Distribution of colony types produced by HBEC with varying phenotypic profiles. HBEC from 5 primary cultures were sorted according to their expression of MUC1 and CALLA and then seeded at low density (800 cells/cm2) in culture and maintained in SF medium. After 12 days, the cultures were fixed, stained with Wrights Giemsa and the colonies scored. The colonies were scored as being composed of cells that express only luminal markers (luminal), myoepithelial markers (myoepithelial), or elements of both (mixed). 158 Figure 40: Representative morphologies of colonies generated from MUC1+/CALLA" (A) and MUCl^CALLA* (B) progenitors cultured at low density culture for 12 days in SF medium. Both of these progenitors were sorted from the same primary culture. Bar =100 um. 160 1*1 Figure 41: Representative dot plot showing the distribution of EpCAM and the retention of Rhl23 among the cells isolated from a short term HBEC culture. The boxed regions show the typical sort gates used to isolate Rhl23d u l l/EpCAM+ and Rhl23 b r i g h t/EpCAM + cell fractions. A similar division in rhodamine fluorescence was used for sorting the EpCAM" fraction. The relative frequencies of the mixed/myoepithelial cell- and luminal cell-restricted progenitors among the different sorted FfBEC subpopulations isolated from 4 primary cultures is shown in the graph below. 162 Rhl23 KXXX MIXED/MYO LUMINAL 60 h Figure 42: Representative dot plot showing the distribution of erbB-2+ and EpCAM* cells among those isolated from a short term FfBEC culture. The quadrant settings reflect the sort gates. The average relative frequencies of mixed/myoepithelial cell- and luminal cell-restricted progenitors among the different sorted FfBEC subpopulations isolated from 5 different primary cultures is shown in the graph below. 164 g>' I" i v% t u r n >' r-i' i«*w" 1 r111 ii1 nnw 1 • • i" •» i'wiiri w tS* iSa ii> iir* E p - C A M MIXED/MYO LUMINAL 100 9 0 8 0 70 co cc g 6 0 z U l 8 5 0 cc a . h-Z UJ 4 0 O tc LU CL 3 0 20 10 0 erbB2-EpCAM- erbB2+EpCAM- erbB2+EpCAM+ HBEC SUBPOPULATION Figure 43: Representative dot plot showing the distribution of E G F R + and E p C A M * cells among those isolated from a short term HBEC culture. The quadrant settings reflect the sort gates. The average relative frequencies of the mixed/myoepithelial cell-and luminal cell-restricted progenitors among the different sorted HBEC subpopulations isolated from 6 primary cultures is shown in the graph below. 166 E p - C A M 0004 MIXED/MYO X^Zii LUMINAL 90 . _ EQFB-EpCAU- EQFB*EpCAM- EQfB*epCAU» EOFR-EpCAM* HBEC SUBPOPULATION U7 Figure 44: Dot plot showing the distribution of BGA + and EpCAM* cells among those isolated from a short term FfBEC culture. The presence of the BGA+/EpC AM* cells was variable in these short term FfBEC cultures. The quadrant settings reflect the typical sort gates. The average relative frequencies of the mixed/myoepithelial cell- and luminal cell-restricted progenitors among the different sorted HBEC subpopulations isolated from 7 primary cultures is shown in the graph below. 168 BGA+EpCAM- BGA-EpCAM- BGA+EpCAM+BGAEpCAM+ HBEC SUBPOPULATION Figure 45: Representative dot plot showing the distribution of a6 integrin* and CALLA* cells among those isolated from a short term FfBEC culture. The boxed areas illustrate typical sort gates for the ct67CALLA\ a6 w e a k /CALLA w e a k (A6wCw), a6*/CALLA* and a67CALLA* cell fractions. The average relative frequencies of the mixed/myoepithelial cell- and luminal cell-restricted progenitors among the different sorted HBEC subpopulations isolated from 6 different primary cultures is shown in the graph below. 170 A6 I N T E G R I N K X X H MIXED/MYO WZZh LUMINAL 100 -90 " 80 -A6-CALLA+ A6+CALLA+ A6wCALLAw A6-CALLA-HBEC SUBPOPULATION ill Figure 46: Influence of NDF on HBEC progenitors. In panel A, CALLA'/EpCAM* cells sorted from 4 primary cultures were seeded at 9 x 103 cells/cm2 into the wells of 96-well plates and maintained in SF medium supplemented with 2.5% FBS and varying (0-50 ng/ml) concentrations of NDF. After 8 days in culture, the relative number of cells in each culture condition was determined using the MTT assay. Results indicate that NDF at all concentrations tested are mildly inhibitory (p<0.05). However, this growth inhibitory effect was not observed on luminal cell-restricted or mixed/myoepithelial progenitors when tested in the HBEC colony forming assay (panel B). In these latter experiments, HBEC from 4 different primary cultures were seeded in SF medium supplemented with 20 ng/ml NDF in the presence of NTH 3T3 cells, and maintained for a further 8-10 days. 172 0 2 20 50 NDF (ng/ml) Figure 47: (A) Influence of EGF on the cloning efficiencies of HBEC progenitors of varying phenotypes. HBEC from 6 different primary cultures were sorted on the basis of EGFR and EpCAM and the progenitor activity in the isolated subpopulations was assessed in the HBEC colony assay in the presence and absence of 10 ng/ml EGF. EGFRTipCAM* progenitors were significantly growth inhibited by withdrawal of EGF from the culture medium (p<0.05), although the growth inhibitory effect of EGF withdrawal on EGFR^pCAM* approaches significance (p = 0.074). When CALLA + sorted HBEC were seeded within a collagen gel in the presence of a HMF feeder in SF medium supplemented with 15% FBS and with or without 10 ng/ml EGF, a striking difference in colony morphology was observed (compare panels B and C). In the absence of EGF, no colony branching occurred, whereas in the presence of EGF, this was extensive. Bar =100 urn. 174 175" CHAPTER V: GENERAL DISCUSSION Cancer theory suggests that it is the stem cells, or their immediate progeny, that are the most common targets leading to naturally occurring cancers. Considering that the vast majority of human mammary tumors exhibit luminal cell markers and not myoepithelial cell markers, it would be reasonable to suggest that the regenerative cells within the mammary epithelium would have a luminal phenotype. Results from assaying the clonogenic potential of immunophenotypically distinct HBEC are consistent with this hypothesis. Two distinct progenitors in the HBEC population were identified here. The first type of progenitor generates colonies of varying sizes composed of cells homogenous in their expression of luminal lineage related proteins. These progenitors are enriched exclusively in populations expressing two luminal cell markers (MUC1 and EpCAM). The second type of progenitor generates colonies composed of cells expressing luminal markers and cells expressing myoepithelial markers. These colonies are characterized by a centrally located cluster of flattened cells expressing typical luminal-specific epitopes and these are surrounded by cells expressing myoepithelial-specific epitopes. Considering the latter are found within the EpCAM* population and that the mixed colonies are composed of EpCAM'/K14+/K19+/SMA7cells surrounding the EpCAM+/K147K19+ cells, it is concluded that it is the cells with a luminal phenotype (or the progenitors that generated these colonies) that also generate the myoepithelial cells and not vice versa. A third type of progenitor, the myoepithelial cell-restricted progenitor, was also detected. Serial passaging studies demonstrate that this progenitor is a relatively mature progenitor descendant from the bipotent progenitor. The phenotypic profile of the different progenitors and the progeny they produce is summarized in Figure 48. Luminal cell-restricted and bipotent progenitors generate colonies in 3-D cultures that have gross morphologies consistent with alveoli and ducts, respectively. Considering that both lobulo-176 alveolar and ductal progenitors have been identified in the mouse mammary gland (Smith 1996), it would be reasonable to suggest that progenitors described here are alveolar and ductal progenitors, respectively. However, it would be incorrect to define these progenitors as stem cells since their in vivo repopulating abilities have not been established. The number of cell doublings obtained with mixed/myoepithelial cell-restricted progenitors is approximately 18, which is well below the value of40 - 50 reported for murine stem cells (Kordon and Smith 1998). Although this suggests that these progenitors are not the true stem cells, it is possible that the culture conditions used to assay these progenitors does not support the survival of