Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Modeling mammary epithelial cell polarization and the role of podocalyxin in breast tumor progression Graves, Marcia Lynn 2008

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

Item Metadata

Download

Media
24-ubc_2008_fall_graves_marcia.pdf [ 27.7MB ]
Metadata
JSON: 24-1.0066552.json
JSON-LD: 24-1.0066552-ld.json
RDF/XML (Pretty): 24-1.0066552-rdf.xml
RDF/JSON: 24-1.0066552-rdf.json
Turtle: 24-1.0066552-turtle.txt
N-Triples: 24-1.0066552-rdf-ntriples.txt
Original Record: 24-1.0066552-source.json
Full Text
24-1.0066552-fulltext.txt
Citation
24-1.0066552.ris

Full Text

MODELING MAMMARY EPITHELIAL CELL POLARIZATION AND THE ROLE OF PODOCALYXIN IN BREAST TUMOR PROGRESSION by MARCIA LYNN GRAVES B.Sc., The University of British Columbia, 2001  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Cell and Developmental Biology)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) August 2008 © Marcia Lynn Graves, 2008  ABSTRACT The mammary gland consists of an organized network of epithelial ducts and lobules. This histoarchitecture can be recapitulated in vitro by culturing mammary epithelial cells as 3D spheroids embedded in a reconstituted basement membrane. I first used this assay to characterize the role of cell-cell and cell-ECM adhesion in the formation and polarization of the apical junction complexes in normal mammary epithelial cells. Cell-cell adhesion alone was sufficient to initiate polarized junction assembly. However, the addition of exogenous ECM generated a spatial polarity signal dependent on laminin-1 and α6 and β1 integrins. This caused clusters of mammary epithelial cells to relocalize the junctional complexes to the center of the spheroid prior to lumen formation.  In ductal breast carcinoma, a critical hallmark is the loss of normal polarized tissue architecture without the induction of an epithelial-to-mesenchymal transformation (EMT). Thus, misregulation of molecules that function as polarity determinants may contribute to ductal tumor progression. Podocalyxin is an anti-adhesive glycoprotein that may be involved, as it is important in epithelial morphogenesis, and its overexpression in clinical breast tumors is associated with poor outcome. Despite this, overexpression of podocalyxin in normal mammary epithelial cells did not disrupt 3D morphogenesis or apicobasal polarity. However, its overexpression in non-metastatic breast tumor cells did perturb the architecture and growth of tumor spheroids in vitro and it facilitated subcutaneous tumor growth in vivo without causing an EMT. ii  Mechanistically, podocalyxin localized to and expanded non-adhesive membrane domains and induced microvillus formation that was dependent on its extracellular domain and Rho GTPase-regulated actin polymerization. Podocalyxin also recruited its intracellular binding partners NHERF-1 and ezrin via its cytoplasmic tail. Strikingly, the formation of this protein complex was not required for microvillus formation. Additionally, podocalyxin delayed cell-cell aggregation and decreased the initial adhesion, spreading and strength of attachment of tumor cells to fibronectin where it restricted β1 integrin localization to the basal/attached domain. These alterations in adhesion possibly contributed to podocalyxin's ability to increase growth factordependent tumor cell migration. Altogether, these data indicate that podocalyxin overexpression may facilitate a ductal tumor-like progression that involves EMT-independent alterations in tissue architecture.  iii  TABLE OF CONTENTS ABSTRACT ........................................................................................................................................ ii LIST OF TABLES........................................................................................................................... viii LIST OF FIGURES ...........................................................................................................................ix LIST OF ABBREVIATIONS......................................................................................................... xii ACKNOWLEDGEMENTS ............................................................................................................xiv CHAPTER 1 : INTRODUCTION ...................................................................................................1 1.1  Basic principles of cell polarity.........................................................................................1  1.2  Epithelial junctions and the establishment of apicobasal polarity ...........................1  1.2.1 1.2.2  1.3  Mammary gland biology and epithelial morphogenesis .........................................12  1.3.1 1.3.2  1.4  Disruption of mammary gland architecture and epithelial cell polarity during breast cancer progression ..................................................................................................................................22  Podocalyxin: A potential regulator of epithelial morphogenesis ...........................28  1.5.1 1.5.2 1.5.3 1.5.4 1.5.5 1.5.6  1.6  Mammary gland development....................................................................................................12 Regulating mammary gland architecture: Insights from 3D in vitro models of mammary morphogenesis .............................................................................................................................19  Breast cancer: Derailing tissue architecture.................................................................21  1.4.1  1.5  Basic structure and function of epithelial cell junctions ............................................................2 Assembly of the apical junction complex during epithelial polarization...................................7  Basic overview of podocalyxin ....................................................................................................29 Tissue profile of podocalyxin expression....................................................................................35 Regulation of podocalyxin expression........................................................................................36 Insights into podocalyxin function: Kidney morphogenesis and podocalyxin knockout analysis ........................................................................................................................................39 Controlling epithelial morphogenesis: Proposed functions of podocalyxin as an anti-adhesin and regulator of apicobasal polarity ..........................................................................................41 Podocalyxin in cancer .................................................................................................................46  Thesis objectives ................................................................................................................50  CHAPTER 2 : MATERIALS AND METHODS........................................................53 2.1  Methods for Chapter 3 .....................................................................................................53  2.1.1 2.1.2 2.1.3 2.1.4  2.2  General cell culture .....................................................................................................................53 3D cell culture model ..................................................................................................................53 Immunocytochemistry and confocal analysis............................................................................55 Cell lysis and Western blotting ..................................................................................................56  Methods for Chapter 4 .....................................................................................................58  2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.2.6 2.2.7 2.2.8 2.2.9  Mice and tissue processing .........................................................................................................58 Immunohistochemistry ...............................................................................................................58 Western blotting..........................................................................................................................59 Semi-quantitative and quantitative RT-PCR ...........................................................................59 Expression vectors and cloning..................................................................................................60 Transfections and generation of stable cell lines .......................................................................61 3D MCF-7 tumor cell spheroid culture.....................................................................................63 Immunocytochemistry and confocal analysis............................................................................64 Scanning electron microscopy ....................................................................................................65  iv  2.2.10  2.3  Methods for Chapter 5 .....................................................................................................66  2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 2.3.6 2.3.7 2.3.8 2.3.9  2.4  Transmission electron microscopy ...........................................................................................65 Sub-cutaneous xenograft model of breast tumorigenesis .........................................................66 Tumor histopathological and immunohistochemical analysis .................................................67 Cell culture, immunocytochemistry and confocal analysis ......................................................67 Western blotting..........................................................................................................................68 Homotypic cell-cell aggregation analysis ..................................................................................68 Single cell attachment and cell spreading assays......................................................................68 Cell-ECM adhesion .....................................................................................................................69 Migration assays .........................................................................................................................70 In vitro growth assays.................................................................................................................71  Methods for Chapter 6 .....................................................................................................73  2.4.1 2.4.2 2.4.3 2.4.4 2.4.5  Mice and tissue processing .........................................................................................................73 Lentiviral shRNA for murine NHERF-1 ..................................................................................74 Podocalyxin mutant analysis .....................................................................................................74 Additional expression vectors/reagents for transient transfection analysis ...........................75 Latrunculin-mediated f-actin disruption...................................................................................76  CHAPTER 3 : REGULATING APICAL POLARIZATION IN A 3D MODEL OF MAMMARY MORPHOGENESIS ...........................................................................81 3.1  Introduction........................................................................................................................81  3.2  Expression and localization of epithelial junction proteins in EpH4 cells............82  3.3  EpH4 cells polarize and functionally differentiate in an-ECM dependent 3D spheroid culture ................................................................................................................87  3.4  Modeling apical orientation of cell junctions in 3D mammary spheroids independent of ECM-dependent cell shape changes................................................90  3.5  ECM-dependent apical orientation occurs prior to functional differentiation ....99  3.6  ECM enhances the basally polarized localization of integrins..............................103  3.7  The ECM component laminin-1 is sufficient to apically re-orient polarity in 3D mammary spheroids.......................................................................................................107  3.8  The α6 and β1 integrin subunits are important to apically re-orient polarity in 3D mammary spheroids.................................................................................................110  3.9  Summary and Conclusions ...........................................................................................114  CHAPTER 4 : CHARACTERIZATION OF PODOCALYXIN EXPRESSION AND FUNCTION IN MAMMARY EPITHELIAL AND BREAST TUMOR CELLS....120 4.1  Introduction......................................................................................................................120  4.2  Podocalyxin expression during mammary gland development in vivo and morphogenesis in vitro ..................................................................................................123  4.2.1 4.2.2 4.2.3 4.2.4  4.3  Podocalyxin is expressed and apically localized in adult mammary glands .........................123 Podocalyxin is not localized in mammary epithelial cells during lactation ..........................124 Podocalyxin protein decreases during ECM-dependent morphogenesis and differentiation in vitro, but is not transcriptionally down-regulated ................................................................130 Podocalyxin protein is secreted during pregnancy/lactation .................................................133  Functional analysis of podocalyxin overexpression in normal mammary epithelial cells and breast tumor cells.........................................................................136  v  4.3.1 4.3.2 4.3.3 4.3.4 4.3.5 4.3.6  4.4  Podocalyxin overexpression does not prevent epithelial cell polarization during normal differentiative mammary morphogenesis.................................................................................136 Podocalyxin overexpression alters the architecture of breast tumor cell spheroids without altering junctional polarity ......................................................................................................143 Podocalyxin causes apical membrane expansion in polarized kidney epithelial cells and breast tumor cells ......................................................................................................................148 Podocalyxin induces robust microvillus formation at free, apical membrane domains .......154 Podocalyxin recruits NHERF-1/ezrin/actin to the apical surface where they co-exist in microvilli....................................................................................................................................157 Ectopic podocalyxin promotes the formation of actin-rich membrane projections in the absence of cell-cell adhesion......................................................................................................163  Summary and Conclusions ...........................................................................................167  CHAPTER 5 : FUNCTION OF PODOCALYXIN IN BREAST CANCER PROGRESSION ......................................................................................................172 5.1  Introduction......................................................................................................................172  5.2  Podocalyxin perturbs epithelial architecture and causes delamination of 2D monolayers without inducing an EMT ......................................................................174  5.3  Podocalyxin perturbs homotypic cell-cell aggregation ..........................................179  5.4  Podocalyxin disrupts cell-ECM adhesion and delays cell spreading ..................183  5.5  Podocalyxin promotes growth factor-induced migration of breast tumor cells……...…………………………………………………………………………… 187  5.6  Podocalyxin defines the free surface domain by restricting β1 integrin localization........................................................................................................................191  5.7  Podocalyxin increases tumor spheroid growth in 3D Culture..............................196  5.8  Podocalyxin may increase breast tumor growth in vivo without disrupting cellcell junctions .....................................................................................................................199  5.9  Summary and Conclusions ...........................................................................................205  CHAPTER 6 : REGULATION OF PODOCALYXIN LOCALIZATION AND FUNCTION IN MAMMARY EPITHELIAL AND BREAST TUMOR CELLS....210 6.1  Introduction......................................................................................................................210  6.2  The role of NHERF-1 in the localization and function of podocalyxin ...............212  6.2.1 6.2.2 6.2.3  6.3  NHERF-1 is expressed and apically localized in luminal mammary epithelial cells in vivo ....................................................................................................................................................213 NHERF-1 is not required for in vitro mammary morphogenesis and apical polarization of podocalyxin in vivo ...................................................................................................................216 The role of NHERF-1 in podocalyxin localization and microvillus formation in breast tumor cells.............................................................................................................................................225  The role of ezrin in podocalyxin localization and microvillus formation in breast tumor cells ........................................................................................................................231  6.3.1 6.3.2 6.3.3  Ezrin is not required for podocalyxin localization in pre-apical membrane domains of single cells.............................................................................................................................................231 Expression of a dominant-negative ezrin mutant does not disrupt apical podocalyxin localization or microvillus formation.......................................................................................234 The extracellular domain of podocalyxin is not required for apical localization but it is required for the induction of microvillus formation ...............................................................242  vi  6.4  The role of actin and regulators of the actin cytoskeleton in podocalyxin localization and microvillus formation ......................................................................249  6.4.1 6.4.2  Podocalyxin-induced microvillus formation requires actin polymerization .........................249 Balanced RhoGTPase and CDC42 activation are required for podocalyxin induced microvillus formation................................................................................................................252  6.5  The ability of podocalyxin mutants to induce microvillus formation and disrupt 3D tumor cell spheroids are positively correlated...................................................253  6.6  Summary and Conclusions ...........................................................................................259  CHAPTER 7 : CONCLUDING REMARKS ...........................................................267 7.1  Summary and Discussion..............................................................................................267  7.2  Potential clinical importance and future directions.................................................275  CHAPTER 8 : REFERENCES..................................................................................281  vii  LIST OF TABLES Table 2-1: List of primary antibodies used in this thesis........................................77 Table 2-2: List of secondary antibodies used in this thesis ....................................79 Table 2-3: Sequences of shRNA oligomers for mouse Podocalyxin and NHERF-1 .............................................................................................................................80  viii  LIST OF FIGURES Figure 1-1: Figure 1-2: Figure 1-3: Figure 1-4: Figure 1-5: Figure 3-1: Figure 3-2: Figure 3-3: Figure 3-4: Figure 3-5: Figure 3-6: Figure 3-7: Figure 3-8: Figure 4-1: Figure 4-2: Figure 4-3: Figure 4-4: Figure 4-5: Figure 4-6:  Figure 4-7:  Schematic of a polarized epithelial cell................................................ 3 Schematic of murine mammary gland development........................14 Genomic organization of CD34 family members ..............................30 Schematic of the podocalyxin/NHERF-1/ezrin protein complex ...33 Podocalyxin is overexpressed in a subset of invasive breast tumors and is an independent prognostic indicator of poor patient outcome ...............................................................................................................48 Expression and localization of junction proteins in 2D monolayer culture of EpH4 cells............................................................................85 EpH4 cells polarize and functionally differentiate in an ECMdependent 3D spheroid culture ..........................................................88 Dissecting mammary spheroid architecture independent of cell shape: Junctional assembly and polarization is ECM-independent, but apical orientation is ECM-dependent ......................................................93 Par3 and Par6 exhibit distinct localization patterns and Par3, but not Par6, is present in a triton insoluble pool at the free surface of mammary spheroids ............................................................................97 The dynamics of ECM-dependent spheroid formation, β-casein production and apical re-orientation over time ..............................100 The α6 integrin subunit becomes exclusively polarized to the basal membrane of ECM-induced 3D mammary spheroids....................105 Laminin-1, but not collagen I, is sufficient to re-orient apical polarity in 3D mammary spheroids................................................................108 Both α6 and β1 integrin are important for ECM-dependant apical junction re-orientation .......................................................................112 Podocalyxin is localized to the apical membrane in luminal mammary epithelial cells...................................................................125 Podocalyxin is not detected in differentiated lobuloalveolar epithelium of the mouse mammary gland during pregnancy and lactation...............................................................................................128 Podocalyxin protein is absent during in vitro differentiation of EpH4 cells but this is not a result of transcriptional down-regulation ....131 Podocayxin is detected within the lumens of primary ducts of mouse mammary glands, and is present in human breast milk ....134 Forced expression of podocalyxin in EpH4 cells does not disrupt mammary epithelial morphology in 2D or functional differentiation in 3D ....................................................................................................138 Forced expression of podocalyxin preferentially localizes to the outer surface of naked clusters and does not alter ECM-dependent signals to re-orient apical polarity in 3D overlay mammary spheroids.............................................................................................141 Forced overexpression of podocalyxin is not sufficient to induce interior lumen formation in 3D breast tumor cell spheroids..........145 ix  Figure 4-8: Podocalyxin causes apical domain expansion in MCF-7 and MDCK cells and recruits its intracellular binding partner NHERF-1 to the apical membrane ................................................................................151 Figure 4-9: Podocalyxin induces robust microvillus formation at the cell surface of MDCK and MCF-7 cells ................................................................155 Figure 4-10: Podocalyxin co-localizes with NHERF-1 and actin where they coexist along individual microvilli.......................................................158 Figure 4-11: Podocalyxin recruits ezrin and actin to microvilli rich apical membranes..........................................................................................161 Figure 4-12: Podocalyxin localizes to the free surface and causes reorganization of apical f-actin in single attached cells ............................................165 Figure 5-1: Podocalyxin mildly perturbs epithelial architecture in 2D monolayers without disruption of cell-cell junction formation or proliferation ........................................................................................177 Figure 5-2: Overexpression of Podocalyxin interferes with homotypic cell-cell aggregation, but does not prevent E-cadherin mediated adhesions in 3D tumor cell spheroids ................................................................181 Figure 5-3: Forced expression of Podocalyxin inhibits initial adhesion, delays spreading, and decreases the strength of adhesion to fibronectin.184 Figure 5-4: Podocalyxin increases the migratory potential of breast tumor cells .............................................................................................................189 Figure 5-5: Free surface-targeted podocalyxin restricts β1 integrin localization .............................................................................................................194 Figure 5-6: Podocalyxin increases growth of breast tumor spheroids..............197 Figure 5-7: Podocalyxin may promote breast tumor growth in vivo.................200 Figure 5-8: Podocalyxin xenografts contain densely packed, proliferative breast tumor cells and display intact cell-cell junctions.............................203 Figure 6-1: NHERF-1 is expressed in luminal mammary epithelia and is enriched at the apical membrane......................................................214 Figure 6-2: NHERF-1 depletion in vitro does not disrupt 3D mammary spheroid formation and apical polarization ....................................................217 Figure 6-3: Lobuloalveolar development is morphologically unchanged in NHERF-1 knockout mice...................................................................221 Figure 6-4: Apical membrane localization of podocalyxin is unaffected in NHERF-1 knockout mice...................................................................223 Figure 6-5: Deletion of podocalyxin’s c-terminal PDZ binding domain is not sufficient to disrupt its apical targeting, but does prevent NHERF-1 recruitment .........................................................................................226 Figure 6-6: The PDZ binding domain of podocalyxin is not required for microvillus formation ........................................................................229 Figure 6-7: Podocalyxin localizes to the pre-apical membrane without ezrin in single cells...........................................................................................232 Figure 6-8: Apical localization of Podocalyxin and NHERF-1 is not dependent on ezrin................................................................................................236 Figure 6-9: The cytoplasmic domain of podocalyxin is not required for its apical targeting or microvilli induction.......................................................239  x  Figure 6-10: Podocalyxin-dependent microvilli formation is not dependent on the sialic acid residues of its extracellular domain..........................243 Figure 6-11: The extracellular domain of podocalyxin is required for microvillus formation.............................................................................................245 Figure 6-12: Deletion of the extracellular domain of podocalyxin does not prevent its apical targeting................................................................247 Figure 6-13: Actin polymerization is required for podocalyxin induced microvillus formation ........................................................................250 Figure 6-14: Expression of dominant negative forms of Rho and CDC42 GTPase disrupt microvilli, but do not affect apical localization of podocalyxin ........................................................................................254 Figure 6-15: Podocalyxin’s ability to form cell surface microvilli correlates with 3D tumor spheroid architectural disruption....................................257 Figure 6-16: Proposed models for podocalyxin-dependent microvillus formation .............................................................................................................264 Figure 7-1: Overall model outlining the consequences of podocalyxin overexpression in MCF-7 breast tumor cells ...................................270  xi  LIST OF ABBREVIATIONS 3D aa AJ AJC APC aPKCλ/ζ bp BSA C3 CDC42 ch CKII ColI CRB DABCO DAPI DMEM DN DTHL EBP50 ECL ECM EDTA EGF EGFR EGFP ER ERM ESC F-12 FACs FAK FBS FVB GFP gp135 GTPase H&E HEV HGEC HRP IF Ig IHC IRES  three dimensional amino acid adherens junctions apical junction complex allophycocyanin atypical protein kinase C lambda/zeta base pair bovine serum albumin clostridium botulinum C3 exoenzyme cell division cycle 42 homolog chicken casein kinase II collagen I crumbs 1,4-diazabicyclo[2.2.2]octane 4’, 6’-diamidino-2-phenylindole Dulbecco’s modified Eagle’s media dominant negative Asp-Thr-His-Leu ezrin binding phosphoprotein of 50 kDa enhanced chemiluminescence extracellular matrix ethylenediaminetetraacetic acid epidermal growth factor epidermal growth factor receptor enhanced green fluorescent protein estrogen receptor ezrin-radixin-moesin embryonic stem cell Ham’s F12 media fluorescence activated cell sorting focal adhesion kinase fetal bovine serum friend leukemia virus B (Fv1b) mouse strain green fluorescent protein glycoprotein 135 guanosine triphosphatase hematoxylin and eosin high endothelial venule human glomerular epithelial cells horse radish peroxidase immunofluorescence immunoglobulin immunohistochemistry internal ribosome entry site xii  kDa KO LamI MDCK MEP Ms NHE NHERF PAGE Par PBS PCR PDGFR PDZ PFA PKC PMSF Podo RAC1 Rb Rho RT-PCR SA SD SDS SEM shRNA SJ TBS TBS-T TEM TER TGF-β TJ UTR WB WT WT1 ZO  kiloDalton knockout Laminin I Madin-Darby canine kidney Myb-Ets transformed progenitor mouse Na+/H+ exchanger Na+/H+ exchanger regulatory factor polyacrylamide gel electrophoresis partitioning defective phosphate buffered saline polymerase chain reaction platelet-derived growth factor receptor PSD-95/Dlg/ZO-1 paraformaldehyde protein kinase C phenylmethanesulfonyl fluoride podocalyxin ras-related C3 botulinum toxin substrate small GTP binding protein rabbit ras homology reverse transcriptase-polymerase chain reaction streptavidin standard deviation sodium dodecyl sulfate scanning electron microscopy small hairpin interfering RNA septate juntion Tris buffered saline Tris buffered saline plus 0.05 % Tween 20 transmission electron microscopy transepithelial resistance transforming growth factor beta tight junctions untranslated region western blot wildtype Wilms’ tumour 1 zonula occludens  xiii  ACKNOWLEDGEMENTS I would like to take this opportunity to acknowledge all the important people that have supported me throughout the course of my academic studies. First, I would like to thank my supervisor, Dr. Cal Roskelley. The time I have spent in his lab has been extraordinary, and a substantial period of personal growth. His mentorship has offered me an unparalleled benchmark to which I have aspired both scientifically and personally, and I will always be grateful for this experience. I am also grateful to have received a tremendous amount of support and helpful advice from my supervisory committee, Dr. Tim O’Connor, Dr. Kelly McNagny, and Dr. Blake Gilks. In addition, The Anatomy (CPS) Department is, and has been very supportive of its Graduate Students. I would particularly like to acknowledge Dr. John Church, Dr. Tim O’Connor, Dr. Wayne Vogl and Roseanne McIndoe for their hard work, encouragement and advice.  The research projects in which I have been involved have allowed me to collaborate with a number of individuals from diverse areas of expertise, all of which have provided me with technical training, scientific guidance and reagents that have greatly enhanced my research endeavors. It is difficult to name all of the important individuals who have contributed to my training and this thesis, but I would like to particularly thank Dr. Chris Mueller, Dr. Kelly McNagny, Dr. Blake Gilks, Dr. Wayne Vogl, Dr. John Stingl, Dr. Julie Nielsen, Poh Tan, Bernhard Lehrnetz, Jane Cipollone, Dr. Sarah McLeod, Dr. Jacky Goetz  xiv  and Spencer Freeman. Also, a very special thanks to Pamela Austin for her time and invaluable expertise with the in vivo mouse experiments.  A large part of what has made working in the Roskelley lab so enjoyable over the years is the people I have had the pleasure of meeting and getting to know personally. I would like to thank the support, and lasting friendships of the past (or honorary) lab members; Dr. Aruna Somasiri, Dr. Jennifer Bonner, Colleen Wu, Dr. Cameron Dehoney, Dr. Ryan Sinotte, Dr. Kenny To, Dr. Lianne McHardy, Arthur Legg and Dr. Julian Guttman for creating such a fun and supportive learning environment early on and throughout my training. I would like to especially acknowledge the current members of the lab; Sarah, Pam, Jane, Stephanie, Spencer, and Holden, for their unwavering support and understanding while I have finished my degree. Also, a special thanks to Greg Handrigan and Kamal Garcha for their support and encouragement during my thesis writing, and to Jane Cipollone for editing parts of this thesis. I have made a number of very important friendships in the lab over the years, and I am especially grateful for Jane, Sarah, Julie, Poh, Jennifer, Colleen, Jacky, Pascal, and Spencer. You have all impacted me dearly, and I thank you.  Lastly, I am indebted to my family. To my Parents, Ed and Angie, To Baba, To Tana and Brian and all the Taylor’s, To Jean and Darrel, To Steve, and the Bennett family, this would not have been possible without your unconditional love, support and encouraging words. And To Dan, for You, I am forever thankful, now and always.  xv  CHAPTER 1 : INTRODUCTION 1.1 Basic principles of cell polarity Polarity within a cell is characterized by an asymmetric spatial distribution of its cellular components, and is critical for most cellular functions. The plasma membrane of a polarized cell is segregated into specialized membrane domains that contain different protein and lipid compositions. Inside the cell, there is an asymmetric subcellular arrangement of organelles and the cytoskeleton which control the precise delivery of the appropriate lipids, membrane-associated proteins and secreted proteins to specific regions of the plasma membrane, thereby defining and maintaining their unique idenitites and functions (Mostov et al., 2003).  Most cells have the capacity to polarize, and they do so by responding to, and interpreting various spatial cues that provide a landmark to orient the cell in a particular direction (Nelson, 2003). For example, polarity can be triggered by cues from the external environmental that include growth factor and chemokine stimulation as well as points of cell-cell and cell-matrix interactions. Multicellular interactions create a higher order level of polarity across the entire tissue, and one of the best examples of polarity that develops within a tissue is the apicobasal polarity axis in polarized epithelia.  1.2 Epithelial junctions and the establishment of apicobasal polarity Epithelial tissues are composed of closely arranged cells that exist either as a single cell layer or as a multi-layer of cells that covers the surface of the body or lines  1  internal cavities and tubes. To regulate tissue homeostasis, sheets of epithelia function to maintan separate physiologically distinct compartments by acting as protective barriers that control the selective exchange of solutes and ions across the epithelial layer (Figure 1-1). This generalized function requires individual epithelial cells within the sheet to establish and maintain a distinct polarity such that regions of each cell are divided into functionally discrete membrane domains (O'Brien et al., 2002). The apical membrane often faces a hollow luminal cavity, while the adhesive basolateral membrane contacts both neighbouring cells and the underlying extracellular matrix (ECM). Therefore, the structure and integrity of this apicobasal axis relies on both cell-cell and cell-matrix junctions.  Epithelial junctions are highly complex networks of numerous and dynamic protein-protein interactions that create platforms where cell adhesion and cell signaling pathways converge. Thus, a basic description of the structure and function of the different junctions found within polarized epithelia is outlined below (Also see Figure 1-1), followed by a more detailed description of how epithelial cells establish apicobasal polarity.  1.2.1  Basic structure and function of epithelial cell junctions  Cell-cell and cell-ECM junctions share a common organization consisting of transmembrane proteins that link to the cytoskeleton through modular, multiprotein scaffolds. The collective linkage of many individual cells to one another and to the underlying ECM, which all connect to the cytoskeleton, allows the entire epithelial sheet to function as a coordinated tissue (Perez-Moreno et al., 2003). Thus, the integrity and structure of cell junctions are critical for its overall function. 2  Figure 1-1: Schematic of a polarized epithelial cell Shown is a schematic of a transporting epithelium that typically lines internal cavities or ducts to separate two biological compartments. Polarized epithelial cells are closely arranged and each cell is sub-divided into apical (pink), lateral (green) and basal (grey) (or collectively known as ‘basolateral’) membrane surfaces. During polarization, the identity of each membrane domain is defined and maintained by conserved classes of polarity proteins. The Crumbs/Pals1/PatJ and Par3/Par6/aPKC help to establish and define the apical membrane domain, while the Scribble/dlg/Lgl complex functions to define the basolateral domain. The structural integrity and apicobasal polarity of the epithelium relies on cell junctions, which anchor the plasma membrane to the cytoskeleton. These include: apical tight junctions and adherens junctions; desmosomes and gap junctions located along the lateral membrane, and cell-matrix junctions such as focal complexes and hemidesmosomes that form at the basal surface.  3  4  There are four main types of cell-cell junctions that form along the lateral membranes of two opposing cells: tight junctions, adherens junctions, desmosomes and gap junctions. The tight junctions and the adherens junctions form between adjacent cells at the apical-most part of the lateral membrane. Together, along with the numerous cytoplasmic scaffolding and regulatory proteins that associate with each junction, they are often collectively referred to as ‘the apical junction complex.’ The localization of desmosomes and gap junctions are less limited, as they are found at multiple points along the entire lateral membrane.  Ultrastructurally, desmosomes resemble patch-like adhesions that form plaques on the inner membrane face to create a linkage to intermediate filaments that helps to maintain structural integrity across the tissue, particularly in epithelia that experiences considerable mechanical stress (Stokes, 2007). Gap junctions form channels between cells that allow for intercellular communication by permitting the transfer of ions and small molecules between adjacent cells within the tissue (McLachlan et al., 2007).  The apical-most tight junction demarcates the apical and lateral membrane domain. Tight junctions help to physically and functionally maintain the identity of each membrane domain by preventing lateral mixing of lipids and proteins between the apical and lateral surfaces, which is often referred to as the ‘fence function’ (Shin et al., 2006). Clusters of integral membrane proteins, primarily of the claudin family, closely interact with cognate receptor pairs in the lateral membrane of opposing cells at ‘kissing points’, and are critical for the formation of a network of sealing strands at the site of apical cell-cell adhesion (Furuse and Tsukita, 2006). Each 5  sealing strand interacts with another strand on the adjacent cell to form a ‘paired tight junction strand’. As a result, the overall network of paired strands creates an impermeable barrier to prevent unwanted diffusion across the epithelium (Furuse and Tsukita, 2006). On the cytoplasmic face of each strand, the cytoplasmic tails of the claudins, as well as the other tight junction-associated transmembrane proteins that include occludin and junctional adhesion molecule (Jam), associate either directly or indirectly with a plethora of PDZ-containing scaffolding proteins such as members of the the zonula occludens family (ie: ZO-1, ZO-2, ZO-3). This creates the formation of large, membrane-associated plaques that link tight junctions to the actin cytoskeleton and to a variety of signaling molecules (Shin et al., 2006).  