such cells. An analogous situation is observed when human epidermal cells are cultured in the absence of an appropriate adhesive substratum (Jones and Watt 1993, Jones et al 1995). Epidermal stem cells with regrafting abilities can be maintained in culture for 2 weeks as long as they are maintained on collagen IV. However, should these cells be placed in suspension culture, they quickly differentiate into squamous keratinocytes. The gross morphology of the colonies generated by the bipotent progenitors with the centrally located luminal marker+ cells and the peripherally located myoepithelial marker+ cells resemble the cell arrangement observed on a cross section of a mammary duct. Our working model is that upon appropriate growth signals, selected progenitors within the luminal cell compartment are recruited to form myoepithelial cells, which in turn migrate into the surrounding extracellular matrix (ductal elongation). Obviously our culture system is static, and there is no negative regulatory signal to tell the progenitors expressing a luminal phenotype to stop producing myoepithelial marker* progeny. As a result there may be an excessive production of myoepithelial marker+ cells. In contrast, our culture conditions do not appear to support the longterm persistence of cells expressing luminal specific proteins nor the generation of luminal cell-restricted progenitors. Thus EpCAM7K14+/SMA" cells rapidly become the predominant 177 phenotype in serially passaged cultures as described in these studies and noted by others (Ethier et al., 1993). One potential negative regulator of this excessive cell proliferation is TGF-P, a factor implicated in pattern formation of ducts in the developing mammary gland (reviewed in Daniel et al 1996). TGF-P is synthesized within the epithelium and is secreted basally to become accumulated within the fibrous extracellular matrix surrounding mature ducts. The role of TGF-P as a negative regulator of HBEC proliferation is supported by two observations. The first is that exogenously added TGF-P inhibits ductal elongation (Silberstein and Daniel 1987), and the second is that TGF-P is absent from the peri-ductal extracellular matrix in regions in which new lateral buds arise from the mature ducts (Silberstein et al 1990). It may be that the secretion, storage and activation of TGF-P is disregulated in the 2- and 3-D cultures described in this thesis. The selection of myoepithelial cells in HBEC cultures is just one example of the difficulties in trying to recapitulate the mammary epithelium in a culture dish. Human mammary development in vivo relies on several different hormonal states to attain full development. Likewise, it would be expected that several different hormonal milieus might be required for appropriate growth and differentiation of HBEC progenitors in vitro. The growth medium used in the experiments described in this thesis contained EGF and high concentrations of insulin, with the latter being mitogenic via its interaction with IGF I receptor (King et al 1980). Since both EGF and IGFs are associated with ductal elongation in vivo, it is not surprising that the majority of colonies observed in the colony assays are classified as being mixed/myoepithelial (i.e., a potential ductal progenitor). 178 Immunohistochemical staining of normal human mammary tissue with monoclonal antibodies specific for K19 has revealed K19" cells to be distributed singly and in small clusters within the luminal cell compartment, particularly within the small ducts and terminal ductal lobular units of the gland (Bartek et al., 1985a, b). These cells also express K18, and thus are argued not to be displaced myoepithelial cells (Bartek et al., 1985a). Bartek and colleagues have postulated that K19" cells are a potential mammary gland stem cell, and that they generate K19+ progeny (Bartek et al., 1985b). The cap cells of the TEB of the developing mammary gland are the leading candidate for a stem cell (Rudland, 1993; Rudland et al., 1997; Williams and Daniel, 1983), although others have argued that this cell is merely a myoepithelial progenitor that paves the way for ductal elongation (Sapino et al., 1993). Data presented here demonstrates that the migration of newly formed K14+/SMA" cells, some of which express K18, appears to be involved in what may be viewed as equivalent to ductal elongation of the mammary tree in vivo. A potential model for explaining this phenomenon is that under the appropriate growth stimulus, K18+/19+ progenitor luminal epithelial cells undergo mitosis to generate K18+19" progeny, not all of which become basally located since asynchronous mitosis is associated with basal cell generation (Ferguson, 1988). Those K18+19" cells that remain within the luminal cell compartment would correspond to those cells observed by Bartek and colleagues, whereas the basally positioned newly formed K14+/18+/19" myoepithelial cells would be responsible for ductal budding and elongation. The K14+/18+ myoepithelial cells that we observe in vitro would then correspond to the K14+/18+ cells that line the distal budding structures of the developing human mammary gland (Rudland 1993a). The concept that basally derived cells are responsible for ductal morphogenesis is further supported by the morphogenic effects of HGF on C A L L A + but not MUCr HBEC cultured within collagen gels (Niranjan et al., 1995). 179 The HBEC colonies described in this thesis exhibit a spectrum of colony phenotypes that range from pure luminal marker* cell colonies to mixed and pure myoepithelial marker* cell colonies. There appears to be no definitive separation between the pure luminal cell colonies and the mixed colonies since colonies exhibiting predominantly luminal cell characteristics often contain one or a few cells producing myoepithelial-specific proteins. Similarly, staining of freshly isolated HBEC for lineage-specific epitopes (e.g., a6 integrin and CALLA) does not yield clearly distinct subpopulations but instead shows a continuum of HBEC expressing increasing levels of these epitopes. Both the luminal cell-restricted and bipotent progenitors are found with the EpCAM* cell fractions although the two can be phenotypically distinguished by subtle differences in expression of a6 integrin, CALLA and in their retention of Rhl23. The relative similarity between these two types of progenitors suggests a common origin for these cells. Conceivably, such a cell could then be directed down a ductal pathway or alveolar pathway, depending on environmental cues or intrinsically determined parameters. A hypothetical model of the relationships between mammary stem cells and their terminally differentiated progeny is shown in Figure 49. Aside from self-renewal, stem cells divide and generate progeny that begin to express luminal-specific markers (left arm) and myoepithelial-specific markers (right arm). The balance of the generation of these lineages, which would be dependent on extrinsic and intrinsic cues, would then determine if lobulo-alveolar or ductal structures are generated. As the cells divide and differentiate, they increase their expression of lineage markers and have decreased proliferative capacity. In vitro studies utilizing the rat mammary epithelial cell line Rama 25 supports the concept of extrinsic control of differentiation since these cells, when cultured in collagen gels, can generate ductal like structures that are composed of both luminal and myoepithelial cells. Furthermore, when this cell line is maintained in the presence of lactogenic hormones it can be induced to synthesize casein (Rudland 1992, reviewed in Rudland etal 1997). 