The adherens junction is an electron dense structure located beneath (ie: basally to) the tight junction and is associated with an organized belt of bundled actin filaments that encircles the apex of the cell, which provides stable adhesive strength between adjacent cells across the tissue (Perez-Moreno et al., 2003). On a molecular level, the junction is formed by calcium dependent homophilic interactions between the extracellular domains of transmembrane E-cadherin molecules. On the cytoplasmic face, E-cadherin molecules form cytoplasmic plaque complexes with β and α catenin that interact in a dynamic fashion with other plaque proteins that can initiate signaling cascades and interact with the actin cytoskeleton. (Weis and Nelson, 2006). Also, the integral membrane protein nectin also localizes to adherens junctions where it interacts with the cytoplasmic plaque protein afadin, which directly binds the actin cytoskeleton. Thus, this latter interaction is likely an  6  important structural linkage connecting the adherens junction to the circumferential actin belt (Gates and Peifer, 2005).  Epithelial tissues are supported at the basal surface by a specialized ECM called the basement membrane. Adhesions form between the basal membrane domain of epithelial cells and the laminin-rich surface of the basement membrane, known as the basal lamina (Sasaki et al., 2004). Thus, specific laminin receptors, many of which are integrins, are critical for cell-ECM adhesions in epithelial tissues. Integrins are transmembrane heterodimers composed of two subunits each from the α and β family of integrin molecules. Laminin receptors include (but are not limited to) α2β1, α3β1, α6β1, and α6β4 integrin dimers (Belkin and Stepp, 2000; Hynes, 1992). Additionally, the glycoprotein dystroglycan binds laminin in muscle, nerve and epithelial tissues. Engagement of all of these receptors by laminin triggers the assembly of a large plaque of scaffolding and signaling molecules on the cytoplasmic side of the basal surface of the cell where it links the junction to the actin cytoskeleton. One exception is the α6β4 receptor, which initiates the formation of hemidesmosomes that uniquely anchor laminin to the intermediate filament cytoskeleton (Litjens et al., 2006).  1.2.2  Assembly of the apical junction complex during epithelial polarization  In multicelluar epithelia, setting up the apicobasal axis involves creating distinct plasma membrane domains and assembling spatially organized cell junctions. A critical discovery in unraveling the mechanisms of cell polarization came from genetic screens in Drosophila and C.elegans (Reviewed in (Goldstein and Macara, 7  2007). Three classes of protein complexes were identified that cooperate in a hierarchical pathway to control epithelial cell polarity (Bilder et al., 2003; Tanentzapf and Tepass, 2003). The subsequent identity of mammalian homologues with similar functions in epithelial cells (and other cell types) suggests that these protein complexes execute a highly conserved program to regulate cell polarity. The ‘Par3/Par6/aPKC’ complex exists at the top of the hierarchy to define the apical membrane domain (Bilder et al., 2003). Par3 is particularly important in assembling members of the tight junction complex, while Par6/aPKC are important in defining the apical plasma membrane surface (Hirose et al., 2002; Martin-Belmonte et al., 2007). Next in the hierarchy, the ‘Crumbs/Pals1/PatJ’ complex is recruited to the apical junction complex by the ‘Par complex’, and while it is not required for junction assembly, it is required to maintain the identity of the apical membrane domain once it is established (Tanentzapf and Tepass, 2003). In contrast, the ‘Scribble/Dlg/Lgl’ complex functions to maintain the identity of the basolateral membrane domain. Altogether, these three protein complexes form the core polarity program that translates an external polarity signal, such as cell-cell contact, into the cell behavioural response to alter and maintain cell asymmetry (Figure 1-1).  How do these polarity complexes work? In general, the two complexes that define the apical domain antagonize the complex that marks the basolateral domain and vice versa creating a “molecular tug of war” amongst the three protein complexes. For example, Lgl is normally localized along the lateral cell membrane. However, if Lgl becomes inappropriately localized near the apical domain, aPKC phosphorylates it and releases it from the membrane. This sends Lgl to the cytoplasm where it is then shuttled back to its proper location at the lateral cell 8  surface (Plant et al., 2003; Yamanaka et al., 2001). There are a series of similar signaling interactions that serve to maintain a mutual spatial exclusion of each of the polarity protein complexes from each other. Together, these signaling interactions act to define and maintain plasma membrane domain asymmetry, and in doing so they correctly position the apical junction complex during epithelial polarization.  A key to understanding the link between cell-cell contact, apical junction assembly and polarity complex signaling came from the finding that the Par3/Par6/aPKC complex is a direct target of the GTPases CDC42 and Rac1 (Joberty et al., 2000; Lin et al., 2000). Both CDC42 and Rac1 can be activated at the site of early primordial cell-cell junctions, and each functions to simulate local actin polymerization which generates filopodial-like projections, followed by more stable lamella during the later stages of junction maturation respectively (Perez-Moreno et al., 2003). The finding that Par6 binds directly to activated CDC42 suggests the following basic sequence of events that initates polarized junction assembly in response to cell-cell contact.  First, E-cadherin and nectin based adhesion between cells creates a lateral cell surface landmark. E-cadherin and nectin initiate apical junction formation through their interactions with the catenins, afadin and ZO-1, which have all been specifically localized to this primordial junction (Asakura et al., 1999). The transmembrane tight junction molecule Jam is also recruited to the initial cell-cell contact site where it interacts with Par3, and is likely involved in its early recruitment. Par3 has been shown to be important for tight junction assembly, and 9  may possibly function to activate local GTPases like CDC42 at the developing adhesion (Ebnet et al., 2001; Goldstein and Macara, 2007; Hirose et al., 2002). Once activated, CDC42 promotes the formation of the zipper-like structure of the primordial junction and recruits and binds directly to Par6. The Par6-CDC42 interaction creates a conformational change in Par6, which activates aPKC and recruits the Crumbs complex (Hurd et al., 2003). As the primordial cell-cell junction matures, another modular protein complex called the exocyst is also recruited to the site of adhesion (Grindstaff et al., 1998). The exocyst is a critical final mediator of polarized protein targeting to the lateral membrane. Thus, the exocyst-directed protein trafficking increases membrane delivery to the basolateral domain of the cell via targeted vesicle docking. As a result, the basolateral membrane surface is extended, which ultimately promotes the spatial separation of the apical and lateral membrane domains to form fully polarized epithelial cells (Grindstaff et al., 1998; Yeaman et al., 2004).  It is conventionally thought that cell-cell adhesion is a required external cue to drive epithelial polarization. However, initial adhesion to either a neighbouring cell or to a basal substrate is each sufficient to generate a modicum of cellular asymmetry. Similar to cell-cell contact, adhesion to ECM can on its own create a basal landmark to initiate some semblance of polarized protein trafficking, which separates nonadhesive and adhesive membrane domains (Rodriguez-Boulan et al., 1983; VegaSalas et al., 1987; Wang et al., 1990a). For example, single epithelial cells in suspension that lack both cell-cell and cell-ECM adhesion display a non-polarized distribution of apical and basolateral membrane proteins at the cell surface (Rodriguez-Boulan et al., 1983). However, single cells attached to a basal 10  substratum (in the absence of cell-cell contact) remarkably displayed polarized trafficking to distinct plasma membrane surfaces such that apically destined proteins were restricted to the free, non-adhesive membrane surface (RodriguezBoulan et al., 1983; Vega-Salas et al., 1987). Interestingly, however, basal proteins were not entirely restricted in single attached cells and instead were localized around the entire plasma membrane. This suggests that cell-ECM interactions on their own initiate a partial asymmetry and that additional polarization cues, like cell-cell adhesion for example, are required before a more complete polarization of the epithelium can be realized.  To build a tissue, the polarity of multiple cells must be coordinated in a 3dimensional (3D) space. This forces the cells to simultaneously integrate both cellcell and cell-ECM cues in order to determine the 3D spatial orientation of each membrane domain. Experimentally, the importance of identifying the co-ordinated directionality of apicobasal polarity in 3D is often overlooked, as most models of epithelial cell polarity have conventionally relied on culturing epithelial cells in 2D monolayers (Bissell et al., 1999; O'Brien et al., 2002). The disadvantage of these conventional cell culture systems is that the cells automatically orient their polarity relative to the flat, and often rigid adhesive substratum. In contrast, when epithelial cells are placed in culture surrounded by a flexible ECM, the cells interact with one another and the ECM to form a polarized 3D architecture that resemble cystic or tubular structures that are more characteristic of simple epithelial tissues in vivo. Utilizing 3D cultures, it has been demonstrated that interactions with the ECM positionally determine the spatial axis along which apicobasal polarity is 11  established (Schwimmer and Ojakian, 1995; Wang et al., 1990b). Importantly, placing the cells in this 3D microenvironmental context can dramatically alter their function that would not otherwise be observed in conventional monolayers. Much of this work was pioneered using the mammary epithelium as an experimental model. As such, the use of this 3D culture model has led to a greater understanding of the factors that regulate normal mammary epithelial structure, and how its structure becomes disorganized during breast cancer progression (This is discussed in greater detail in sections 1.3.2 and 1.4).  1.3 Mammary gland biology and epithelial morphogenesis Compared to most organs, the development of the mammary gland is unique. During embryogenesis, most tissues undergo complete differentiation and patterning that dictates their basic structure throughout adult life. Although mammary gland development is initiated in the embryo, breast tissue primarily develops post-natally and continues to change shape through adulthood. This unique feature establishes the mammary gland as a powerful and accessible developmental model system to study epithelial structure and morphogenesis.  1.3.1  Mammary gland development  Mammary gland development is fundamentally controlled by hormonal fluctuations that occur during different female reproductive stages. Thus, development occurs in distinct phases, and with each pregnancy, the mammary gland is subjected to a complex cycle of cell proliferation, tissue remodeling,  12  elongation, differentiation and apoptosis. Rodent models have emerged as primary model systems to further our understanding of mammary gland biology. Therefore, mammary gland structure and development as it occurs in the mouse will be described in some detail below (Figure 1-2).  In the mouse, there are five pairs (#1-#5) of mammary stromal fat pads located subcutaneously, and each of the glands within the fat pads connects to an external nipple. The pairs of mammary glands (left and right) are located in specific regions that follow two roughly symmetrical lines along the ventral body wall. There is one pair within the left and right cervical regions (#1), two pairs located in the thorax (#2, and 3), and one in the abdomen (#4) that connects to the gland located within the inguinal region (#5) (Richert et al., 2000). Of note, the fourth abdominal mammary gland is the largest and most easily dissected and so is most frequently used as a model for histological analysis during development or studies that involve transplantation and/or genetic manipulation.  During embryonic development, a small group of specified ectodermal cells emerge as a mammary epithelial bud and begins to invade and elongate from the nipple into the surrounding stroma of the mammary fat pad to form a rudimentary ductal structure (Robinson, 2007). This rudimentary duct is present at birth, and for the first three weeks the structure of the mammary gland remains rather quiescent, only moderately growing in proportion to the growth of the animal. At puberty however, an influx of hormones induces a growth spurt, and the mammary gland begins a dramatic change in overall structure.  13  Figure 1-2: Schematic of murine mammary gland development Left: Shown is a schematic indicating the overall tissue architectural changes that occur within the murine mammary gland during development and throughout cycles of pregnancy, lactation and involution. The light pink oval represents the stromal fat pad and the dark pink lines indicate the ductal epithelium that invades into the fat pad from the external nipple (depicted on the left side of the fat pad).  Right: The right panels show a more detailed schematic of the terminal end bud of the growing mammary ductal tree, and the mammary alveolus (contained within terminal ductal lobular units (TDLU) in humans.)  14  15  Puberty in the mouse begins around 3-4 weeks post-natally, and is accompanied by a dramatic increase in circulating estrogen which provides indispensable signals for ductal outgrowth, since mammary glands of estrogen receptor α knockout females fail to develop past the rudimentary glandular structure (Couse et al., 1995; Korach et al., 1996). During puberty, the mammary ducts undergo numerous cycles of branching and elongation that ultimately extend and fill the entire mammary fat pad by approximately 10-12 weeks of age (Richert et al., 2000).  The ductal growth and branching morphogenesis that occurs during puberty is driven by the terminal end bud (TEB), a group of highly proliferative cells that support the outward growth and movement of the growing duct. The TEB’s are located at the tip of growing ducts and consist of multiple cell layers. The outer layer is made up of highly proliferative, pluripotent cap cells which, as the duct grows, gives rise to multiple layers of inner body cells (Howlin et al., 2006). During ductal elongation, the trailing edge of the cap cell layer differentiates into myoepithelial cells while the inner body cells closest to the cap cells give rise to the inner luminal epithelial cells (Visvader and Lindeman, 2006). The remaining innermost body cells undergo apoptosis to form a hollow lumen inside the developing duct. Ductal branching occurs either at lateral points along the growing duct, or by bifurcation of TEB’s until the ductal tree reaches the edge of the mammary fat pad signifying the end of the pubertal stage (Fata et al., 2004). Once this point is reached, the TEB’s regress and the tips of the ductal tree form alveolar buds. At each estrous cycle throughout adulthood, the secretion of ovarian hormones initiates the development of additional alveolar buds to form rudimentary adult alveoli (Hennighausen and Robinson, 1998). These terminal 16  structures are the functional units capable of differentiating into milk secreting alveoli during pregnancy. It is worth noting that the alveolar buds in the mouse are functionally equivalent to the structures in humans known as the terminal ductal lobular units (TDLU) (Hovey and Trott, 2004). However, there are subtle structural differences. Specifically, the TDLU are larger, more sophisticated structures that consist of several alveoli clustered around a single terminal duct with smaller laterally branched ductules.  In the resting adult mammary gland, the lumens of the terminal duct and the alveoli are lined by a single continuous layer of polarized luminal epithelial cells. Beneath these luminal epithelial cells lies a discontinuous layer of myoepithelial cells. Because of this discontinuity, both the luminal and myoepithelial cells rest upon a basement membrane that separates the ductal epithelium from the surrounding stroma, which is largely filled with large adipocytes interspersed with stromal fibroblasts (Richert et al., 2000). This thin, specialized basement membrane layer is crucial for orchestrating the architecture of the developed mammary gland (as discussed below in section 1.3.2).  Within the mammary epithelium, not all of the cells appear to be fully differentiated into either luminal or myoepithelial lineages. Accumulating evidence suggests the presence of mammary progenitor cells, which are thought to be basally positioned within the luminal epithelial cell layer. Indeed, rare populations of mouse mammary stem cells have been recently isolated whereby a single isolated cell was capable of reconstituting an entire mammary gland comprising all mature epithelial cell types, and was also capable of self-renewal upon serial 17  transplantation (Shackleton et al., 2006; Stingl et al., 2006). One hypothesis is that the presence of bipotent progenitors and/or multipotent stem cells supports the dramatic proliferative bursts and tissue renewal that occur throughout mammary development, particularly through cycles of pregnancy.  During the first stage of pregnancy, the mammary epithelium undergoes massive proliferation and differentiation in response to increased circulating hormones such as progesterone (Lydon et al., 1995). This is associated with an increase in lateral ductal branching and the development of new alveolar buds, and together this expanded ductal tree displaces the surrounding adipocytic stroma. At this stage, the alvelor buds differentiate into functional alveoli, a process commonly referred to as lobuloaveolar differentation (Hennighausen and Robinson, 1998). By late pregnancy, the lobuloalveloar compartment fills the entire gland and the differentiated alveoli are capable of milk production in response to lactogenic hormones such as prolactin (Ormandy et al., 1997).  During lactation, the luminal epithelial cells change shape, from cuboidal to very flat, as the inner lumens fill with milk secretions and become distended. The surrounding myoepithelial cells are instructed to contract, which functions to squeeze the mammary ducts and expel milk along the ductal tree to the external nipple. In the post-lactational female, the mammary gland is dramatically remodeled, a process termed involution. During involution, the secretory alveolar epithelium collapses and the cells undergo apoptosis. Importantly, at this stage the mammary stromal compartment re-emerges (including the reappearance of adipocytes and basement membrane remodeling), which is instrumental in re18  arranging the architecture of the ductal epithelium. The remodeling of both the epithelial and stromal compartments continues until the entire gland closely resembles the dormant pre-pregnant adult mammary gland.  1.3.2  Regulating mammary gland architecture: Insights from 3D in vitro models of mammary morphogenesis  Regulation of mammary gland development and the structural organization of the mammary epithelium critically depend on instructive cues from the local mammary microenvironment. A prominent example is the signaling provided to the mammary epithelial cells by the underlying basement membrane (BM).  The importance of the basement membrane became apparent in ex vivo cultures of normal mammary epithelial cells from pre-lactating mice. When preparations of primary mammary epithelia were separated from the BM and placed in a tissue culture dish, the cells lost their glandular architecture, became flattened and failed to respond to lactogenic hormones to produce and secrete milk proteins. Seminal studies showed that if, instead, these primary mammary epithelial cells were placed on floating collagen gels, this flexible microenvironment stimulated the cells to deposit their own basement membrane. The cells then organized into 3D polarized spheroids that closely resembled normal mammary acini (Emerman and Pitelka, 1977; Streuli and Bissell, 1990).  Since the intital findings using floating collagen gels over 30 years ago, studies using both primary cells as well as numerous cell lines derived from the normal 19  mammary epithelium have demonstrated that the basement membrane component laminin is a critical instructive cue that regulates the morphological changes required to create polarized acini (Barcellos-Hoff et al., 1989) (Aggeler et al., 1991). Additionally, mammary epithelial cells maintained in this 3D microenvironment retain their functional capacity to differentiate, as they are able to synthesize milk proteins in response to lactogenic hormones and secrete them in a polarized fashion (Barcellos-Hoff et al., 1989; Streuli et al., 1995). In fact, adhesion-dependent signaling downstream of laminin-β1 integrin engagement directly coordinates with prolactin receptor signaling and Stat5 activation to promote the transcription of milk protein encoding genes (Faraldo et al., 2002; Streuli et al., 1991; Streuli et al., 1995). Importantly, this was the first demonstration that tissue-specific gene expression could be induced by the ECM (Streuli et al., 1991).  The mechanisms downstream of laminin that dictate normal mammary architecture and apicobasal polarity are not yet well defined. It seems clear, however, that it likely requires multiple laminin receptors (Muschler et al., 1999). Loss-of-function studies targeting dystroglycan displayed defects in the ability of normal mammary epithelial cells to form polarized acini (Weir et al., 2006). This is likely due to dystroglycan’s ability to cluster laminin molecules at the extracellular surface. Also, when non-functional mutants of β4 integrin that lack the ability to from hemidesmosomes were expressed in mammary epithelial cells, they failed to form properly organized acini, and were associated with an increased sensitivity to apoptotic stimuli (Weaver et al., 2002). This is in keeping with the notion that loss of anchorage from the basement membrane triggers apoptosis in normal mammary  20  epithelia (Boudreau et al., 1995). Likewise, targeted deletion of β1 integrin in the mouse mammary gland caused the basal surface of luminal epithelial cells to detach from the basement membrane, which caused a decrease in cell survival and contributed to an overall proliferation defect during mammary gland development (Li et al., 2005). Intriguingly, these studies provide a link between β1 integrins and the control of proliferation, which could have important implications in breast cancer where loss of anchorage dependence results in multi-layering and luminal filling of the glandular epithelium (Debnath and Brugge, 2005).  1.4 Breast cancer: Derailing tissue architecture Breast cancer is a heterogeneous disease, both genotypically and phenotypically, which makes classifying the many different types of tumors difficult. Traditional pathological classification schemes typically utilize various clinical and histopathological features such as tumor size, cellular arrangement, nuclear pleomorphism, and the degree of architectural disorder within the tissue (VargoGogola and Rosen, 2007). Among the incredible diversity found in breast cancer, two major forms can be easily distinguished and are referred to as “ductal” or “lobular.” Based on histopathological criteria, invasive ductal carcinoma (IDC) is the most commonly diagnosed breast cancer. In contrast, lobular breast cancer is an aggressive, but more rare subtype that can be easily distinguished from ductal tumors as it more frequently loses expression of E-cadherin and a hallmark of its progression involves the single-file invasion of E-cadherin negative tumor cells into the surrounding stroma (Cleton-Jansen, 2002; Yoder et al., 2007).  21  More recently, the phenotypic diversity seen in breast cancer has been mapped to specific gene expression profiles enabling a new type of “molecular” classification scheme for breast cancer (Perou et al., 2000). Remarkably, this identified five reproducible subtypes: luminal A, luminal B, ErbB2 (HER2/neu), basal, and normal breast-like. Together with traditional pathology, this molecular classification scheme is a useful tool to determine patient prognosis. Regardless of the new and improved ways to define breast cancer, often the most reliable prognostic marker to predict metastatic breast cancer is high tumor grade, which is defined by a high degree of disrupted tissue architecture (Vargo-Gogola and Rosen, 2007).  1.4.1  Disruption of mammary gland architecture and epithelial cell polarity during breast cancer progression  It is thought that the well-differentiated, polarized organization of the mammary gland collectively functions as a tumor suppressor. This correlates with epidemiological evidence, which suggests that the increased differentiation in the mammary gland that occurs as a result of pregnancy may be protective against breast cancer (Russo et al., 2005). Importantly, when breast tumor cells are compared to normal mammary epithelial cells in 3D culture conditions, the breast tumor cells can be phenotypically and behaviourly distinguished from their normal counterparts as they fail to form organized, polar 3D acini (Petersen et al., 1992). In contrast, the tumor cell-generated structures display a chronically disrupted polarity, fail to respond to the growth suppressive cues of the ECM, and thus continue to proliferate into large, disordered tumor cell clusters (Petersen et al., 1992). This suggests that a chronic disruption of cell polarity may be a fundamental difference that distinguishes normal mammary epithelial cells from breast tumor 22  cells. In support of this, many different cues and molecular determinants of apicobasal polarity are disrupted in breast cancer. Furthermore, there is increasing evidence to suggest that specific defects in polarity pathways actively intersect with growth factor signaling pathways in breast tumor cells. Conversely, there is also evidence that oncogenic growth factor signaling itself can directly target polarity modulators to cooperatively promote tumor progression.  1.4.1.1 Disruption of external polarity cues in breast cancer: Cell-cell junctions E-cadherin mediated adhesions are a crucial landmark to set up epithelial cell polarity, and loss of E-cadherin can cause morphological changes to the epithelium that resemble an epithelial to mesenchymal transformation (EMT). Inappropriate EMT is thought to be a key mechanism that contributes to invasion and metastasis of many solid tumors (Birchmeier and Birchmeier, 1995). Mutational inactivation of the E-cadherin gene, CDH1, is a frequent event in lobular forms of breast cancer (Berx et al., 1995). Also experimental evidence suggests that indirect repression of Ecadherin, either by matrix metalloproteinase digestion of its extracellular domain (Lochter et al., 1997) or by upregulation of its transcriptional repressors (Yang et al., 2004), can induce an EMT and promote experimental breast tumor metastasis. However, it is important to consider that E-cadherin loss is primarily affiliated with lobular breast carcinomas which only account for ~5-10% of diagnosed breast tumors. In contrast, the large majority of breast tumors, most of which are ductal, have altered tissue polarity, but retain E-cadherin expression. Thus, alternative  23  mechanisms are likely important to disrupt epithelial polarity and architecture in the majority of breast tumors.  In ductal breast cancer, the levels of tight junction molecules are altered and could contribute to breast tumor architectural disruption. Specifically, reduced levels of both ZO-1 and Claudin-1 have been reported in numerous breast tumor cell lines, and the loss of ZO-1 in primary tumors correlated with decreased glandular differentiation (Hoover et al., 1998; Kramer et al., 2000). Also, it has been shown that claudin-7 is reduced in both invasive lobular and invasive ductal carcinoma, and that this loss is highly correlated with the disorganized architecture typical of highgrade lesions (Kominsky et al., 2003). Interestingly, loss of claudin-7 was also associated with high-grade ductal carcinoma in situ (DCIS) when compared to low grade DCIS, where its expression was found to be comparable to adjacent normal mammary tissue. This suggests that deregulation of tight junctions may contribute to the tissue architectural disruption of cohesive tumor cells before they become locally invasive. However, direct experimental evidence regarding the true functional significance of decreased tight junction molecules in breast cancer is currently lacking.  1.4.1.2 Disruption of external polarity cues in breast cancer: Cell-ECM junctions The surrounding stroma is a key factor in dictating normal mammary epithelial architecture, thus it is not surprising that there are multiple ways in which breast tumor cell-ECM interactions are disrupted in breast cancer. For example, the 24  activity and/or expression of integrins are frequently altered in breast cancer (Hood and Cheresh, 2002). Insight into how modulating integrins affects the breast tumor phenotype has come from 3D culture models, which aim to either rescue polarity defects in tumor cells and force them to resemble normal mammary acini, or to manipulate the normal cells to induce the tumorigenic architecture.  Seminal studies from Mina Bissell and colleagues have focused on interrogating the different properties between a non-malignant and malignant mammary cell line that were each derived from the same parental cell population (Reviewed in (Nelson and Bissell, 2006). Morphologically, the non-malignant cells form small, polarized acini in 3D culture similar to normal mammary epithelial cells. In contrast, the malignant cells form highly disorganized non-polar clusters and hyperproliferate, which resemble primary breast tumor cells in culture (Petersen et al., 1992). A key difference found between these two cell types is that the malignant variant showed significantly higher β1 integrin and EGFR expression (Wang et al., 1998; Weaver et al., 1997). Functional blocking of β1 integrin in the malignant cells reverted their phenotype in 3D culture causing apicobasal polarity to be restored. Remarkably, β1 inhibition also blocked the hyperproliferation of these tumor cells by downmodulating EGFR signaling activity, an effect that occurred in 3D, but not 2D, cultures (Wang et al., 1998; Weaver et al., 1997). This suggests that integrins are under strict control in normal cells and that in the right microenvironmental context, they cooperate with growth factor signaling to modulate developmental decisions whether to proliferate or undergo polarized morphogenesis. However, chronic, uncontrolled β1 integrin signaling in breast tumor cells appears to override  25  the decision to polarize which causes architectural disorder that perpetually converges with growth factor signaling to promote hyperproliferation. Importantly, when this chronic β1 integrin signaling is switched off, this was sufficient to control proliferation, restore tissue morphogenesis, and reduce tumorigenesis (Weaver et al., 1997).  One mechanism that can lead to hyperactive β1 integrin signaling in breast tumors is alterations within the stroma itself. The composition of breast tumor stroma is different compared to its normal counterpart. For example, tumor associated myoepithelial cells lose their ability to produce laminin 1 (Liu et al., 2005), and the ECM surrounding the tumor can be persistently degraded and remodeled due to the aberrant expression of matrix metalloproteinases. Overall, the tumor stroma is more rigid compared to the stroma of the normal breast (Paszek et al., 2005). Interestingly, increased ECM rigidity causes non-malignant mammary cells to become contractile, which disrupts polarized morphology, clusters integrins at the cell surface and enhances EGFR signaling pathways (Paszek et al., 2005). Thus, alterations in the tumor stroma itself can be an upstream factor that influences the integrin-driven alterations in polarity and proliferation of breast tumor cells by triggering contractile cell tension.  From these studies, it is clear that alterations in polarity disruption and hyperproliferation are linked in breast tumor cells. An important question is whether the alterations in polarity are just merely a consequence of hyperproliferation. Bissell and colleagues have shown that, similar to the  26  phenotypic reversion observed with blocking β1 integrin, attenuated PI3K signaling also restored tissue polarity and growth control in the malignant cells (Liu et al., 2004). While PI3K mediates signaling downstream of both β1 integrin and the EGFR, these authors demonstrated that separate effector pathways downstream of PI3K control hyperproliferation and polarity independently. Specifically, Akt activation is responsible for increased proliferation while Rac1 activation independently causes the polarity defects and together these effects promote the emergence of the disorganized and hyperproliferative tumor phenotype. Interestingly, Rac1 is directly involved in driving apicobasal polarity via its association with the Par3/Par6/aPKC complex that is critical for the assembly of the apical junction complex (Lin et al., 2000). This provides evidence that alterations in cell polarity are not just a consequence of breast tumor hyperplasia, but instead contribute to breast cancer progression.  The notion that the disruption of polarity may actively promote breast tumor progression is strengthened by recent data uncovering the tumorigeneic phenotype associated with HER2/neu overexpression. It has been shown that chronic activation of HER2/neu homodimers in normal mammary epithelial cells induces hyperproliferation and disrupts the polarized architecture of 3D acini to form abnormal, multilobular structures that resemble non-invasive lesions (Muthuswamy et al., 2001). Hyperactivation of HER2/neu directly targets and disrupts members of the Par3/Par6/aPKC polarity complex. As a result, an aberrant protein complex forms between HER2/neu and Par6/aPKC, which leads to the generation of disorganized, multilobular structures. However, if the  27  Par6/aPKC interaction is lost, epithelial cell polarity is restored, but hyperproliferation and luminal filing of the acini continues (Aranda et al., 2006). This indicates that oncogenic HER2/neu signaling can actively dysregulate polarity regulatory proteins in a proliferation-independent pathway to modulate glandular organization of mammary epithelial cells during tumorigenesis (Walker and Brugge, 2006).  The data described above suggests that many different regulators of epithelial polarity that execute changes in epithelial morphogenesis could be targeted during breast tumorigenesis with functional consequences. Dysregulation of such polarity regulators could provide a molecular basis that contributes to high tumor grade, and could represent novel markers of aggressive disease. In this thesis I examined the functional role of podocalyxin, an ‘apical polarity determinant’ in normal mammary morphogenesis and breast tumor progression.  1.5 Podocalyxin: A potential regulator of epithelial morphogenesis In 1984, Marilyn Farquhar’s research group identified podocalyxin as the major molecular component that makes up the cell surface glycocalyx that coats podocytes within the kidney glomerulus (podocyte glycocalyx protein; Kerjaschki et al., 1984). Both biochemical and genetic studies have determined that podocalyxin acts as a critical regulator of the specialized architecture characteristic to differentiated podocytes (Doyonnas et al., 2001). It is now recognized that podocalyxin is not exclusively expressed in the kidney glomerulus, but is expressed on a number of specialized cell types, and accumulating evidence suggests that it 28  may play a more generalized role in orchestrating cell architecture, adhesion, polarity and morphogenesis. Importantly, podocalyxin overexpression has been observed in a number of aggressive tumor types including breast, prostate, and leukemia among others, but its functional contribution to cancer progression is poorly understood.  1.5.1  Basic overview of podocalyxin  Podocalyxin is a cell surface molecule closely related to the CD34 family of proteins based on their similar amino acid sequences, protein structure and genomic organization. Each protein within this gene family, which includes podocalyxin, CD34 and endoglycan, is encoded from a gene containing eight exons, and individual exons correspondingly encode strikingly similar protein motifs (Nielsen et al., 2002; Figure 1-3).  Like CD34 and endoglycan, podocalyxin is a single-pass transmembrane sialomucin. The N-terminal extracellular region of podocalyxin contains a large mucin domain (~250 a.a.), a small globular domain consisting of two paired cysteine residues, and a juxtamembrane stalk region (Kershaw et al., 1997a; McNagny et al., 1997). Based on amino acid sequence, podocalyxin’s core protein backbone has a predicted mass of approximately 50-55kDa. However, the mucin domain contains several putative sites of N-linked glycosylation, and is rich in serine/threonine residues that undergo extensive O-linked glycosylation and sialylation during its biosynthetic processing (Kershaw et al., 1997a; McNagny et al., 1997; Takeda et al., 2000). These extensive post-translational modifications cause the  29  Figure 1-3: Genomic organization of CD34 family members All three CD34 family members display similar genomic structure and encode for strikingly similar protein motifs: Purple: Blue: Green: Yellow: Orange: Red:  Signal peptide Mucin domain Globular motif Extracellular stalk domain Transmembrane domain Cytoplasmic tail  The numbers refer to intron sizes in kilobase pairs. Reproduced with kind permission of S. Karger AG, Basel (Nielsen et al., 2002).  30  31  apparent protein mass of podocalyxin to range between 140kDa-200kDa, and renders the extracellular region to be bulky and highly negatively charged, which is an important characteristic common to most sialomucins expressed on cell surfaces.  The extracellular domain of podocalyxin, while structurally similar, has very little sequence homology across species. In contrast, the cytoplasmic tail of podocalyxin is highly conserved with ~95% a.a. identity between rat, rabbit and human (Kershaw et al., 1997a; Li et al., 2002; McNagny et al., 1997; Miettinen et al., 1999). This 75 a.a. intracellular domain contains putative phosphorylation sites for protein kinase C (PKC) and casein kinase II (CKII) as well as a C-terminal PDZ-binding motif consisting of the sequence DTHL (Asp-Thr-His-Leu) (Doyonnas et al., 2001; Kershaw et al., 1997a; McNagny et al., 1997). Also, the cytoplasmic domain of podocalyxin can be anchored to the actin cytoskeleton through ezrin. Ezrin is a member of the ezrin-radixin-moesin (ERM) family of actin binding proteins, and can either directly bind to an intracellular juxtamembrane region of podocalyxin (Nielsen et al., 2007; Schmieder et al., 2004), or it can indirectly bind through PDZ domain-containing NHERF scaffolding proteins that have been shown to interact with podocalyxin’s C-terminal PDZ binding motif (Li et al., 2002; Orlando et al., 2001; Takeda et al., 2001; Tan et al., 2006); (Figure 1-4; and discussed in more detail in Section 1.5.5).  32  Figure 1-4: Schematic of the podocalyxin/NHERF-1/ezrin protein complex The schematic of podocalyxin structure was modified from Nielsen et al., 2002. The blue regions represent the extracellular mucin domain. Within this domain, the black circles are potential N-linked carbohydrates, the horizontal lines are potential O-linked carbohydrates, and the triangles are potential sialic acid residues. The green region is the paired cysteine globular domain, the yellow region is the stalk, orange is the transmembrane region and red is the highly conserved cytoplasmic tail. Ezrin can bind either directly to podocalyxin via a juxtamembrane ezrin binding site, or through NHERF proteins. The C-terminal DTHL sequence preferentially binds to the second PDZ domain of NHERF, and ezrin interacts with its carboxy ERM binding motif which links podocalyxin to the actin cytoskeleton.  33  34  1.5.2  Tissue profile of podocalyxin expression  Podocalyxin is highly expressed in many diverse cell types and it is typically localized at their membrane surfaces. Recent experiments in mice have shown that podocalyxin is expressed in all three germ layers during embryogenesis (Doyonnas et al., 2005). It is found in the first intraembryonic mesoderm, at the apical face of the neuroepithelium in the primitive ectoderm, and by the primitive endoderm surrounding the embryo. Later in fetal development, podocalyxin is highly expressed along the luminal linings of the vasculature and the mesothelium that lines the inner surface of body cavities in both mice and avians (Doyonnas et al., 2005; McNagny et al., 1997). Podocalyxin is also a surface marker of the earliest detectable hematopoietic progenitors (Doyonnas et al., 2005; McNagny et al., 1997), and hemangioblasts (Hara et al., 1999), which are the cells that give rise to both hematopoietc cells and endothelial cells. However, in more differentiated hematopoietic cells, podocalyxin is only expressed on platelets, their precursors, megakaryocytes, and by erythroid cells under anemic conditions (Doyonnas et al., 2005; Miettinen et al., 1999).  Podocalyxin was first identified over twenty years ago as the most abundant glycoprotein present on the surface of kidney podocytes (Kerjaschki et al., 1984). Around the same time, podocalyxin was also detected lining the surface of vascular endothelia both in the kidney (Sawada et al., 1986), and in several extra-renal tissues (Horvat et al., 1986). More recently it has been shown that podocalyxin is expressed in a subset of neurons both during brain development and throughout adulthood (Vitureira et al., 2005). In the adult, podocalyxin expression is found on 35  the luminal surface of mesothelial cells that cover the surface of various internal organs (Doyonnas et al., 2001; Doyonnas et al., 2005; McNagny et al., 1997), and it is also localized at the free surface of the epithelium that covers the surface of the ovary (Cipollone et al., 2006), which itself is a modified mesothelium. Interestingly, the overall level of podocalyxin transcript is reported to be substantially higher in murine newborn ovaries compared to ovaries from adult mice (Herrera et al., 2005). However, it is not known which cell types in the newborn ovary display increased podocalyxin expression as the localization of podocalyxin transcript or protein has not been determined. Thus the significance of this higher overall expression during ovarian development is not known. Additionally, podocalyxin is expressed in other reproductive tissues, and localizes to the apical surface of the Mullerian ductderived epithelium of the oviducts and endometrium (Cipollone et al., 2006) and is also expressed at the apical surface of luminal mammary epithelial cells (Somasiri et al., 2004; Chapter 4).  1.5.3  Regulation of podocalyxin expression  Although the podocalyxin gene has been cloned in multiple species including human (Kershaw et al., 1997a), mouse (Hara et al., 1999), rabbit (Kershaw et al., 1995), rat (Miettinen et al., 1999; Takeda et al., 2000), dog (Cheng et al., 2005; Meder et al., 2005) and chicken (McNagny et al., 1997), very little is known about the regulatory elements and upstream factors that control its gene transcription. The human podocalyxin gene (Podxl) has been mapped to chromosome 7q32-q33 (Kershaw et al., 1997b), and putative upstream regulatory regions have been isolated (Kershaw et al., 1997a; Palmer et al., 2001). Sequence analysis of the 5’ 36  regulatory region (1297 bp upstream of the transcriptional start site and an additional 231 bp of 5’ untranslated region) have identified potential binding sites for the transcription factors AP2 and NFκB as well as numerous SP1 sites, three GATA-1 sites, and three WT-1 sites (Butta et al., 2006). Using reporter analysis, these authors showed that there is maximal promoter activity between base pairs 364 and -111 relative to the transcriptional start site, and the transcription factor SP1 interacts with this region and is critical for its activation. Interestingly, sequential inclusion of the remainder of the upstream 5’ flanking sequence up to -1297 caused varying levels of promoter inhibition in three different cell lines tested. Other features described for the podocalyxin promoter include the lack of TATA or CAAT boxes and a high CG content (CpG islands), which along with control by SP1 binding elements, are all characteristics of housekeeping genes (Butta et al., 2006). Thus, it has been suggested that SP1 may provide a constitutive level of expression and that additional tissue specific regulation may contribute to restricting the pattern of podocalyxin expression, which could possibly be mediated by the upstream regions and/or variations in methylation of the abundant CpG islands present in the Podxl promoter.  Indeed, it is likely that the normal regulation of podocalyxin transcription and expression is tissue specific. For example, several studies have implicated WT1 in regulating the expression of podocalyxin. While WT1 is expressed in many tissues during mammalian embryonic development, its expression in the adult is mainly restricted to the kidney (Roberts, 2005). Thus WT1 presumably plays a larger role in regulating normal podocalyxin expression in the glomerular epithelium compared to the other tissues that express podocalyxin. In support of this, the expression of 37  WT1 in the developing kidney temporally and spatially coincides with the induction of podocalyxin expression in vivo (Armstrong et al., 1993; Palmer et al., 2001). Also, WT1 physically binds the podocalyxin promoter in vitro, and the inducible expression of WT1 in rat embryonic kidney cells leads to a dramatic increase in Podxl promoter activity and podocalyxin transcript (Palmer et al., 2001; Stanhope-Baker et al., 2004). Intriguingly, WT1 is often highly upregulated in adult tumors originating from many different tissues that do not normally express WT1 including colorectal (Oji et al., 2003), breast (Loeb et al., 2001; Silberstein et al., 1997), and brain tumors (Dennis et al., 2002; Oji et al., 2004). Thus, abnormal WT1 expression may induce tumor-associated increases in podocalyxin expression, but this possibility has not yet been formally tested.  It has also been reported that podocalyxin expression is negatively regulated by p53. cDNA microarray analysis of cells with inactive p53 compared to those with restored p53 function identified podocalyxin as a downstream target of functional p53 (Stanhope-Baker et al., 2004). Reporter assays and quantitative RT-PCR confirmed that restoration of p53 caused a decrease in Podxl promoter activity and podocalyxin transcript levels (Stanhope-Baker et al., 2004). Since inactivating mutations of p53 occur at a very high frequency in multiple tumor types, this could lead to inappropriate upregulation of podocalyxin expression in tumors associated with the loss of p53. However, direct interaction between the Podxl promoter and p53 has not been reported. Thus further studies are required to understand the mechanisms underlying p53-dependent repression of podocalyxin expression.  Using a gene array-based screen, it was recently discovered that podocalyxin 38  expression is highly upregulated in response to erythropoietin (Epo) in erythroblast progenitors (Sathyanarayana et al., 2007). Although the exact mechanisms are unclear, under these conditions, podocalyxin upregulation is mediated through cytokine signaling downstream of the transmembrane Epo receptor and requires phosphorylated Stat5 at Y343. Sequence analysis verified the presence of predicted Stat binding elements in the murine Podxl promoter, although direct Stat5 binding has not been shown. Stat5-dependent regulation of podocalyxin expression could be of particular importance in the mammary gland given that Stat5 activation translates numerous growth factor and hormone signals, and depending on reproductive status, can regulate mammary epithelial cell proliferation, alveolar differentiation and milk-protein production (Barash, 2006). Also, Stat5 promotes mammary tumorigenesis in some contexts, as hemizygous loss of Stat5 in the SV40T antigen transgenic model of breast cancer causes a significant reduction in the percentage of tumor-bearing mice, as well as decreased tumor size, delayed first tumor appearance, and increased apoptotic indices (Ren et al., 2002). Altogether, aberrant expression of WT-1, increased Stat5 activity, and the loss of p53 could lead to upregulated podocalyxin in the mammary epithelium in a pathologic setting.  1.5.4  Insights into podocalyxin function: Kidney morphogenesis and podocalyxin knockout analysis  To date, the majority of the proposed functions of podocalyxin have emerged from its role during renal development and the study of nephrotic disease models. Early in glomerular development, podocytes are undifferentiated and resemble a typical epithelium with appropriate junctions at the lateral membrane between cells and 39  polarized tight junction complexes located beneath the apical membrane. Interestingly, when podocalyxin expression is initiated in these cells, this coincides with a radical change in podocyte cell morphology (Schnabel et al., 1989). Podocalyxin is targeted to the apical surface, and the cells begin to extend a dramatic network of surface membrane projections (called foot processes) that closely interdigitate with the foot processes from neighbouring podocytes. During this process, initial adherens and tight junctions between cells remodel to form a specialized junction (called the slit diaphragm) at the basal surface which bridge two adjacent foot processes. In mature podocytes, podocalyxin decorates the entire apical surface of the foot process membrane with its highly glycosylated and negatively charged extracellular domain located within the open urinary space (Sawada et al., 1986).  Although podocalyxin was not discovered until 1984, it has long been recognized that a sialic-acid rich glycocalyx coats the apical surface of the glomerular epithelial cells (Michael et al., 1970). Early in vivo functional studies that involved chemical neutralization of this “epithelial polyanion” severely disrupted podocyte architecture. As a result, podocyte foot processes collapsed, and adjacent cells fused together with typical junctions between them. As a result, the overall appearance of charge-neutralized podocytes resemble the undifferentiated epithelium early in development (Kerjaschki, 1978; Seiler et al., 1977). It was later determined that podocalyxin resembled the histochemically characterized “epithelial polyanion” and was likely the key molecular component of the negatively charged glomerular glycocalyx (Kerjaschki et al., 1984). This suggested that disruption of podocalyxin itself could directly cause the architectural defects in the glomerlur epithelium. 40  Indeed, deletion of the podocalyxin gene in mice confirms this hypothesis, since kidneys from these mice also fail to form the characteristic architecture of mature podocytes (Doyonnas et al., 2001). Similar to the neutralization studies, the glomerular epithelium in podocalyxin knockout mice do not form foot processes, but instead both adherens junctions and tight junctions are maintained between cells and the podocyte basal membrane remains in firm contact with the glomerular basement membrane (Doyonnas et al., 2001). Therefore, this suggests that podocalyxin functions as the key molecular regulator of podocyte architecture during kidney morphogenesis.  1.5.5  Controlling epithelial morphogenesis: Proposed functions of podocalyxin as an anti-adhesin and regulator of apicobasal polarity  Genetic loss of function studies in mice determined that podocalyxin is essential for podocyte morphogenesis, and indicates that podocalyxin could alter multiple aspects of cell adhesion, including both cell-cell and cell-matrix interactions. (Doyonnas et al., 2001). Further evidence and mechanistic details of podocalyxin’s function as an anti-adhesin have come from in vitro studies.  The expression of podocalyxin can be modulated in human glomerular epithelial cell lines (HGEC) by altering the level of glucose in the culture media (Economou et al., 2004). High glucose levels significantly reduce podocalyxin expression, which likely contributes to the frequent foot process effacement that occurs in diabetic nephropathy (Kanwar et al., 1996; Kanwar et al., 1997). Interestingly, HGEC expressing physiologically normal levels of podocalyxin showed a decreased ability 41  to adhere to ECM components compared to podocalyxin-depleted cells cultured in low glucose conditions. Also, this low-level adhesion to both collagen IV and laminin could be increased when podocalyxin was functionally blocked using an inhibitory antibody (Economou et al., 2004). This supports the hypothesis that podocalyxin inhibits glomerular epithelial cell-matrix interactions. This implies that in vivo, podocalyxin’s anti-adhesiveness could contribute to the functionally important arrangement of discontinuous adhesions between the podocyte basal surface and the underlying glomerular basement membrane (Economou et al., 2004).  Direct functional evidence to support podocalyxin’s ability to block cell-cell adhesion first came from gain of function studies done in Marilyn Farquhar’s lab (Takeda et al., 2000). Ectopic expression of podocalyxin in MDCK kidney epithelial cells caused an expression level-dependent decrease in cell-cell aggregation. The observed inhibition in cell-cell aggregation was reversed by sialidase treatment, which supports that negative charge repulsion by podocalyxin’s extracellular domain contributes to its anti-adhesive characteristics. The authors also suggested that podocalyxin overexpression decreases the integrity of tight junctions between cells, as they observed a decrease in transepithelial resistance (TER) and a subtle modulation of cell junction molecule localization in monolayer culture. This provided the first gain of function evidence that supported podocalyxin’s function as an anti-adhesive molecule that possibly “repels” adjacent cell membranes, which could explain its role in maintaining open slit diaphragms between podocytes.  42  The physicochemical properties of the extracellular domain clearly seems to play a role in the anti-adhesive nature of podocalyxin. (Dekan et al., 1991; Takeda et al., 2000). However, the function of the highly conserved intracellular domain was not yet known. Marilyn Farquhar and colleagues postulated that the cytoplasmic tail of podocalyxin might also contribute to the unique organization of podocytes by interacting with the actin cytoskeleton. They found that the c-terminus of podocalyxin directly interacts with NHERF-2, which interacts with ezrin (Orlando et al., 2001; Takeda et al., 2001) Ezrin is a member of the ERM family of adaptors that are known to bind actin filaments and link membrane proteins to the cytoskeletal network (For a review see Bretscher et al., 2000). Importantly, they showed that the integrity of the podocalyxin/NHERF/ezrin complex is disrupted in nephrotic disease models where the architecture of podocytes is dramatically altered (Takeda et al., 2001). This suggests that, in addition to the biophysical properties of its extracellular domain, podocalyxin’s linkage to the actin cytoskeleton through its cytoplasmic domain may also be functionally important in controlling podocyte morphology.  It has been recently shown that podocalyxin might play a more general role in controlling epithelial cell morphology that may have important functional implications beyond kidney development. Podocalyxin was recently identified as the human orthologue of canine gp135 (Meder et al., 2005), the prototypical apical marker of MDCK epithelial cells (Ojakian and Schwimmer, 1988). These authors observed that, unlike most membrane proteins that are typically restricted to a particular membrane domain only after cell-cell junctions are formed, podocalyxin exclusively localizes to the free surface in single cells attached to an ECM 43  substratum (Meder et al., 2005). This implies that free surface podocalyxin localization signifies an early stage of apicobasal polarity as it helps define the membrane domain that is destined to become the apical surface of fully polarized epithelia. Additionally, Meder et al. (2005) also localized NHERF-2 to this “preapical” domain in single cells. They proposed that the podocalyxin/NHERF-2 interaction, which is likely connected to the actin cytoskeleton via ezrin, may contribute to the formation of an early apical scaffold that acts as either a signaling platform and/or a trafficking docking site to promote epithelial polarization.  In support of this, a stable knockdown of podocalyxin caused a “flattening” of MDCK monolayers that appeared to lack a defined apical membrane domain (Meder et al., 2005). Furthermore, in 3D spheroid culture MDCK cells with depleted podocalyxin displayed defects in apical lumen formation (Meder et al., 2005) and HGF-dependent tubulogenesis, which was dependent on its cytoplasmic domain (Cheng et al., 2005). Taken together, these findings suggest that podocalyxin and its ability to link to the actin cytoskeleton is potentially part of a general mechanism that defines the apical membrane domain during epithelial morphogenesis, and as a consequence, it may be an important regulator of de novo lumen formation.  It was suggested that NHERF-2 might participate with podocalyxin in defining the “pre-apical” membrane domain in polarizing epithelia (Meder et al., 2005). While the expression of NHERF-2 is mainly restricted to the kidney glomerulus (Orlando et al., 2001), another NHERF family member, NHERF-1, is more widely distributed in extrarenal tissues. Like NHERF-2, we and others have demonstrated that NHERF-1 is also a bona fide binding partner for the C-terminal tail of podocalyxin 44  (Li et al., 2002; Takeda et al., 2001; Tan et al., 2006). Thus, a podocalyxin/NHERF-1 interaction is likely more relevant in regulating epithelial morphogenesis in more diverse tissue types. For example, NHERF-1 is expressed in the mammary gland and expression of both podocalyxin and NHERF-1 are upregulated in response to estrogen levels (Ediger et al., 1999; Ediger et al., 2002; Ginger et al., 2001). This supports the notion that the podocalyxin/NHERF-1/ezrin complex may be relevant in regulating epithelial architecture and morphogenesis in the mammary gland particularly during estrogen dependent proliferation that is critical for ductal elongation and branching during puberty.  The interaction between podocalyxin and NHERF-1 could have additional functional implications. NHERF-1 is a 50kDa cytosolic protein that contains two tandem PDZ motifs and a C-terminal domain that can directly interact with ERM family members, thus it is also known as EBP50 (ERM-binding phosphoprotein 50) (Reczek et al., 1997). Functionally, NHERF-1 was first described as a specific regulator of transmembrane sodium proton (Na+/H+) exchangers (Weinman et al., 2000). However, it is becoming increasingly clear that NHERF-1 is an extremely promiscuous scaffolding protein that can interact with numerous different transmembrane molecules at the cell surface that include G-protein coupled receptors (ex: β2-adrenergic receptor), receptor tyrosine kinases (ex: EGFR and PDGFR), and other ion channels (ex: CFTR). Additionally, NHERF-1 can interact with intracellular signaling molecules such as β-catenin and PTEN (NHERF-1 interactions are reviewed in Weinman et al., 2005). Because of its ability to selfoligomerize and bind to the ERM cytoskeletal linkers, it is believed that NHERF-1  45  can facilitate the formation of actin-associated oligomeric scaffolds that anchors and clusters their membrane ligands at the cell surface (Lau and Hall, 2001; Weinman et al., 2005). For example, direct NHERF-1 interaction with both the PDGFR and EGFR has been shown to stabilize the receptors at the membrane and potentiate their downstream signaling (Lazar et al., 2004; Maudsley et al., 2000). In other scenarios, however, NHERF-1 has been shown to target certain transmembrane binding partners for internalization, thus dictating selective cell surface retention and trafficking of resident membrane proteins (Shenolikar and Weinman, 2001). Ultimately, these diverse functions reported for NHERF-1 could link podocalyxin to multiple signaling networks, which could functionally contribute to podocalyxin’s role in potentially orchestrating epithelial polarization and morphogenesis.  1.5.6  Podocalyxin in cancer  It is likely that anti-adhesive molecules with the potential to modulate cell polarity, cell morphology and tissue architecture may contribute to cancer progression when they are dysregulated. Podocalyxin has been shown to be functionally important for all of the aforementioned processes, and as such, has been implicated in numerous malignancies. These include: testicular germ cell carcinomas (Schopperle et al., 2003), prostate cancer (Casey et al., 2006; Sizemore et al., 2007), leukemia (Kelley et al., 2005), hepatocellular carcinoma (Heukamp et al., 2006), and pancreatic adenocarcinoma (Ney et al., 2007). Additionally, our group has shown that podocalyxin is also highly expressed in serous ovarian carcinomas (Cipollone J.A., Graves M.L., Roskelley, C.D. submitted to Clin.Can.Res.), and it is overexpressed in a distinct subset of invasive breast carcinomas (Somasiri et al., 2004).  46  Using a tissue microarray (TMA) linked to long-term patient outcome data (Makretsov et al., 2003), we assessed levels of podocalyxin expression in 272 human breast carcinomas and classified them based on the strength of immunohistochemical staining. Sixty percent of the invasive breast cancer cases did not display detectable podocalyxin staining and were assigned a score of ‘0’. The remaining forty percent of breast tumors expressed podocalyxin to varying degrees, and were scored accordingly in groups ‘1-3’. Importantly, the group of breast tumor cases that showed the highest level of podocalyxin expression (Group 3) was significantly associated with poor patient outcome (Figure 1-5). High podocalyxin also significantly correlated with tumors that were of high grade, were negative for the estrogen receptor, and had lost p53 expression. Taken together, these observations indicate that podocalyxin overexpression is associated with particularly aggressive forms of invasive breast cancer. Strikingly, we found that podocalyxin overexpression as a marker on its own could predict poor outcome better than other independent prognostic indicators, such as lymph node status and HER2/neu overexpression. Thus, podocalyxin overexpression identifies a unique subpopulation of invasive breast tumors with an increased potential to undergo progression.  47  Figure 1-5: Podocalyxin is overexpressed in a subset of invasive breast tumors and is an independent prognostic indicator of poor patient outcome A) Using tissue microarray (TMA) technology, podocalyxin expression was analyzed on 272 invasive breast tumors with linked patient outcome data. The intensity of the podocalyxin positive staining was scored as follows: No staining=‘0’; Weak=‘1’; Moderate=‘2’; Strong=‘3’. A score of ‘3’ was given if there was intense staining in more than 50% of the cells. Representative tissues from each score are shown.  B) Disease-specific survival analysis indicated that tumors with strong podocalyxin staining (Group 3) had a significantly poorer outcome when compared to “no” (Group 0) or “low” (Groups 1 and 2) Podocalyxin staining categories.  48  49  While only a small proportion of breast tumor cases displayed high podocalyxin expression, this was not the case for the analyzed cases of ovarian carcinomas. We found that podocalyxin is expressed in a majority (67%) of ovarian carcinomas on a 541 case TMA (Cipollone J.A., Graves M.L., Roskelley, C.D. submitted to Clin.Can.Res.). Podocalyxin positivity was greatest in high-grade serous tumors, which form the most common and lethal ovarian carcinoma subtype. Interestingly, in podocalyxin-positive serous tumors, membrane localized podocalyxin was an indicator of poor outcome. Importantly, the combined presence of NHERF-1 and ezrin in membrane-localized podocalyxin positive tumors was associated with an even worse overall patient survival. While the small sample size limited our ability to perform this kind of correlative analysis in the breast tumor cases with high podocalyxin overexpression, the data from the ovarian TMA analysis suggests that the cell surface localized podocalyxin/NHERF-1/ezrin complex may be important in cancer progression. Given these clinical correlations, and the fact that both NHERF-1 and ezrin have, on their own, been implicated in breast cancer, it is important to determine the potential functional significance of both podocalyxin overexpression and the role of the ‘podocalyxin complex’ on breast tumor formation and/or progression.  1.6 Thesis objectives For normal function, the mammary gland exists in a highly ordered structure that is under tight regulatory control. When the signals that regulate and define its normal context are disrupted, this can lead to architectural disorder, which increases the risk of tumor formation. Ultimately, it is critical to understand the regulatory 50  mechanisms that govern normal mammary structure in order to understand how these signals go awry in cancer (Bissell and Radisky, 2001). Thus, I characterized a unique 3D culture model of mammary epithelial morphogenesis using the normal mouse mammary epithelial cell line, EpH4 (Chapter 3). This model aimed to isolate the individual contributions that cell-cell and cell-matrix adhesions have on regulating the polarized architecture of 3D mammary spheroids. Using this model I determined that basement membrane-mediated integrin dependent signals can relocalize polarity determinants and tight junctions prior to lumen formation without disrupting tight junctions themselves. Thus, this model can now be used to interrogate the important mechanisms for normal mammary morphogenesis, and to test candidate genes that may functionally contribute to the loss of architecture that is frequently observed in breast cancer.  The anti-adhesive molecule podocalyxin is a putative regulator of polarity and tissue architecture that is overexpressed in invasive breast cancers that are associated with poor patient outcome. Therefore, I tested the functional significance of this overexpression by force expressing podocalyxin into normal mammary epithelial cells and in non-metastatic breast tumor cells to assess whether podocalyxin promoted architectural disruption, cell adhesion defects and tumorigenicity in these cells (Chapters 4 and 5).  Podocalyxin is localized to, and functions at the apical cell surface where it forms a complex with NHERF-1 and ezrin that is linked to the actin cytoskeleton. It is possible that this multi-protein complex is important for the membrane localization of podocalyxin, which is likely a prerequisite for its function as an anti-adhesin and 51  regulator of epithelial morphology. Thus, I studied the role of NHERF-1, ezrin and f-actin on the targeting of podocalyxin to the apical plasma membrane (Chapter 6).  52  CHAPTER 2 : MATERIALS AND METHODS 2.1 Methods for Chapter 3 2.1.1  General cell culture  The EpH4 cell line was isolated as a homogeneous epithelial sub-population of the IM-2 mouse mammary cell line. In brief, the IM-2 cell line originated from primary mouse mammary cells (isolated from the #4 abdominal mammary fat pads of midpregnant BALB/c mice) that underwent spontaneous immortalization in culture (Reichmann et al., 1989). IM-2 cells contain a heterogeneous mixture of mammary epithelial cells and stromal fibroblasts, so were subsequently separated into clonal populations with either epithelial or fibroblastic morphologies (Reichmann et al., 1989). The resulting 31E epithelial clone was re-named Ep1 (Reichmann et al., 1992) and was subsequently infected with the retroviral empty vector pHMW; the resulting clonal population resistant to hygromycin that was obtained is the EpH4 cell line (Fialka et al., 1996) which was used in these studies.  EpH4 cells were routinely maintained in DMEM/F12 (Sigma, St. Louis, MO) supplemented with 5% Fetal Bovine Serum (FBS; Invitrogen, Carlsbad, CA), insulin (5 µg/ml; Sigma) and gentamycin (50 µg/ml; Sigma) in a stable atmosphere at 37°C and 5% CO2.  2.1.2  3D cell culture model  To analyze functional differentiation and cell junction dynamics, EpH4 cells were cultured in serum-free differentiation media (DMEM/F12 supplemented with insulin (5 µg/ml; Sigma), hydrocortisone (1 µg/ml; Sigma), gentamycin (50µg/ml; 53  Sigma), and prolactin (3 µg/ml; Sigma) and maintained as 2D monolayers or 3D clusters/spheroids as described below.  To generate ‘naked’ 3D cell clusters in the absence of exogenously added ECM, EpH4 cells were plated onto tissue culture dishes pre-coated with polyhydroxyethylmethacrylate (PolyHEMA; 5mg/ml; Sigma). PolyHEMA is a simple polymer, that when hydrated, forms a hydrogel-like plastic coating on top of the tissue culture-treated plastic; a method which has been described as an effective means to prevent cell attachment (Folkman and Moscona, 1978). Thus, cells that are normally adherent can be maintained and analyzed in suspension cultures.  For ECM-induced 3D mammary spheroid formation, two different methods were employed. For “On Gel” spheroid formation, tissue culture dishes or sterile coverslips were pre-coated with a 1:1 mixture of DMEM/F12 and MatrigelTM (BD Biosciences, Mississauga, ON) (v/v) and incubated at 37°C for 30 minutes-1 hour to allow the ECM to form a gel. EpH4 cells were trypsinized, resuspended in differentiation media and plated onto the pre-formed Matrigel layer at a concentration of 1.5-2.0 x 105 cells/ml.  For “ECM-Overlay” spheroid formation, EpH4 cells were first pre-clustered on PolyHEMA for 6 hours at 37°C followed by gentle mixing, 1:1 (v/v), with a 5% Matrigel solution (diluted in differentiation media) and plated onto dishes/coverslips pre-coated with a thin layer of 2.5% Matrigel. To analyze the effects of individual purified ECM components in the overlay assay, pre-clustered  54  EpH4 cells were mixed, 1:1 (v/v), with either collagen I or a laminin-1 at a final concentration of 1mg/ml and 100µg/ml respectively and plated onto ECM-specific pre-coated coverslips.  For function-blocking antibody experiments, EpH4 cells (5 x 105 cells/ml) were preclustered on PolyHEMA for 4 hours in the presence of isotype control or blocking antibodies against either the α6 integrin subunit (clone GoH3; BD Biosciences) or the β1 integrin subunit (clone AIIB2; BD Biosciences) at a final concentration of 10µg/ml. Alternatively, both blocking antibodies were added together. Preclustered, cells pre-treated with antibodies were then overlayed with Matrigel that also contained the appropriate blocking antibodies, maintained as described above for 24 hours, and processed for immunocytochemistry.  2.1.3  Immunocytochemistry and confocal analysis  For immunostaining, EpH4 cells maintained as 2D monolayers or 3D-ECM-treated spheroids were cultured on glass coverslips for the indicated times, and subsequently fixed in ice cold methanol at –20°C for 20 minutes. To analyze 'naked' cell clusters a 200µl of suspension of cells pre-clustered on polyHEMA was plated as a single large droplet onto 18mm glass coverslips, incubated at 37°C for 2-4 hours to allow clusters to tether to the glass, and immediately submersed in fixative.  Fixed cells were rehydrated in 1xPBS and blocked with 1%BSA/10% normal growth serum (NGS; Jackson ImmunoResearch) in PBS for 30 minutes at room temperature. For multi-immunolabelling, cells were simultaneously incubated with the indicated primary antibodies for 1 hour at room temperature (See Table 2.1). For 55  primary antibody detection, the cells were washed in 1xPBS and subsequently labeled with appropriate fluorescently conjugated secondary antibodies (See Table 2.2), and nuclei were counterstained with 4’6-Diamidino-2-phenylindole (DAPI; Sigma). The coverslips were mounted in glycerol containing the anti-fade diazabicyclo[2.2.2]octane (DABCO; Sigma), and labeled cells were imaged using the Olympus FV1000 confocal microscope (Olympus, Center Valley, PA) using the 60x immersion oil objective (NA=1.40). Final images were analyzed and processed using the FV1000 Fluoview and Adobe Photoshop 8.0 software. For co-localization analysis between two fluorophores, Pearson’s correlation coefficients, which reflect the degree of linear relationship between the two fluorophores across the image (ie. +1.0 = a perfect positive correlation), were calculated on single optical slices using the FV1000 Fluoview software.  2.1.4  Cell lysis and Western blotting  For differentiative β-casein analysis, EpH4 cell clusters and spheroids were enzymatically treated with the neutral protease solution Dispase (BD Biosciences) to digest the ECM at 37°C for 30 minutes. Dispase-released cells were then rinsed, pelleted and lysed on ice in RIPA lysis buffer (150 mM NaCl, 50 mM Tris pH 7.4, 5 mM EDTA, 1.0% NP-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulphate) containing the following protease and phosphatase inhibitors: 2 µg/ml aprotinin (Sigma), 1 µg/ml leupeptin (Sigma), 1 µg/ml pepstatin A (Sigma), 100 µM PMSF (Sigma), 10 mM Sodium fluoride (Sigma), 2.5 mM EDTA. After incubation for 15 minutes on ice, lysates were cleared by centrifugation, protein concentration was determined using the BCA assay (Pierce, Rockford, IL) and equal amounts of  56  protein were separated by SDS-PAGE followed by transfer onto PVDF membranes. Membranes were then blocked (4% BSA, 1% NGS in TBS-T), and Western blotting was performed by incubating membranes with the indicated primary antibodies (see Table 2.1) at 4°C overnight, followed by binding of the appropriate HRPconjugated secondary antibody (see Table 2.2) that was visualized using enhanced chemiluminesence substrate (ECL; Amersham Biosciences; Piscataway, NJ).  For analysis of adhesion/junction molecules naked clusters and ECM-overlaid spheroids were scraped directly into ice-cold 1xPBS supplemented with protease inhibitors and 10mM EDTA followed by incubation at 4°C on a vertical rocker for 45 minutes to liquify the added ECM in a protease-independent manner. Dissociated clusters and spheroids were then harvested by centrifugation at 4°C and lysed in RIPA buffer for total protein analysis. To compare soluble and cytoskeletal protein fractions, dissociated spheroids were lysed in a 1% triton cytoskeletal stabilization buffer (1% Triton x-100, 50mM NaCl, 300mM sucrose in 1xPBS with protease inhibitors) and centrifuged. After removal of the soluble fraction, the insoluble pellet was further lysed in SDS lysis buffer (1% SDS, 10mM Tris-HCl, 2mM EDTA, protease inhibitors; cytoskeletal fraction). Protein quantification and Western Blotting of the soluble and insoluble (cytoskeletal) fractions were then performed as described above.  57  2.2 Methods for Chapter 4 2.2.1  Mice and tissue processing  To analyze the immunolocalization and expression profile of podocalyxin in the mammary gland, FVB mice at different developmental stages were obtained from the BC Cancer Agency in collaboration with Dr. John Stingl and Dr. Connie Eaves (BC Cancer Agency, Vancouver, BC). The abdominal (#4) mammary glands were excised and either lysed in RIPA buffer for total protein analysis, or fixed in 10% buffered formalin and paraffin embedded by Wax-it Histology Services Inc. (Vancouver, BC). Mouse kidneys were similarly processed as a positive control for podocalyxin expression.  Paraffin-embedded blocks of normal human mammary gland tissue were obtained in collaboration with Dr. JoAnne Emerman and Darcy Wilkinson (Department of Anatomy, UBC). The tissue originated from a 44 year old female patient who had undergone mammary reduction surgery at Vancouver General Hospital.  2.2.2  Immunohistochemistry  Formalin-fixed and paraffin-embedded tumor specimens were sectioned (thickness=5.0 µm), deparaffinized, and stained with H&E using standard procedures. For immunohistochemical analysis, deparaffinized tissue was antigenretrieved in heated citrate-buffer, blocked for endogenous peroxidase, and incubated with a primary antibody against mouse podocalyxin (1:100; R&D Systems). Antibody binding was visualized using a horseradish peroxidase-labelled polymer (DAKO EnVisionTM+ System, Troy, MI), developed with Nova RedTM 58  (Vector Laboratories, Burlingame, CA) and followed by Mayer’s hematoxylin counterstaining.  2.2.3  Western blotting  Mouse mammary tissue or subconfluent cell monolayers were either homogenized or scraped directly into RIPA lysis buffer (For recipe, see Methods section 2.1.4), incubated on ice for 15 minutes and cleared by centrifugation. Protein quantification, SDS PAGE, transfer to PVDF membrane and Western blotting were all performed as described in section 2.1.4.  2.2.4  Semi-quantitative and quantitative RT-PCR  For semi-quantitative RT-PCR, total RNA was extracted with Trizol (Invitrogen, Carlsbad CA) and cDNA was synthesized using the Thermoscript RT-PCR system (Invitrogen). The primer sequences used for murine Podocalyxin transcript amplification were 5’-GAGGATTTGTGCACTCTACATGTG-3’ (Forward); 5’TACTCGAGTGGGTTGTCATGGTAACC -3’ (Reverse) which generates a 352 bp fragment. To control for the relative quality and quantity of RNA samples, primers for β-actin were used which generates a 510bp fragment (Wong et al., 1999). Both sets of primers span intron boundaries to ensure mRNA-specific amplification.  For Quantitative RT-PCR, total RNA was extracted using the RNAeasy RNA isolation kit (Qiagen Inc. Mississauga, ON) and cDNA was generated using the High Capacity cDNA archival kit (Applied Biosystems, Foster City, CA) according to manufacturers recommendations. Quantitative PCR was performed using 59  standard protocols using the TaqMan Gene Expression Assay for the detection of murine Podocalyxin tranascript, and normalized to 18S rRNA (Applied Biosystems).  2.2.5  Expression vectors and cloning  Dr. Julie Nielsen, (a former PhD student in Dr. Kelly McNagny’s laboratory) engineered the pIRES2-EGFP vector (Clontech; Mountain View, CA) for expression of mouse podocalyxin. Briefly, full length murine podxl cDNA (a generous gift from Dr. David Kershaw, University of Michigan) was excised from the pBluescript SK (Stratagene; La Jolla, CA) cloning vector via the SacII and BglII sites and subcloned into pIRES2-EGFP between SacII and BamHI restriction sites. This vector was successfully used to ectopically express podocalyxin in human MCF-7 and canine MDCK cells.  Due to poor protein expression driven from the CMV promoter present in the pIRES2-EGFP expression vector in mouse EpH4 cells, I excised the podxl cDNA (without the 3’UTR region) from pIRES2-EGFP, ligated it into the TOPO cloning vector (Invitrogen), and subsequently subcloned it into the pβ-actin (pβ) expression vector at the EcoRI restriction site of its Multiple Cloning Site (MCS). The pβ vector, (a generous gift from Dr. Philip Leder, Harvard University), was derived from the pcDNA3.1 expression vector backbone and it contains a CMV enhancer region and chicken β-actin promoter which together, is capable of successfully driving stable constitutive transgene expression in EpH4 cells (Pinkas and Leder, 2002).  60  2.2.6 Transfections and generation of stable cell lines 2.2.6.1 General cell culture The MCF-7 human breast tumor cell line was isolated from the pleural effusion of a woman with metastatic breast cancer (Soule et al., 1973). MCF-7 cells are morphologically epithelial and they express the estrogen receptor. Thus, they have been used extensively over the years as a model of well-differentiated breast tumor cells. The MDCK epithelial cell line was derived from a kidney of a normal adult female cocker spaniel (Madin et al., 1957), and is believed to be primarily distal renal tubule in origin. MDCK cells express a full complement of adherens and tight junction proteins, which they assemble in a polarized function, thus are commonly used as a model of epithelial morphogenesis.  Both MCF-7 and MDCK cells were routinely cultured in a stable atmosphere of 5%CO2 at 37°C in DMEM/F12 supplemented with 5% FBS and gentamycin (50 µg/ml; Sigma). The same culture media was used for EpH4 cells, but the media was additionally supplemented with insulin (5 µg/ml; Sigma) (as described in section 2.1.1).  