180 Future Directions Although HBEC exhibiting luminal and myoepithelial cell characteristics can be observed when maintained in the presence of SF medium and a NTH 3T3 feeder, this culture environment is still problematic in that those expressing luminal characteristics are quickly lost in vitro, and cells exhibiting abnormal morphologies (e.g., elevated masses) are observed. Steps to rectify this situation could include testing the influence of CDM6 (luminal cell) growth medium (Pechoux et al 1999) and low calcium medium (Berthon et al 1992) on the clonogenicity of HBEC progenitors. The ultrastructural morphology of the colonies generated within the collagen gels did not resemble mammary epithelium in vivo. Part of this is related to the fact that the FfMF feeders used in these experiments do not support maintenance of luminal HBEC. Human alveoli have been recapitulated when luminal HBEC are seeded within Matrigel (Petersen et al 1992); however, the formation of clonal ductal structures has not been recapitulated using Matrigel or any other substrate. To try to achieve this, HBEC could be maintained in the presence of NTH 3T3 feeders and in the presence of a more refined growth medium (e.g., CDM6) and collagen matrix. The collagen gels used in the experiments described here were relatively crude preparations with considerable batch to batch variability. Studies examining megakaryocyte progenitor activity have revealed that alternate commercial sources of collagen, as well as collagen isolated from different species (i.e., bovine and human) profoundly influence the cloning properties of these progenitors (personal communication with Dr. Cindy Miller, StemCell Technologies). Considering the role of laminin in HBEC differentiation, collagen titrated with Matrigel may provide the appropriate extracellular environment for HBEC growth and differentiation. It is likely that a single culture medium will not induce the appropriate 181 growth and differentiation of ductal progenitors. Instead HBEC progenitors may have to be maintained in a growth medium that promotes the formation of myoepithelial cells (e.g., the SF medium described here), followed by a switch to a medium that induces myoepithelial cell differentiation and the synthesis of basement membrane proteins (e.g., a low growth factor medium; Petersen and van Deurs 1988). A critical point in these experiments may be that the formation of the K14+/SMA' cells from bipotent progenitors would have to be limited so as not to exhaust any luminal cell progenitors produced. It is possible that this may have been achieved in the collagen gel cultures analyzed for this thesis. Recent reports have demonstrated that ER is expressed by non-proliferating cells within the human mammary epithelium (Clarke et al 1997, Russo et al 1999). Recently, a collaboration has been initiated with Dr. Leigh Murphy at the University of Manitoba to determine the distribution of ERct and ERP among different breast cell populations. The ER-a and -P status in the different flow-sorted cells are being analyzed by long range reverse transcription-polymerase chain reaction analysis which detects all deleted ER variant mRNAs at a frequency relative to their initial mRNA representation in the unamplified sample (Leygue et al 1996, Fasco 1997). Preliminary results indicate that ERP mRNA is present in all cell populations (both stromal and epithelial), and that the majority of ERP mRNA detected is the variant form. Variable results have been obtained with ERcc analysis, and any trends are difficult to comment on since only 6 samples have been analyzed to date. An important next step in the investigation of mammary stem cells is to test the repopulating abilities ofthe different progenitors in an in vivo model system. To address this, this laboratory is initiating experiments in which epithelial organoids and HBEC enriched progenitor fractions of adult human mammary tissue would be engrafted along with a suitable extracellular matrix 182 under the renal capsule of athymic rodent hosts. A similar type of xenograft was recently performed using a recombination of adult human prostatic epithelium and mouse prostatic mesenchyme (Hayward et al 1998). Should colonies exhibiting appropriate epithelial cell morphology be observed, engraftment of limiting dilutions of HBEC could allow for the quantitation of the progenitors in the input population. Serial transplantation of existing colonies into new hosts would then permit detection of stem cell self-renewal. There is increasing evidence that there is a pool of stem cells distributed throughout different tissues and organs that have the capacity to differentiate into specific lineages of cells. For example, there exists a population of cells within muscle which can by isolated by flow cytometry and can reconstitute the hematopoietic system in irradiated hosts (Gussoni et al 1999). A similar cell type has also been identified within neural tissue (Bjornson et al 1999). Experiments to examine if a similar cell type exists within the adult human mammary epithelium would therefore be of considerable interest also. 183 Figure 48: Schematic diagram showing the growth and differentiation potentials ofthe different HBEC progenitors. 184 L u m i n a l c e l l - r e s t r i c t e d p r o g e n i t o r ( A l v e o l a r p r o g e n i t o r ? ) M U C 1 + / C A L L A 7 E p - C A M + / E G F R " 1 0 + / N E U + / B G A / R h 1 2 3 b r i g h V a 6 - t 0 ± B i p o t e n t p r o g e n i t o r ( D u c t a l p r o g e n i t o r ? ) MUCr t o ± /CALLA ± t o + / E p - C A M + / E G F R + / N E U + / B G A - l o + / R h l 2 3 d u l l / a 6 ± t o + L u m i n a l c e l l - r e s t r i c t e d c o l o n i e s M U C l + / E p - C A M + / K 6 " 1 0 * / K 1 4 7 K 8 & 1 8 + / K 1 9 + / B G A " M i x e d c o l o n i e s c o m p o s e d o f : L u m i n a l c e l l s M U C l ± / E p - C A M + / K 6 + / K 1 4 7 K 8 & 1 8 + / K 1 9 + & m y o e p i t h e l i a l c e l l s M U C 1 7 E p - C A M " t o ± / K 6 " t o + / K 8 & 1 8 ± / K . 1 9 -M y o e p i t h e l i a l c e l l - r e s t r i c t e d c o l o n i e s M U C 1 7 E p - C A M 7 K 6 " K 1 4 + / K 8 & 1 8 7 K 1 9 / B G A + 18^  Figure 49: Hypothetical relationship between mammary stem cells and their differentiated progeny. 186 STEM C E L L LOBULO-ALVEOLAR DEVELOPMENT 7 / O E p C A M + K8/18+ K19+ R h l 2 3 d u 1 1 E p C A M + R h l 2 3 bright K8/18+ K19+ K14" Casein" O Lactogenic hormones E p C A M R h l 2 3 b r i g h t K8/18+ K19+ K14" Casein* TERMINALLY DIFFERENTIATED ALVEOLAR CELLS DUCTAL DEVELOPMENT E p C A M " K8/18* K19" S M A " K14+ • E p C A M " K8/18" S M A " R h l 2 3 b r i g h t K14+ E p C A M " K8/18" S M A + R h l 2 3 b r i g h t K14+ TERMINALLY DIFFERENTIATED MYOEPITHELIAL CELLS 187 REFERENCES Albanes D. and Winick M. (1988) Are cell number and cell proliferation risk factors for cancer. Journal of the National Cancer Institute 80: 772-775. Anbazahagan R., Bartek J., Monaghan P. and Gusterson B.A. (1991) Growth and development of the human infant breast. American Journal of Anatomy 192: 407-417. Anbazhagan R. and Gusterson B.A. (1992) (letter) Lancet 339: 1056. Anbazhagan R. and Gusterson B.A. (1994) Prenatal factors may influence predisposition to breast cancer (commentary). European Journal of Cancer 30A: 1-3. Anbazhagan R., Osin P.P., Bartkova J., Nathan B., Lane E.B. and Gusterson B.A. (1998) The development of epithelial phenotypes in the human fetal and infant breast. Journal of Pathology 184: 197-206. Anderson T.J., FergusonD.J.P. and Raab G M . (1982) Cell turnover in the "resting" human breast: influence of parity, contraceptive pill, age and laterality. British Journal of Cancer. 46: 376-382. Bailey M.E., Brown R.W., Mody DR., Cagle P. and Ramzy I. (1996) Ber-EP4 for differentiating adenocarcinoma from reactive and neoplastic mesothelial cells in serous effusions. Comparison to carcinoembryonic antigen, B72.3 and Leu-Ml. Acta Cytology 40: 1212-1216. Barcellos-Hoff MH. Aggeler J. Ram TG. Bissell MJ. (1989) Functional differentiation and alveolar morphogenesis of primary mammary cultures on reconstituted basement membrane. Development 105: 223-35. Bartek J., Taylor-Papadimitriou J., Miller N. and Millis R. (1985a) Patterns of expression of keratin 19 as detected with monoclonal antibodies in human breast tissues and tumors. International Journal of Cancer 36:299-306. Bartek J., Durban E.M., Hallowes R.C. and Taylor-Papadimitriou J. (1985b) A subclass of luminal epithelial cells in the human mammary gland, defined by antibodies to cytokeratins. Journal of Cell Science 75: 17-33. Bartek J., Bartkova J. and Taylor-Papadimitriou J. (1990) Keratin 19 expression in the adult and developing human mammary gland. Histochemical Journal 22: 537-544. Bergstraesser L.M. and Weitzman S.A. (1993) Culture of normal and malignant primary human mammary epithelial cells in a physiological manner simulates in vivo growth patterns and allows discrimination of cell type. Cancer Research 53: 2644-2654. Berthon P., Pancine G , DeCremoux P., Roseto A , Gespach C. and Calvo F. (1992) Characterization of normal breast epithelial cells in primary culture: Differentiation and growth factor receptor studies. In Vitro Cellular & Developmental Biology 28 A: 716-724. 