2.2.6.2 Overexpression of podocalyxin Parental MCF-7 and MDCK cells were transfected with either control pIRES2-EGFP or pIRES2-EGFP-msPodo using the lipofection reagent DMRIE-C according to Manufacturer’s instructions (Invitrogen). EpH4 cells were transfected with the pβmsPodo expression vector (or the pβ empty vector as a control) using  61  Lipofectamine 2000 (Invitrogen). Initial stable populations were genetically selected and maintained using 400 µg/ml G418.  To enrich for stable transfectants expressing significant amounts of ectopic murine podocalyxin, the stable populations were subjected to two independent rounds of fluorescent automated cell sorting (FACs) after indirect immunofluorescent labeling of murine podocalyxin using a species-specific antibody. Cells transfected with the empty vector were used as a negative control and were also collected by FACs to control for any phenotypic changes resulting from transfection and the cell sorting process. Briefly, cells were gently trypsinized and resuspended in rat anti-mouse PCLP-1 antibody (1µg/ml; MBL, Nagoya, Japan), which was then labeled with a biotinylated anti-rat IgM+IgG secondary antibody (1:250; Southern Biotech, Birmingham, AL) followed by streptavidin-allophycocyanin (SAv-APC; 1:250, BD Biosciences). SAv-APC positive cells were then collected by Andy Johnson using the FACs Vantage cell sorter at the Biomedical Research Center’s FACs Facility, expanded in culture, and maintained under genetic selection (400 µg/ml G418).  2.2.6.3  Lentiviral shRNA for murine podocalyxin  Predicted shRNA sequences designed for targeted knockdown of murine Podocalyxin were identified using PSI Oligomaker v1.5 freeware (http://web.mit.edu/jacks-lab/protocols/pSico.html). Three individual shRNA oligomers (Table 2.3) were each cloned into the GFP-containing pLL3.7 lentiviral vector downstream of the U6 RNA promoter into HpaI and XhoI sites, and positive clones were selected and verified by sequencing. 62  To produce lentiviral particles, 293T cells were co-transfected with 10µg of pLL3.7 and the appropriate packaging plasmids (3.5µg of pVSVg, 3.5µg of pRSV-Rev, 6.5µg of pMDLgag/pol) by calcium phosphate transfection. Lentiviral containing media was collected 36 hours post-transfection and transferred to sub-confluent EpH4 cells seeded the day before. The virus-containing media was replaced with regular growth media after 48 hours and incubated for an additional 48 hours. Cells were then harvested and bulk FAC sorted for GFP expression. The resulting heterogeneous population was then sorted a second time for the top 2% of cells expressing the highest level of GFP. These cells were then harvested and ectopic mouse Podocalyxin expression was assessed at both the mRNA and protein levels.  2.2.7  3D MCF-7 tumor cell spheroid culture  To generate tumor cell spheroids, 1.5x105 MCF-7 stable transfectants (empty control vector or podocalyxin expressing) were plated onto 18 mm glass coverslips that were pre-coated with a pre-gelled layer of Matrigel (diluted 1:1 in DMEM/F12 (v/v) for a total of 100µl). Within 12 hours, the cells typically aggregated and formed 3D spheroids on top of the ECM. These spheroids were then cultured for 5 additional days in serum-free DMEM/F12 supplemented with gentamycin (50µg/ml).  63  2.2.8  Immunocytochemistry and confocal analysis  For immunostaining analysis, cells were cultured as monolayers, naked clusters, or ECM-induced spheroids as described in section 2.1.2 (for EpH4 cells) or as indicated in the results section of Chapter 4. Briefly, cells in these various states were cultured on glass coverslips and fixed in ice-cold methanol at –20°C for 20 minutes. Immunocytochemistry and confocal analysis was performed as described in section 2.1.3.  For specific Z-series confocal analyses of MCF-7 and MDCK cells, 1.0x105 cells were plated on glass coverslips and allowed to grow to confluency (typically 3-5 days). To preserve optimal phallodin binding and anti-NHERF-1 antibody recognition, coverslips were subsequently fixed in 4% paraformaldehyde at room temperature for 20 minutes, rinsed in 1XPBS+0.1MGlycine for 10 minutes, followed by 3X5 minute rinses in 1XPBS. Cells were then permeabilized using 0.1%triton/1XPBS solution for 10 minutes at room temperature. At this point, coverslips were processed for immunocytochemistry generally as described in section 2.1.3.  For the great majority of the Z-axis imaging and 3D reconstruction analyses, multiple incremental images were collected along the Z-axis using an Olympus FV1000 confocal microscope equipped with a 60X objective (N.A.=1.40) (Olympus, Center Valley, PA). However, for detailed confocal analysis of the apical microvilli, images were collected using the 100X oil immersion objective lens (N.A.=1.42). All Z-axis step sizes were chosen according to the Nyquist criterion (ie: sampling 2.3X the optimal spatial frequency) in order to preserve the 3D spatial resolution of the  64  sample. As a result, I typically ‘oversampled’ in the Z-azis and chose a step size at least 1/2.3 of the optimal sampling interval within the optical image as limited by the N.A. of the objective and the wavelength of light utilized. This optimal sampling interval value was calculated by the FV1000 FluoView software.  Single slice Z-axis images were analyzed and processed using the FV1000 FluoView, and Adobe Photoshop 8.0 software. The 3D reconstructed and rotated images were generated from confocal stacks and re-sliced down the Z-axis every 5µM along the Y-axis using FluoView1000 imaging software. The resulting images and movies were subsequently modified using ImagePro Plus Discovery 3D (Media Cybernetics).  2.2.9  Scanning electron microscopy  Cells were grown on glass cover slips and fixed using 2.5% glutaraldehyde with 1% tannic acid in 0.1 M cacodylate buffer. SEM processing was performed by Derrick Horne in the BioImaging Facility at the University of British Columbia. Briefly, samples were post-fixed with buffered 1% osmium tetroxide, dehydrated in a graded series of ethanols, and critical point dried. Images were then collected using a Hitachi S4700 FESEM, and photos were arranged using Adobe Photoshop and Adobe Illustrator software.  2.2.10 Transmission electron microscopy Cells were grown on filters (1µm pore size; BD Biosciences) and fixed for 1 hour in 1.5 % glutaraldehyde and 1.5 % paraformaldehyde in 0.1 M sodium cacodylate 65  buffer (pH 7.3). The filters were then washed with 0.1 M sodium cacodylate buffer, post-fixed for 30 minutes on ice in buffered 1 % osmium tetroxide, washed again with distilled water and stained en bloc for 30 minutes with 1 % uranyl acetate. Samples were then dehydrated through a graded series of ethanols, infiltrated with propylene oxide and Polybed which was then polymerized for 24 hours at 60°C. Thin sections were prepared, stained with uranyl acetate and lead citrate, and viewed and photographed on a Philips 300 electron microscope operated at 60 kV by A. Wayne Vogl at the University of British Columbia. Negatives were scanned into digital format, contrast adjusted using the Image Adjustments tool and figures were generated using Adobe Photoshop and Adobe Illustrator software.  2.3 Methods for Chapter 5 2.3.1  Sub-cutaneous xenograft model of breast tumorigenesis  1 day prior to tumor cell injections, 17-β-estradiol tablets (60-day release; IRA, Sarasota, FL) were implanted sub-cutaneously (s.c.) into the cervical scapular space of twelve-week-old female Rag 2M mice (Taconic). Subconfluent, estrogendependent MCF-7 vector control cells, and MCF-7 cells stably overexpressing wild type mouse podocalyxin were cultured without antibiotics for 24 hours. Cells were then harvested and mice were inoculated s.c. with 1.0 x 107 tumor cells. Once palpable tumor masses formed, tumor measurements were performed three times per week, and volumes were calculated using the formula 0.52[length (mm)] x [width (mm)] x [height (mm)]. After 80-110 days post-inoculation, mice were sacrificed, and tumors were excised, weighed and processed for histological analysis. 66  2.3.2  Tumor histopathological and immunohistochemical analysis  Formalin-fixed and paraffin-embedded tumor specimens were sectioned, deparaffinized, and stained with H&E using standard procedures. For immunohistochemical analysis, deparaffinized tissue was antigen-retrieved in heated citrate-buffer, blocked for endogenous peroxidase, and incubated with the indicated primary antibody (see Table 2.2). Antibody binding was visualized using a horseradish peroxidase-labelled polymer (DAKO EnVisionTM+ System, Troy, MI), developed with Nova RedTM (Vector Laboratories, Burlingame, CA) and counterstained with Mayer’s hematoxylin.  2.3.3  Cell culture, immunocytochemistry and confocal analysis  For indirect immunolocalization of ectopic mouse podocalyxin and various adhesion molecules in monolayer culture, 5 x 104 cells/ml cells were plated onto 3µm pore transwell filters and allowed to form a confluent monolayer. The filters were then fixed in ice-cold methanol for 20 minutes at –20°C, and immunostaining was performed as generally described in section 2.1.3. The filters were cut out of the transwell insert and mounted in glycerol containing the anti-fade diazabicyclo[2.2.2]octane (DABCO; Sigma), and labeled cells were imaged along the Z-axis as described in section 2.2.8.  For indirect immunolocalization of MCF-7 cells in 3D culture, tumor cell spheroids of MCF-7 cells were generated as described in section 2.2.7 on glass coverslips for 5 days and fixed in ice cold absolute methanol for immunocytochemistry and confocal imaging analysis as described in section 2.1.3. 67  2.3.4  Western blotting  The FACs enriched, stably-transfected populations of MCF-7 cells (vector control or wt podocalyxin expressing) were grown as monolayers and scraped on ice into RIPA lysis buffer. Western blotting of cell RIPA lysates was performed as described in section 2.1.4.  2.3.5  Homotypic cell-cell aggregation analysis  To retain intact cell surface molecules during passaging, cells were isolated using an enzyme free cell-dissociation buffer (Invitrogen) and resuspended in calcium-free DMEM/F12 supplemented with either 1mM CaCl2 or 1mM EDTA. Cells were then passed through an 18-gauge needle to ensure that single cell suspensions were generated. 5x105 single cells were then plated in triplicate in 6 well plates preblocked with 1%BSA and rotated continuously at 80 rpm at 37°C. Phase micrographs of three fields at 10x magnification were taken for each replicate at the indicated time points, and the number of single cells that had failed to aggregate was quantified.  2.3.6  Single cell attachment and cell spreading assays  Single cell suspensions were plated on glass coverslips coated with bovine fibronectin (5µg/cm2) and allowed to attach and spread for the indicated times. Cells in suspension or attached cells were then fixed in 4% paraformaldehyde, permeabilized with 0.1% triton/PBS, and immunostained for ectopic podocalyxin, 68  together with either β1 integrin or f-actin using rhodamine-phalloidin. Confocal microscopy was carried out as described in section 2.1.3 and stacks consisting of 0.2µm Z-step images were obtained such that either XZ vertical images of attached cells, or XY images specifically at the apical and basal surface could be generated. For cell spreading quantification, the total area of the basal surface of individual cells was calculated using ImagePro 3DS v6.0 software.  2.3.7  Cell-ECM adhesion  Cells were removed from the culture dish with cell dissociation buffer and resuspended at 2.0 x 105 cells/ml in adhesion assay buffer (Dulbecco’s modified Eagle’s medium (DMEM) calcium-free base media supplemented with 1mM CaCl2, 0.1% BSA). Cells were then fluorescently labeled with 25µM 5chloromethylfluorescein (CMFDA, Molecular Probes-Invitrogen) for 30 minutes at 37°C, washed twice in adhesion assay buffer, incubated at 37°C for an additional 30 minutes and resuspended in fresh buffer. 2.0 X 104 cells were then plated into individual wells of a 96 well plate and pre-coated and blocked with 5µg/cm2 bovine fibronectin (Sigma) and 1%BSA in PBS, respectively. Total fluorescence of dyeloaded cells was measured (excitation 485nm/530nm emission filters) to represent the total number of cells plated. After the indicated times, the plate was gently washed three times with adhesion assay buffer and a second fluorescence emission was obtained of the remaining adherent cells. Each experiment was performed in triplicate, and the data is expressed as the average ratio of the fluorescent emission from the adherent cells : total fluorescent emission (X 100; mean +/- SD).  69  To test the adhesive strength of cell-ECM interactions, 2.0 X 104 cells were plated as single cells onto low (0.5µg/cm2) and high concentrations of fibronectin (5µg/cm2) in triplicate in a 96 well plate and allowed to adhere for 4 hours. The regular growth media was switched to 1xPBS and incubated at 37°C for 10 minutes to facilitate cellECM dissociation of loosely adherent cells. The PBS was removed and the remaining strongly adherent cells were fixed and stained with crystal violet. The average number of strongly adherent cells per well were then counted, and representative fields were photographed with a Nikon Coolpix 5400 digital camera attached to a Nikon TMS inverted microscope. Each experiment was carried out in triplicate and presented as the average number of adherent cells +/-SD.  2.3.8  Migration assays  To assess migration, control MCF-7 cells and MCF7 cells stably overexpressing podocalyxin were subjected to scratch assays. Cells were grown to confluence, serum starved for 16 hours, scratched using a sterile 20-200µl pipet tip and they were incubated with either 100 ng/ml EGF (PeproTech Inc., Rocky Hill, NJ), 5% fetal bovine serum, or they were left untreated for 48 hours. Phase micrographs were taken at the indicated times with a Nikon TMS phase Microscope equipped with a Nikon Coolpix digital camera to capture the migratory infilling of the scratch/wound.  The migratory behavior of individual cells was imaged and quantified using a phagokinetic track migration assay on glass coverslips coated first with bovine  70  fibronectin (5µg/cm2) and 1.0µm fluorescent microspheres (excitation 350nm, 0.1% solution; Molecular Probes-Invitrogen). Cells were resuspended in DMEM/F12, or media supplemented with EGF (100ng/ml; PeproTech Inc.), and 2.0X104 single cells in serum-free media or media supplemented with 100 ng/ml EGF were plated onto the lawn of microspheres over fibronectin and cultured for 24 hours. Coverslips were then fixed in 4% paraformaldehyde, permeabilized with 0.1% triton in PBS, and immunostained for ectopic podocalyxin and labeled for f-actin using rhodamine-phalloidin. Individual cells and disturbances to the microsphere lawn were then imaged using a 20x objective on the Olympus FV1000 confocal microscope. Cell migration was quantified by measuring the length of phagokinetic tracks, which were detected as negative images in the fluorescent bead background using the Olympus Fluoview software.  2.3.9  In vitro growth assays  Cell proliferation over time in 2D monolayer culture was determined using an MTT colorimetric assay for cell growth and viability (Mosmann, 1983). Briefly, cells were resuspended in either serum containing or serum free regular growth media, and 5 X 103 cells were plated in triplicate in individual wells of a 96 well plate (one plate per time point). On the indicated days, cells were incubated with MTT (3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) for 4 hours at 37°C. Cells were lysed in DMSO and the solublized reaction product was quantified using a multi-well scanning spectrophotometer.  71  To assess anchorage independent growth, 5 X 103 single cells were suspended in growth media containing 0.7% agarose and plated onto a solidified layer of growth media/1.0% agarose. Fresh growth media was added every four days and the cultures were fixed and stained with crystal violet after 4 weeks. Stained cultures were then imaged by dark field microscopy, colonies greater than 100µm in diameter were counted and individual colony size was determined using an ocular cytometer.  To assess the growth of clonally-derived MCF-7 tumor spheroids, a suspension of 3x104 single cells in 50µl of serum-free DMEM/F12 medium supplemented with gentamycin (50µg/ml) was mixed 1:1 with Matrigel and spread onto glass coverslips. The cell/matrigel mixture was incubated at 37°C for 30 minutes to allow the gel to solidify, and then supplemented with serum free DMEM/F12 and refreshed every two days for seven days. Cultures were then fixed in ice-cold methanol and processed for immunocytochemistry to detect ectopic mouse podocalyxin and the proliferation nuclear marker Ki-67. Tumor spheroids were imaged using the Olympus FV1000 confocal microscope and the longest axis of multiple (at least 70) tumor spheroids were quantified using the FV1000 Fluoview software.  72  2.4 Methods for Chapter 6 2.4.1  Mice and tissue processing  For NHERF-1 expression/localization analysis, WT FVB mice were obtained and #4 abdominal mammary glands, as well as kidneys for control tissue, were excised and processed for immunohistochemistry as described in section 2.2.1 and 2.2.2.  2.4.1.1 NHERF-1 knockout mice The NHERF-1 knockout mice were generously provided by Dr. Edward Weinman (Shenolikar et al., 2002); University of Maryland School of Medicine, Baltimore, MD) and housed at the Biomedical Research Center (BRC) Animal Facility. Mice at varying developmental stages of NHERF-1 -/-, NHERF +/-, and litter-mate wild type (WT) female mice were obtained in collaboration with Poh Tan (Dr. Kelly McNagny’s laboratory, BRC). The genotype of each animal was confirmed by PCR, which was performed by Taka Murakami (BRC). Mammary glands were excised and either lysed in RIPA buffer for total protein analysis, processed for immunohistochemistry (as described in sections 2.2.1 and 2.2.2), or processed for whole mount analysis.  For whole mount analysis, mammary glands were spread onto histological glass slides and submerged in Carnoy’s fixative (75% Ethanol, 25% Acetic Acid) for 4 hours at room temperature. Glands were rehydrated, fixed overnight in Carmine Alum (Stem Cell; Vancouver, BC), destained for 2 hours (2% HCl in 70% Ethanol), dehydrated in graded ethanol, cleared in xylene and mounted for imaging using a dissecting microscope.  73  2.4.2  Lentiviral shRNA for murine NHERF-1  Predicted shRNA sequences designed for targeted knockdown of murine NHERF-1 were identified using PSI Oligomaker v1.5 freeware (http://web.mit.edu/jackslab/protocols/pSico.html). Three individual shRNA oligomers (see Table 2.4) were each cloned into the pLL3.7 lentiviral vector downstream of the U6 RNA promoter into HpaI and XhoI sites, and positive clones were selected and verified by sequencing. Lentiviral infections of EpH4 cells were performed as described in section 2.2.6.3.  To assess the role of NHERF-1 depletion in 3D mammary spheroid formation apical polarization of tight junctions and lumen formation, EpH4 with stable shRNA NHERF-1 knockdown and control cells were plated on a layer of pre-gelled Matrigel (“On Gel”, described in section 2.1.2), fixed and immunostained for the tight junction protein ZO-1. Confocal analysis was then performed as described in section 2.1.3.  2.4.3  Podocalyxin mutant analysis  Dr. Julie Nielsen (Biomedical Research Centre) engineered the pIRES2-EGFP vector (Clontech) for ectopic expression of wild type chicken podocalyxin (WTchPodo; original cDNA from pcDNA3.1; (McNagny et al., 1997)). She also generated mutants of chicken podocalyxin lacking the extracellular domain (ΔEC; an Nterminal flag tag was added for membrane surface detection), C-terminal deletion  74  of the cytoplasmic domain downstream of the juxtamembrane CCHQRF sequence (ΔTAIL), or deletion of just the terminal PDZ binding domain (ΔDTHL) (Nielsen et al., 2007). MCF-7 cells were transfected with each mutant, selected under G418 (400 µg/ml), FAC sorted as single cells for positive transgene expression, and cells were grown with continued G418 genetic selection to generate stable clonal populations (Nielsen et al., 2007). Confocal, SEM and TEM analyses of the podocalyxin mutant expressing cells were performed as described in Chapter 4 Methods.  2.4.4  Additional expression vectors/reagents for transient transfection analysis  To assess the functional importance of ezrin, I utilized an expression construct containing a vsv-tagged N-terminal fragment of ezrin (1-309aa) lacking the Cterminal actin binding region (Algrain et al., 1993). This dominant negative form of ezrin (termed DN-ezrin) was a generous gift from Dr. Monique Arpin (Institut Curie-UMR, Paris, France). Stable MCF-7 transfectants (vector control and podocalyxin expressing) were transiently transfected with this construct and were either maintained as monolayers or trypsinized and replated to allow the cells to reestablish an epithelial monolayer. Cells in either condition were fixed in 4% paraformaldehyde 36 hours post-transfection and processed for immunocytochemistry and Z-axis confocal imaging as described in section 2.2.8.  To analyze the functional requirement of the GTPases RhoA and CDC42, podocalyxin expressing MCF-7 cells were transiently transfected with the dominant negative mutants RhoA-N19 (Ridley and Hall, 1992) and CDC42-N17 (Nobes and 75  Hall, 1995). These mutants exert a dominant negative effect by interfering with guanine nucleotide exchange from GDP to GTP. Alternatively, inhibition of all Rho GTPase isoforms was also performed by introducing the C3 exoenzyme form Clostridium botulinum (10µg/ml; Calbiochem, San Diego, CA) by transfection (Renshaw et al., 1996). Cells were fixed in 4% paraformaldehyde 24-36 hours posttransfection/post-treatment and processed for immunocytochemistry and Z-axis confocal imaging as described in section 2.2.8.  2.4.5  Latrunculin-mediated f-actin disruption  To analyze whether the maintenance of apically localized podocalyxin is dependent on the actin cytoskeleton, MCF-7 cells were grown as monolayers and treated with 10µM Latrunculin A (Sigma) for 15 minutes at room temperature to block actin polymerization. The treated monolayers were rinsed, fixed in 4% paraformaldehyde, immunostained and imaged by confocal microscopy as described section 2.2.8.  76  Table 2-1: List of primary antibodies used in this thesis (page 1 of 2) Primary Antibody (Species of Immunogen)  Species Raised/ Antibody Type  Applications: dilution and final concentration (if known)  Actin  MsIgG2A  WB: 1:1000 (From ascites)  Alpha(α)6-integrin (Rt)  RtIgG2A  Atypical PKCζ  Rb (antiserum)  IF: 1:100; 1µg/ml FB: 1:10; 10µg/ml IF: 1:500  Beta(β)1-integrin (Rt)  RtIgG2A  Beta(β)1-integrin (Hu)  MsIgG  IF: 1:100; 1µg/ml FB: 1:10; 10µg/ml IF: 1:100; 5µg/ml  Beta(β)-casein (Ms)  Ms  WB: 1:1000  Beta(β)-catenin  MsIgG1  Claudin-1 (Hu)  RbIg  E-cadherin (Hu)  MsIgG2A  ERK1/2  Rb  WB: 1:1000; 0.25µg/ml IF: 1:100; 2.5µg/ml WB: 1:1000 IF: 1:100 WB: 1:1000; 0.25µg/ml IF: 1:500; 0.5µg/ml IHC: 1:2000; 0.125µg/ml WB: 1:3000  Ezrin  Ms IgG1  IF: 1:50  FAKpY397  Rb  Flag  Mouse  IF: 1:500 WB: 1:1000 IF: 1:800;  GFP  RbIgG  IF: 1:200; 5µg/ml  Keratin  Rb (Antiserum)  IF: 1:200  Ki-67 (Hu)  RbIgG  Myc  MsIgG  IF: 1:200 IHC: 1:200 IF: 1:500  NHERF-1 (Hu)  RbIgG  Occludin (Hu)  MsIgG1-κ  Par3 (Ms)  RbIgG1  Par6 (Hu)  Rb (Antiserum)  WB: 1:1000; 1µg/ml IF: 1:1000; 1µg/ml IHC: 1:2000; 0.5µg/ml WB: 1:1000; 0.5µg/ml IF: 1:100; 5µg/ml IF: 1:500; 0.5µg/ml WB: 1:1000; 1µg/ml IF: 1:50  Par6 (Hu)  RbIgG  WB: 1:1000  Podocalyxin (Hu)  Ms  WB: 1:1000 IHC: 1:100  Podocalyxin (Ms)  Rt IgG2B  WB: 1:1000; 0.1µg/ml IF: 1:100; 1µg/ml IHC: 1:100; 1µg/ml  Source  Sigma (St. Louis, MO) BD Biosciences (Mississauga, ON) Millipore-Upstate (Billerica, MA) BD Biosciences (Mississauga, ON) R&D systems (Minneapolis, MN) M. Bissell Lab BD Biosciences (Mississauga, ON) Invitrogen-Zymed (Carlsbad, CA) BD Biosciences (Mississauga, ON) Santa Cruz Biotechnology Inc. (Santa Cruz, CA) Abcam (Cambridge, MA) Cellular Signaling Sigma (St. Louis, MO) Invitrogen-Molecular Probes (Carlsbad, CA) Dako (Missassauga, ON) LabVision (Freemont, CA) Clone 9E10 Purified from Hybridoma (M.Gold, Vancouver, BC) Abcam (Cambridge, MA) Invitrogen-Zymed (Carlsbad, CA) Millipore-Upstate (Billerica, MA) BC32AP From S. Ohno (Yokohama, Japan) Santa Cruz Biotechnology Inc. (Santa Cruz, CA) Clone 3D3 Hybridoma Supernatent D. Kershaw R&D systems (Minneapolis, MN)  77  Table 2-1: List of primary antibodies used in this thesis (page 2 of 2) Podocalyxin (Ms)  RtIgG1  FC: 1:100; 1µg/ml  RhoA GTPase (Hu)  Ms IgG1  WB: 1:1000  Vimentin  Mouse  IF: 1:100  VSV G-Glycoprotein  MsIgG1 (Ascites)  IF: 1:1000  ZO-1  RtIgG  ZO-1 (hu)  RbIg  WB: 1:1000 IF: 1:100 WB: 1:1000; 0.25µg/ml IF: 1:100; 2.5µg/ml  MBL (Nagoya, Japan) Santa Cruz Biotechnology Inc. (Santa Cruz, CA) Sigma (St. Louis, MO) Sigma (St. Louis, MO) Millipore-Chemicon (Billerica, MA) Invitrogen-Zymed (Carlsbad, CA)  Hu: Human Ms: Mouse Rt: Rat WB: Western Blot IF: Immunofluoresence FC: Flow cytometry IHC: Immunohistochemistry (Formalin/paraffin processed)  78  Table 2-2: List of secondary antibodies used in this thesis Secondary Antibody, Direct Conjugate or Toxin Labels Goat anti-Ms IgG H&L  Conjugate  Applications: dilution and final concentration (if known) IF: 1:100; 10µg/ml  Goat anti-Ms IgG H&L  Alexa Fluor 488 Alexa Fluor 488 Alexa Fluor 568 Alexa Fluor 488 Alexa Fluor 568 Alexa Fluor 488 Alexa Fluor 568 Alexa Fluor 647 HRP  Goat anti-Rb IgG H&L  HRP  WB: 1:5000; 0.08 µg/ml  Goat anti-Rt IgG H&L  HRP  WB: 1:5000; 0.08 µg/ml  Goat anti-Rt F(ab’)2 Fragment Goat anti-Rt IgM+IgG  Biotin  IHC: 1:250; 2.7µg/ml  Biotin  FC: 1:200  Straptavidin  APC  FC: 1:250  Ms anti-hu β1 integrin  FITC  FC: 1:10  Phalloidin  Rhodamine  IF: 1:400  4’6-Diamidino-2phenylindole (Dapi)  Unconjugated  IF: 1:10000  Goat anti-Ms IgG H&L *highly cross-adsorbed Goat anti-Ms IgG H&L Goat anti-Rb IgG H&L Goat anti-Rb IgG H&L Goat anti-Rt IgG H&L Goat anti-Rt IgG H&L Goat anti-Rt IgG H&L  IF: 1:100; 10µg/ml IF: 1:100; 10µg/ml IF: 1:100; 10µg/ml IF: 1:100; 10µg/ml IF: 1:100; 10µg/ml IF: 1:100; 10µg/ml IF: 1:100; 10µg/ml WB: 1:5000; 0.08 µg/ml  Source  Invitrogen-Molecular Probes (Carlsbad, CA) Invitrogen-Molecular Probes (Carlsbad, CA) Invitrogen-Molecular Probes (Carlsbad, CA) Invitrogen-Molecular Probes (Carlsbad, CA) Invitrogen-Molecular Probes (Carlsbad, CA) Invitrogen-Molecular Probes (Carlsbad, CA) Invitrogen-Molecular Probes (Carlsbad, CA) Invitrogen-Molecular Probes (Carlsbad, CA) Jackson Immunoresearch Inc. (West Grove, PA) Jackson Immunoresearch Inc. (West Grove, Jackson Immunoresearch Inc. (West Grove, Biosource International Camarillo, CA Southern Biotech (Birmingham, AL) Invitrogen-Molecular Probes (Carlsbad, CA) IOTEST, Immunotech Marseille, France Invitrogen-Molecular Probes (Carlsbad, CA) Sigma (St. Louis, MO)  Hu: Human Ms: Mouse Rt: Rat WB: Western Blot IF: Immunofluoresence FC: Flow cytometry IHC: Immunohistochemistry (Formalin/paraffin processed)  79  Table 2-3: Sequences of shRNA oligomers for mouse Podocalyxin and NHERF-1  Sequence Knockdown Name shRNA-Podo-1 *Primer #2603  shRNA-Podo-2 *Primer #3293  shRNA-Podo-3 *Primer #4512  shRNA-NH-1 *Primer #1681  shRNA-NH-2 *Primer #1792  shRNA-NH-3 *Primer #1899  _________SENSE ________ ___LOOP___ ______ANTISENSE________  5’- TGAGACTGGCCTATCATTTATTCAAGAGATAAATCATAGGCCAGTCTCTTTTTTC -3’ 3’- ACTCTGACGGCATAGTAAATAAGTTCTCTATTTACTATCCGGTCAGAGAAAAAAGAGCT -5’  5’- TGTATTCTTGTGGTATAAGTTTCAAGAGAACTTATACCACAAGAATACTTTTTTC -3’ 3’- ACATAAGAACACCATATTCAAAGTTCTCTTGAATATGGTGTTCTTATGAAAAAAGAGCT -5’  5’- TGAATGTAAATGTCTATTTATTCAAGAGATAAATAGACATTTACATTCTTTTTTC -3’ 3’- ACTTACATTTACAGATAAATAAGTTCTCTATTTATCTGTAAATGTAAGAAAAAAGAGCT -5’  5’- TGCATCTTGGGT TCATTTGATTCAAGAGATCAAATGAACCCAAGATGCTTTTTTC -3’ 3’- ACGTAGAACCCAAGTAAACTAAGTTCTCTAGTTTACTTGGGTTCTACGAAAAAAGAGCT -5’  5’- TGCAATGGCCTCATCCTTAATTCAAGAGATTAAGGATGAGGCCATTGCTTTTTTC -3’ 3’- ACGTTACCGGAGTAGGAATTAAGTTCTCTAATTCCTACTCCGGTAACGAAAAAAGAGCT–5’  5’- TGGAAATGCCTTCAGAAATTTTCAAGAGAAATTTCTGAAGGCATTTCCTTTTTTC -3’ 3’- ACCTTTACGGAAGTCTTTAAAAGTTCTCTTTAAAGACGTCCGTAAAGGAAAAAAGAGCT -5’  80  CHAPTER 3 : REGULATING APICAL POLARIZATION IN A 3D MODEL OF MAMMARY MORPHOGENESIS 3.1 Introduction A major disadvantage to modeling epithelial cell polarity in typical 2D in vitro monolayers is that the epithelial cells interact in an environment where the rigid 2D substratum creates an overwhelming spatial cue in which the cells are artificially instructed to orient apicobasal polarity. In contrast, culturing epithelial cells surrounded by ECM creates a microenvironment where cells are forced to translate physiologically relevant spatial cues. As a result, the cells interact with one another to form a polarized 3D architecture that resemble cyst or tubule structures characteristic to simple epithelial tissues in vivo. Thus, the ability to model tissuelike organization of epithelial cells when grown inside a 3D ECM creates a powerful in vitro tool to identify the molecular signals that signify epithelial architecture.  Cultured mammary epithelial cells provide an attractive model to address the mechanisms that control epithelial morphogenesis and apicobasal integrity in 3D. When these cells are grown in vitro in the presence of ECM, they undergo glandular morphogenesis and form 3D spheroids that recapitulate the alveolar structures present in the mammary gland. These in vitro 3D spheroids not only establish apicobasal polarity and cavitate to form a central lumen, but also functionally differentiate in response to lactogenic hormones. Current 3D mammary morphogenesis models, however, have limitations in the ability to characterize the required signals for polarized junction formation and apicobasal polarity, 81  particularly in long-term culture models that involve proliferation of a single cell embedded in ECM. In these models (using MCF-10A cells, for example) polarized spheroid formation occurs over the course of many days and lumen formation depends on apoptotic clearance of cells that are not in contact with the surrounding ECM (Debnath et al., 2002). Thus, the specific signaling pathways required for generating a polarized architecture are difficult to independently isolate within the context of such a complex and long-term model; one which establishes apicobasal polarity while concurrently integrating a balance of proliferation and apoptosis.  This study describes a unique 3D model of mammary epithelial spheroid organization that separates junctional polarization, orientation of apicobasal polarity, and functional differentiation independently of cell growth. Using this model I demonstrate that cell-cell interactions alone, in the absence of signals from the ECM, are sufficient to initiate the assembly of apical polarity complexes in multi-cell clusters. However, cellular interactions with the ECM are required to orient those junctions appropriately prior to lumen formation.  3.2 Expression and localization of epithelial junction proteins in EpH4 cells I chose the EpH4 mouse mammary epithelial cell line as the basis of this model for four primary reasons. First, EpH4 cells in culture maintain a very uniform epithelial morphology with little phenotypic variability between experiments. Second, grown as 2D monolayers, EpH4 cells have also been well characterized as a highly polarized cell line that form lateral adherens junctions and functional tight junctions 82  at cell-cell contacts (Ikenouchi et al., 2007). Third, EpH4 cells are capable of forming functionally differentiated 3D spheroids in response to ECM and lactogenic hormones (Somasiri et al., 2000). Fourth, although the MCF10a human mammary cell line is more commonly used as a model of polarized mammary spheroid architecture, these spheroids only form when generated from a single cell embedded in ECM (Debnath et al., 2003), and surprisingly the resulting spheroids do not form tight junctions between cells (Fogg et al., 2005; Underwood et al., 2006).  To confirm the epithelial characteristics of EpH4 cells, I first analyzed the localization and expression of a panel of cell junction molecules. In 2D culture, the adherens junction proteins E-cadherin and β-catenin, and the tight junction molecules claudin-1 and ZO-1 are all present at sites of cell-cell interaction (Figure 3-1A). Vertical optical slices also confirmed that these cells form polarized junctions on a 2D substratum, since the tight junction proteins are localized apical to the adherens junction proteins (Figure 3-1B; arrows). It is notable, however, that in these conditions the cells lack a pronounced cuboidal cell shape and instead lie extremely flat.  It has recently been shown that Par3, Par6 and aPKC interact with one another to form an apical polarity complex in mammalian epithelial cells (Joberty et al., 2000; Lin et al., 2000). For example, in fully polarized MDCK cells Par3 has been shown to co-localize with tight junctions just beneath the apical membrane (Hirose et al., 2002; Hurd et al., 2003; Yamanaka et al., 2006). However, the localization patterns of the Par complex components have not yet been determined in EpH4 cells.  83  In monolayer culture, I found that Par6 was enriched at lateral cell-cell contacts (Figure 3-1A). Additionally, some Par6 appeared diffusely localized in the cytoplasm. In contrast, Par3, which was also found at sites of cell-cell interaction, specifically co-localized with the tight junction component ZO-1 at apical sites (compare Par6 with Par3, Figure 3-1B). Thus, 2D analysis suggested that there might be a physical separation of members of the Par complex in EpH4 cells.  It was previously shown that the α6 and β1 integrin subunits contribute to the laminin-dependent differentiation of the scp2 mouse mammary epithelial cell line (Muschler et al., 1999). Therefore, as a prelude to my own functional studies, I assessed the localization of these integrin subunits in EpH4 cells. In monolayer culture, both subunits were present and enriched at the basolateral membrane domain (Figure 3-1A).  84  Figure 3-1: Expression and localization of junction proteins in 2D monolayer culture of EpH4 cells EpH4 mouse mammary epithelial cells were maintained as confluent 2D monolayers in serum-free medium supplemented with lactogenic hormones ('differentiation media’; see Materials and Methods section 2.1.2 for details) for five days, fixed, stained for immunofluorescence and imaged by confocal microscopy. A) In XY-axis images adherens junction proteins (β-catenin, E-cadherin), tight junction proteins (ZO-1, claudin-1), polarity complex proteins (Par3, Par6) and cellECM junctions (α6 and β1 integrin) were all localized primarily at cell-cell borders. Scale=10µm. B) Z-axis images (apical/free surface upward) reveal that EpH4 monolayers were very flat, but polarized. The tight junction molecule ZO-1 was specifically localized in discrete puncta apical to the adherens junction protein β-catenin (upper panel). Additionally, ZO-1 colocalized in apical bars with the polarity complex molecule Par3 (lower panel). These images are representative of at least three separate experiments. Scale=5µm.  85  86  3.3 EpH4 cells polarize and functionally differentiate in an-ECM dependent 3D spheroid culture The flat morphology of the EpH4 cells cultured on a 2D substratum does not accurately reflect the morphology of mammary epithelial cells in vivo. Furthermore, cultured in this 2D context they also fail to undergo functional differentiation. Even in response to lactogenic hormones, cells cultured as 2D monolayers do not produce the milk protein β-casein (Figure 3-2A). In contrast, when EpH4 cells are cultured on top of a gelled layer of ECM (Matrigel), the cells cluster together, reorganize to form 3D cell spheroids with a central lumen (Figure 3-2A; arrow), and are capable of producing milk proteins (Figure 3-2A). This ECM-dependent differentiation has been demonstrated in both primary mammary epithelial cells (Barcellos-Hoff et al., 1989; Streuli et al., 1991) and in other functional mammary cell lines such as the mouse mammary cell lines HC-11 (Ball et al., 1988), CID-9 (Schmidhauser et al., 1992) and a clonal derivative of CID-9 cells designated the scp2 cell line (Roskelley et al., 1994).  Confocal optical sections through the center of differentiated EpH4 spheroids demonstrate that there was a dramatic polarization of the α6 integrin subunit to the basal membrane domain in response to ECM engagement (Figure 3-2B). However, the basal distribution of the β1 subunit is less prominent. As expected, the adherens junction proteins β-catenin and E-cadherin were localized along lateral sites of cellcell contact throughout the spheroid. In contrast, the tight junction proteins were localized apically where they caged the central lumen that forms in these spheroids. 87  Figure 3-2: EpH4 cells polarize and functionally differentiate in an ECMdependent 3D spheroid culture EpH4 cells were maintained as 2D monolayers on tissue culture plastic or as 3D spheroids “On Gel” (ie. placed upon a pre-gelled layer of Matrigel) for 5 days in differentiation medium (See Methods Section 2.1.2). β-casein expression and junctional protein localization were then assessed. All images shown are representative of at least 3 experiments. A) Left panel: Phase microscopic images of EpH4 cell monolayers on tissue culture plastic and spheroids 'On Gel'. The latter structures contained a phase dark central lumen (arrows and see below). Scale= 40µm. Right panel: Western blot analysis revealed that differentiation medium containing lactogenic hormones (insulin, hydrocortisone and prolactin) was able to induce βcasein expression in 'On Gel' spheroids, but not in flat 2D monolayers. The blot was re-probed for ERK1/2 as a protein loading control. B) Confocal images taken through the center of 'On Gel' spheroids show that while adherens junction proteins (β-catenin, E-cadherin) localized along the entire length of basolateral membranes between cells, tight junction proteins (ZO-1 and claudin1) and polarity complex proteins (Par3, Par6) were all localized exclusively at central, apical domains that surrounded a central luminal space. The ECM junction complex protein α6 integrin was localized exclusively along the outer, basal edge of the sphere, and the ECM junction complex protein β1 integrin was localized basally and laterally. Thus, the 'On Gel' spheroids were well polarized. Scale=10µm.  88  89  Importantly, both Par3 and Par6 also localized at these apical membrane domains, which is the first demonstration of the endogenous polarization of these polarity complex molecules in mammary epithelial 3D spheroids. Taken together, these observations indicate that EpH4 cells form organized functional spheroids with a central lumen and highly polarized cell junctions in response to exogenously added ECM. Thus, I reasoned that these cells would be amenable to in depth investigations designed to model the individual steps of junctional polarization in a 3D context.  3.4 Modeling apical orientation of cell junctions in 3D mammary spheroids independent of ECM-dependent cell shape changes Flat monolayers of cultured mammary epithelial cells are not capable of producing milk proteins in response to lactogenic hormones (Figure 3-2A; (Muschler et al., 1999; Roskelley et al., 1994; Streuli et al., 1991)). However, when these flat monolayers are overlaid with ECM, the cells engage with the matrix and undergo a morphogenic cell rounding process to ultimately form differentiated spheroids that are then capable of milk production (Muschler et al., 1999; Roskelley et al., 1994; Streuli et al., 1991; Streuli et al., 1995). This suggests that the surrounding ECM provides multiple signals; first the matrix instructs the cells to change shape, to remodel interactions with neighboring cells to form multi-cell spheroids, and it also signals the cells to differentiate and produce milk proteins.  To dissect out these processes initiated by the addition of ECM, I chose to model the morphogenic cell rounding by pre-clustering EpH4 cells in suspension cultures in 90  the absence of ECM (referred to as ‘naked clusters’; See Methods, section 2.1.2). Naked clusters of EpH4 cells form 3D ‘spheroid-like’ multi-cell clusters that can be subsequently treated with exogenously added ECM (referred to as ECM-overlaid clusters). To this end, this method effectively models the ECM-dependent effects on mammary epithelial cells independent of its ability to induce changes in cell shape. When assayed for functional differentiation, naked clusters of EpH4 cells were not capable of β-casein production, while ECM-overlaid clusters were (Figure 3-3A). Therefore, changes in cell shape alone are not sufficient to induce β-casein production. Instead, as has been demonstrated in the scp2 mammary epithelial cell line (Roskelley et al., 1994), EpH4 cells absolutely require additional signals from the ECM for this process.  