188 Bissell MJ., Weaver V.M., Lelievre S.A., Wang F., Petersen O.W. and Schmeichel KL. (1999) Tissue structure, nuclear organization, and gene expression in normal and malignant breast. Cancer Research. 59:1757-1763. Bjornson C.R.R., RietzeR.L., ReynoldsB.A., Magli M.C. and Vescovi A L . (1999) Turning brain into blood: A hematopoietic fate adopted by adult neural stem cells in vivo. Science 283: 534-537. Bocker W., Bier B., Freytag G., Brommelkamp B., Jarasch E.-D., Edel G., Dockhorn-Dworniczak B. and Schmid K.W. (1992a) An immunohistochemical study of the breast using antibodies to basal and luminal keratins, alpha-smooth muscle actin, vimentin, collagen IV and laminin. Part I: normal breast and benign proliferative lesions. Virchows Archiv A [Pathological Anatomy and Histopathology] 421: 315-322. Bocker W., Bier B., Freytag G., Brommelkamp B., Jarasch E.D., Edel G., Dockhorn-Dworniczak B. and Schmid K.W. (1992b) An immunohistochemical study of the breast using antibodies to basal and luminal keratins, alpha-smooth muscle actin, vimentin, collagen IV and laminin. Part II: Epitheliosis and ductal carcinoma in situ. Virchow Archiv [Pathological Anatomy & Histopathology] 421: 323-330. Brisken C , Park S., Vass T., Lydon J.P., O'Malley B. and Weinberg RA. (1998) A paracrine role for the epithelial progesterone receptor in mammary gland development. Proceedings of the National Academy of Sciences U.S.A. 95: 5076-5081. Brisken C , Kaur S., Chavarria T.E., Binart N , Sutherland R.L., Weinberg R.A., Kelly P.A. and Ormandy C.J. (1999) Prolactin controls mammary gland development via direct and indirect mechanisms. Developmental Biology 210: 96-106. Burchell J., Durbin H. and Taylor-Papadimitriou J. (1983) Complexity of expression of antigenic determinants recognized by monoclonal antibodies HMFG-1 and HMFG-2 in normal and malignant human mammary epithelial cells. The Journal of Immunology 131: 508-513. Burchell J., Gendler S., Taylor-Papadimitriou J., Girling A., Lewis A., Mullis R. and Lamport D. (1987) Development and characterization of breast cancer reactive monoclonal antibodies directed to the core protein of the human milk mucin. Cancer Research 47: 5476-5482. Cailleau R., Young R., Olive M. and W.J.J. Reeves (1974) Breast tumor cell lines from pleural effusions. Journal of the National Cancer Institute 53: 661-666. Cardiff RD. and Wellings S.R. (1999) The comparative pathology of human and mouse mammary glands. Journal of Mammary Gland Biology and Neoplasia 4: 105-122. Ceriani R.L., Thompson K., Peterson J.A. and Abraham S. (1977) Surface determination antigens on human mammary epithelial cells carried on the human milk fat globule. Proceedings of the National Academy of Sciences, U.S.A. 74: 582-586. Chaudhaury P.M. and RoninsonIB. (1991)Expression and activity of P-glycoprotein, a multidrug efflux pump, in human hematopoietic stem cells. Cell 66: 85-94. 189 Chen C L . and Meites J. (1970) Effects of estrogen and progesterone on serum and pituitary prolactin levels in ovariectomized rats. Endocrinology 86: 503-505. Chepko G. and Smith G.H. (1997) Three division-competent, structurally-distinct cell populations contribute to murine mammary epithelial renewal. Tissue and Cell 29: 239-253. Clarke C , Titley J., Davies S. and O'Hare M.J. (1994) An immunomagnetic separation method using superparamagnetic (MACS) beads for large-scale purification of human mammary luminal and myoepithelial cells. Epithelial Cell Biology 3: 38-46. Clarke R.B., Howell A. and Anderson E. (1997) Estrogen sensitivity of normal human breast tissue in vivo and implanted into athymic nude mice: Analysis of the relationship between estrogen-induced proliferation and progesterone receptor expression. Breast Cancer Research and Treatment 45: 121-133. Close M.J., Howlett A.R., Roskelley CD. , Desprez P.Y., Bailey N., Rowning B., Teng C.T., Stampfer MR. and YaswenP. (1997) Lactoferrin expression in mammary epithelial cells is mediated by changes in cell shape and actin cytoskeleton. Journal of Cell Science 110: 2861-2871. Coleman S., Silberstein G. B. and Danial C. W. (1988) Ductal morphogenesis in the mouse mammary gland: Evidence supporting a role for epidermal growth factor. Developmental Biology 127, 304-315. Cooper N.L., Bardy P., Bacani J., Kuusk U., Dougherty G. J., Eaves C. J. and Emerman J.T. (1998) Correlation of CD44 expression with proliferative activity of normal human breast epithelial cells in culture. Breast Cancer Research and Treatment 50: 143-153. Croce M.V., Colussi A G . , Price MR. and Segal-Eiras A. (1997) Expression of tumor associated antigens in normal, benign and malignant human epithelial tissue: a comparative immunohistochemical study. Anticancer Research 17: 4287-4292. Cunha G., Young P., Horn Y., Cooke P., Taylor J.and Lubahn D. (1997) Elucidation of a role for stromal steroid hormone receptors in mammary gland growth and development using tissue recombinations. Journal of Mammary Gland Biology and Neoplasia 2: 393-402. Dairkee SH, Blayney CM, Smith HS, Hackett AJ (1985) Monoclonal antibody that defines human myoepithelium. Proc Natl Acad Sci USA 82: 7409-7413. Dairkee S.H., Blayney-Moore C M . , Smith HS. and Hackett A.J. (1986) Concurrent expression of basal and luminal epithelial markers in cultures of normal human breast analyzed using monoclonal antibodies. Differentiation 32: 93-100. Dairkee S.H., Ljung B.M., Smith H. and Hackett A. (1987) Immunolocalization of a human basal epithelium specific keratin in benign and malignant breast disease. Breast Cancer Research & Treatment. 10:11-20. 190 Dairkee S.H, DengG., StampferMR., WaldmanFM. and SmithH.S. (1995) Selective cell culture of primary breast carcinoma. Cancer Research 55: 2516-2519. Daniel C.W., DeOme K B . , Young J.T., Blair PB. and Faulkin L.J. (1968) The in vivo life span of normal and preneoplastic mouse mammary glands: A serial transplantation study. Proceedings of the National Academy of Sciences U.S.A. 61: 52-60. Daniel C.W., Young L.J.T., Medina D. and DeOme K.B. (1971) The influence of mammogenic hormones on serially transplanted mouse mammary gland. Experimental Gerentology 6: 95-101. Daniel C.W., Silberstein G.B and Strickland P (1987) Direct action of 17f3-estradiol on mouse mammary ducts analyzed by sustained release implants and steroid autoradiography. Cancer Research 47: 6052-6057. Daniel C.W., Robinson S. and Silberstein G.B. (1996) The role of TGF-p in patterning and growth of the mammary ductal tree. Journal of Mammary Gland Biology and Neoplasia 1: 331-341. Darcy K M . , Black J.D., Hahm H A . and Ip M M . (1991) Mammary organoids from immature virgin rats undergo ductal and alveolar morphogenesis when grown within a reconstituted basement membrane. Experimental Cell Research 196: 49-65. Darcy K M . , Shoemaker ST., Lee P.-P.H, Vaughan M M . , Black J. and Ip M M . (1995) Prolactin and epidermal growth factor regulation of the proliferation, morphogenesis, and functional differentiation of normal rat mammary epithelial cells in three dimensional primary culture. Journal of Cellular Physiology 163: 346-364. Delahaye M., van der Ham F. and van der Kwast T H . (1997) Complementary value of five carcinoma markers for the diagnosis of malignant mesothelioma, adenocarcinoma metastasis, and reactive mesothelium is serous effusions. Diagnostic Cytopathology 17: 115-120. Dulbecco R., Henehan M. and Armstrong B. (1982) Cell types and morphogenesis in the mammary gland. Proceedings of the National Academy of Sciences U.S.A. 79: 7346-7350. Dunbar M.E., Young P., Zhang J.P., McCaughern-Carucci J.F., Lanske B., Orloff J.J., Karaplis A, Cunha G. and Wysolmerski J.J. (1998). Developmental Biology 203: 75-89. Dunbar M.E. and Wysolmerski J.J. (1999) Parathyroid hormone-related protein: A developmental regulatory molecule necessary for mammary gland development. Journal of Mammary Gland Biology and Neoplasia 4: 21-34. Dundas SR., Ormerod M.G., Gusterson B.A. and O'Hare M.J. (1991) Characterization of luminal and basal cells flow-sorted from the adult rat mammary parenchyma. Journal of Cell Science 100: 459-471. Eaves C.J., Cashman J.D., and Eaves A C . (1991) Methodology of long-term culture of human hematopoietic cells. Journal of Tissue Culture Methods 13: 55-62. 