When analyzed for cell junctions, naked clusters of EpH4 cells lacked a central lumen, and throughout the cluster, cells interacted with one another to form adherens junctions marked by E-cadherin and β-catenin localization at all sites of cell-cell contact (Figure 3-3B). Strikingly, the tight junction proteins claudin-1 and ZO-1 displayed a very restricted, punctate localization exclusively between cells that lined the exterior of the naked cluster. Furthermore, adherens junction proteins were found excluded from the free surface membrane domain (Figure 3-3B; arrowheads), and displayed very little co-localization with tight junction markers that lined the exterior of the cell cluster (Figure-3-3B; arrows). This suggests that even in the absence of ECM signals, naked clusters of EpH4 cells in suspension may assemble polarized cell-cell junctions and establish defined membrane domains, but that the “apical” pole is directed towards the exterior free surface.  91  In response to exogenously added ECM, however, cell clusters dramatically reorganized and formed a single cell layer that enclosed a luminal space. In fact, these ECM-overlaid spheroids strongly resembled the acinar structures that formed when EpH4 cells were directly placed “On gel” (Figure 3-2B). While the adherens junction proteins E-cadherin and β-catenin localized at lateral cell-cell contacts within ECM-overlaid spheroids, importantly, the tight junction proteins were now exclusively localized between cells at restricted puncta towards the interior of the spheroid and lined the newly established luminal free surface. Taken together, this suggested that even in the absence of ECM, naked clusters of EpH4 cells might have a hard-wired program to polarize apical cell junction complexes towards free surface membranes, which in this condition, are directed towards the culture medium (exterior polarity). In contrast, cell-ECM interactions provide a spatial signal to reorganize the cell cluster, generate a central lumen, and re-orient the direction of apicobasal polarity towards the luminal free surface membrane domain (interior polarity), as quantified in Figure 3-3C.  92  Figure 3-3: Dissecting mammary spheroid architecture independent of cell shape: Junctional assembly and polarization is ECM-independent, but apical orientation is ECM-dependent A) EpH4 cells were cultured in suspension to form pre-rounded cell clusters for 8 hours and were either re-plated onto a non-adhesive substrata (PolyHEMA coated tissue culture dishes) and maintained as ‘naked clusters’ (no ECM added) or treated with exogenously added ECM to form ‘ECM Overlay Spheroids’ (See Methods Section 2.1.2). After 5 days, representative phase micrographs demonstrated that EpH4 cells formed multi-cell clusters in suspension culture (naked clusters). Suspension clusters overlaid with ECM (ECM-Overlay Spheroids) formed organized 3D spheroids with smooth outer edges. Also note that after 5 days overlaid spheroids, but not naked clusters differentiated and expressed β-casein. The quantity of ERK1/2 was determined as a protein loading control. Scale=40µm. B) Naked clusters and ECM overlaid spheroids were immunostained and subjected to single slice confocal microscopy (0.23µm thickness) through the center of the structure. In naked clusters, adherens junction proteins (E-cadherin, β-catenin) were localized between cells throughout the cluster but were absent from the free, outer surface of the cluster (arrowheads). Tight junction proteins (claudin 1; ZO-1) were localized between cells, but concentrated at the exterior outer edge of the cluster (arrows), and showed little co-localization with adherens junction proteins. In ECM overlaid spheroids, adherens junction proteins were localized between cells that surrounded a central lumen, while tight junction proteins were localized between cells exclusively at the surface of the cell that faced the interior central lumen. Scale=10µm. 93  C) The orientation of polarity in naked clusters and ECM overlaid spheroids was quantified as a percent of spheroids that displayed the tight junction marker ZO-1 around the exterior of the 3D cell cluster/spheroid or towards an interior center lumen. Greater than 100 spheroids per condition were counted per experiment, and the graph depicts the averages obtained across three independent experiments +/SD.  94  95  During MDCK epithelial cell polarization in 2D culture after a 'calcium switch,' (ie. an artificial protocol where cell-cell junctions are disrupted and then allowed to reform) it is thought that pre-existing cell-ECM adhesions and newly formed cell-cell junctions together provide spatial cues for the sub-apical targeting of a polarity complex containing both Par3 and Par6 (Yamanaka et al., 2001). Later, when the MDCK cells are fully polarized Par3 dissociates from the complex and associates with tight junctions (Yamanaka et al., 2006). Interestingly, in naked clusters of EpH4 cells that are not in contact with the ECM, Par3 and Par6 did not co-localize (Figure 3-4). Par6 was localized along the entire exteriorly directed apical membrane domain while Par3 was specifically co-localized with ZO-1 at, presumably, sites of tight junction formation (Figure 3-4A). In ECM-overlaid EpH4 clusters with a central lumen the same separation applied; Par6 was localized along the entire internally directed apical membrane domain while Par3 was specifically localized with ZO-1 at tight junctions. These data further suggested that tight junctions and apical membrane domains may form in the absence of cell-ECM cues and that the latter cues serve to reorient these structures prior to central lumen formation. This was further investigated with time course experiments (see Figure 3-5 below).  Par 3 and Par6 also showed a differential susceptibility to Triton X-100 detergent extraction. A significant proportion of Par3, like the great majority of ZO-1, was present in the triton-insoluble fraction (ie. unextractable due to cytoskeletal association) while the great majority of Par6 was triton-soluble (ie. extractable due to the lack of cytoskeletal association). Strikingly, this differential extraction between Par3 and Par6, and implied difference in cytoskeletal association existed in both naked, (externally polarized) clusters and ECM-overlaid (internally polarized) 96  Figure 3-4: Par3 and Par6 exhibit distinct localization patterns and Par3, but not Par6, is present in a triton insoluble pool at the free surface of mammary spheroids A) EpH4 cells were cultured as naked clusters in suspension or as ECM overlaid spheroids and cultured for 5 days. Clusters/spheroids were then fixed and dual immunolabelled for Par3/ZO-1 or Par6/ZO-1. To analyze the localization of triton insoluble proteins, ECM-overlayed spheroids were treated with a buffer containing 1% triton x-100 for 15 minutes, and soluble proteins were washed away prior to fixation and immunostaining. Naked clusters of EpH4 cells were too weakly tethered to glass coverslips, and treatment with the triton buffer proved technically difficult for immunolocalization studies, so they were only analyzed biochemically. Single slice confocal images shown are representative of at least three experiments. Scale =10µm. B) EpH4 cells were cultured as described above, and sequentially lysed in a 1% triton-containing buffer and an SDS buffer to separate triton soluble and insoluble fractions respectively. Each fraction was then analyzed by SDS-PAGE followed by western blotting for the presence of ZO-1 and Par proteins. In both rounded clusters and ECM-overlaid spheroids, ZO-1 existed exclusively in the triton insoluble pool. Par3 was found in both fractions, although the 100kDa Par3 isoform was only detected in the insoluble pool. In contrast, Par6 was only associated with the soluble fraction. The data shown is representative of two independent experiments.  97  98  clusters (Figure 3-4B). This phenomenon could also be imaged in the ECM-overlaid clusters (Figure 3-4A, right panels). These observations further suggest that Par3, but not Par6, associates with cytoskeletally-linked tight junction complexes, and that these polarized tight junction complexes can assemble without ECMdependent directional orientation.  3.5 ECM-dependent apical orientation occurs prior to functional differentiation Naked clusters exhibited an exterior polarity and did not produce β-casein. However, clusters that were treated for 5 days with exogenously added ECM reoriented their polarity such that apical membrane domains and apically localized junctions re-localized towards an interior central lumen and signaled the cells to functionally differentiate (Figure 3-3 and 3-4). To more thoroughly investigate these responses to ECM addition I next assessed β-casein induction, lumen formation and spheroid re-polarization over time.  When EpH4 cells were overlaid with ECM, β-casein began to accumulate in the cells after three days and this accumulation was maximal after 5 days (Figure 3-5A). The induction was robust, as significant quantities were observable by Western blotting of whole cell lysates containing only 10 µg of total protein. In contrast, no β-casein was observable in naked cluster whole cell lysates containing 20 µg total protein. Thus, this provided further support that EpH4 cells absolutely required ECM for βcasein production.  99  Figure 3-5: The dynamics of ECM-dependent spheroid formation, β-casein production and apical re-orientation over time A) Western blot analysis of EpH4 cells cultured as either rounded clusters or as ECM overlaid spheroids for the indicated times, shows that the production of βcasein occurs only in the presence of exogenously added ECM, and is initiated at Day 3. The membrane was re-probed for ERK1/2 as a protein loading control. B) EpH4 cells were pre-rounded on PolyHEMA and treated with exogenously added ECM for the indicated times and visualized by phase microscopy. Lumen formation is clearly observed by Day3, and lumens are visible in most spheroids thereafter (arrowheads). The images shown are representative micrographs of at least three independent experiments. Scale=40µm. C) ECM overlaid spheroids were assayed for apical re-orientation by the localization of the tight junction marker ZO-1. Multiple confocal sections were projected in order to visualize the location of all of the immunolabelled ZO-1 throughout the spheroids. For ease of visualization, the approximate basal surface of the projected spheroids is marked with a white dotted line. Even at the early time points, ZO-1 appears to cage nascent lumens towards the center of the spheroid, and by Day 3, shows strong localization at cell-cell interaction sites exclusively surrounding a central apical lumen. Scale=10µm. D) Single confocal slices through the center of ECM overlaid spheroids triple labeled for ZO-1 (red), Par3 (green) and nuclei (blue) show that ZO-1 and Par3 continue to co-localize throughout the dynamics of apical re-orientation. Pearson’s correlation coefficients between the two channels (red and green) are shown. All images shown are representative of three individual experiments. Scale=10µm. 100  101  Lumen formation was monitored by live phase microscopy. Naked EpH4 cell clusters appeared uniformly phase bright with scalloped outer edges. The same was true of cell clusters that had been overlaid with ECM for 12 hr or 1 day (Figure 35B). After 2 days the clusters became smooth edged and some spheres had noticeably phase dark centers. These phase dark centers, which are indicative of lumen formation (see below) were very prominent after 3 days (Figure 3-5B, arrowheads) and expanded in size after 5 and 7 days.  ECM-mediated repolarization was monitored by assessing the tight junctionassociated protein ZO-1 (Figure 3-5C). Strikingly, at the early 12 hour time point ZO-1 was no longer localized in puncta at the outer free surface of the cluster (ie. compare 12 hr treatment with the naked cluster in Figure 3-4A). Instead ZO-1 relocalized to cell-cell junctions at the site of multiple nascent lumens towards the center of the re-organizing spheroid. After 1 and 2 days, the majority of the ZO-1 had re-positioned to the interior of the spheroid and the multiple small lumens observed earlier, were replaced by the appearance of a single, larger central lumen. This rapid re-organization, which occurs before β-casein induction, suggests that multiple regions of the spheroid re-orient their polarity and establish new apical free surfaces in response to ECM addition. Furthermore, it appears that multiple free apical surfaces coalesce into one free surface to form the single central lumen at Day 2/3 that were also observable by phase microscopy.  Interestingly, single optical sections through the center of ECM overlaid spheroids revealed that ZO-1 and the polarity protein Par 3 remain co-localized throughout the re-orientation process (Figure 3-5D; Pearson's correlation coefficients greater 102  than 0.75 indicate a high degree of co-localization). These observations suggest that, rather than disassembling and reassembling, tight junctions may re-orient as intact complexes.  In summary, upon the addition of ECM, naked EpH4 cell clusters were able to quickly re-orient their apicobasal polarity and initiate central lumen formation. Additionally, this new spheroid architecture was firmly established well before the production of β-casein. Thus, appropriately polarized mammary epithelial cell spheroid morphogenesis precedes lactational differentiation.  3.6 ECM enhances the basally polarized localization of integrins We next sought to determine the localization of basal components in the EpH4 polarization model. Previously, it was determined that the α6 and β1integrin components both contribute to the ECM dependent differentiation of mouse mammary epithelial cells (Muschler et al., 1999; Streuli et al., 1995). Therefore, we assessed the localization of these integrin subunits before and after the addition of the ECM overlay to EpH4 cells (Figure 3-6).  In naked EpH4 cell clusters with externally directed apical polarization (ie. see ZO-1 in left panels of Figure 3-6), both α6 and β1 integrin subunits were found within the clusters at sites of cell-cell interaction. Interestingly, both integrin subunits were found absent from the exterior free surface membrane (Figure 3-6; arrowheads). This supports the hypothesis that even in the absence of ECM, the exterior free surface membrane is distinct from basolateral cell membranes in these conditions. 103  In EpH4 cell clusters that were overlaid with ECM for 5 days (Figure 3-6 right panels) ZO-1 moved centrally to cage the central lumen, while the vast majority of the α6 integrin subunit moved to the external, now basal, surface of the spheroid (presumably to interact with ligands within the overlay). The relocalization of the β1 integrin subunit was more equivocal. While a proportion of the β1 integrin did relocate to the outer, basal surface of the ECM-overlaid spheroid, the majority of the β1 was located laterally, between cells. Importantly, there was complete segregation between the internally localized ZO-1 at the apical membrane and the basal (or basolateral) distribution of the integrin subunits in the overlaid clusters.  104  Figure 3-6: The α6 integrin subunit becomes exclusively polarized to the basal membrane of ECM-induced 3D mammary spheroids EpH4 cells were cultured in the same manner as described in Figure 3-3A, and were fixed and analyzed for the localization of β1 and α6 integrin. In the absence of ECM, rounded clusters of EpH4 cells display lateral localization of both integrin subunits, but their localization is excluded from the free surface (arrowheads). In response to ECM however, α6 integrin dramatically polarizes to the cell-ECM interface (arrows), while β1 integrin remains localized along the basolateral membrane. Scale=10µm.  105  106  3.7  The ECM component laminin-1 is sufficient to apically re-orient polarity in 3D mammary spheroids  It has previously been shown that multiple laminin receptors are important for both milk protein production and alveolar morphogenesis by mammary epithelial cells (Muschler et al., 1999; Streuli et al., 1991) (Streuli et al., 1995). Moreover, laminin-1 is an important component of the endogenously deposited basement membrane that interacts with mammary epithelial cells in vivo (Slade et al., 1999). Although Matrigel, which was used in the overlay experiments described above, is rich in laminin-1, it also contains a number of additional ECM components. Therefore, we overlaid naked EpH4 cell clusters with purified laminin-1 and determined that it was sufficient to induce robust apicobasal re-orientation of both ZO-1 and α6 integrin within 24 hours (ie. well before β-casein induction; Figure 3-7A). In contrast, EpH4 cells overlaid with purified collagen I did not reorient and ZO-1 remained on the outer surface (Figure 3-7B). Furthermore, the ability of laminin-1 to re-orient polarity was dominant. Specifically, if cells were first overlaid with collagen I for 24 hr and then subsequently overlaid with laminin-1 robust reorientation occurred (Figure 3-7C and quantification in D). These observations suggest that the ECM-dependant signals required for re-organizing mammary spheroid architecture are in part translated by laminin specific receptors.  107  Figure 3-7: Laminin-1, but not collagen I, is sufficient to re-orient apical polarity in 3D mammary spheroids A and B) Pre-rounded EpH4 cells were overlaid with either collagen I (A) or laminin-1 (B) and assessed for ZO-1 and α6 integrin localization. Collagen I produced spheroids with an inverted apicobasal polarity similar to that which is observed in cell clusters in the absence of ECM. In contrast, laminin-1 was sufficient to re-orient ZO-1 towards the center and polarize α6 integrin at the basal surface. C) EpH4 cells overlaid with collagen I for 24 hours were subsequently overlaid with laminin-1 for an additional 24 hours. The presence of laminin-1 could effectively rescue the percentage of spheroids that displayed an apical/luminal localization of ZO-1 as quantified in D). Scale=5µm. D) The orientation of polarity in overlaid spheroids was quantified as described in Figure 3-3C.  108  109  3.8 The α 6 and β 1 integrin subunits are important to apically reorient polarity in 3D mammary spheroids In response to laminin-rich ECM, EpH4 cell clusters dramatically re-position α6 integrin along the basal membrane at the cell-ECM interface. This suggests that the α6 integrin subunit is a likely candidate for transduction of the polarity reorientation signal downstream of ECM engagement. The α6 integrin subunit heterodimerizes with the β1 integrin subunit and together they act as receptors for laminin-1 (Mercurio, 1995). To test the requirement of these specific integrin subunits in re-orienting apical polarity, EpH4 cells were pre-clustered in the presence of α6 and/or β1 integrin function blocking antibodies and polarity reorientation after the addition of a basement membrane ECM overlay was assessed (Figure 3-8).  First, it is worth noting that neither the α6 or the β1 integrin function blocking antibody, or a combination of the two, prevented EpH4 cells from forming naked clusters in suspension culture (data not shown) suggesting that these integrin subunits are not required for cell-cell adhesion during the formation of multi-cell clusters, which is in contrast to other models of 3D epithelial spheroid reorganization (Yu et al., 2007b).  Second, the great majority of the 'no antibody' and the IgG control spheroids exhibited either a full or partial re-orientation of internal/apical ZO-1, and  110  external/basal β-dystroglycan (which was used in these experiments because the use of the integrin-directed function blocking antibodies interfered with integrin imaging) in response to the ECM overlay (Figure 3-8A). In contrast, both the α6 and β1 integrin function blocking antibodies alone each had a noticeable deleterious effect on apical/basal re-orientation (Figure 3-8B). These deleterious effects were enhanced when both α6 and β1 integrin function blocking antibodies were added together. This increased efficacy was clear from an examination of individual spheroids (Figure 3-8B) and from the scoring of multiple spheroids (Figure 3-8C). Interestingly, even when individual integrin-blocked spheroids were able to undergo a modicum of polarity re-orientation they most often formed small, multiple lumina rather than single central lumens. This suggests that blocking α6 and/or β1 integrin interferes with the signals downstream of laminin engagement that drive re-orientation of apical junctional polarity towards a nascent interior central lumen during ECM-dependent organization of mammary acini.  111  Figure 3-8: Both α6 and β1 integrin are important for ECM-dependant apical junction re-orientation In control conditions (A), EpH4 cells were pre-rounded in the absence (a’) or presence of control Rat IgG antibodies at a concentration equivalent to the individual integrin blocking conditions (a’’; 10µg/ml) or the combination integrin block (a’’’; 20µg/ml). To block functional integrins (B), EpH4 cells were prerounded in the presence of α6 or β1 integrin blocking antibodies individually (b’, b’’; 10µg/ml) or in combination (b’’’; 20µg/ml final concentration). After 4 hours, clusters were overlaid with ECM and polarity was assessed after 24 hours, and quantified based on the localization of ZO-1 from single slice confocal images (C). As a control for the presence of the blocking antibody, an example of the surface localization of the blocking or control antibodies at the time of polarity assessment is shown (D). Data shown is from one representative experiment of three. Scale =10µm.  112  113  3.9 Summary and Conclusions Conventional studies aimed at elucidating how epithelial cells establish apicobasal polarity often focus on cells cultured in vitro as 2D monolayers in which the basal membrane domain automatically forms at the point of cell contact with a rigid substratum, while the apical membrane forms on the free surface that faces the culture medium. In contrast, groups of epithelial cells in vivo are required to collectively integrate multiple signals to define membrane domain identity, establish apicobasal polarity and organize to form polarized 3D structures in the correct orientation. While it is clear that signals from neighbouring cells and signals from the surrounding microenvironment are both required for epithelial morphogenesis, the individual functions of cell-cell and cell-ECM interaction independently of one another in a 3D context are not well defined.  In the work presented here, I have characterized a unique 3D in vitro culture model of mammary epithelial morphogenesis that attempts to uncouple the role of cell-cell interactions and cell-ECM interactions in the assembly of polarized cell junctions, the orientation of apicobasal polarity and lumen formation. I demonstrate that in the absence of ECM, cell-cell interactions within suspended clusters of EpH4 cells (referred to as ‘naked’ clusters) were sufficient to assemble and polarize apical junctions between cells, but that the direction of the apical pole was oriented towards the exterior free surface membrane. In contrast, cell-ECM interactions (in ECM-overlaid clusters), acted as a directional cue to rapidly reorganize the multicell clusters and re-orient the apical membrane and apical junction complexes towards the interior of the reorganized cluster. Once apical polarity was reoriented  114  towards the interior of the spheroid, this subsequently lead to lumen formation and differentiative milk protein expression.  The initial formation of naked clusters is likely dependent on homophilic Ecadherin interaction upon cell-cell contact. Importantly, E-cadherin mediated adhesion between cells has been shown to provide a critical landmark that initiates membrane identity determination. For instance, forced expression of E-cadherin in non-polarized fibroblasts is sufficient to trigger membrane domain segregation and exclusive targeting of basolateral membrane proteins (McNeill and Nelson, 1990). Furthermore, early seminal studies using 2D monolayers of MDCK cells have shown that there is a close relationship between E-cadherin mediated adhesion and tight junction formation, since functional blocking of E-cadherin prevents tight junction re-assembly and polarization during junction recovery after a calcium switch (Gumbiner and Simons, 1986). Thus it seems likely that E-cadherin could be facilitating both polarized membrane domain segregation and functional tight junction assembly. However, these results were obtained in cells cultured in 2D, so it is difficult to determine whether this occurred independently of cell-substratum interactions.  To address this, EpH4 cells were cultured in the absence of ECM as suspended ‘naked’ clusters. In these conditions, I found that ECM was not required to segregate apical and basolateral membrane domains and assemble polarized junctional complexes that separated them. This was evident by the localization of apical junction components ZO-1, claudin-1 and Par3 at precise points between cells within these ‘naked clusters’ that are devoid of ECM. In addition, basolateral 115  proteins such as the adherens junction proteins E-cadherin and β-catenin, as well as the integrin subunits α6 and β1 were primarily excluded from the free surface. In this scenario however, the “apical” non-adhesive surface was polarized towards the exterior surface of the cluster and the basolateral membrane domain occupied all points of cell-cell contact. This implies that cell-cell adhesion may indeed be sufficient to signal polarized trafficking and the assembly of junctional complexes to elicit membrane domain identity in these cells.  Similar to the model described for EpH4 cells, MDCK kidney epithelial cells also form multi-cell clusters in suspension culture (Wang et al., 1990a, b). In support of the results I have shown for EpH4 cells, it has been previously shown that cell-cell contact between MDCK cells in suspended clusters is sufficient to cause exclusive segregation of both apical and basolateral and membrane domains (Wang et al., 1990a). These clusters also display polarized tight junction proteins between cells at the external periphery. However, there is an important caveat to the MDCK model. Suspension clusters of MDCK cells typically deposit the basement membrane components collagen IV and laminin I into a luminal space (Wang et al., 1990a). Thus it is important to realize that even in detached suspension cultures, the MDCK model poses significant challenges in identifying polarization signals downstream of cell-cell adhesion independently of potential cell-ECM interactions. In contrast, naked clusters of EpH4 cells form in the complete absence of ECM. As a result, the EpH4 cell model offers a unique advantage over the more commonly used MDCK 3D polarization model, since ECM signals can be entirely regulated exogenously.  116  To this end, the ECM-overlay technique (ie:adding exogenous ECM to naked clusters of EpH4), isolates the specific effects cell-ECM interactions have on the 3D architecture of naked cell clusters. I found that in response to ECM, the cells within the cluster quickly re-organized to re-orient the direction of polarity, as evident by the movement of the tight junction component ZO-1 from the exterior of the cluster towards the interior of the cluster. This process required a laminin-I rich matrix, and proper re-organization was dependent on both α6 and β1 integrin. Interestingly, collagen I was not sufficient to signal apicobasal re-orientation in EpH4 cells. This is in contrast with the MDCK cell model, since in response to a collagen overlay, clusters of MDCK cells form internally polarized 3D spheroids in a β1 integrin-Rac1 dependent manner (O'Brien et al., 2001; Ojakian and Schwimmer, 1994; Schwimmer and Ojakian, 1995; Yu et al., 2005). Inhibiting either β1 integrin or Rac1 function in MDCK spheroids results in an inversion of the apical pole from the interior towards the exterior surface, and is accompanied by a failure of spheroids to form a central lumen. Interestingly, these signals lie upstream of the ability of MDCK cells to deposit and organize extracellular laminin, as the inverted polarity and lack of lumen formation can be rescued with exogenously added laminin in either β1 integrin or Rac1 deficient spheroids (O'Brien et al., 2001; Yu et al., 2005). Taken together this provides further support for the importance of laminin in determining apicobasal orientation, but suggests that the EpH4 cell model (in which the cells do not deposit their own laminin) may be more amenable to teasing out the required signaling specifically downstream of laminin engagement.  117  An interesting aspect of this study was the observation that a portion of ZO-1 and Par3 co-localized throughout the ECM-driven reorientation process. It has been suggested that interior lumen formation of 3D epithelial structures may require endocytosis of small parts of the outer plasma membrane to form a vacuole inside the cell called the vacuolar apical compartment (VAC) (Mostov and MartinBelmonte, 2006). These vacuoles reportedly can fuse together to produce small transient lumina between cells, and/or may be actively transported to form the apical membrane that faces an internal hollow compartment (Vega-Salas et al., 1987; Vega-Salas et al., 1988). Thus lumen formation may occur as a result of VAC exocytosis. VAC’s have been described during MDCK spheroid formation (MartinBelmonte et al., 2007; Wang et al., 1994), and importantly, this idea has been confirmed recently in vivo in the development of blood vessel architecture in zebrafish (Kamei et al., 2006). It is unclear however, how junction complexes are relocated during spheroid polarization and lumen formation. The continued association observed between Par3 and ZO-1 suggest the intriguing possibility that apical junction complexes may not require complete disassembly (and subsequent reassembly) when the direction of polarity is re-organized. Perhaps intact junction complexes are internalized in VAC compartments and delivered to their new location by transcytosis. There is accumulating evidence to support that apical junction complexes are actively internalized, but it is not clear which pathway(s) are involved (macropinocytosis, clathrin-dependent, and/or lipid raft-dependent) particularly in 3D epithelial morphogenesis (Ivanov et al., 2005). It will be of interest to test the requirement of these different internalization pathways for the reorganization of apical junctions as well as for interior lumen formation.  118  Of note, a recent report has shown that Par3 acts independently of the Par6/aPKC complex in establishing tight junction formation, while the Par6/aPKC complex was critical for de novo lumen formation in MDCK spheroids (Martin-Belmonte et al., 2007). Thus, it will be interesting to test the requirement of Par3 for the interior movement of ZO-1 in responses to ECM in EpH4 cells. The observed differences in the specific distribution of Par3 compared to the Par6 complex lend support to the potential divergent roles of these Par proteins. Overall, the EpH4 model described in this Chapter provides an attractive system to investigate these possibilities in an effort to further understand epithelial polarization in a 3D context.  119  CHAPTER 4 : CHARACTERIZATION OF PODOCALYXIN EXPRESSION AND FUNCTION IN MAMMARY EPITHELIAL AND BREAST TUMOR CELLS 4.1 Introduction Podocalyxin is a sialomucin molecule expressed at the cell surface of a number of specialized epithelial cells. At the outset of this study, podocalyxin had been primarily characterized as an anti-adhesin that is critical for the unique cell architecture of differentiated kidney podocytes. In collaboration with Dr. Kelly McNagny, Dr. Blake Gilks and Dr. David Huntsman, our lab had determined that podocalyxin is also moderately expressed at the apical surface of luminal mammary epithelial cells. Additionally, we found that high overexpression and mislocalization of this molecule occurs in a distinct subset of invasive breast carcinomas, and that this overexpression correlated with poor patient outcome independently of common prognostic indicators such as lymph node metastasis and HER2/neu amplification (Somasiri et al., 2004). This suggested that podocalyxin overexpression might be useful as a clinical biomarker to identify a unique group of invasive breast tumors that are likely to progress to metastatic disease.  The clinical significance of podocalyxin overexpression observed in invasive breast cancer served as a rationale to determine the biological role of podocalyxin in the mammary gland. Therefore, I sought to further characterize the normal expression  120  pattern of podocalyxin during mammary gland development, and to determine the biological consequences of forced overexpression of podocalyxin in both normal mammary epithelial cells and breast tumor cells.  At the time, I formulated the following hypotheses: 1) As the mammary epithelium undergoes dramatic morphological changes during post-natal development that are driven, at least in part, by adhesive changes, the expression of podocalyxin may be dynamically regulated during this process.  2) Overexpression of podocalyxin may alter normal mammary epithelial cell adhesion in vitro, and disrupt the organization of polarized cell junction complexes during mammary morphogenesis.  3) Overexpression of podocalyxin may disrupt epithelial cell adhesion and architecture of non-invasive breast tumor cells, which could function to increase their motility and tumorigenic potential (I also address this hypothesis in more detail in Chapter 5).  Based on these hypotheses, I initially expected that Podocalyxin overexpression would target the integrity of cell-cell junctions of both mammary epithelial and breast tumor cells, which would result in decreased cell-cell adhesion and a disruption of cell polarity. However, as the work presented in this chapter shows, podocalyxin may not directly impact cell-cell junction complexes. Instead, based on the data generated by the experiments performed, I learned important general  121  insights about podocalyxin’s function specifically at the apical membrane of epithelial cells.  Overall I found that podocalyxin is preferentially targeted to any part of the cell surface not in contact with a neighbouring cell or a basal substratum. Furthermore, podocalyxin's presence along the membrane may play a general role in defining the structure of free surface, ‘non-adhesive’ membrane domains. Interestingly, I found that overexpression of podocalyxin enlarged these free membrane surfaces, but did not interfere with the polarity of established cell-cell junctions, or mammary morphogenesis. Importantly, I found that while podocalyxin did not interfere with the overall architecture of normal mammary epithelial cells, it did perturb the surface morphology of 3D breast tumor spheroids in cells that have already lost the ability to respond appropriately to polarity cues provided by a surrounding basement membrane ECM. A detailed examination of the cell surface of these tumor cells revealed that overexpression of podocalyxin induced actin-rich microvillus formation at the apical membrane, and caused a significant restructuring of the actin cytoskeleton even in single cells. Podocalyxin also recruited the scaffolding protein NHERF-1 and the ERM-family member ezrin to this expanded apical surface, and these three proteins were highly co-localized together with actin along individual microvilli. Taken together, this suggests that podocalyxin may play an organizational role in modulating the structure of nonadhesive membrane surfaces. Furthermore, this function may have important consequences particularly in breast tumor cells that have lost polarity, such that their free surface membranes are no longer restricted to lining interiorly polarized lumens. 122  4.2 Podocalyxin expression during mammary gland development in vivo and morphogenesis in vitro As a prelude to in vitro functional overexpression studies, I sought to determine the expression pattern of podocalyxin during different stages of mouse mammary gland development. During kidney development, podocalyxin is a critical molecular regulator of the unique architectural changes that occur during podocyte differentiation. Specifically, podocalyxin is required for the complex morphogenic remodeling and cell junctional rearrangements that result in slit diaphragm formation during the generation of the urinary filtration apparatus (Doyonnas et al., 2001). Since the development of the mammary epithelium and the adult cycles of pregnancy, lactation and involution all involve significant morphological remodeling; I sought to characterize podocalyxin expression and localization at different stages of these processes in vivo. I then examined podocalyxin expression and localization using the in vitro 3D model of ECM-dependent EpH4 mammary epithelial morphogenesis described previously in Chapter 3.  4.2.1  Podocalyxin is expressed and apically localized in adult mammary glands  To determine the expression pattern of endogenous podocalyxin in virgin mammary glands, I obtained tissue specimens from 4 and 12 week old FVB mice and performed immunohistochemistry (See Methods Section 2.2.1 and 2.2.2). For immunostaining specificity, kidney sections prepared from 12 week old adult mice were used as a positive control. As expected (Kerjaschki et al., 1984), the latter 123  sections showed a strong localization of endogenous mouse podocalyxin specifically in the glomerular epithelium (Figure 4-1A; arrowheads) and the surrounding vasculature (Figure 4-1A; arrows), but not in the adjacent renal tubules. In the virgin mammary glands, podocalyxin exclusively stained the apical surface of luminal ductal epithelium both in the pubertal gland, and in the resting adult gland (Figure 4-1B). This apical staining was also observed in the luminal epithelium of the normal human breast (Figure 4-1C; (Somasiri et al., 2004)). Thus, the glycoprotein podocalyxin is a bona fide marker of the apical membrane domain in mammary luminal epithelial cells.  4.2.2  Podocalyxin is not localized in mammary epithelial cells during lactation  The mammary gland undergoes dramatic changes in tissue architecture throughout pregnancy cycles. To determine whether the expression pattern and/or localization of podocalyxin is altered during this process, I performed immunohistochemical analysis on mammary tissue obtained from female mice at different stages of pregnancy, lactation and involution (Figure 4-2A-C). Although podocalyxin was present in the apical portion of luminal epithelial cells at Day 7.5 of pregnancy (Figure 4-2A’; arrowhead), by mid-pregnancy (Day 14.5), podocalyxin was scarcely visible in the alveolar epithelium, and was only strongly observed continuously along the surface of vascular endothelia (Figure 4-2A’’; arrow).  124  Figure 4-1: Podocalyxin is localized to the apical membrane in luminal mammary epithelial cells Immunohistochemistry was performed to detect Podocalyxin expression and localization in mouse and human tissue. A) Podocalyxin was strictly localized to the glomerular epithelium (arrowheads) and lining the apical surface of vascular endothelium (arrows) in adult mouse kidney sections. Equal concentrations of a control IgG antibody showed no immunoreactivity confirming the specificity of the podocalyxin staining with the anti-mouse podocalyxin antibody (also used in panel B; R&D Systems). B) Mammary glands from 4 week and 12 week old FVB mice were excised, processed and analyzed for podocalyxin immunolocalization. In both the 4 week and 12 week old mouse mammary glands, podocalyxin primarily localized to the apical surface of luminal mammary epithelial cells (see enlarged inset). C) Apical localization of podocalyxin was also observed in the luminal epithelium of human mammary gland tissue obtained from a breast reduction patient. Immunolocalization of human podocalyxin was detected with the anti-human podocalyxin mouse monoclonal antibody (clone 3D3; a kind gift from Dr. David Kershaw) and a mouse serum control was used to control for antibody specificity and showed no immunoreactivity. Scale = 100µm  125  126  During lactation, the mammary gland becomes a fully differentiated secretory organ, whereby the alveolar epithelium secretes copious amounts of milk and lipids as evident by visible fat droplets inside the enlarged luminal spaces (Figure 4-2B’; arrow). Similar to the mid-pregnant stage, neither at the onset of lactation (Figure 41B’) nor at later stages (Figure 4-B’’), was podocalyxin present within the mammary epithelium based on immunohistochemical staining. To cease lactation, the suckling pups were removed from the mother’s cage, which effectively triggers the mammary gland to involute and undergo massive restructuring in order to return to a resting state. By Day 4 of this process, the large majority of the secretory epithelium had collapsed, and the ductal epithelium appeared to be in a remodeling phase. Only trace levels of podocalyxin were visible at this stage, however it was not clear if it was localized at the membrane surface (Figure 4-2C’; arrowhead). At a later stage of involution, the presence of podocalyxin was once again more prominent and appeared to have regained an apical localization in some small remodeled ducts (Figure 4-2C’’, arrowheads). Taken together, this analysis confirms that in the resting mammary gland, podocalyxin primarily localizes to the apical surface of luminal cells, but raises the possibility that podocalyxin expression may decrease during lobuloaveolar differentiation that occurs during pregnancy and lactation.  