191 Edwards P A W . and Brooks LM. (1984) Antigenic subsets of human breast epithelial cells distinguished by monoclonal antibodies. Journal of Histochemistry and Cytochemistry 32: 531-537. Edwards P.A.W., Brooks I.M., Bunnage H.J., Foster A V , Ellison M L . and O'Hare M.J. (1986) Clonal analysis of expression of epithelial antigens in cultures of normal human breast. Journal of Cell Science 80: 91-101. Emerman J.T. and Pitelka D.R. (1977) Maintenance and induction of morphological differentiation in dissociated mammary epithelium on floating collagen membranes. In Vitro 13: 316-328. Emerman J.T. Enami J. Pitelka D.R. Nandi S. (1977) Hormonal effects on intracellular and secreted casein in cultures of mouse mammary epithelial cells on floating collagen membranes. Proceedings of the National Academy of Sciences U.S.A. 74:4466-4470. Emerman J.T. and Vogl A.W. (1986) Cell size and shape changes in the myoepithelium of the mammary gland during differentiation. The Anatomical Record 216: 405-415. Emerman J.T. and Wilkinson D.A. (1990) Routine culturing of normal, dysplastic and malignant human mammary epithelial cells from small tissue samples. In Vitro Cellular & Developmental Biology 26: 1186-1194. Ethier S.P., Mahacek M L . , Gullick W.J., Frank T.J. and Weber B.L. (1993) Differential isolation of normal luminal mammary epithelial cells and breast cancer cells from primary and metastatic sites using selective media. Cancer Research. 53: 627-635. Ethier S.P. (1996) Human breast cancer cell lines as models of growth regulation and disease progression. Journal of Mammary Gland Biology and Neoplasia 1: 111-121. Fasco M. (1997) Quantitation of estrogen receptor mRNA and its alternatively spliced mRNAs in breast tumor cells and tissues. Analytical Biochemistry 245: 167-178. Ferguson D.J.P. (1985) Ultrastructural characterization of the proliferative (stem?) cells within the parenchyma of the normal "resting" breast. Virchows Archiv [Pathol Anat] 407: 379-385. Ferguson D.J.P. (1988) An ultrastructural study of mitosis and cytokinesis in normal "resting" human breast. Cell and Tissue Research 252: 581-587. Feur E.J., Wun L M . Boring C C , Flanders D., Timmel M.J. and Tong T. (1993) The lifelong risk of developing breast cancer. Journal of the National Cancer Institute 85: 892-897. Fiorentini S., Matczak E., Gallo R C , Reitz M.S., Keydar I. and Watkins B.A. (1997) Humanization of an antibody recognizing a breast cancer specific epitope by CDR-grafting. Immunotechnology 3: 45-59. Flynn D., Yang J. and Nandi S. (1982) Growth and differentiation of primary cultures of mouse mammary epithelium embedded in collagen gel. Differentiation 22: 191-194. 192 Foster C.S., Edwards P.A.W., Dinsdale E.A. and Neville A M . (1982) Monoclonal antibodies to the human mammary gland. Virchows Arch. [A] 394: 279-293. Foster C.S., Smith C. A., Dinsdale E., Monaghan P. and Neville A M . (1983) Human mammary gland morphogenesis in vitro: The growth and differentiation of normal breast epithelium in collagen gel cultures defined by electron microscopy, monoclonal antibodies, and autoradiography. Developmental Biology 96L 197-216. Fournier D. V., Weber E., Hoeffken M., Bauer M., Kubli F. and Barth V. (1980) Growth rate of 147 mammary carcinomas. Cancer 45: 2198-2207. Fox SB., Fawcett J., Jackson D.G., Collins I., Gatter K.C., Harris A.L., Gearing A. and Simmons D.L. (1994) Normal human tissues, in addition to some tumors, express multiple different CD44 isoforms. Cancer Research 54: 4539-4546. Fujii H., Marsh C , Cairns P., Sidransky D and Gabrielson E. (1996) Genetic divergence in the clonal evolution of breast cancer. Cancer Research 56: 1493-1497. Gaffney E.V., Polanowski F.P., Blackburn S.E., Lambiase J.T. and Burke R E . (1976) Cultures of normal human mammary cells. Cell Differentiation 5: 69-81. Geschwind N. and Galaburda A.M. (1985) Cerebral lateralization, biological mechanisms, association, and pathology, I: a hypothesis and a program for research. Archives of Neurology 42: 428-459. Going J.J., Anderson T.J., Battersby S. and Maclntyre C C A . (1988) Proliferation and secretory activity in the human breast during natural and artificial menstrual cycles. American Journal of Pathology 130: 193-204. Gomm J.J, Browne P.J., Coope R C , Bansal G.S., Yiangou C , Johnston C.L., Mason R. and Coombes R C . (1997a) A paracrine role for myoepithelial cell-derived FGF2 in the normal human breast. Experimental Cell Research 234: 165-173. Gomm J.J., Coope R.C., Browne P.J. and Coombes R C . (1997b) Separated human breast epithelial and myoepithelial cells have different growth requirements in vitro but can reconstitute normal breast lobuloalveolar structure. Journal of Cellular Physiology 171: 11-19. Gorska AE. Joseph H. Derynck R. Moses HL. Serra R. (1998) Dominant-negative interference of the transforming growth factor beta type II receptor in mammary gland epithelium results in alveolar hyperplasia and differentiation in virgin mice. Cell Growth & Differentiation. 9: 229-38. Guelstein V.I., Tchypyscheva T. A., Ermilova V.D., Litvinova L.V., Troyanovsky SM. and Bannikov G.A. (1988) Monoclonal antibody mapping of keratins 8 and 17 and of vimentin in normal human mammary gland, benign tumors, dysplasias and breast cancer. International Journal of Cancer 42: 147-153. Gugliotta P., Sapino A., Macri L., Skalli O., Gabbiani G. and Bussolati G. (1988) Specific demonstration of myoepithelial cells by anti-alpha smooth muscle actin antibody. Journal of Histochemistry and Cytochemistry 36: 659-663. 193 GussoniE, Soneoka Y., Strickland CD. , BuzneyE.A, KhanMK, Flint A.F., KunkelLM. and Mulligan R.G. (1999) Dystrophin expression in the mdx mouse restored by stem cell transplantation. Nature 401: 390-394. Gusterson B.A., Monaghan P., Mahendran R, Ellis J., O'Hare M.J. (1986) Identification of myoepithelial cells in human and rat breasts by anti-common acute lymphoblastic leukemia antigen antibody A12. Journal of the National Cancer Institute 77: 343-349. Hammond S L , Ham R G . and Stampfer MR. (1984) Serum-free growth of human mammary epithelial cells: Rapid clonal growth in defined medium and extended serial passage with pituitary extract. Proceedings of the National Academy of Sciences U.S.A. 81: 5435-5439. Haslam S.Z (1986) Mammary fibroblasts influence on normal mouse mammary epithelial cell responses to estrogen in vitro. Cancer Research 46: 310-316. Haslam S.Z. (1988) Acquisition of estrogen-dependent progesterone receptors by normal mouse mammary gland. Ontogeny of mammary progesterone receptors. Journal of Steroid Biochemistry 31:9-13. Hayward S.W., Haughney P C , RosenM.A., GreulichKM., Weier H.-U.G, DahiyaR. and Cunha G.R. (1998) Interactions between adult human prostatic epithelium and rat urogenital sinus mesenchyme in a tissue recombination model. Differentiation 63: 131-140. Heatley M , Maxwell P., Whiteside C. and Toner P. (1995) Cytokeratin intermediate filament expression in benign and malignant breast disease. Journal of Clinical Pathology 48: 26-32. Hirai Y., Lochter A., Galosy S, Koshida S, Niwa S. and Bissell M.J. (1998) Epimorphin functions as a key morphoregulator for mammary epithelial cells. Journal of Cell Biology 140: 159-169. Holmgren J. (1981) Actions of cholera toxin and the prevention and treatment of cholera. Nature 292: 413-417. Hoshino K (1965) Brown fat as a stroma for mammary growth. Anat Record 151: 364. Ip M M . and Darcy K.M. (1996) Three-dimensional mammary primary culture. Journal of Mammary Gland Biology and Neoplasia 1:91-110. Imagawa W., Yang J., Guzman R. and Nandi S. (1994) Control of mammary gland development. In: The Physiology of Reproduction. Knobil E. and Neill J.D. (eds), Raven Press Ltd., New York, ppl033-1063. Johnston C S , Shpall E.J , Williams S, Hami L , Jones R.B., Bast R C , Ceriani R L . and Franklin W. (1992) Detection of minimal residual breast cancer in bone marrow. In: Advances in Bone Marrow Purging and Processing. Worthington-White D A , Gee AP. and Gross S (eds), Wiley-Liss, New York, pp 637-642. 194 Jones F.E., Jerry D.J., Guarine B.C., Andrews G C . and Stern D.F. (1996) Heregulin induces in vivo proliferation and differentiation of mammary epithelium into secretory lobulo-alveoli. Cell Growth and Differentiation 7: 1031-1038. Jones PH. and Watt F.M. (1993) Separation of human epidermal stem cells from transit amplifying cells on the basis of differences in integrin function and expression. Cell 73: 713-724. Jones PH. , Harper S. and Watt F.M. (1995) Stem cell patterning and fate in human epidermis. Cell 80: 83-93. Joshi K., Smith J. A., Perusinghe N. and Monoghan P. (1986) Cell proliferation in human mammary epithelium. Differential contribution by epithelial and myoepithelial cells. American Journal of Pathology 124: 199-206. Kao C-Y, Nomata K, Oakly CS, Welsch CW and Chang C-C (1995) Two types of normal human breast epithelial cells derived from reduction mammoplasties: Phenotypic characterization and response to SV40 transfection. Carcinogenesis 16: 531-538. Kao C.-Y., Oakley C.S., Welsch C.W. and Chang C.-C. (1997) Growth requirements and neoplastic transformation of two types of normal human breast epithelial cells derived from reduction mammoplasty. In Vitro Cellular and Developmental Biology - Animal 33:282-288. Karsten U, Papsdorf F, Pauly A, Vojtesek B, Moll R, Lane EB, Clausen H, Stosiek P, Kasper M (1993) Subtypes of non-transformed human mammary epithelial cells cultured in vitro: Histo-blood group antigen H type 2 defines basal cell-derived cells. Differentiation 54: 55-66. Kellokumpu-Lehrinen P., Johansson R M . and Pelimiemi L.J. (1987) Ultrastructure of human fetal mammary gland. Anatomical Record 218: 66-72. Kelsey J.L. and Berkowitz GS. (1988) Breast cancer epidemiology. Cancer Research 48: 5615-5623. King G.L., Khan C.R., Rechler M.M. and Nissley S.P. (1980) Direct demonstration of separate receptors for growth and metabolic activities of insulin and multiplication-stimulating activity (an insulin like growth factor) using antibodies to the insulin receptor. Journal of Clinical Investigation 66: 130-140. KordonEC, McKnight R.A., Jhappan C , HennighausenL., Merlino G. and Smith G H . (1995) Ectopic TGF-P 1 expression in the secretory mammary epithelium induces early senescence of the epithelial stem cell population. Developmental Biology 168: 47-61. Kordon E C and Smith G H . (1998) An entire functional mammary gland may comprise the progeny from a single cell. Development 125: 1921-1930. KoukoulisG.K., VirtanenI, HorhonenM, LaitinenL., Quaranta V. and Gould V E . (1991) Immunohistochemical localization of integrins in the normal, hyperplastic, and neoplastic breast. American Journal of Pathology 139: 787-799. 195 Kratochwil K. and Schwartz P. (1976) Tissue interactions in androgen response of embryonic mammary rudiment of mouse: Identification of target tissue for testosterone. Proceedings of the National Academy of Sciences U.S.A. 73: 4041-4044. Krege J.H, Hodgin J.B, Couse J.F, Enmark E , Warner M , Mahler J.F, Sar M , Koarch K S , Gustafsson J.A. and Smithies O. (1998) Generation and reproductive phenotypes of mice lacking estrogen receptor beta. Proceedings of the National Academy of Sciences U.S.A. 95: 15 677-15 682. Lad P.M., Nielsen T B , Preston M.S. and Rodbell M. (1980) The role of guanine nucleotide exchange reaction in the regulation of the P-adrenergic receptor and in the actions of catecholamines and cholera toxin on adenylate cyclase in Turkey erythrocyte membrane. Journal of Biological Chemistry 255: 988-995. Latza U , Niedobitek G , Schwarting R, Nekarda H. and Stein H. (1990) Ber-EP4: new monoclonal antibody which distinguishes epithelia from mesothelia. Journal of Clinical Pathology 43: 213-219. Lee J.-S, Pauli K , Alvarez M , Hose C , Monks A , Grever M , Fojo A T . and Bates SE. (1994) Rhodamine efflux patterns predict p-glycoprotein substrates in the National Cancer Institute drug screen. Molecular Pharmacology 46: 627-638. Lemoine F . M , Humphries R.K, Abraham S.D.M, Krystal G , and Eaves C.J. (1988) Partial characterization of a novel stromal cell-derived pre-B-cell growth factor active on normal and immortalized pre-B cells. Experimental Hematology 16: 718-726. Levay-Young B K , Bandyopadhyay GK. and Nandi S. (1987) Lineolic acid, but not Cortisol, stimulates accumulation of casein by mouse mammary epithelial cells in serum-free collagen gel culture. Proceedings of the National Academy of Sciences U.S.A. 84: 8448-8452. Leygue E , Huang A , Murphy L and Watson P. (1996) Prevalence of estrogen receptor variant messenger RNAs in human breast cancer. Cancer Research 56: 4324-4327. Lieberman M . E , Maurer R.A. and Gorski J. (1978) Estrogen control of prolactin synthesis in vitro. Proceedings of the National Academy of Sciences U.S.A. 75: 5946-5949. Litvinov S. V , Velders M.P, Bakker H A M , Fleuren G.J. and Warnaar S.O. (1994) EpCAM: a human epithelial antigen is a homophilic cell-cell adhesion molecule. Journal of Cell Biology 125:437-446. Longacre T A . and Bartow SA. (1986) A correlative morphologic study of human breast and endometrium in the menstrual cycle. American Journal of Surgical Pathology 10: 382-393. Lydon J.P, DeMayo F.J, Conneely O M . and CMalley B.W. (1996) Reproductive phenotypes of the progesterone receptor null mutant mouse. Journal of Steroid Biochemistry and Molecular Biology 56: 67-77. Masters J.R.W, Drife J.O, Scaribrick J.J. (1977) Cyclic variation of DNA synthesis in human breast epithelium. Journal of the National Cancer Institute 58: 1263-1265. 196 Moll R., Franke WW., Schiller D.L. (1982) The catalogue of human cytokeratins: patterns of expression in normal epithelial, tumors and cultured cells. Cell 31: 11-24. Momburg F., Moldenhauer G., Hammerling G.J. and Moller P. (1987) Immunohistochemical study of the expression of a M, 34,000 human epithelium-specific surface glycoprotein in normal and malignant tissues. Cancer Research 47: 2883-2891. Monoghan P, Perusinghe N.P., Cowen P. and Gusterson B.A. (1990) Peripubertal human breast development. The Anatomical Record 226: 501-508. Mosmann T. (1983) Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxicity assays. Journal of Immunological Methods 65: 55-63. Neville M . C , Allen J.C. and Watters C. (1983) The mechanism of milk secretion. In: Lactation. Neville M.C. andNeifert MR. (eds), Plenum Press, New York, NY, pp 49-102. O'Hare M.J., Ormerod M.G., Monaghan P., Lane E.B., Gusterson B.A. (1991) Characterization in vitro of luminal and myoepithelial cells isolated from the human mammary gland by cell sorting. Differentiation 46: 209-221. Ormerod E.J. and Rudland P.S. (1984) Cellular composition and organization of ductal buds in developing rat mammary glands: evidence for morphological intermediates between epithelial and myoepithelial cells. American Journal of Anatomy 170: 631-52. Osborne MP. (1991)Breast development and anatomy. In: Harris JR., Hellman S., Henderson I.C, Kinne D.W. (eds) Breast Diseases, JB. Lippincott Company, Philadelphia, pp 1-13. Pandis N., Heim S., Bardi G., Limon J., Mandahl N. and Mitelman F. (1992) Improved technique for short-term culture and cytogenetic analysis of human breast cancer. Genes, Chromosomes and Cancer 5:14-20. Patton S., Gendler S.J. and Spicer AP. (1995) The epithelial mucin, MUCl, of milk, mammary gland and other tissues. Biochim. Biophys. Acta 1241:407-423. Pechoux C , Gudjonsson T., Ronnov-Jessen L., Bissell M.J. and Petersen O.W. (1999) Human mammary luminal cells contain progenitors to myoepithelial cells. Developmental Biology 206: 88-99. Perusinghe N.P., Monaghan P., O'Hare M.J., Ashley S. and Gusterson B.A. (1992) Effects of growth factors on proliferation of basal and luminal cells in human breast epithelial explants in serum-free culture. In Vitro Cellular and Developmental Biology 28A: 90-96. Petersen O., Hoyer P. and van Deurs B. (1987) Frequency and distribution of estrogen receptor-positive cells in normal, non-lactating human breast tissue. Cancer Research 47: 5748-5751. Petersen O.W. and van Deurs B. (1987) Preservation of defined phenotypic traits in short-term cultured human breast carcinoma derived epithelial cells. Cancer Research 47: 856-866. 197 Petersen O.W, van Deurs B. (1988) Growth factor control of myoepithelial-cell differentiation in cultures of human mammary gland. Differentiation: 197-215. Petersen O.W, Ronnov-JessenL, Howlett AR. and Bissell M.J. (1992) Interaction with basement membrane serves to rapidly distinguish growth and differentiation pattern of normal and malignant human breast epithelial cells. Proceedings of the National Academy of Sciences U.S.A. 89: 9064-9068. Ponka P. Beaumont C. Richardson DR. (1998) Function and regulation of transferrin and ferritin. Seminars in Hematology. 