127  Figure 4-2: Podocalyxin is not detected in differentiated lobuloalveolar epithelium of the mouse mammary gland during pregnancy and lactation Immunohistochemistry was performed to detect Podocalyxin immunolocalization on FVB mouse mammary glands at different stages of (A) pregnancy, (B) lactation and (C) involution. Single arrows indicate positive immunoreactivity for podocalyxin in adjacent vascular endothelium as an internal control. A) Podocalyxin was detected at the apical surface of luminal mammary epithelial cells at Day 7.5 of pregnancy (A’), but was no longer detected in glands obtained from mid-pregnant mice (A’’; Day 14.5). B) Mammary glands obtained from post-partum mice were in a secreting state as evident by fat droplets inside expanded lumens (B’; double arrow). No podocalyxin protein was detectable within the mammary luminal epithelium during lactation at either an early (B’; Day1) or late stage (B”; Day 9). C) Trace levels of podocalyxin immunopositivity were detectable in involuting mammary glands at Day4 (C’; arrowhead), and appeared more prominent in small luminal ducts at Day 14 of involution (C”; arrowhead). Scale = 100µm.  128  129  4.2.3  Podocalyxin protein decreases during ECM-dependent morphogenesis and differentiation in vitro, but is not transcriptionally down-regulated  Given the observed loss of podocalyxin in the mammary epithelium during lobulolveolar morphogenesis and differentiation in vivo, I sought to determine whether this downregulation was occurring at the transcriptional level. Here I took advantage of the 3D, ECM-dependent in vitro model of mammary epithelial morphogenesis using EpH4 cells that not only recapitulates the morphogenesis and differentiation of mammary epithelial cells that occurs during pregnancy/lactation in vivo, but also is amenable to a precise quantification of protein and RNA analysis of the mammary epithelial cells themselves (ie. such quantification is very difficult in vivo as the proportion of epithelium to stroma varies greatly during pregnancy and lactation).  In the presence of lactogenic hormones, I found that EpH4 cells cultured as undifferentiated 2D monolayers expressed podocalyxin protein (Figure 4-3A). However, when EpH4 cells were placed on a gelled layer of Matrigel (On Gel) long enough to undergo alveolar morphogenesis and functionally differentiate and express β-casein, podocalyxin protein was no longer present in the cells (Figure 43A). This result agreed with the observations in vivo, that podocalyxin protein is lost during lactational mammary epithelial differentiation.  130  Figure 4-3: Podocalyxin protein is absent during in vitro differentiation of EpH4 cells but this is not a result of transcriptional down-regulation A) EpH4 cells were cultured for 5 days in serum free differentiation media (see Materials and Methods Section 2.1.2 for details) as either flat 2D monolayers or as “On Gel” differentiated mammary spheroids grown on a gelled layer of Matrigel and lysed for western blot analysis. Only the EpH4 cells cultured as mammary spheroids were capable of functional differentiation as evident by the presence of the milk protein β-casein. While podocalyxin was present in EpH4 cells cultured as 2D monolayers, podocalyxin protein was no longer detected in differentiated spheroids. Data shown is representative of three independent experiments. B) Total RNA was isolated (using Trizol) from EpH4 cells cultured in serum free differentiation media for 5 days as either flat 2D monolayers or as “On Gel” differentiated mammary spheroids. Semi-quantitative RT-PCR revealed that podocalyxin transcript expression is present in both 2D monolayers and in differentiated 3D spheroids where protein was not detected (see panel A). Primers for β-actin were used as a control for both the relative quality and quantity of RNA. Data shown is representative of three independent experiments. C) Total RNA suitable for quantitative RT-PCR was isolated using the RNeasy RNA isolation kit from EpH4 cells cultured as either 2D monolayers or “On Gel” differentiated spheroids. In accordance with the semi-quantitative RT-PCR analysis (B), there was no statistical difference in podocalyxin transcript expression between EpH4 cells cultured as 2D monolayers or 3D mammary spheroids as determined by a student’s t-test (p=0.42). Data shown is representative of two independent experiments. 131  132  Surprisingly, however, when I analyzed the levels of podocalyxin mRNA present in EpH4 cells, I found that podocalyxin transcript was present in cells cultured as 2D monolayers and as lactationally-differentiated 3D spheroids. This suggested that although podocalyxin protein appeared to be absent during functional lobuloalveolar differentiation in vivo and in vitro, this difference may not be caused by transcriptional down-regulation.  4.2.4  Podocalyxin protein is secreted during pregnancy/lactation  During lobuloalveolar differentiation, the mammary epithelium becomes capable of secreting milk apicially into a central lumen. Although milk ejection is a hormonally regulated process initiated at parturition, several reports have indicated that milk production and secretion products are present in the mammary gland during pregnancy (Bauman et al., 2006). Interestingly, closer examination of longitudinal sections through mammary ducts in the pregnant mouse gland revealed strong podocalyxin immunopositivity within ductal lumens, while still absent within adjacent alveolar epithelium as initially observed (Figure 4-4A). Thus, podocalyxin protein may be lost in mammary secretions. Additionally, podocalyxin protein was found to be present in human milk (Figure 4-4B). Taken together, these observations suggest that podocalyxin protein may not be detectable in the epithelium during mammary gland differentiation, not due to a decrease in expression, but because it may simply be lost from the apical membrane during the secretory phase of differentiation. Of note, this is, to my knowledge, the first report to identify the apical glycoprotein podocalyxin as a molecular component present in milk.  133  Figure 4-4: Podocayxin is detected within the lumens of primary ducts of mouse mammary glands, and is present in human breast milk A) Mouse mammary glands obtained from pregnant female mice at Day 7.5 and Day 14.5 (the same samples from Figure 4-2A) were sectioned further to reveal longitudinal cross-sections through large primary ducts. Immunohistochemistry was performed to detect podocalyxin expression and strong immunoreactivity was observed inside the lumens of mammary ducts at both stages of pregnancy (arrows). Scale = 100µm. B) A sample of human milk was obtained and analyzed for the presence of podocalyxin by SDS-PAGE and western blotting. Podocalyxin protein was present in both skim and cream human milk fractions. Of note, the form of podocalyxin present in the milk sample ran at a higher molecular weight, potentially indicating different glycosylation compared to the predominant form present in the MDA-MB 231 human breast tumor cell line used as a positive control for podocalyxin expression.  134  135  4.3 Functional analysis of podocalyxin overexpression in normal mammary epithelial cells and breast tumor cells Prior to this work, there was only one published study analyzing the functional effects of podocalyxin overexpression in vitro. In that study, Marilyn Farquhar’s group overexpressed podocalyxin in the MDCK kidney epithelial cell line and found that podocalyxin inhibited cell-cell aggregation in suspension and modified cell junctions in epithelial monolayer culture (Takeda et al., 2000). To build upon podocalyxin’s proposed functional role in regulating epithelial cell morphology, I overexpressed podocalyxin in EpH4 normal mouse mammary epithelial cells and the non-invasive MCF-7 human breast tumor cell line.  4.3.1  Podocalyxin overexpression does not prevent epithelial cell polarization during normal differentiative mammary morphogenesis  Normal EpH4 cells were transfected with either the vector control or murine podocalyxin cDNA, and the resulting stable populations were FAC sorted for high levels of podocalyxin overexpression (See Materials and Methods Section 2.6.2). These sorted pooled populations were then analyzed for total expression of podocalyxin by western blot analysis (Figure 4-5A). Clearly, EpH4 cells force expressing podocalyxin were able to glycosylate the core protein as it formed a molecule of approximately 145kD. Morphological analysis indicated that  136  podocalyxin overexpression did not cause any marked disruption to epithelial monolayers in 2D culture, or 3D spheroid formation on reconstituted basement membrane gels (Figure 4-5B). Additionally, podocalyxin overexpression did not impede ECM-dependent differentiation as β-casein production was unchanged compared to parental EpH4 (WT) or the vector control EpH4 population (Figure 45C).  137  Figure 4-5: Forced expression of podocalyxin in EpH4 cells does not disrupt mammary epithelial morphology in 2D or functional differentiation in 3D Parental EpH4 cells were transfected with murine podocalyxin (pβ-podo) or control plasmid (pβ) and stable populations were generated (see Materials and Methods section 2.2.6.2 for more details) and analyzed for functional differentiation. Data shown is representative of at least three independent experiments. A) Subconfluent vector control cells (Control) and podocalyxin overexpressing EpH4 cells (Podocalyxin) were lysed and analyzed for podocalyxin expression. While control cells had low levels of endogenous mouse podocalyxin, there was a significant increase in mouse podocalyxin expression in the population containing the ectopic transgene. B) Control and Podocalyxin overexpressing EpH4 cells were cultured in differentiation media for 5 days either as 2D flat monolayers or as 3D “On Gel” mammary spheroids. Representative phase micrographs show little differences in morphology in monolayer culture (compare left panels). Similarly, the 3D mammary spheroids in both Control and Podocalyxin cells were similar in size, shape, and phase-dark central lumens were observed in spheroids from both cell populations (compare right panels; arrows). Scale = 40µm. C) Parental EpH4 (WT), Control and Podocalyxin cells were cultured as described in panel (B) and lysed for western blotting analysis. Both the Parental and Control EpH4 cells were able to functionally differentiate, and the presence of podocalyxin overexpression did not appear to greatly alter this process as evident by unchanged production of the milk protein β-casein. Equal protein loading was determined by re-probing the membrane for β-actin. 138  139  I also investigated whether forced expression of podocalyxin altered 3D spheroid polarization and apical re-organization using the ECM overlay assay (See Materials and Methods section 2.1.2 and Chapter 3). Like control cells, in the absence of ECM, EpH4 cells overexpressing podocalyxin were able to adhere to one another and form naked clusters in suspension culture (Figure 4-6A). It is worth noting, however, that I did observe a mild delay in initial multi-cell cluster formation, which suggests that podocalyxin may initially interfere with cell-cell interactions in static suspension culture, but in these conditions, the inhibition appeared to be quickly overcome. (I address this ‘anti-adhesion’ effect in suspension more rigorously in Chapter 5). Once naked clusters formed, podocalyxin overexpression did not appear to affect the localization of basolateral E-cadherin (not shown) or the polarized distribution of ZO-1, which was oriented towards the exterior of the cluster (Figure 4-6A). Interestingly, when assessed for podocalyxin localization, cells that lined the outside of naked clusters displayed strong podocalyxin localization along what appeared to be ‘bulging’ free surface membrane domains exclusively localized apical to the tight junctions between cells (Figure 4-6B; double arrow). Additionally, cells that were embedded within the interior of the cluster (with all membrane surfaces in contact with neighbouring cells) showed little membrane localization of podocalyxin. Instead, podocalyxin often appeared improperly targeted and intracellular in these cells where no free membranous surface was present (Figure 4-6A, arrow).  When naked cell clusters were overlaid with ECM, even at 12 hours the large majority of podocalyxin no longer localized to the outer membrane of the cluster where cells were in contact with the surrounding ECM (Figure 4-6B). Instead, 140  Figure 4-6: Forced expression of podocalyxin preferentially localizes to the outer surface of naked clusters and does not alter ECM-dependent signals to re-orient apical polarity in 3D overlay mammary spheroids A) Vector control and podocalyxin expressing EpH4 cells were cultured as naked clusters in serum free differentiation media for 5 days and were processed for immunocytochemistry. Single confocal slices through the center of the clusters revealed that ZO-1 localized to discrete puncta between cells around the exterior of the cluster in both cell types. While the control cells showed low levels of endogenous mouse podocalyxin (arrowhead), the forced expressed podocalyxin was found in the cytoplasm of cells that lacked a free surface membrane domain (arrow), and strongly localized to bulging free surfaces facing the culture medium (double arrow). Scale = 10µm. B) Vector control and podocalyxin expressing EpH4 cells were pre-rounded and overlayed with ECM (Matrigel) for the indicated times and were assayed for apical re-orientation by the localization of the tight junction marker ZO-1. Single confocal slices through the reorganizing spheroids showed polarization of ZO-1 towards the interior of the spheroid in both Control and Podocalyxin cells. Forced expression of podocalyxin did not disrupt this process. In contrast, it readily relocated to nascent apical membrane domains in the interior of the spheroid. Scale = 10µm.  141  142  podocalyxin appeared disorganized and often diffuse in the cytoplasm. After 1 Day, it was clear that ectopic podocalyxin did not prevent the ECM overlay-mediated internalization and apical reorganization of ZO-1. Interestingly, podocalyxin overexpression may have actually promoted the biogenesis of the interior apical membranes, as interior lumen formation and expansion appeared to occur more rapidly in these clusters in comparison to controls (Figure 4-6B; Day 1 to Day 3), but this observation was difficult to quantify. Force expressed podocalyxin also localized to these nascent internal apical membranes and these membrane domains were observed to be properly separated by tight junctions between cells marked by ZO-1 localization, which resembled control spheroids. Altogether, this analysis suggests that podocalyxin overexpression did not prevent tight junction formation; instead it preferentially localized to free surface membrane domains exclusively apical to the tight junction complex. Furthermore, podocalyxin overexpression did not interfere with ECM-dependent signals controlling polarity orientation during normal mammary spheroid morphogenesis. Instead, it may have facilitated it.  4.3.2  Podocalyxin overexpression alters the architecture of breast tumor cell spheroids without altering junctional polarity  Upon ECM overlay, naked clusters of normal mammary epithelial cells overexpressing podocalyxin rapidly organized into properly polarized 3D structures, and readily formed new internal free surface membranes and generated a central lumen equally as well, if not better, than control cells (Figure 4-6). This surprising observation lead me to ask whether podocalyxin overexpression would facilitate de novo lumen formation in 3D breast tumor cell spheroids. 143  When breast tumor cells are cultured in the presence of ECM, they often do not form the characteristic alveolar architecture typical of normal mammary epithelial cells. For example, when non-invasive MCF-7 human tumor cells are cultured in Matrigel, they aggregate to form solid spheroids that do not polarize appropriately and form a central lumen (Kirshner et al., 2003; Wang et al., 2002). Thus, I utilized this model to test the ability of podocalyxin to induce the formation of new free surface membranes and promote lumen formation.  MCF-7 cells are also amenable to overexpression studies because they express low levels of endogenous human podocalyxin (Somasiri et al., 2004). MCF-7 cells were stably transfected with the murine cDNA of Podocalyxin or control vector, FAC sorted to enrich the transgenic population, and the resulting cells was assessed for total expression of ectopically expressed mouse podocalyxin with a species specific monoclonal antibody (Figure 4-7A). As expected, when the control cells were cultured on reconstituted basement membrane gels (Matrigel) they formed solid 3D structures that lacked a central lumen (Figure 4-7B; left panel). Interestingly, ZO1 primarily localized between cells around the exterior of the tumor spheroids suggesting that unlike normal mammary epithelial cells, MCF-7 breast tumor cells were incapable of responding to architectural and polarity cues from the lamininrich basement membrane. Thus, MCF-7 cells exhibited an inherent polarity cue defect in 3D culture.  144  Figure 4-7: Forced overexpression of podocalyxin is not sufficient to induce interior lumen formation in 3D breast tumor cell spheroids MCF-7 breast tumor cells were transfected with either murine podocalyxin (pIRESpodo) or the empty vector (pIRES) as a control and the resulting stable populations were plated on a gelled layer of Matrigel to assess potential lumen formation and orientation of polarity as determined by the localization of the tight junction protein ZO-1. A) Subconfluent vector control cells (Control) and podocalyxin overexpressing MCF-7 cells (Podocalyxin) were lysed and assessed for mouse podocalyxin expression using a species specific monoclonal antibody that only recognizes the ectopic mouse transgene. B) Single confocal sections through the center of Control MCF-7 breast tumor cell spheroids showed no indication of lumen formation, and ZO-1 localized around the exterior surface (left panel). Forced expression of podocalyxin did not induce interior lumen formation, and ZO-1 localization was also exteriorly polarized (right panel). However, the tumor cell spheroids that formed with the podocalyxin transfectants typically appeared less cohesive compared to MCF-7 control cells (compare left and right insets) and exhibited bulging outer free surface membranes where the ectopic podocalyxin had localized. The images shown are representative of at least three independent experiments. C) The polarity of the tumor cell spheroids were quantified on the basis of exterior or interiorly localized ZO-1 localization. The majority of tumor cell spheroids exhibited an external polarity, which was not affected by podocalyxin  145  overexpression. A small percentage of spheroids exhibited both internally and externally localized ZO-1. The quantitative data shown is the averages obtained across three independent experiments +/- SD.  146  147  MCF-7 cells overexpressing podocalyxin also aggregated and formed 3D structures on basement membrane gels (Figure 4-7B; right panel). Forced podocalyxin expression in these cells did not induce de novo central lumen formation and it also failed to alter the abnormal, outward orientation of tight junctions, as ZO-1 was primarily localized around the outside of the tumor cell spheroid in a manner similar to control cell spheroids (Figure 4-7B and quantified in Figure 4-7C). Importantly, however, podocalyxin overexpressing MCF-7 cells consistently formed less compact, more irregularly shaped aggregates, such that the surface of the spheroid often appeared ‘ragged’ and multi-lobular compared to the smooth edges of control spheroids (Figure 4-7B; compare phase insets of representative spheroids in live culture). Furthermore, the ectopically expressed podocalyxin localized to these ‘ragged’ bulging outer membranes of the spheroid. Taken together, this suggests that while podocalyxin was not sufficient to instruct cells to dramatically reorganize to form new free surface membranes, it induced a unique morphogenic perturbation to 3D tumor spheroid architecture without influencing junctional polarity. Given the localization of podocalyxin, I hypothesized that it may exert these morphological effects by influencing the structure of the free membrane surface.  4.3.3  Podocalyxin causes apical membrane expansion in polarized kidney epithelial cells and breast tumor cells  I next sought to take a closer examination of the morphogenic effects of podocalyxin specifically at the cell surface in MCF-7 cells. MCF-7 cells maintained in monolayer  148  culture form a cuboidal epithelial monolayer with very discernable apical, lateral and basal cell surfaces. This makes them amenable to an image-based analysis of distinct cell surface membrane domains using 3D confocal microscopy. To confirm selected findings in a second cell type, I also stably transfected murine podocalyxin into the MDCK kidney epithelial cell line and confirmed ectopic transgene expression by Western blotting (Figure 4-8A). Like MCF-7 cells, MDCK cells also form cuboidal epithelial monolayers (Figure 4-8B) and are suitable for this type of imaging analysis. Since podocalyxin has been shown to link indirectly to the actin cytoskeleton (Orlando et al., 2001; Takeda et al., 2001), I first analyzed the consequence of podocalyxin overexpression on f-actin localization at membrane surfaces. Both MCF-7 and MDCK transfectants were maintained in monolayer culture until they reached confluency, and were fixed and dual labeled for ectopic podocalyxin and factin. Confocal images along the Z-axis showed that control cells formed very uniform epithelial monolayers and f-actin was evenly distributed along the apical and basolateral membranes. Both MDCK cell and MCF-7 cells that expressed ectopic podocalyxin showed a distinct apical expansion above surrounding nonexpressing cells. Podocalyxin localized to this expanded free surface membrane and showed potential co-localization with f-actin at the apical surface (Figure 4-8B; orange/yellow at apical membrane in merged image).  We and others have shown that podocalyxin interacts with the PDZ domaincontaining NHERF adaptor proteins via a PDZ-binding domain at the C-terminus of the cytoplasmic domain of podocalyxin (Li et al., 2002; Orlando et al., 2001; 149  Takeda et al., 2001; Tan et al., 2006). NHERF-1 molecules act as promiscuous scaffolding proteins (Weinman et al., 2005) and can connect podocalyxin to the actin cytoskeleton through ezrin (Orlando et al., 2001; Takeda et al., 2001). To assess the localization of NHERF-1, confluent monolayers of MCF-7 cells were immunolabelled for ectopic podocalyxin and endogenous NHERF-1 then analyzed by confocal microscopy (Figure 4-8C). 3D confocal stacks were reconstructed and rotated to clearly reveal the surface of the stained monolayer, while displaying a single slice view along the Z-axis (as outlined in the 'blue box' at the front of the merged images). MCF-7 control cells exhibited a relatively flat apical, free surface, and displayed some endogenousNHERF-1 protein near that surface. However, the large majority of NHERF-1 appeared diffuse in the cytoplasm (Figure 4-8C; left panel). In contrast, MCF-7 podocalyxin overexpressing cells bulged apically out of the monolayer and podocalyxin appeared concentrated in discrete puncta at the apical surface of the bulging membranes (Figure 4-8C; right panel, arrow). Importantly, in cells overexpressing ectopic podocalyxin there was a dramatic increase in apically localized endogenous NHERF-1 and a concomitant decrease in cytoplasmic staining of the scaffolding protein. At the bulging apical membrane surface, the two molecules were highly co-localized (Figure 4-8C; merged image; Also see Supplemental Movie 4-1). In addition, the ability of podocalyxin to alter NHERF-1 localization was not due to changes in endogenous NHERF-1 expression levels (Figure 4-8D). Thus, podocalyxin expanded free, apical membrane domains and it recruited its binding partner, NHERF-1, to these domains.  150  Figure 4-8: Podocalyxin causes apical domain expansion in MCF-7 and MDCK cells and recruits its intracellular binding partner NHERF-1 to the apical membrane Similar to the MCF-7 cells, parental MDCK cells were stably transfected with either murine podocalyxin (pIRES-podo) or the empty vector (pIRES) as a control (as in Figure 4-7). A) Whole cell lysates of the MDCK transfectants (Control and Podocalyxin) were assessed for ectopic mouse podocalyxin expression by western blotting using a species-specific antibody. B) MCF-7 and MDCK transfectants were grown as 2D monolayers and triple labeled for ectopic mouse podocalyxin, f-actin and nuclei. Single confocal images along the Z-axis showed that control cells (left panels) exhibited a relatively flat morphology as evident by the uniform localization of f-actin beneath the apical surface. In contrast, cells with overexpressed podocalyxin displayed moderately perturbed f-actin localization and disrupted morphology as apparent by bulging apical membrane domains (arrows) where podocalyxin localized. Scale = 10µm. C) 3D reconstructed confocal images of MCF-7 transfectants triple labeled for ectopic mouse podocalyxin, its intracellular binding partner NHERF-1, and nuclei. (Of note, immunolocalization of endogenous NHERF-1 proved difficult in the canine MDCK cells, thus this analysis could only be carried out in the human MCF7 cells). Control MCF-7 cells showed partially apical, but diffuse localization of NHERF-1, while podocalyxin expressing cells recruited NHERF-1 where they colocalized at the podocalyxin-induced expanded apical membranes. Dimensions of blue rectangles shown in the merged images = 118 x 26 µm. 151  D) Whole cell lysates from subconfluent populations of the MCF-7 and MDCK transfectants (Control and Podocalyxin) were analyzed for endogenous NHERF-1 by western blotting. Although forced expression of podocalyxin appeared to alter NHERF-1 localization in MCF-7 cells (C), this was not due to altered protein expression levels as the amount of NHERF-1 appeared unchanged in both cell types.  152  153  4.3.4  Podocalyxin induces robust microvillus formation at free, apical membrane domains  NHERF-1 is normally expressed in polarized epithelia, and is found in a variety of different tissues (Reczek et al., 1997). For example, NHERF-1 is expressed at the apical surface of intestinal epithelial cells and exists in a complex with ERM proteins in brush border membranes. In NHERF-1 knockout mice, however, ERM proteins are no longer localized to the apical membrane and the intestinal epithelium display dramatic defects in brush border microvilli (Morales et al., 2004). Since I observed a striking recruitment of NHERF-1 to apical membrane domains and alterations to the actin cytoskeleton in podocalyxin expressing cells, I reasoned that podocalyxin might be causing ultrastructural changes to the cell surface membrane.  In collaboration with Dr. Julie Nielsen, Dr. Wayne Vogl and the UBC Bioimaging facility, we examined the free apical surface of transfected cell monolayers by scanning and transmission electron microscopy (Figure 4-9; right and left panels respectively). Although vector control-transfected cells displayed some surface membrane protrusions, there was a striking increase in the number of protrusions  154  Figure 4-9: Podocalyxin induces robust microvillus formation at the cell surface of MDCK and MCF-7 cells MDCK (A, B, E, F) and MCF-7 (C, D, G, H) cells transfected with murine Podocalyxin (B, D, F, H) or empty vector (A, C, E, G) and stable cell populations were examined by TEM (A-D) and SEM (E-H). A-D) TEM images of the apical surface along the vertical Z-axis are shown. While the control cells displayed some membrane surface projections, forced expression of podocalyxin resulted in a significant increase in microvilli (B and D; arrows). Of note, a number of microvilli are visible as small membrane-bound circles as they are seen in cross-section. TEM imaging was performed by Dr. Wayne Vogl. Scale = 1µm. E-H) Representative SEM images of the apical cell surface (along XY-axis) are shown. In both MDCK and MCF-7 cells, podocalyxin overexpression induced a striking increase in microvilli which are seen as thin membrane projections. SEM imaging was performed by Dr. Julie Nielsen and Derrick Horne. Scale = 2µm. I) Microvilli in six SEM fields (50 µm2 ) were enumerated and graphed. Average number of microvilli per field are shown and error bars represent standard deviation. Student T-tests were performed to show statistically significant differences between Control and Podocalyxin-transfected cells, with p<0.002 in both cell types. Quantification was performed by Dr. Julie Nielsen, and the data shown is representative of two independent experiments.  155  156  that highly resembled cell surface microvilli in both MDCK and MCF-7 cells transfected with podocalyxin. Quantification from SEM images revealed a 2-3 fold increase in microvillus numbers in cells overexpressing podocalyxin (Figure 4-9I). Furthermore, high magnification TEM revealed that individual protrusive structures contained all the structural hallmarks of typical microvilli including the presence of actin filaments in the microvillar core (Nielsen et al., 2007). Thus, we conclude that ectopic expression of podocalyxin is sufficient to induce the formation of microvilli.  4.3.5  Podocalyxin recruits NHERF-1/ezrin/actin to the apical surface where they co-exist in microvilli  In Figure 4-8, I showed that podocalyxin actively recruited NHERF-1 to the apical membrane domain. To determine the specific localization of this complex in relation to the microvillar projections, I performed high-resolution confocal microscopy at the cell surface of the MCF-7 cells transfected with podocalyxin. Cells triple immunolabeled for ectopic podocalyxin, endogenous NHERF-1 and f-actin showed that all three molecules were complexed together and co-localized along the length of individual microvilli (Figure 4-10).  Podocalyxin and NHERF-1 both interact with the actin-binding, ERM family member ezrin (Orlando et al., 2001; Schmieder et al., 2004). Since ezrin is an important component of microvilli (Bretscher, 1983), I reasoned that podocalyxin might also recruit ezrin to the microvilli-rich free apical membrane surface. Highresolution confocal analysis revealed that in control MCF-7 cells lacking ectopic  157  Figure 4-10: Podocalyxin co-localizes with NHERF-1 and actin where they co-exist along individual microvilli MCF-7 cells transfected with podocalyxin were triple-labeled for ectopic mouse podocalyxin, endogenous NHERF-1 and f-actin. Single high-resolution confocal images at the apical surface of cells overexpressing podocalyxin revealed that podocalyxin was specifically localized along the length of individual microvilli (blue, top left) where it highly co-localized with NHERF-1 (green, top right) and factin (red, bottom left). The white within the merged image (bottom right) indicates significant co-localization between all three labeled molecules, which suggests their co-existence in a protein complex in microvilli. Scale = 2µm.  158  159  podocalyxin expression, endogenous ezrin localized to small discrete puncta at the cell surface and did not show significant co-localization with f-actin. In fact, much of the localized f-actin was associated with cell borders (Figure 4-11; left panel arrows) presumably at junctional complexes. In contrast, in the cells expressing podocalyxin, both ezrin and f-actin were enriched at the apical membrane and they were both dramatically co-localized with podocalyxin in microvillar structures (Figure 4-11; right panel). Taken together, these findings indicated that forced podocalyxin expression induced the formation of apical microvilli membrane projections, perhaps by nucleating the formation of multiprotein complexes containing, at the least, NHERF-1, ezrin and f-actin.  160  Figure 4-11: Podocalyxin recruits ezrin and actin to microvilli rich apical membranes MCF-7 cells transfected with either murine podocalyxin or control vector were triple-labeled for ectopic mouse podocalyxin, the microvilli associated protein ezrin, and f-actin. Images shown are 1.0µm thick confocal projections of the apical cell surface. The control cells showed ezrin localized to small puncta, with some colocalization with f-actin. However, a significant portion of f-actin was highly concentrated at the sub-apical, lateral cell-cell contact site in these cells (arrows). In contrast, in cells expressing podocalyxin, both ezrin and f-actin were highly enriched at the microvilli-rich apical surface and displayed a high degree of colocalization (apparent by the white labeling in the triple-merged image). Scale = 5µm.  161  162  4.3.6  Ectopic podocalyxin promotes the formation of actin-rich membrane projections in the absence of cell-cell adhesion  During epithelial polarization, it has been conventionally thought that cell-cell adhesion is an important initiating signal required to define membrane asymmetry and polarized reorganization of the cytoskeleton (Nelson, 2003). However, it has been recently shown that single epithelial cells can polarize in the absence of cellcell contact, which can result in the formation of a polarized microvilli-rich actin cap (Baas et al., 2004). Interestingly, gp135, the canine homologue of podocalyxin has been shown to localize to the ‘pre-apical’ surface of single attached cells where it appears to contribute to the formation of membrane domain asymmetry even in the absence of cell-cell contacts (Meder et al., 2005). To substantiate and extend these findings, I sought to determine whether podocalyxin could alter the actin cytoskeleton in single cells where cell-cell junctions were absent.  Single cell suspensions of both control and podocalyxin expressing MCF-7 cells were allowed to attach to glass coverslips for 2 hours. They were then fixed, labelled for ectopic podocalyxin and f-actin, and subjected to high-resolution confocal microscopy (Figure 4-12). Images along the Z-axis of control cells showed that the actin cytoskeleton was evenly distributed at the cell periphery (Figure 4-12; top left) and a significant amount of f-actin localized at the basal surface where the cell was attached and spread on the substratum (Figure 4-12; bottom left). A confocal projection of the free apical surface revealed that the actin cytoskeleton in  163  control cells was enriched in a number of small punctate structures but appeared largely diffuse at the cell cortex (Figure 4-12; center left).  Similar to what has been demonstrated for gp135 (Meder et al., 2005), I observed that ectopically expressed podocalyxin exclusively localized to the free, apical surface membrane in single substratum-attached MCF-7 cells (Figure 4-12; top right). Strikingly, the apical surface displayed a dramatic increase in f-actin rich puncta, many of which co-localized with podocalyxin (Figure 4-12; middle right). Furthermore, this enrichment of f-actin at the free, apical surface was accompanied by a concomitant decrease in basally distributed f-actin (Figure 4-12; bottom right). Taken together, this suggests that free surface localization of podocalyxin does not require cell-cell contact, and that it may be functionally capable of restructuring the actin cytoskeleton in single cells.  164  Figure 4-12: Podocalyxin localizes to the free surface and causes reorganization of apical f-actin in single attached cells MCF-7 cells transfected with either murine podocalyxin or control vector were gently trypsinized, re-suspended into single cells, plated onto glass coverslips as single cells for 2 hours, and subsequently fixed and labeled for mouse podocalyxin and f-actin.  Top panels left and right: Confocal images along the Z-axis of control cells showed that f-actin was evenly distributed around the cell and was highly enriched along the basal membrane (Top left panel). In the Z-axis view of cells overexpressing podocalyxin, the transgene exclusively localized to the free membrane surface (Top right panel).  Middle and Bottom panels: A 2µm thick projected image along the XY plane at the apical surface of control cells revealed that f-actin was minimally organized into discrete puncta beneath the apical membrane (Middle left panel). In cells overexpressing podocalyxin, f-actin appeared as condensed punctate structures at the apical surface (Middle right panel). Single slice images along the XY plane at the basal surface showed that while control cells showed abundant f-actin and a highly spread basal surface area (Bottom left panel), podocalyxin overexpressing cells showed reduced spreading and significantly less f-actin at the basal surface (Bottom right panel). Scale = 5µm.  165  166  4.4 Summary and Conclusions At the outset of this study I hypothesized that podocalyxin may directly influence mammary epithelial morphogenesis by disrupting cell-cell junctions which would then destabilize epithelial cell polarity. This hypothesis was based on data generated by Marilyn Farquhar's group in MDCK cells (Takeda et al., 2000) and our initial observation that high podocalyxin expression was associated with increased tumor grade (ie. architectural disruption) in human breast carcinoma (Somasiri et al., 2004). However, the findings reported here led me to conclude that podocalyxin overexpression alone does not prevent cell junction formation or alter their polarized localization. Instead, I found that podocalyxin preferentially localizes to, and influences, the structure of non-adhesive free, apical membrane surfaces. In support of this, overexpression of podocalyxin caused an expansion of free apical membranes, actively recruited its binding partners NHERF-1 and ezrin to those surfaces and it reorganized the actin cytoskeleton to form cell surface microvillar membrane projections. Additionally, the data presented here also suggests that the phenotypic outcome of podocalyxin overexpression may by modulated in breast tumor cells that have already deviated from their ability to respond to appropriate architectural cues provided by the basement membrane ECM. Thus, in a tumorigenic context high podocalyxin levels may have deleterious effects.  I found that Podocalyxin was consistently targeted to unattached free cell surface membranes, and this occurred irrespective of the inherent polarity of the cell. For example, the free surface of normal mammary epithelial cells is polarized towards 167  an interior luminal cavity. So it is not surprising that podocalyxin localized primarily to the apical membrane in the resting mammary gland in vivo. Furthermore, when podocalyxin was expressed in EpH4 cells it lined the outside free surface of naked clusters, and upon ECM-overlay, podocalyxin promptly relocated towards the interior membrane that lined the newly formed lumen. Importantly, despite the suggested anti-adhesive properties of podocalyxin, there was no indication that podocalyxin overexpression interfered with the organization of cell-cell junctions, as podocalyxin was invariably located apical to cell junctions and did not disrupt their localization.  It is intriguing to consider that podocalyxin may be functionally important in either the formation or maintenance of luminal cavities. While I found that forced podocalyxin expression was not sufficient to be the initiating signal to drive interior lumen formation in tumor cells that do not inherently form lumens, this does not rule out that podocalyxin may be involved in the morphogenesis of epithelial tissues that require lumen formation. This may possibly occur through the antiadhesive properties of it's highly charged, glycosylated extracellular domain. For example, studies in podocalyxin knockout animals suggest that podocalyxin prevents opposing mesothelial membrane surfaces from inappropriately adhering to one another during development (Doyonnas et al., 2001). Thus, it is tempting to speculate that podocalyxin may be functionally important in sustaining a nonadhesive apical surface to avoid lumen collapse. Furthermore, targeted deletion of gp135 in MDCK cells results in lumen defects in 3D cyst cultures (Meder et al., 2005), but whether this would also occur in mammary epithelial cells requires further investigation. To address this, I have generated constructs to stably express 168  shRNA sequences targeting murine podocalyxin in EpH4 cells and will assess the ability of these cells to generate lumens in the absence of podocalyxin, and these experiments are ongoing.  In 3D basement membrane gel culture, MCF-7 breast tumor cells differed from normal EpH4 cells in that they exhibited an inherent polarity defect. Specifically, apical tight junctions did not polarize towards the interior of MCF-7 cell spheroids. Instead, they localized between cells around the spheroid's exterior surface at the cell-ECM interface. While ectopic podocalyxin did not influence the polarization of the tumor cell spheroids, it did cause a noticeable disruption to their overall architecture particularly at the edges of tumor spheroids where podocalyxin was itself inappropriately localized. This result suggested that when podocalyxin localization is not restricted to interior free surface membranes, such as the case in MCF-7 3D spheroids that cannot polarize appropriately, it might contribute to increased tumor grade, which is often characterized by increased architectural disruption. Importantly, this may be of particular relevance in the unique aspects of ductal breast tumor progression where significant architecture disruption can occur without the loss of cell-cell junctions (This is addressed in more detail in Chapter 5).  In addition to localizing at free cell surfaces, podocalyxin also expanded free, apical plasma membrane domains. Ultrastructural analysis indicated that this expansion was associated with increased microvillus formation. Podocalyxin also recruited NHERF-1 and ezrin to the expanded apical membrane domain where all three molecules were co-localized together, presumably in a complex (Tan et al., 2006; Nielsen et al., 2007). As NHERF-1 and ezrin connect podocalyxin to the actin 169  cytoskeleton, their apical domain recruitment may be functionally important for the free surface targeting of podocalyxin, and its subsequent function at the apical membrane, such as its ability to induce microvillus formation. This question is addressed in more detail in Chapter 6.  The ability of podocalyxin to re-direct the localization of NHERF-1 may have additional important consequences. NHERF-1 has been shown to directly interact with a wide variety of ligands including ion transporters (ex: NHE3), G-protein coupled receptors (ex: β2-adrenergic receptor), receptor tyrosine kinases (ex: EGFR and PDGFR), as well as intracellular signaling molecules such as β-catenin and PTEN (Reviewed in (Weinman et al., 2005). The fact that podocalyxin is a potent inducer of NHERF-1 recruitment to microvilli may suggest a new link between microvillus formation and these diverse NHERF-1-associated signaling cascades. Regardless, the dramatic ability of podocalyxin to regulate the distribution of NHERF-1 to the apical membrane could likely affect NHERF-1’s overall activity as a signaling scaffold.  The ability of podocalyxin to reorganize the actin cytoskeleton towards the cell surface it occupies and away from the cell-ECM interface could also have important implications for breast tumor cell motility. Importantly, I found that podocalyxin could alter the actin cytoskeleton in both the presence and the absence of cell-cell junctions, which indicates that podocalyxin could potentially influence the collective movement of tumor cell clusters as well as the migration of individual tumor cells. This is addressed in greater detail in Chapter 5.  170  Overall, the data presented in this chapter suggests that the presence of podocalyxin at the cell surface may act as a landmark to define non-adhesive cell surfaces. Altogether I found that overexpression of podocalyxin expands this membrane domain and has the ability to recruit NHERF-1, ezrin and reorganize the actin cytoskeleton. This implies that there may be important functional consequences when podocalyxin is overexpressed and coating the membrane surface of breast tumor cells.  171  CHAPTER 5 : FUNCTION OF PODOCALYXIN IN BREAST CANCER PROGRESSION 5.1 Introduction During ductal breast cancer progression, a major pathological hallmark is the loss of normal breast tissue architecture. Despite the emergence of highly disorganized tumor cells that have clearly lost normal epithelial tissue structure, breast tumor cells of ductal origin most often retain E-cadherin mediated cell-cell adhesions (Gamallo et al., 1993). This is in stark contrast to many other epithelially-derived cancers where E-cadherin, along with a number of additional epithelial markers, are routinely lost and tumor cells are thought to undergo a morphological transition to resemble a highly motile and invasive mesenchymal cell; a process known as an epithelial to mesenchymal transition (EMT) (Birchmeier and Birchmeier, 1995; Thiery, 2002). It is clear that the molecular mechanisms that drive a complete EMT are likely critical regulators in the progression of lobular breast cancer (Berx et al., 1995; Yang et al., 2004), but it has not yet been determined if the same mechanisms are important in ductal breast carcinoma, which is the most common breast cancer subtype (Cleton-Jansen, 2002).  Despite the retention of epithelial characteristics observed in the large majority of primary ductal breast carcinomas, the tumors cells are often clinically invasive, and collectively migrate into the surrounding stroma as cohesive cell clusters. The precise mechanism by which a cohesive mass of tumor cells gains the ability to invade the local microenvironment is not well understood, but likely requires a 172  complex interplay of cell-cell adhesion, epithelial plasticity at the migrating front of the invading tumor, and dynamic interactions with the surrounding ECM (Friedl and Wolf, 2003). It is critical, therefore, to identify alternative molecular regulators that may contribute to this mode of EMT independent tumor progression, which appears to be most relevant in ductal carcinomas.  Podocalyxin is an anti-adhesion molecule that is highly overexpressed in a subset of invasive human breast tumors (Somasiri et al., 2004), and more recently has been associated with a number of additional aggressive tumor subtypes (Casey et al., 2006; Heukamp et al., 2006; Kelley et al., 2005; Ney et al., 2007; Schopperle et al., 2003; Sizemore et al., 2007). In Chapter 4 I demonstrated that podocalyxin overexpression alters the architecture of breast tumor cells in vitro, such that it dramatically influences actin dynamics at the apical membrane, and leads to robust microvillus formation and expansion of this free, non-adhesive surface (Nielsen et al., 2007). However, it is not clear whether its ability to alter epithelial architecture and modulate the actin cytoskeleton plays a functional role in breast tumor progression.  In this chapter I show that podocalyxin overexpression perturbs the architecture and tumorigenicity of MCF-7 breast tumor cells in the absence of an EMT. In vitro analysis confirms that podocalyxin does not significantly alter cell junctions in polarized monolayers or 3D tumor spheroids, but is capable of perturbing tumor cell aggregation. I also demonstrate that podocalyxin overexpression decreases the strength of cell-ECM interactions, drives the asymmetric localization of β1 integrins, increases growth factor-dependent tumor cell migration and promotes the growth 173  of 3D tumor spheroids. Furthermore, this is the first study to test the role of podocalyxin tumorigenesis in vivo. I found that overexpressed podocalyxin has the potential to increase breast tumor burden in vivo, but without the loss of cell-cell junctions. This suggests that podocalyxin may function to promote breast tumor progression by perturbing the actin cytoskeleton, cell adhesion and tumor histoarchitecture in an EMT-independent manner, which is characteristic of ductal breast carcinoma progression in humans.  5.2 Podocalyxin perturbs epithelial architecture and causes delamination of 2D monolayers without inducing an EMT In Chapter 4, I observed that podocalyxin overexpression in either normal mammary epithelial cells or MCF-7 breast tumor cells did not appear to directly alter the assembly of polarized cell-cell junctions. Instead, ectopic podocalyxin localized on the apical side of cell-cell junctions and that it modulated the structure of apical membrane domains on free cellular surfaces. This led me to consider the possibility that podocalyxin may be functionally involved in breast tumor progression by uniquely modulating cell adhesion-mediated changes in breast tumor cell morphology without inducing a full blown EMT. Therefore, I carefully assessed the consequences of podocalyxin overexpression on a number of aspects of cell adhesion.  Despite the anti-adhesive properties attributed to podocalyxin (Takeda et al., 2000), at sub-confluent conditions, MCF-7 breast tumor cells over expressing podocalyxin  174  in monolayer culture formed tightly packed epithelial cell islands similar to control cells (Figure 5-1A). Although podocalyxin preferentially localized to the apical membrane, and along the free surface of cells at the edge of the epithelial islands, there was no evidence that overexpression of podocalyxin was inducing a transition towards a mesenchymal morphology (Figure 5-1A, arrow). Instead, I observed a more subtle perturbation to the epithelium. Particularly in sub-confluent conditions, I observed that podocalyxin overexpression caused the cells to appear as bulging cell clusters that would often delaminate from the culture dish (Figure 51A, arrowheads). This agreed with my previous findings that podocalyxin overexpression caused an expansion of the apical membrane in epithelial monolayers (see Chapter 4).  Importantly, it did not appear that the podocalyxin-mediated expansion of the apical membrane surface was linked to cell-cell junction disruption in the monolayers. Confocal projections along the XY plane showed that, similar to controls, cells overexpressing podocalyxin displayed E-cadherin appropriately localized at cell-cell borders (Figure 5-1B). Similar results were obtained for another adherens junction protein, β-catenin, as well as the tight junction proteins ZO-1 and occludin (data not shown). Also, total expression of the junctional molecules Ecadherin and ZO-1 appeared unchanged in podocalyxin overexpressing cells (Figure 5-1B; right panel). Confocal images along the Z-axis showed that E-cadherin was localized along the entire basolateral membrane between cells as is expected for adherens junctions, and that overexpression of podocalyxin did not discernibly alter its localization (Figure 5-1C; top panel). Consistent with my previous results, the control cells exhibited a uniform architecture at the apical surface, while 175  overexpression of podocalyxin caused a significant expansion of the apical membrane domain. This bulging of the free surface caused the location of the tight junction protein ZO-1 to significantly vary within the vertical plane but did not cause a disruption and/or a loss of the apical tight junctional bars (Figure 5-1C; middle panel). Therefore, the overexpression of podocalyxin does not prevent cellcell junction formation in monolayer culture as would occur if the cells had undergone an EMT.  The loss of epithelial keratin intermediate filaments and the appearance of vimentin filaments, is a classical hallmark of an EMT (Thierry et al., 2002). Importantly, keratin expression is retained in the podocalyxin overexpressing cells. However, the perturbed localization pattern of keratin filaments reflects the expanded apical domain induced by podocalyxin overexpression (Figure 5-1C; bottom panel). Furthermore, these distinct morphological alterations to the apical surface are occurring independently of any changes in cell proliferation in monolayer culture (Figure 5-1D). Taken together, these data indicate that podocalyxin overexpression alters the architecture and morphology of cell monolayers without disrupting cellcell junctions or inducing an EMT in epithelial breast tumor cells.  176  Figure 5-1: Podocalyxin mildly perturbs epithelial architecture in 2D monolayers without disruption of cell-cell junction formation or proliferation A) Phase micrographs of sparse cultures show that podocalyxin expressing MCF-7 cells maintain an epithelial morphology and show no signs of an epithelial-tomesenchymal transformation (EMT). Scale=50µm. A confocal XZ vertical image reveals that podocalyxin is preferentially targeted to the apical membrane domain and localizes along the entire free membrane surface of cells. Scale=10µm. B) Podocalyxin overexpression does not affect the localization of E-cadherin by confocal analysis or total expression of either E-cadherin or the tight junction protein ZO-1. Scale=10µm. C) Confocal XZ vertical images show that podocalyxin alters the architecture of the apical surface and does not induce an EMT. Basolateral localization of E-cadherin is relatively unaffected in podocalyxin expressing cells, and tight junctions are also intact as marked by discrete puncta of localized ZO-1, but the location of the junction varies within the vertical plane. Podocalyxin does not cause a loss of epithelial keratin filaments, but alters their localization at the apical surface Scale=10µm. D) Podocalyxin does not alter the growth rate of MCF-7 breast tumor cells in 2D regular growth conditions, nor does it allow for proliferation in the absence of serum. Each experiment was performed in triplicate and the data is expressed relative to Day 1 (mean +/-SD, n=3).  177  178  5.3 Podocalyxin perturbs phonotypic cell-cell aggregation In kidney epithelial (ie. MDCK) cells, it has been shown that podocalyxin has antiadhesive functions that are mediated, at least in part, by the highly glycosylated and overall net negative charge of its extracellular domain (Takeda et al., 2000). Podocalyxin freely diffuses along the entire surface of cells maintained, unattached, in suspension (Meder et al., 2005), but it rapidly polarizes and becomes restricted to the free membrane surface following cell attachment to a substratum (Figure 4-12; (Meder et al., 2005). Thus, I reasoned that, unlike the situation in attached cell monolayers, podocalyxin might perturb cell-cell adhesion in suspension when its localization is unrestricted. To test this, I performed short-term aggregation assays and assessed the ability of the cells (in a single cell suspension) to form cohesive cell aggregates over time while under continuous agitation (Figure 5-2A). In this assay, vector control MCF-7 cells quickly began to form large cell aggregates and after 1 hour very few cells remained as single cells. This aggregation was mediated in large part by calcium-dependent adhesion molecules (ie. cadherins) as it could be prevented by the addition of EDTA. In contrast, even in the absence of EDTA, the large majority of podocalyxin overexpressing cells failed to form such large cell aggregates, and although loose cell aggregates began to form, many single cells remained, even after 2 hours. This suggests that while apically restricted podocalyxin may not be sufficient to disrupt cell-cell junctions in polarized epithelia, the unrestricted membrane localization of podocalyxin appears to be capable of perturbing de novo homotypic cell-cell aggregation.  179  To further test the tentative conclusion that podocalyxin perturbs homotypic cellcell aggregation, I plated single cell suspensions onto a pre-gelled layer of Matrigel and allowed the cells to aggregate into ECM-facilitated, 3D tumor cell spheroids. After 2 days, control cells readily contacted one another and formed compact cell clusters (Figure 5-2B; phase contrast images, left panel). In contrast, cells overexpressing podocalyxin aggregated more slowly and formed very loosely organized spheroids (Figure 5-2B; phase contrast images, right panel), which suggests that homotypic aggregation was, at the very least, less efficient in these cells under these conditions. Furthermore, the disorganized morphology in podocalyxin overexpressing clusters persisted even after 5 days on Matrigel (Figure 5-2B; bottom panel; also refer to Figure 4-7B). Interestingly, within these five-day clusters, podocalyxin overexpression did not inhibit E-cadherin-mediated interactions between cells. This suggests that, in keeping with the ability of podocalyxin to perturb cell-cell aggregation, podocalyxin interferes with initial cohesive spheroid formation, but, once spheroids have formed, it is not sufficient to block cell junction formation in either 2D or 3D adherent culture. Interestingly, however, unlike the situation in normal mammary epithelial cells, MCF-7 breast tumor cells were unable to reorient podocalyxin centrally in response to contact with the basement membrane ECM (Figure 5-2B; external podocalyxin localization). We believe that this is an inherent 3D polarity defect in the MCF-7 cells as both the controls and podocalyxin overexpressors did not form a central lumen (Figure 5-2B; apparent by E-cadherin staining).  180  Figure 5-2: Overexpression of Podocalyxin interferes with homotypic cell-cell aggregation, but does not prevent E-cadherin mediated adhesions in 3D tumor cell spheroids A) The affect of podocalyxin overexpression on cell-cell aggregation in suspension was assessed over a two-hour time course and analyzed by phase microscopy. The extent of aggregation was quantified by the average number of remaining single cells within three separate fields visualized with a 10x objective at each given time point (mean cell number +/- SD, n=4). Representative phase micrographs taken at each time point are shown. Scale=50µm.  B) When single cell suspensions are plated on top of a reconstituted ECM gel, control cells aggregate and form uniform cohesive tumor cell spheroids in 2 days. In contrast, podocalyxin blocks this initial spheroid formation. Representative phase micrographs are shown; Scale=50µm. Regardless, podocalyxin does not alter adherens junctions in the 3D tumor cell clusters as indicated by the appropriate localization of E-cadherin at sites of cell-cell contact within both the control and podocalyxin expressing tumor cell spheroids. Confocal imaging was performed by Dr. Aruna Somasiri. Scale=10µm.  181  182  5.4 Podocalyxin disrupts cell-ECM adhesion and delays cell spreading Since it has been consistently observed that podocalyxin overexpression promotes delamination of MCF-7 cells in monolayer culture ((Somasiri et al., 2004); Figure 51A), I reasoned that it might also be interfering with cell adhesion to the substratum. Indeed, I found that podocalyxin expressing cells did not adhere to fibronectin as effectively as control cells (Figure 5-3A). Furthermore, the cells that did adhere, displayed a delay in their ability to spread at the basal, attached surface (Figure 5-3B). This suggests that podocalyxin overexpression interferes with both the initial adhesion to ECM substrata and that it delays post-adhesion cell spreading.  In Chapter 4, I demonstrated that even in single adherent cells, overexpression of podocalyxin was capable of restructuring the actin cytoskeleton such that f-actin was enriched in apical puncta at the free surface and diminished at the basal surface compared to control cells (Figure 4-12). As a result, I hypothesized that this redistribution of actin might affect the strength of established cell-substratum adhesions. To test this, single cells were plated on fibronectin and allowed to adhere for four hours and were then subjected to a standard de-adhesion assay (See Materials and Method Section 2.3.7). Compared to controls, cells overexpressing podocalyxin more readily disengaged from both low and high-density fibronectin substrata (Figure 5-3C). Taken together, these data suggest that even in single cells (ie. in the absence of cell-cell junctions), podocalyxin may facilitate the recruitment of actin rich complexes to the apical membrane at the expense of forming actin183  Figure 5-3: Forced expression of Podocalyxin inhibits initial adhesion, delays spreading, and decreases the strength of adhesion to fibronectin A) Initial adhesion to a fibronectin substrate (5µg/cm2) was less effective in podocalyxin expressing cells compared to control cells. Each experiment (n=3) was performed in triplicate, and the data shown from one representative experiment is expressed as the average ratio of the fluorescent emission from the adherent cells : total fluorescent emission (X 100; mean +/- SD). Student T-tests were performed to determine statistical significance between control and podocalyxin expressing cells. Representative phase micrographs are shown. Scale=50µm.  B) Cell spreading on a fibronectin substrate (5µg/cm2) was less effective in podocalyxin expressing cells compared to control cells. For cell spreading quantification, the total area of the basal surface of individual cells (>60 cells were quantified per experiment, n=2) was calculated using ImagePro 3DS v6.0 software (mean +/- SD). Student T-tests were performed to determine statistical significance between control and podocalyxin expressing cells, and at all time points, p<0.001. Please note that in order to visualize the entire basal perimeter the images acquired of basal f-actin were oversaturated during collection. Scale=5µm.  C) The relative strength of cell adhesion to two concentrations of fibronectin (FN) was assessed in a standard deadhesion assay (See Materials and Methods, Section 2.3.7) and showed that the strength of adhesion was threshold dependent and that fewer podocalyxin expressing cells remained adherent post-dissociation. Each 184  experiment was done in triplicate and quantified (mean +/-SD, n=2). Student Ttests were performed to determine statistical significance between control and podocalyxin expressing cells at each time point. Representative bright field micrographs of adherent cells present post-dissociation are shown and compared to equivalent control wells that were not subject to dissociation to represent the total number of plated cells.  185  186  enriched adhesions at the basal surface, and this may in part contribute to podocalyxin’s ability to modulate cell-ECM adhesions.  5.5 Podocalyxin promotes growth factor-induced migration of breast tumor cells Tumor cell migration is a highly dynamic process that requires reorganization of the actin cytoskeleton, as well as coordinated adhesion and de-adhesion to the underlying ECM (Friedl and Wolf, 2003). As forced expression of podocalyxin modulated the actin cytoskeleton (Chapter 4) and interfered with cell-ECM adhesion (Figure 5-3), I next asked if podocalyxin overexpression facilitates enhanced breast tumor cell motility.  I first assessed cell migration across “wounded” monolayers of MCF-7 cells on a tissue culture plastic substratum (Figure 5-4A). A sterile pipet tip was drawn across confluent monolayers and the cells were monitored over time for their ability to migrate and close the wound. In serum-free, un-stimulated conditions control MCF7 cells did not break away from the wounded edge. Instead, they slowly migrated as a collective sheet of cells, but did not close the wound after 48 hours. The cells overexpressing podocalyxin also migrated collectively and failed to completely close the wound. However, I did observe numerous differences between podocalyxin overexpressors and control cells. For example, the wounded edge of podocalyxin overexpressing monolayers readily lifted off the tissue culture plastic substratum and cell aggregates often became refractile and sometimes disengaged from the dish (Figure 5-4A). While this was also occasionally observed in control  187  cells, especially at later time points (Figure 5-4A; arrowhead), it was a much more consistent characteristic of podocalyxin overexpressors. Some of the disengaged clusters appeared to reattach within the wound space, and thus may have contributed to the wound healing process. This supports the notion that podocalyxin overexpression does not, on its own, promote elongated (ie. mesenchymal) tumor cell motility.  MCF-7 cells are responsive to epidermal growth factor (EGF) (Mueller et al., 1994) and EGF dependent signaling has been shown to promote breast tumor cell motility and invasion (Holbro and Hynes, 2004). Therefore, I postulated that the ability of podocalyxin overexpression to alter adhesion and the actin cytoskeleton may help “prime” growth factor-stimulated MCF-7 cell motility. This notion was first tested by carrying out wound assays in the presence of EGF. Under these conditions, control cells responded to the EGF stimulation and displayed an increased ability to close the wound but were still unable to completely close it within 48 hours. In contrast, the podocalyxin overexpressing cells, which continued to display the characteristics of edge lifting and aggregate release when treated with EGF, collectively migrated and completely closed the wound within 48 hours (Figure 54A, bottom panel). Importantly, I found that EGF stimulation alone was not sufficient to sustain proliferation in these cells, so the ability of the podocalyxin expressing cells to close the wound is likely independent on cell growth (Figure 54B).  188  Figure 5-4: Podocalyxin increases the migratory potential of breast tumor cells A) MCF-7 control cell and podocalyxin overexpressing cells were subjected to a standard wounding assay in monolayer culture under non-stimulated (serum free DMEM/F12) or growth factor stimulated conditions (EGF 100ng/ml in DMEM/F12). The ability of the cells to close the wound/scratch was monitored by phase microscopy and photomicrographs at the same approximate area of the wound are shown. There was very little difference in wound closure in nonstimulated conditions, but the podocalyxin expressing cells were able to close the wound more readily in response to EGF compared to controls. Scale=50µm.  B) Podocalyxin does not alter the growth rate of MCF-7 breast tumor cells in 2D regular growth conditions, nor does it allow for proliferation in the EGF stimulated conditions. Thus, the differences in wound closure observed in the podocalyxin expressing cells are not due to changes in proliferation. Each experiment was performed in triplicate and the data is expressed relative to Day 1 (mean +/-SD, n=3).  C) Phagokinetic track assays display the motile properties of single cells. Under non-stimulated conditions neither control cells, nor podocalyxin expressing cells were very motile within 24 hours. However, podocalyxin overexpressing cells responded to EGF stimulation and showed an increase in the number of motile cells compared to control cells. Migrated distance was calculated for individual cells (n>70) and the number of cells quantified was graphed categorically. Data from one representative experiment of three is shown. Scale=60µm. 189  190  To further assess the ability of podocalyxin overexpression to 'prime' EGFdependent motility, a phagokinetic track assay was used to visualize the distance and direction traveled by individual motile cells. More specifically, single cells at low density were plated onto a lawn of fluorescent microspheres over fibronectin and cultured for 24 hours with and without EGF stimulation. Using this assay, I found that both the control and podocalyxin overexpressing cells exhibited very little productive motility of greater than 10 µm in serum-free, un-stimulated conditions (Figure 5-4C; top panels). In EGF-treated conditions however, compared to control cells, there was a larger proportion of podocalyxin overexpressing cells with an increased ability to migrate, as indicated by the clearance of beads trailing behind motile cells. Interestingly, the motile podocalyxin expressing cells exhibited a round (non-spread) morphology and ectopic podocalyxin often appeared clustered or polarized to one specific membrane region in response to EGF. This suggests that podocalyxin localization may be modulated in a growth factor dependent manner, and that overexpression of podocalyxin may cooperate with growth factor signaling to facilitate EMT-independent breast tumor cell motility.  5.6 Podocalyxin defines the free surface domain by restricting β 1 integrin localization Disrupting the balance of appropriate integrin-mediated cell-ECM interactions can have profound consequences on the architecture, growth and survival of breast tumor cells (Chrenek et al., 2001; Paszek et al., 2005). For example, β1 integrins are often upregulated in breast cancers, and aberrant signals downstream of β1 integrin  191  can cooperate with growth factor receptors to derail growth arrest and disrupt the architecture of spatially organized breast tumor spheroids grown in reconstituted basement membrane gels (Park et al., 2006; Wang et al., 1998; Weaver et al., 1997). Given that podocalyxin modulates cell-ECM adhesion and growth factor induced motility on ECM, I next asked what consequence podocalyxin overexpression has on β1 integrins, spheroid architecture and MCF-7 breast tumor cell growth in a 3D context.  The expression and localization of β1 integrin was first assessed in single cells. In the absence of ECM ligand engagement, β1 integrins and podocalyxin co-mingled along the entire membrane of single cells in suspension (Figure 5-5A). When single control cells were allowed to attach to a rigid fibronectin substratum, they spread over a two-hour period. However, despite the fact that this attachment is β1 integrin-dependent (Mauro et al., 1999), a considerable proportion of the total β1 integrin remained localized along the free, unattached surface of control cells (Figure 5-5B and C). β1 integrin engagement to ECM ligands such as fibronectin triggers the autophosphorylation of FAK at Y397 during focal adhesion complex assembly as it clearly did in control cells here (Figure 5-5D). Despite the significantly less efficient initial cell-ECM attachment in podocalyxin overexpressing cells (Figure 5-3), the total levels (by Western blot, data not shown) and localization of pY397FAK were unaffected in podocalyxin-overexpressing cells, although FAK-clustering was more spatially-restricted, presumably due to the decreased cell spreading (Figure 5-5D). This suggests that once the cells are attached, integrin signaling is initiated in podocalyxin-overexpressing breast tumor  192  cells attached to a rigid 2D ECM substratum. There was, however, a redistribution of the total β1 integrin in attached podocalyxin overexpressors. While total surface levels of β1 integrin were unchanged (Figure 5-5E), podocalyxin mutually exclusively occupied the free, apical membrane surface, and therefore dramatically restricted the localization of β1 integrin to the basal, attached surface. This suggests that podocalyxin modulates the localization of β1 integrins forcing its asymmetric distribution in response to ECM ligand ligation. This mutually exclusive distribution of podocalyxin and β1 integrin was particularly striking when the cells were cultured on reconstituted basement-membrane gels (Figure 5-3F). Unlike the situation with normal mammary epithelial cells (see Chapter 3), the inherent polarity defect of MCF-7 tumor cells resulted in large membrane domains in which β1 integrins were abnormally excluded at the cell-ECM interface (Figure 5-5F; arrows) or abnormally located centrally, away from the cell-ECM interface (Figure 5-5F; arrowheads). Given the importance of integrin signaling in breast epithelial structure and function, this podocalyxin-driven asymmetry could have profound effects on important breast tumor phenotypes.  193  Figure 5-5: Free surface-targeted podocalyxin restricts β1 integrin localization A-D) MCF-7 control and podocalyxin overexpressing cells were analyzed for localization of ectopic podocalyxin and β1 integrin. In suspension cells, podocalyxin and β1 integrin are both distributed at the cell surface (A). Upon attachment to a fibronectin substrate, podocalyxin restricted the free surface localization of β1 integrin (B and C) without disrupting β1 signaling at the basal surface as the level of phosphorylated FAK at Y397 was unaltered (D). Scale=5µm.  E) MCF-7 control and podocalyxin overexpressing cells were non-enzymatically dissociated from monolayer culture, and subsequently labeled with FITC conjugated-anti-β1 integrin antibody and analyzed by FACs. This showed that the redistribution of β1 integrin was not a result of altered expression of β1 at the cell surface.  F) MCF-7 control and podocalyxin overexpressing cells were cultured on a pregelled layer of ECM (Matrigel) for 5 days and analyzed for ectopic podocalyxin and β1 integrin localization. Podocalyxin and β1 integrin were found to be mutually exclusive in their localization. In control spheroids β1 integrin was localized at all membrane surfaces, while podocalyxin restricted β1 to the membrane domains that it did not occupy. This was particularly evident in this tumor spheroid cross-section that exhibited both interiorly and exteriorly-directed polarity. Scale=10µm.  194  195  5.7 Podocalyxin increases tumor spheroid growth in 3D Culture β1 integrin signaling is required for the ECM-dependent growth and survival of MCF-7 breast tumor cells in 3D culture (Park et al., 2006; Wang et al., 2002). As podocalyxin modulates the localization of β1 integrin in 3D spheroids, I reasoned that podocalyxin overexpression might affect the growth of 3D tumor cell spheroids. To test this, single cells were embedded in reconstituted ECM and clonal spheroids were cultivated for 7 days. Control MCF-7 cells formed consistently small, cohesive spheroids under these conditions over this period of time. However, despite their small size, a considerable proportion of the cells within these control spheroids were in a proliferative state as indicated by the presence of the nuclear Ki-67 (Figure 5-6A; left panel). Importantly, the podocalyxin-overexpressing cells consistently formed larger, proliferative colonies consisting of a greater number of cells (Figure 5-6A; right panels).  It has been shown that aberrant signals downstream of β1 integrin converge with growth factor signals to elicit ECM-dependent and independent growth (Wang et al., 1998). Indeed, I found that podocalyxin increased both the average size and number of anchorage independent colonies grown in soft agar (Figure 5-6B). Taken together, this implies that podocalyxin may perturb the interplay of both ECMdependent and ECM-independent proliferation signals, which results in enhanced spheroid growth and anchorage independent survival, all while not preventing cellcell junction-dependent spheroid formation.  196  Figure 5-6: Podocalyxin increases growth of breast tumor spheroids A) The ECM-dependent growth potential of individual MCF-7 cells overexpressing podocalyxin was tested by embedding single cells in reconstituted matrix (Matrigel). Compared to control cells, there was a significant increase in tumor spheroid growth in podocalyxin overexpressing cells as evident by the larger spheroids that grew up after 7 days. Tumor spheroids were immunolabeled for ectopic podocalyxin and the proliferation nuclear marker Ki-67. At least 70 tumor spheroids (per cell type) were imaged using the Olympus FV1000 confocal microscope and the longest axis of each tumor spheroids was quantified using the FV1000 Fluoview software. A student T-test was performed to determine statistical significance between control and podocalyxin expressing cells, p<0.001. Data shown is representative of three independent experiments. Scale=10µm.  B) Anchorage independent survival of MCF-7 control and podocalyxin overexpressing cells was assessed by embedding single cells in soft agar and assessing colony formation. After 4 weeks, the podocalyxin overexpressing cells displayed increased growth and colony size in soft agar. A student T-test was performed to determine statistical significance between control and podocalyxin expressing cells, p<0.01. Data shown is representative of three independent experiments. Scale=1mm.  197  198  5.8 Podocalyxin may increase breast tumor growth in vivo without disrupting cell-cell junctions I next determined if podocalyxin affects 3D breast tumor cell growth in vivo. Three independent groups of immunocompromised mice supplemented with slow release β-estradiol pellets were injected sub-cutaneously (s.c.) with either control MCF-7 cells or MCF-7 cells overexpressing podocalyxin and assessed for tumor formation and tumor growth (Figure 5-7;A-C; MCF-7 growth in vivo is normally estrogendependent). Measurable tumor masses formed in all mice at approximately the same time regardless of the cell type. However, despite differences between experiments, overall, the tumors generated by podocalyxin overexpressing MCF-7 cells grew to a larger size with increasing time. Furthermore, in two of three experiments these size differences became most apparent more than 40 days after injection at a time when the supplemental slow-release estradiol pellets that were inserted to facilitate MCF-7 cell survival began to become exhausted (Manufacturer's specifications). While not entirely conclusive at this point, these observations suggest that podocalyxin overexpression may provide a growth advantage in sub-cutaneous MCF-7 cell-derived tumors that appears to become most pronounced under conditions of estrogen independence. To confirm that the tumors formed by podocalyxin overexpressing MCF-7 cells continued to express the ectopic transgene, the tumors were processed for immunohistochemistry. Indeed, murine podocalyxin was expressed in the human tumor xenografts and it exhibited a strong membrane association suggesting that it was also functional (Figure 5-8A). In control cell xenografts mouse podocalyxin was only observed in 199  Figure 5-7: Podocalyxin may promote breast tumor growth in vivo Three individual trials were conducted to determine the effects of podocalyxin overexpression on sub-cutaneous (s.c.) tumor growth of MCF-7 breast tumor cells. (A-Trial 1, n=8; B-Trial 2, n=7; C-Trial 3 n=6).  For each trial, 1.0 x 107 MCF-7 control or podocalyxin overexpressing cells were injected s.c. into Rag2M immunocompromised mice 1 day after estrogen pellet implantation. Once tumors were palpable, tumor volumes were measured with calipers and the average volumes are shown graphically (+/- SEM). Although there was variation among the three trials, overall the trend suggests that overexpression of podocalyxin provides an advantage to MCF-7 tumor growth in vivo.  200  201  vascular endothelial cells as was expected of the endogenous gene (Figure 5-8A; arrowhead). Ki-67 labeling confirmed that both the control and podocalyxin tumors contained proliferative tumor cells (Figure 5-8B). Importantly, E-cadherin exhibited strong membrane positivity in both control and podocalyxin xenografts (Figure 58C). Therefore, I tentatively conclude that podocalyxin overexpression, on its own, increased sub-cutaneous breast tumor growth in vivo without initiating an overt EMT. Despite these observations, however, no noticeable differences in tumor grade were noted between the control and podocalyxin-overexpressing cells under these conditions.  202  Figure 5-8: Podocalyxin xenografts contain densely packed, proliferative breast tumor cells and display intact cell-cell junctions A) Immunohistochemical staining of murine podocalyxin showed minimal immunoreactivity in the control xenograft except lining blood vessels. Representative tumor sections (from trial #1) of the podocalyxin xenografts showed expression of the ectopic transgene throughout the tumor. Scale=50µm.  B) Immunohistochemical staining of the nuclear proliferation marker Ki-67 showed that both control and podocalyxin xenograft tumors contained actively proliferating tumor cells. Scale=50µm.  C) Immunohistochemical staining revealed that E-cadherin expression and membranous localization was retained in the podocalyxin xenograft tumors. Scale=50µm.  203  204  5.9 Summary and Conclusions Primary breast cancers that arise from the ductal epithelium consist of hyperproliferative tumor cells that retain epithelial characteristics but are able to survive and persist independent of normal growth suppressive cues from the microenvironment, including those that normally regulate apical-basal polarity. As a result, the breast tumor cells proliferate and fill the luminal space, which results in a severe disruption of the normal mammary ductal architecture. When tumors progress from a carcinoma in situ to an invasive carcinoma, ductal tumor cells often become collectively motile and locally invasive without undergoing an EMT. Understanding molecular regulators that not only functionally contribute to this process, but also are readily identifiable as predictive markers for aggressive disease will be critical for improving diagnosis and treatment.  