35: 35-54. Potten C S , Watson R.J, Williams G.T, Tickle S, Roberts S A , Harris M. and Howell A. (1988) The effect of age and menstrual cycle upon proliferative activity of the normal human breast. British Journal of Cancer 58: 163-170. Purkis P E , Steel J.B, Mackenzie I C , Nathrath W.B.J, Leigh IM. and Lane E.B. (1990) Antibody markers of basal cells in complex epithelia. Journal of Cell Science 97: 39-50. Richards J , Larson L , Yang J , Guzman R, Tomooka Y , Osborn R, Imagawa W , Nandi S. (1983) Method for culturing mammary epithelial cells in a rat tail collagen gel matrix. Journal of Tissue Culture Methods 8: 31-36. Robinson S D , Silberstein G B , Roberts A B , Flanders K.C. and C.W. Daniel (1991) Regulated expression and growth inhibitory effects of transforming growth factor beta isoforms in mouse mammary gland development. Development 113: 867-878. Robinson G.W. and Hennighausen L. (1997) Inhibins and activins regulate mammary epithelial cell differentiation through mesenchymal-epithelial interactions. Development 124: 2701-2718. Ronot X , Paillasson S. and Muirhead K. A. (1990) Assessment of cell viability in mammalian cell lines. In: Guide to Flow Cytometry Methods. Grogan W.M. and Collins J.M. (eds), M. Dekker, New York, pp. 177-192. Roskelley C D , Desprez PY. and Bissell M.J. (1994) Extracellular matrix-dependent tissue-specific gene expression in mammary epithelial cells requires both physical and biochemical signal transduction. Proceedings of the National Academy of Sciences U.S. A. 91: 12378-12382. Rudland P S , Paterson F.C, Twiston Davies A C . and Warburton M.J. (1983) Retinoid-specific induction of differentiation and reduction of the DNA synthesis rate and tumor-forming ability of a stem cell line from a rat mammary tumor. Journal of the National Cancer Institute 70: 949-958. Rudland P S , Paterson F.C, Monaghan P. Twiston Davies A C . and Warburton M.J. (1986) Isolation and properties of rat cell lines morphologically intermediate between cultured mammary epithelial and myoepithelial-like cells. Developmental Biology 113: 388-405. Rudland PS, Hughes C M (1989) Immunocytochemical identification of cell types in human mammary gland: Variations in cellular markers are dependent on glandular topography and differentiation. J Histochem Cytochem 37: 1087-1100. 198 Rudland P.S., Ollerhead G. and Barraclough R. (1989) Isolation of simian vims 40-transformed human mammary epithelial stem cell lines that can differentiate to myoepithelial-like cells in culture and in vivo. Developmental Biology 136: 167-180. Rudland PS (1991) Histochemical organisation and cellular composition of ductal buds in developing human breast: evidence of cytochemical intermediate between epithelial and myoepithelial cells. JHistochem Cytochem 39: 1471-1484. Rudland P.S. and Hughes C M . (1991) Bindings of the lectins Griffonia simplicifolia-l and pokeweed mitogen mark discreet stages of myoepithelial-like differentiation of cell lines from the rat mammary gland. Journal of Cellular Physiology 146: 222-233. Rudland P.S., Ollerhead G.E. and Platt-Higgnis A M . (1991) Morphogenetic behaviour of Simian virus 40-transformed human mammary stem cell lines on collagen gels. In Vitro Cellular and Developmental Biology 27A: 103-112. Rudland P.S., Leinster S.J., Winstanley J., Green B., Atkinson M. and Zakhour H.D. (1993) Immunological identification of cell types in benign and malignant breast disease: Variations in cell markers accompany the malignant state. Journal of Histochemistry and Cytochemistry 41: 543-553. Rudland P.S., Barraclough R., Fernig D G . and Smith J.A. (1997) Mammary stem cells in normal development and cancer. In: Stem Cells. Potten C. (ed), Academic Press Ltd, San Diego, CA, pp 147-232. Russo J. and Russo I.H. (1980) Influence of differentiation and cell kinetics on the susceptibility of the rat mammary gland to carcinogenesis. Cancer Research 40: 2677-2687. Russo J. and Russo I.H. (1987) Development of the human mammary gland. In: The Mammary Gland. Development, Regulation and Function. Neville M.C. and Daniel C.W. (eds), Plenum Publishing Corp, New York, pp 67-93. Russo J., Mills M.J., Moussalli M.J. and Russo I.H. (1989) Influence of human breast development on the growth properties of primary cultures. In Vitro Cellular and Developmental Biology 25: 643-649. Russo J., Rivera R. and Russo I.H. (1992) Influence of age and parity on the development of the human breast. Breast Cancer Research and Treatment 23: 211-218. Russo J., Ao X., Grill C. and Russo I.H. (1999) Pattern of distribution of cells positive for estrogen receptor a and progesterone receptor in relation to proliferating cells in the mammary gland. Breast Cancer Research and Treatment 53: 217-227. Sanderson T.A., Wen P.Y. and Lemay M. (1992) Reversed cerebral asymmetry in women with breast cancer. Lancet 339: 523-524. Sapi E. and Kacinski B.M. (1999) The role of CSF-1 in normal and neoplastic breast physiology. Proceedings of the Society for Experimental Biology and Medicine 220: 1-8. 199 Sapino A , Macri L , GugliottaP. and Bussolati G. (1990) Immunocytochemical identification of proliferating cell types in mouse mammary gland. The Journal of Histochemistry and Cytochemistry 38: 1541-1547. Sapino A , Macri L , GugliottaP, Pacchioni D , Liu Y.-J , Medina D. and Bussolati G. (1993) Immunophenotypic properties and estrogen dependency of budding cell structures in the developing mouse mammary gland. Differentiation 55: 13-18. Schaefer, Custer RP. and Sorof S. (1983) Induction of squamous metaplasia: Requirement for cell multiplication, and competition with lobuloalveolar development in cultured mammary glands. Differentiation 25: 185-192. Schaefer F .V, Custer RP. and Sorof S. (1984) Persistence of precursor cells of squamous metaplasia in preneoplastic mammary outgrowth lines from mice. Journal of the National Cancer Institute 72: 185-189. Sharrock C.E.M, Kaminski E. and Man E. (1990) Limiting dilution analysis of human T cells: a useful clinical tool. Immunology Today 11: 281-286. Shyamala G , Yang X , Silberstein G B , Barcellos-Hoff M.H. and Dale E. (1998) Transgenic mice carrying an imbalance in the native ratio of A to B forms of progesterone receptor exhibit developmental abnormalities in mammary gland. Proceedings of the National Academy of Sciences U.S.A. 95: 696-701. Silberstein G B . and Daniel C.W. (1987) Reversible inhibition of mammary gland growth by transforming growth factor-beta. Science 237: 291-293. Silberstein G.B, Strickland P, Coleman S. and Daniel C.W. (1990) Epithelium-dependent extracellular matrix synthesis in transforming growth factor-beta 1-growth-inhibition mouse mammary gland. Journal of Cell Biology 110: 2209-2219. Silberstein G.B, Flanders K . C , Roberts A B . and Daniel C.W. (1992) Regulation of mammary morphogenesis: Evidence for extracellular matrix-mediated inhibition of ductal budding by transforming growth factor beta 1. Developmental Biology 152: 354-362. Simon B , Podolsky D.K, Moldenhauer G , Isselbacher K.J,Gattoni-Celli S. and Brand S.J. (1990) Epithelial glycoprotein is a member of a family of epithelial cell surface antigens homologous to nidogen, a matrix adhesion protein. Proceedings of the National Academy of Sciences U.S.A. 87:2755-2759. Skalli O , Ropraz P, Trzeciak A , Benzonana G , Gillessen D. and Gabbiani G. (1986) A monoclonal antibody against a-smooth muscle actin: a new probe for smooth muscle differentiation. Journal of Cell Biology 103: 2787-2796. Slamon D.J, Clark G M , Wong S.G, Levin W.J, Ullrich A. and McGuire W.L. (1987) Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 235: 177-182. 200 Slamon D.J., Godolphin W., Jones L. A., Holt J.A., Wong S.G., Keith D.E., Levin W.J., Stuart S.G., Udove J., Ullrich S.G. and Press M.F. (1989) Studies of the HER-2/neu proto-oncogene in human breast and ovarian cancer. Science 244: 707-712. Sloane J. and Ormerod M.G. (1981) Distribution of epithelial membrane antigen in normal and neoplastic tissues and its value in diagnostic tumor pathology. Cancer 47:1786-1795. Smith C. A., Monoghan P. and Neville A.M. (1984). Basal clear cells of the normal human breast. Virchows Archiv [Pathol Anat] 402: 319-329. Smith C. A., Bunnage H.J., Monoghan P. and Neville A.M. (1987) Human breast epithelium in vitro: The re-expression of structural and functional cellular differentiation in long term culture. Cell and Tissue Research 247: 433-440. Smith G H . and MedinaD. (1988) A morphologically distinct candidate for an epithelial stem cell in the mouse mammary gland. Journal of Cell Science 89: 173-183. Smith G.H., Mehrel T. and Roop D.R. (1990) Differential keratin gene expression in developing, differentiating, preneoplastic, and neoplastic mouse mammary epithelium. Cell Growth and Differentiation 1: 161-170. Smith G H . (1996) Experimental mammary epithelial morphogenesis in an in vivo model: Evidence for distinct cellular progenitors of the ductal and lobular phenotype. Breast Cancer Research and Treatment 39: 21-31. Smith H.S., Lan S., Ceriani R., Hackett AJ. and Stampfer MR. (1981) Clonal proliferation of cultured nonmalignant and malignant human breast epithelia. Cancer Research 41: 4637-4643. Smith J.A., Barraclough R., Fernig D G . and Rudland P.S. (1989) Identification of alpha transforming growth factor as a possible local trophic agent for the mammary gland. Journal of Cellular Physiology. 141: 362-370. Snedeker S. M., Brown C. F. and DiAugustine R. P. (1992) Expression and functional properties of transforming growth factor a and epidermal growth factor during mouse mammary gland ductal morphogenesis. Proceedings of the National Academy of Sciences USA 88, 276-280. Stampfer M., Hallowes R.C. and Hackett AJ. (1980) Growth of normal human mammary cells in culture. In Vitro 16: 415-425. Stampfer M R , Yaswen P., Alhadeff M. and Hosoda J. (1993) TGF beta induction of extracellular matrix associated proteins in normal and transformed human mammary epithelial cells in culture is independent of growth effects. Journal of Cellular Physiology 155: 210-221. Strange R., Li F., Saurer S., Burkhardt A. and Friis RR. (1992) Apoptotic cell death and tissue remodelling during mouse mammary gland involution. Development 115: 49-58. Streuli C H . and Bissell M.J. (1990) Expression of extracellular matrix components is regulated by substratum. Journal of Cell Biology 110: 1405-1415. 201 Streuli C.H., Bailey N. and Bissell M.J. (1991) Control of mammary epithelial cell differentiation: Basement membrane induces tissue-specific gene expression in the absence of cell-cell interaction and morphological polarity. Journal of Cell Biology 115: 1383-1395. Sutherland H.J., Eaves C.J, Lansdorp P.M., Thacker J.D., and Hogge D.E. (1991) Differential regulation of primitive human hematopoietic cells in long-term cultures maintained on genetically engineered murine stromal cells. Blood 78: 666-672. Sympson C.J, Talhouk R.S., Alexander C M , Chin J.R, Clift S M , Bissel M.J. and Werb Z. (1994) Targeted expression of stromelysin-1 in mammary gland provides evidence for a role of proteinases in branching morphogenesis and the requirement for an intact basement membrane for tissue-specific gene expression. Journal of Cell Biology 125: 681-692. Taylor-Papadimitriou J , Shearer M. and Stoker M.G.P. (1977) Growth requirements of human mammary epithelial cells in culture. International Journal of Cancer 20: 903-908. Taylor-Papadimitriou J , Purkis and Fentiman IS. (1980) Cholera toxin and analogues of cyclic AMP stimulate the growth of cultured human mammary epithelial cells. Journal of Cellular Physiology 102: 317-321. Taylor-Papadimitriou J. and Lane E. B. (1987) Keratin expression in the mammary gland. In The Mammary Gland: Development, Regulation and Function. Neville M.C. and Daniel C.W. (eds), Plenum Publishing Corp, New York, pp 181-215. Taylor-Papadimitriou J , Stampfer M , Bartek J , Lewis A , Boshell M , Lane E.B. and Leigh LM. (1989) Keratin expression in human mammary epithelial cells cultured from normal and malignant tissue: Relation to in vivo phenotypes and influence of medium. J. Cell Science 94:403-413. Tepperman J. and Tepperman H M . (1987) Metabolic and Endocrine Physiology. Year Book Medical Publishers Inc, Chigaco, IL. Tobon H. and Salazar H. (1974) Ultrastructure of the human mammary gland. I. Development of the fetal gland throughout gestation. Journal of Clinical Endocrinology and Metabolism 39: 443-456. Tobon H. and Salazar H. (1975) Ultrastructure of the human mammary gland. II. Postpartum lactogenesis. Journal of Clinical Endocrinology & Metabolism. 40: 834-44. TokunagaM, Norman J E , Asano M , Tokuoka S, Ezaki H , Nishimori I and Tsuji Y (1979) Malignant breast tumors among atomic bomb survivors, Hiroshima and Nagasaki, 1950-74. Journal of the National Cancer Institute 62: 1347-1359. Tokunaga M , Land C . E , Yamamoto T , Asano M , Toluoka S, Ezaki H. and Nishimori I. (1987) Incidence of female breast cancer among atomic bomb survivors, Hiroshima and Nagasaki, 1950-1980. Radiation Research 112: 243-272. Tonelli Q.J. and Sorof S. (1982) Induction of biochemical differentiation in three-dimensional collagen cultures of mammary epithelial cells from virgin mice. Differentiation 22: 195-200. 202 Trichopoulos D. (1990) Does breast cancer originate in utero? Lancet 335: 939-940. Tsai Y . C , Lu Y , Nichols P.W, Zlotnikov G , Jones P.A. and Smith HS. (1996) Contiguous patches of normal human mammary epithelium derived from a single stem cell: implications for breast carcinogenesis. Cancer Research 56: 402-404. ValveriusE, Walker-JonesD, Bates S E , Stampfer M R , ClarkR, McCormickF, Dickson R.B. and Lippman M.E. (1989) Production of and responsiveness to transforming growth factor beta in normal and oncogene transformed human mammary epithelial cells. Cancer Research 49: 6269-6274. Vogel P.M., Georgiade N G , Fetter BF. and McCarty KS. (1981) The correlation of histologic changes in the human breast with the menstrual cycle. American Journal of Pathology 104: 23-34. Vorherr H. (1974) The development on the female breast. In: The Breast: Morphology, physiology and lactation. Vorherr H. (ed), Academic Press, New York, N.Y, pp 1-18. Walden P.D, Ruan W , Feldman M. and Kleinberg D L . (1998) Evidence that the mammary fat pad mediates the action of growth hormone in mammary gland development. Endocrinology 139: 659-662. Wasner G, Hennerman I. and Kratochwil K. (1983) Ontogeny of mesenchymal androgen receptors in the embryonic mouse mammary gland. Endocrinology 113: 1771-1780. Wellings S.R, Jensen H M . and Marcum R G . (1975) An atlas of subgross pathology of the human breast with special reference to possible precancerous lesions. Journal of the National Cancer Institute 55:231-273. Wiesen J.F, Young P, Werb Z and Cunha G.R. (1998) Signalling throughout the stromal epidermal growth factor receptor is necessary for mammary ductal development. Development 126: 335-344. Williams J M . and Daniel C.W. (1983)Mammary ductal elongation: Differentiation of myoepithelium and basal lamina during branching morphogenesis. Developmental Biology 97: 274-290. Witty J.P, Wright J.H. and Matrisian L.M. (1995) Matrix metalloproteinases are expressed during ductal and alveolar mammary morphogenesis, and misregulation of stromelysin-1 in transgenic mice induces unscheduled alveolar development. Molecular Biology of the Cell 6: 1287-1303. Woodcock-Mitchell J , Eichner R, Nelson W G , Sun T.T. (1982) Immunolocalization of keratin polypeptides in human epidermis using monoclonal antibodies. Journal of Cell Biology 95: 580-588. Wysolmerski J.J, Philbrick W M , Dunbar M . E , Lanske B , Dronenberg H , Karaplis A. and Broadus A.E. (1998) Rescue of the parathryoid hormone-related protein knockout mouse of the 203 parathryoid hormone-related protein knockout mouse demonstrates that parathryoid hormone-related protein is essential for mammary gland development. Development 125: 1285-1294. Xing P X , Prenzoska J , Quelch K. and McKenzie I.F.C. (1992) Second generation anti-MUCl peptide monoclonal antibodies. Cancer Research 52: 2310-2317. Yang J , Richards J , Guzman R, Imagawa W. and Nandi S. (1980) Sustained growth in primary culture of normal mammary epithelial cells embedded in collagen gels. Proceedings of the National Academy of Sciences U.S.A. 77: 2088-2092. Yang, Y , Spitzer E , MeyerD, SachsM, Niemann C , HartmannG, Weidner K M , Birchmeier C. and Birchmeier W. (1995) Sequential requirement of hepatocyte growth factor and neuregulin in the morphogenesis and differentiation of the mammary gland. Journal of Cell Biology 131: 215-226. Zeps N , Dawkins H.J.S, Papadimitriou J.M, Redmond SL. and Walters M.-N.I. (1996) Detection of a population of long-lived cells in mammary epithelium of the mouse. Cell and Tissue Research 286: 525-536. 204 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0099500/manifest

Comment

Related Items