The results presented in this Chapter indicate that the transmembrane sialomucin podocalyxin, an independent prognostic indicator of poor outcome in invasive breast cancer, functions to perturb breast tumor cell architecture without the loss of epithelial morphology, which led to their enhanced motile potential and increased growth and survival. Importantly, this study represents the first attempt to assess the potential functional importance of podocalyxin overexpression on breast tumorigenesis in vivo.  205  The functional consequences of podocalyxin overexpression are multi-faceted and they are influenced by the molecule's localization within the cell, as well as the cell’s microenvironmental context. Thus, in non-polarized single cells maintained in suspension, podocalyxin is evenly dispersed along the entire membrane surface. In this context, the bulky, negatively charged extracellular domain coats the surface of the membrane to function as a ”Molecular Teflon” which likely contributes to the anti-adhesive function reported for this molecule (Takeda et al., 2000). Indeed, I found that overexpression of podocalyxin impaired the ability of well-differentiated MCF-7 breast tumor cells in suspension to undergo homotypic cell-cell aggregation (Figure 5-2A), and it decreased cell-ECM adhesion (Figure 5-3). In addition, podocalyxin overexpression perturbed 3D tumor spheroid formation (Figure 5-2B), which is dependent on both cell-cell and cell-ECM interactions. However, in the presence of an instructive cue to orient apical-basal polarity either through cell-cell contact, or contact with a substratum, podocalyxin was not sufficient to prevent the ultimate assembly of adherens and tight junctions either in 2D monolayers or 3D spheroids. These observations, coupled with the data presented in Chapter 4 suggests that podocalyxin, on it's own, is not capable of disrupting cell junction formation and cell polarization.  Polarity signals generated by cellular attachment to an ECM substratum direct podocalyxin to be targeted exclusively to the free membrane surface. As a result, in attached cells, podocalyxin is excluded from the basal membrane (Figure 4-12; Figure 5-4B and C). In the absence of cell-cell contact, podocalyxin’s restricted localization at the free surface of adherent cells is thought to pre-determine what will become the apical membrane during epithelial polarization (Meder et al., 2005). 206  Importantly, overexpression of podocalyxin expands this free surface in breast tumor cells. As a consequence, I found that podocalyxin defines the free surface it occupies in part by excluding β1 integrin receptors from this membrane domain. This was most dramatically observed when the tumor cells were cultured in a 3D microenvironment and were forced to attempt to integrate both cell-cell and cellECM-dependent polarity cues to spatially organize into tumor cell spheroids. In this context, where, unlike normal mammary epithelial cells, there was an inherent polarity defect such that the podocalyxin localization sometimes became dominant in these cells, and β1 integrins were localized away from the cell-ECM interface. Given the importance of appropriate β1 integrin signaling in breast morphogenesis and growth suppression, it is possible that the “localization dominance” of podocalyxin may have contributed to the observed increases in anchorage independent survival and 3D spheroid proliferation that did not occur in 2D monolayer culture. Regardless, it is likely that these phenotypic changes initiated by podocalyxin overexpression in a 3D context may have contributed to the molecule's ability to increase tumor growth in vivo.  An interesting aspect of this study was the potential of podocalyxin to promote growth factor mediated migration. I observed that EGF signaling altered the localization of podocalyxin, and increased the motile potential of individual breast tumor cells. There is a precedent for this altered membrane localization, as IL-3dependent cytokine signaling enhances the clustering of podocalyxin into a polarized cap on the surface of hematopoietic progenitor cells (Tan et al., 2006). It has also been recently shown that podocalyxin overexpression increased the  207  migration of prostate cancer cells in vitro, and that ezrin signaling contributed to this process (Sizemore et al., 2007). Ezrin is known to facilitate the formation of membrane protrusions (Algrain et al., 1993), and recent evidence supports that ezrin cooperates with other signaling factors to promote tumor cell migration (Crepaldi et al., 1997; Elliott et al., 2004; Ng et al., 2001). Importantly, ezrin has been implicated specifically in amoeboid (non-elongated) migration, and polarizes to one side of rounded migrating tumor cells (Sahai and Marshall, 2003). Furthermore, ezrin is rapidly phosphorylated and activated downstream of EGF signaling (Berryman et al., 1995). One possibility is that a podocalyxin-mediated modulation of NHERF-1 localization could facilitate EGF-dependent ezrin activation, as NHERF-1 is known to couple directly to the EGF-R and modulate it's signaling outputs (Lazar et al., 2004). Altogether, it will be important to determine whether ezrin and/or NHERF-1 plays a role in coordinating podocalyxin-enhanced migration in breast tumor cells. Interestingly, membrane localization of ezrin in breast tumors is associated with metastatic disease and poor patient outcome (Sarrio et al., 2006).  It is intriguing to consider that podocalyxin plays a general role in defining the free surface membrane domain and as a consequence may control the localization of other surface molecules, as I observed with β1 integrin (Figure 5-5). In keeping with this idea, it could be thought that overexpression of podocalyxin could serve to “over-apicalize” tumor cells by apically recruiting the NHERF-1/ezrin/actin complex, (which may in turn effect numerous NHERF-1/ezrin dependent processes), while presumably marginalizing the basolateral domain and restricting the localization of basolateral membrane proteins. By altering entire membrane 208  domains, and as a consequence, altering the localization of numerous classes of proteins important for cell adhesion, proliferation and migration, it is possible that podocalyxin overexpression could impinge upon all these important processes.  The multiple effects podocalyxin overexpression has on breast tumor cells are certainly complex, and the specific mechanism by which it exerts these effects is not clear. One common denominator is that podocalyxin seems to exert its functional effects at the membrane surface. This suggests that the mechanisms that underlie the targeting of podocalyxin to the appropriate membrane domain are very likely critical for its function and potential consequences in breast tumor progression. Therefore, I addressed the issue of podocalyxin targeting in Chapter 6.  209  CHAPTER 6 : REGULATION OF PODOCALYXIN LOCALIZATION AND FUNCTION IN MAMMARY EPITHELIAL AND BREAST TUMOR CELLS  6.1 Introduction High expression of podocalyxin clinically correlates with poor prognosis in invasive breast cancer (Somasiri et al., 2004) and it is highly expressed in the poor prognosis serous ovarian carcinoma sub-type (Cipollone et al., 2006 - Proc. Amer Assoc. Cancer Research 47a: 5017). Interestingly, we found that specific detection of free surface, membrane-localized podocalyxin was an indicator of poor outcome in the latter tumors (Cipollone J.A., Graves M.L., Roskelley, C.D. submitted to Clin.Can.Res.). This suggests that membrane localization of podocalyxin may be a critical pre-requisite for its functional consequences in tumor progression.  In support of this, when it is overexpressed in breast tumor cells podocalyxin localizes to the free surface, expands this non-adhesive apical membrane, recruits the scaffolding protein NHERF-1, the ERM family member ezrin, and actively remodels the actin cytoskeleton resulting in robust microvilli formation (Chapter 4; (Nielsen et al., 2007). Additionally, I found that free surface expression of podocalyxin inhibited cell-cell aggregation, cell-ECM adhesion and perturbed the architecture and growth of breast tumor cell spheroids in 3D culture (Chapter 5). This further implies that the functional capacity of podocalyxin to alter breast 210  tumor cell morphology and potentially increase breast tumorigenicity may rely on its targeting to the free membrane surface. However, how podocalyxin is targeted to and maintained at the free cell surface, which presumably dictates its function at the membrane, is not well understood.  Membrane-targeted podocalyxin exists in a protein complex with NHERF-1 and ezrin, which connects podocalyxin to the actin cytoskeleton. Thus, NHERF-1 and ezrin are both strong candidates to be regulators of the localization and function of podocalyxin. Along with podocalyxin, NHERF-1 binds the C-terminus of a variety of membrane receptors via its two tandem PDZ domains and these interactions have been shown to be important for the proper localization and function of many of its membrane ligands (reviewed in (Weinman et al., 2005). Furthermore, direct NHERF-1 interaction has been shown to be a critical regulator of both the trafficking and stabilization of many membrane receptors at the cell surface (Cao et al., 1999; Hernando et al., 2002; Lazar et al., 2004; Moyer et al., 2000). Also, both NHERF-1 and ezrin are found to be enriched in, and may contribute to the formation of, apical microvilli of polarized epithelia from numerous tissues (Bonilha et al., 1999; Crepaldi et al., 1997; Morales et al., 2004; Reczek et al., 1997).  I found that ectopic expression of podocalyxin was capable of recruiting both NHERF-1 and ezrin to the apical membrane where they co-localized together at the free surface (Chapter 4; (Nielsen et al., 2007). It is not yet clear, however, whether the NHERF-1/ezrin complex is absolutely required for either the apical localization of podocalyxin or the functional consequences of its overexpression. Therefore, in this chapter I examine the role of individual members of the podocalyxin/NHERF211  1/ezrin/actin protein complex in regulating the localization of podocalyxin and its ability to generate microvillar structures at the apical, free cell surface.  6.2 The role of NHERF-1 in the localization and function of podocalyxin To address the role of the Podocalyxin/NHERF-1 interaction in the free-surface targeting and function of Podocalyxin, I attempted three approaches:  1) A stable knockdown of NHERF-1 was generated in EpH4 normal mammary epithelial cells. The aim was to subsequently introduce ectopic podocalyxin into the NHERF-1 depleted population and assess its ability to polarize to the luminal cell surface using the 3D model of mammary morphogenesis.  2) The role of NHERF-1 in the apical localization of podocalyxin was determined in the mammary gland. This was carried out by determining the localization of podocalyxin in mammary glands excised from NHERF-1 whole animal knockout mice  3) The ability of podocalyxin and NHERF-1 to directly interact was disrupted by deleting the C-terminal PDZ interaction site (DTHL) on podocalyxin’s cytoplasmic tail that is required for direct NHERF-1 binding. This C- terminal truncation mutant was force-expressed in MCF-7 breast tumor cells and its ability to localize to the free surface and induce microvillus formation was assessed. 212  6.2.1  NHERF-1 is expressed and apically localized in luminal mammary epithelial cells in vivo  NHERF-1 was first identified in renal apical brush borders of the proximal tubule (Weinman et al., 1993; Weinman et al., 1995), thus I utilized mouse kidney sections as a positive control for NHERF-1 immunopositivity. I found that NHERF-1 was diffusely localized throughout the cytoplasm of tubular epithelial cells and was highly enriched along the apical membrane surface (Figure 6-1A; arrows). The absence of glomerular staining indicates the isoform-specificity of the antibody, as this tissue is known to express large amounts of NHERF-2 (Takeda et al., 2001). Additionally, a control IgG showed minimal immunoreactivity throughout the tissue.  NHERF-1 was present in the luminal epithelium of 4 week and 12 week old mouse mammary glands and the immunopositivity was strongest at the apical membrane (Figure 6-1B; see insets). NHERF-1 was also localized at the luminal surface in human mammary tissue and was highly immunonegative in the surrounding stroma. This agreed with previous findings that NHERF-1 is highly expressed in normal human mammary tissue (Ediger et al., 1999) where it is primarily membrane localized (Stemmer-Rachamimov et al., 2001). Thus, this analysis indicated that both podocalyxin (see Figure 4-1) and NHERF-1 (Figure 6-1) are expressed in, and co-localized to the apical membrane of luminal mammary epithelial cells in vivo.  213  Figure 6-1: NHERF-1 is expressed in luminal mammary epithelia and is enriched at the apical membrane Immunohistochemistry was performed to detect NHERF-1 expression in mouse and human tissue. A) NHERF-1 was found diffusely distributed in the cytoplasm of tubular kidney epithelial cells, and highly enriched at the apical membrane (arrows). Also, NHERF1 was absent from the neighbouring glomerular epithelium (arrowheads). As a negative control, equal concentrations of a control IgG antibody showed minimal reactivity. Scale=50µm.  B) Mammary glands from 4 week and 12 week old FVB mice were excised, processed and analyzed for NHERF-1 immunolocalization. In both the 4 week and 12 week old mouse mammary glands, NHERF-1 was localized throughout the cytoplasm of the luminal epithelium, but was particularly enriched at the apical surface (see enlarged inset). Scale=50µm.  214  215  6.2.2  NHERF-1 is not required for in vitro mammary morphogenesis and apical polarization of podocalyxin in vivo  Normal mouse mammary epithelial EpH4 cells express NHERF-1 (See Figure 6-2A; control cells). In addition, the localization of podocalyxin can be manipulated in these cells using the 3D ECM-overlay model of mammary morphogenesis Figure 46). Therefore, I chose to address the consequences of NHERF-1 depletion in these cells.  To generate EpH4 cells in which NHERF-1 was stably depleted, I constructed lentiviral vectors and generated virus containing three different shRNA sequences termed NH1-1, NH1-2, NH1-3. EpH4 cells were infected with virus and subjected to two rounds of FAC sorting for GFP (a viral marker) expression followed by genetic selection to generate stable, pooled populations. To control for the lentiviral infection stable cell lines were generated after infection with viruses containing the vector alone (pLL3.7 vector) or a control shRNA sequence that targets the Luciferase gene (shRNA-Luc). Western blot analysis indicated that the vector control - and shRNA-Luc - infected cells continued to express high levels of NHERF-1 (Figure 6-2A; lanes 1 and 2), while all three NHERF-1 shRNA infected cell populations showed a significant reduction in the total NHERF-1 protein (Figure 62A; lanes 3, 4, 5). Importantly, Western blotting also indicated that the NHERF-1 knockdown did not lead to a compensatory upregulation of NHERF-2 expression, which would have made data interpretation difficult (Figure 6-2A). Thus, these cell  216  Figure 6-2: NHERF-1 depletion in vitro does not disrupt 3D mammary spheroid formation and apical polarization A) EpH4 mammary epithelial cells were infected with pLL3.7-GFP containing viruses (empty vector as a control) to stably express either shRNA’s targeting the luciferase gene (shRNA-Luc), or one of three different shRNA’s targeting NHERF-1 (shRNA-NH1-1; shRNA-NH1-2; shRNA-NH1-3). After two independent rounds of FAC sorting for GFP expression followed by genetic selection, stable cells were lysed for western blot analysis. All three NHERF-1 shRNA expressing populations showed a significant decrease in NHERF-1 expression compared to control cells. NHERF-2 was not expressed in these cells.  B) Control cells (WT, Vector, shRNA-Luc) and the NHERF-1 depleted EpH4 cells were cultured on a gelled layer of ECM (Matrigel) for 5 days in the presence of lactogenic hormones. The NHERF-1 depleted cells were capable of 3D mammary spheroid formation and the tight junction protein ZO-1 was apically localized in the knockdown cells similar to controls. Scale=10µm.  217  218  populations were amenable to test the specific consequences of NHERF-1 loss on the apical targeting of podocalyxin in a model of mammary epithelial morphogenesis.  Unfortunately, I experienced technical difficulties in trying to obtain ectopic expression of podocalyxin in the NHERF-1 depleted EpH4 cells. After numerous unsuccessful attempts to obtain a stable population of double transfectants, I tried to transiently express podocalyxin in the NHERF-1 knockdown cells (by means of lipofection and nucleoporation) but was not able to achieve a high enough efficiency of expression to confidently assess its localization in the ECM-overlay model. Despite not being able to analyze the effect of NHERF-1 loss on the polarization of podocalyxin in these cells, I did determine that the loss of NHERF-1 did not appear to affect the formation of 3D polarized spheroids in response to ECM, as movement of the apical tight junction protein ZO-1 was not significantly altered compared to control spheroids (Figure 6-2B; compare top and bottom panels). Thus the loss of NHERF-1 does not appear to affect in vitro 3D ECMdependent mammary morphogenesis.  The above described difficulties in testing the role of NHERF-1 in the localization of podocalyxin using the EpH4 in vitro model lead me to seek out an alternative approach. Thus, I began a collaboration with Poh Tan and Dr.Kelly McNagny (BRC, UBC) who had obtained whole animal NHERF-1 knockout mice generously provided by Dr. Edward Weinman (University of Maryland School of Medicine, Baltimore, MD). In addition to assessing podocalyxin localization, to my knowledge, the mammary glands of NHERF-1 knockout animals had not been 219  previously characterized. Therefore, I performed whole mount analyses to determine the overall structure of the mammary glands.  Mammary glands from mid-pregnant NHERF-1 knockout animals displayed similar alveolar outgrowth and histoarchitecture compared to both heterozygous and age matched wild type control animals (Figure 6-3A). In addition, the differentiative production of the milk protein β-casein was not significantly altered in the transgenic glands (Figure 6-3B). This analysis agreed with the in vitro data obtained from NHERF-1 depleted EpH4 cells (Figure 6-2), and suggests that the loss of NHERF-1 does not significantly interfere with lobuloalveolar morphogenesis or lactational differentiation during pregnancy.  Whole mount analysis of pubertal and 12 week adult (data not shown) virgin mammary glands of NHERF-1 knockout mice indicated that there were no dramatic differences in the outgrowth or organization of the ductal tree (Figure 6-4A). Additionally, podocalyxin was apically targeted on the luminal surface of the ductul epithelial cells of the NHERF-1 knockout glands (Figure 6-4B) and this polarized localization was indistinguishable from that observed in control wild type animals. This surprising result suggests that interactions with the NHERF-1 scaffolding protein are not required for the free surface localization of podocalyxin in luminal mammary epithelial cells.  220  Figure 6-3: Lobuloalveolar development is morphologically unchanged in NHERF-1 knockout mice A) Mammary glands (#4, right abdominal) from NHERF-1 -/-, NHERF-1 +/- and aged matched wild type (WT) controls were excised from mid-pregnant (Day 15) animals. Mammary glands were fixed and stained with carmine alum for whole mount analysis. Representative brightfield images from a dissecting microscope (6X and 12X magnification) are shown to display the overall structure of the mammary glands. There was no significant difference in lobuloalveolar morphogenesis in the NHERF-1 -/- animals compared to controls.  B) Mammary glands (#4, left abdominal) from NHERF-1 -/- and aged matched WT animals were excised and homogenized/lysed in RIPA buffer for western analysis. The levels of milk protein production (β-casein) were relatively unaltered in NHERF-1-/- animals compared to WT animals.  221  222  Figure 6-4: Apical membrane localization of podocalyxin is unaffected in NHERF-1 knockout mice A) Mammary glands (#4, right abdominal) from NHERF-1 -/-, NHERF-1 +/- and wild type (WT) littermates were excised from 6 week old virgin mice. Mammary glands were fixed and stained with carmine alum for whole mount analysis. Representative brightfield images from a dissecting microscope (6X and 12X magnification) are shown to display the overall structure of the mammary glands. There was no significant difference in branching morphogenesis during pubertal development in the NHERF-1 -/- animals compared to controls.  B) Mammary glands (#4, left abdominal) from NHERF-1 -/-, NHERF-1 +/- and wild type (WT) littermates were excised from 6 week old virgin mice and processed for immunohistochemistry. Similar to WT and NHERF-1+/- animals, podocalyxin was exclusively localized to the apical membrane of luminal mammary epithelial in the NHERF-/- animals. Scale=100µm.  223  224  6.2.3  The role of NHERF-1 in podocalyxin localization and microvillus formation in breast tumor cells  The C-terminal tail of podocalyxin contains a putative PDZ-binding domain (DTHL) that binds NHERF-1 (Tan and Graves, et al., 2006). In collaboration with Julie Nielsen and Dr. Kelly McNagny I helped determine the functional importance of this interaction using a series of mutants of the chicken podocalyxin gene that Julie generated and stably force expressed in MCF-7 cells (See Materials and Methods Section 2.3.3; (Nielsen et al., 2007). Full length chicken podocalyxin (WT chPodocalyxin) was transported to the plasma membrane of MCF-7 cells in suspension as assessed by FACs analysis (Figure 6-5A). Additionally, confocal microscopy indicated that WT chPodocalyxin was targeted at the apical surface of attached MCF-7 cells in monolayer culture (Figure 6-5B). Furthermore, apical WT chPodocalyxin displayed the same punctate localization that was observed when WT mouse Podocalyxin was force expressed in these cells (see Figure 4-8, above). Finally, the WT chPodocalyxin recruited endogenous NHERF-1 to the apical domain of the cells where the two molecules co-localized (Figure 6-5B; compare top panel and middle panel). Thus, in all respects assayed so far, chicken and mouse podocalyxin act very similarly in MCF-7 cells.  Forced expression of a chicken podocalyxin mutant lacking the conserved Cterminal PDZ binding domain (ΔDTHL chPodocalyxin) failed to recruit NHERF-1 to the apical cell surface (Figure 6-5B; bottom panel). Importantly, deletion of this  225  Figure 6-5: Deletion of podocalyxin’s c-terminal PDZ binding domain is not sufficient to disrupt its apical targeting, but does prevent NHERF-1 recruitment MCF-7 cells were transfected with the expression vector pIRES2-EGFP (vector control) to overexpress either full-length wild type chicken podocalyxin, or a mutant molecule lacking the C-terminal PDZ binding motif DTHL. Transfected cells were genetically selected, FAC sorted for podocalyxin expression into single cell clones and the resulting clonal populations were used for these experiments. (Cells were generated by Dr. Julie Nielsen, BRC (Nielsen et al., 2007)).  A) MCF-7 control cells and cells force expressing either WT chPodocalyxin or ΔDTHL chPodocalyxin were immunolabelled to detect chPodocalyxin at the cell surface and analyzed by FACs. As expected control were negative for chPodocalyxin, and both the mutant and WT forms of podocalyxin were expressed at the cell surface. FACs analysis was performed by Dr. Julie Nielsen.  B) Monolayers of control MCF-7 cells or those force expressing either WT chPodocalyxin or the ΔDTHL chPodocalyxin mutant were immunolabeled for ectopic chPodocalyxin and endogenous NHERF-1. High-resolution confocal images at the apical surface show that the WT molecule localized in a punctate pattern at the cell surface, which highly co-localized with NHERF-1 (Pearson’s coefficient=0.9). The mutant ΔDTHL chPodocalyxin was also localized in a punctate pattern at the cell surface, but it did not co-localize with NHERF-1 (Pearson’s coefficient -0.1). The levels of NHERF-1 beneath the apical membrane in these cells resembled control cells. Scale=5µm. 226  227  domain did not prevent the apical targeting of the overwhelming majority of the ectopic podocalyxin. These findings led to two important conclusions. First, podocalyxin recruits NHERF-1 to the apical domain in a DTHL C-terminal motifdependent manner. Second, podocalyxin targeting to the apical membrane does not require NHERF-1 binding. The conclusion is consistent with the finding that podocalyxin was targeted to the apical domain of luminal mammary epithelia in NHERF-1 knockout mice (see above).  The ΔDTHL chPodocalyxin mutant was expressed at the cell surface and it displayed the same punctate localization pattern as the full-length, wild type molecule (Figure 6-5B; compare middle and bottom panels). At the ultrastructural level, WT chPodocalyxin caused a significant increase in microvillar membrane surface projections compared to control cells (Figure 6-6; compare top and middle panels, which further conformed that the chicken protein is acting similarly to the mouse protein. Strikingly, however, expression of the ΔDTHL chicken mutant also formed numerous microvilli (Figure 6-6; bottom panel). This indicated that the interaction between podocalyxin and NHERF-1 is, surprisingly, not required for the induction of microvilli at the cell surface.  228  Figure 6-6: The PDZ binding domain of podocalyxin is not required for microvillus formation Monolayers of control MCF-7 cells or cells force expressing either WT chPodocalyxin or the ΔDTHL chPodocalyxin mutant were fixed and processed for TEM (Z-axis; left panels) and SEM (XY-axis; right panels) imaging of the apical surface. The vector control cells displayed very few membrane projections at the apical surface (top panels). In contrast, both the WT chPodocalyxin and the ΔDTHL chPodocalyxin mutant overexpressing cells showed a dramatic increase in the presence of microvilli at the cell surface (middle and bottom panels). TEM imaging was performed by Dr. A. Wayne Vogl, UBC. Scale=1µm. SEM imaging was performed by Dr. Julie Nielsen. Scale=2µm.  229  230  6.3 The role of ezrin in podocalyxin localization and microvillus formation in breast tumor cells At the cell surface, podocalyxin is linked to ezrin and the actin cytoskeleton indirectly through its interaction with NHERF-1 which binds ezrin directly. However, the juxtamembrane region of podocalyxin’s cytoplasmic tail is also capable of binding ezrin directly (Schmieder et al., 2004). Thus, I reasoned that this direct interaction might bypass the need for NHERF-1 for podocalyxin localization and function.  6.3.1  Ezrin is not required for podocalyxin localization in pre-apical membrane domains of single cells  I previously observed that ectopically expressed wild type podocalyxin localized to the free surface of single substratum-attached MCF-7 cells (Figure 4-12). As a prelude to a functional analysis of ezrin, I assessed the ability of podocalyxin to recruit endogenous ezrin to this 'pre-apical' domain. Ezrin was enriched in membrane ruffles at the basal surface of single, attached control cells (Figure 6-7; left panels). This agreed with previous studies in which ezrin was found to remain basal in single cells (Algrain et al., 1993; Prag et al., 2007).  In single, attached cells, force expressed wild type podocalyxin was unable to recruit endogenous ezrin to the pre-apical domain (Figure 6-7; right panels). Instead, similar to control cells, the majority of the ezrin remained at the basal 231  Figure 6-7: Podocalyxin localizes to the pre-apical membrane without ezrin in single cells MCF-7 cells transfected with either murine podocalyxin or control vector were gently trypsinized, resuspended into single cells, plated onto glass coverslips as single cells for 1 hour, and subsequently fixed and labeled for mouse podocalyxin and endogenous ezrin.  Top panels left and right: Confocal projections of the entire Z-stack showed that, in control cells, ezrin was primarily localized to the ruffled membrane at the basal surface. In podocalyxin overexpressing cells, ezrin was also localized at the basal surface, and did not co-segregate with ectopic podocalyxin, which was localized at the free surface membrane and was excluded from the basal domain. Scale=10µm.  Bottom panels left and right: Confocal images along the Z-axis of control cells showed that ezrin was localized exclusively at the basal surface. In the Z-axis view of cells overexpressing podocalyxin, the transgene localized to the free membrane surface and ezrin was mainly localized at the basal domain.  232  233  surface. This suggested that ectopic podocalyxin is capable of moving to the preapical domain without interacting with ezrin, either indirectly or directly. Interestingly, this observation also tentatively suggested the possibility that cell-cell contact may be required to recruit ezrin to the apical membrane domain, and that the podocalyxin-dependent actin reorganization observed in single cells that was noted previously (Figure 4-12) might not require the ezrin-actin linkage at the free surface.  6.3.2  Expression of a dominant-negative ezrin mutant does not disrupt apical podocalyxin localization or microvillus formation  To more definitively determine whether or not ezrin plays a role in the apical targeting of podocalyxin and its ability to induce microvillus formation, MCF-7 cells overexpressing wild type podocalyxin were transiently transfected with a dominant negative form of human ezrin (termed DN-ezrin). This N-terminal mutant (aa. 1309) acts by binding to and masking its effector ligands throughout the cell because it lacks its C-terminal actin-binding domain (Algrain et al., 1993). Thus, as expected, DN-ezrin was evenly distributed along the membrane periphery and throughout the cytoplasm when it was expressed in pre-formed, cell junction-containing, polarized control MCF-7 cell monolayers (Figure 6-8A and B). In MCF-7 cells overexpressing wild type podocalyxin, the DN-ezrin mutant also localized at the cell periphery, but appeared enriched at the apical surface where it highly colocalized with ectopic podocalyxin and endogenous NHERF-1, which was also apically recruited as previously described (Figure 6-8C and D). Importantly, podocalyxin maintained a punctate distribution at the cell surface indicative of the presence of cell surface microvilli. Taken together these data suggested that the DN234  ezrin mutant incorporated into the podocalyxin/NHERF-1 complex, but it did not disrupt the apical localization of this complex or the podocalyxin-dependent microvilli formation. This was an unexpected result, given that the recruitment of the DN-ezrin mutant should have blocked apical actin recruitment to the podocalyxin complex either via direct or indirect (ie. via the NHERF-1 scaffold) interaction with podocalyxin.  In order to examine the potential role of ezrin for the apical localization of podocalyxin during the establishment of epithelial cell polarity, DN-ezrin was expressed in cells prior to monolayer formation. In brief, cells were transiently transfected with DN-ezrin, trypsinized and re-plated, and small, adherent cell islands with cell junctions were analyzed for podocalyxin localization. Under these conditions, DN-ezrin once again localized along the entire membrane periphery occupying apical, lateral and basal surfaces (Figure 6-8F; far left panel). There was a small amount of non-apical podocalyxin that was localized laterally in DN-ezrin expressing cells. However, the great majority of podocalyxin and NHERF-1 continued to localize to the free, apical membrane surface of the polarizing cells (Figure 6-8E and F; white indicates co-localization of all three molecules). Therefore, ezrin does not appear to be critical for the apical targeting of podocalyxin or the podocalyxin-mediated recruitment of NHERF-1.  235  Figure 6-8: Apical localization of Podocalyxin and NHERF-1 is not dependent on ezrin  A-D) Established monolayers of MCF-7 control cells (A and B) or cells stably expressing ectopic murine podocalyxin (C and D) were transiently transfected with dominant negative (DN) ezrin and were processed for immunocytochemistry 36 hours post-transfection. Control cells showed that the DN-ezrin was evenly distributed at the plasma membrane (A and B). Cells overexpressing podocalyxin showed that the DN-ezrin molecule appeared to have incorporated into the podocalyxin/NHERF-1 complex at the apical membrane, and did not disrupt their localization.  E and F) MCF-7 cells stably overexpressing podocalyxin were transiently transfected with DN-ezrin, gently trypsinized and replated to allow the cells to reattach and establish polarity. In this context, DN-ezrin highly co-localized with podocalyxin and NHERF-1 at the apical surface (E; merged image), and did not overtly disrupt podocalyxin targeting, or NHERF-1 recruitment compared to neighbouring untransfected cells as an internal control. Scale =5µm.  236  237  To confirm that the podocalyxin-ezrin interaction was likely not involved in either its apical localization or podocalyxin-dependent microvilli formation, MCF-7 cells were transfected with a chicken podocalyxin mutant lacking the entire cytoplasmic domain except for the extreme proximal juxtamembrane sequence 'HQRF' which was retained as a membrane anchor (Figure 6-9; termed chPodocalyxin Δtail). Previously we determined that this mutant is unable to interact with ezrin, either directly (Nielsen et al., 2007) or indirectly given that it cannot recruit (Figure 6-9A) or bind to NHERF-1 (Tan and Graves et al., 2006).  Cell surface FACs analysis and confocal imaging of the apical cell surface demonstrated that that the chPodocalyxinΔTail mutant was properly targeted to the free surface similar of MCF-7 cells (Figure 6-9A). Also, when the apical plasma membranes of these cells were analyzed by SEM and TEM, the chPodocalyxin Δtail mutant induced microvillus formation (Figure 6-9B). Therefore, especially when coupled with the data described above, these results suggest that neither the NHERF-1 interaction nor the ezrin interaction is required for apical podocalyxin targeting or podocalyxin-induced microvillus formation.  238  Figure 6-9: The cytoplasmic domain of podocalyxin is not required for its apical targeting or microvilli induction MCF-7 cells were transfected with the expression vector pIRES2-EGFP to express a mutant podocalyxin molecule lacking the entire cytoplasmic domain, with the exception of the juxtamembrane HQRF sequence (chPodocalyxin ΔTAIL). Stable cells were generated as described (Figure 6-5; Materials and Methods section 2.4.3). Cells were generated by Dr. Julie Nielsen, BRC (Nielsen et al., 2007).  A) MCF-7 control cells and cells stably overexpressing either WT chPodocalyxin or ΔTAIL chPodocalyxin were immunolabelled to detect chPodocalyxin at the cell surface and analyzed by FACs. Control cells were negative for chPodocalyxin, and both the mutant and WT forms of podocalyxin were expressed at the cell surface. FACs analysis was performed by Dr. Julie Nielsen. Monolayers of each cell type were immunolabeled for ectopic chPodocalyxin and endogenous NHERF-1. Highresolution confocal images at the apical surface show that the WT molecule localized in a punctate pattern at the cell surface, which highly co-localized with NHERF-1 (Pearson’s coefficient=0.9). The mutant ΔTAIL chPodocalyxin was also localized in a punctate pattern at the cell surface, but it did not co-localize with NHERF-1 (Pearson’s coefficient -0.1). The levels of NHERF-1 beneath the apical membrane in these cells resembled control cells. Scale=5µm. B) Monolayers of control MCF-7 cells or cells stably overexpressing either WT chPodocalyxin or the ΔTAIL chPodocalyxin mutant were fixed and processed for TEM (Z-axis; left panels) and SEM (XY-axis; right panels) imaging of the apical  239  surface. The vector control cells displayed very few membrane projections at the apical surface (top panels). In contrast, both the WT chPodocalyxin and the ΔTAIL chPodocalyxin mutant overexpressing cells showed a dramatic increase in the presence of microvilli at the cell surface (middle and bottom panels). TEM imaging was performed by Dr. A. Wayne Vogl, UBC. Scale=1µm. SEM imaging was performed by Dr. Julie Nielsen. Scale=2µm.  240  241  6.3.3  The extracellular domain of podocalyxin is not required for apical localization but it is required for the induction of microvillus formation  The extracellular domain of podocalyxin is subject to extensive O-linked glycosylation and terminal sialylation which contributes to the overall high negative charge of this domain (Dekan et al., 1991). In the kidney, removal of this negative charge disrupts podocyte foot process architecture (Seiler et al., 1977; Takeda et al., 2001). To test the requirement of these terminal sialic acid residues for podocalyxin-induced microvilli formation, monolayers of MCF-7 control cells overexpressing the full-length wild type podocalyxin were treated with neuraminidase which caused podocalyxin to migrate more slowly by SDS-PAGE (Figure 6-10A) which is typical when heavily sialylated glycoproteins become less negatively charged due to neuraminidase digestion (Dekan et al., 1991; Takeda et al., 2000). However, neuraminidase-mediated de-sialylation had no noticeable effect on the podocalyxin-mediated microvillus induction (Figure 6-10B).  In an attempt to determine the function of the extracellular region of podocalyxin more precisely, a chicken mutant, termed chPodocalyxin ΔEC, lacking the complete extracellular domain was expressed in MCF-7 cells. While it was appropriately localized to the apical membrane, the chPodocalyxin ΔEC mutant did not induce microvillus formation (Figure 6-11; Figure 6-12A and B; FACs expression and confocal imaging at the apical surface). Importantly, however, the chPodocalyxin ΔEC mutant did recruit endogenous NHERF-1 to the apical domain (Figure 6-12B; 242  Figure 6-10: Podocalyxin-dependent microvilli formation is not dependent on the sialic acid residues of its extracellular domain A) MCF-7 control cells or cells stably overexpressing podocalyxin were gently trypsinized and treated with neuraminidase in suspension to remove sialic acid residues from Podocalyxin’s extracellular domain. Cells were lysed and analyzed by western blotting to confirm the efficacy of neuraminidase treatment.  B) Monolayers of MCF-7 control cells or cells stably overexpressing podocalyxin were treated with neuraminidase to remove sialic acid residues from podocalyxin’s extracellular domain, and microvillus formation was assessed by SEM. Scale bar: 1 µm. Representative micrographs are shown of three independent experiments. Similar results were obtained if cells were treated with neuraminidase before plating and monolayer formation. Also, results from these experiments were confirmed with both the chicken and murine forms of WT podocalyxin. Scale bar=1 µm. SEM was performed by Dr. Julie Nielsen.  243  244  Figure 6-11: The extracellular domain of podocalyxin is required for microvillus formation Monolayers of control MCF-7 cells or cells stably overexpressing either WT chPodocalyxin or the ΔEC chPodocalyxin mutant were fixed and processed for TEM (Z-axis; left panels) and SEM (XY-axis; right panels) imaging of the apical surface. The vector control cells displayed very few membrane projections at the apical surface (top panels), while WT chPodocalyxin overexpressing cells showed a dramatic increase in the presence of microvilli at the cell surface (middle panels). Strikingly, expression of the ΔEC chPodocalyxin mutant did not result in microvilli formation (bottom panels). TEM imaging was performed by Dr. A. Wayne Vogl, UBC. Scale=1µm. SEM imaging was performed by Dr. Julie Nielsen. Scale=2µm.  245  246  Figure 6-12: Deletion of the extracellular domain of podocalyxin does not prevent its apical targeting A) MCF-7 control cells and cells stably overexpressing either WT chPodocalyxin or ΔEC chPodocalyxin were immunolabelled to detect chPodocalyxin at the cell surface and analyzed by FACs. The ΔEC mutant contains an extracellular Flag tag to detect surface expression. Control cells were negative for chPodocalyxin, and both the mutant and WT forms of podocalyxin were expressed at the cell surface. The FACs analysis was performed by Dr. Julie Nielsen.  B) Monolayers of each cell type were immunolabeled for ectopic chPodocalyxin and endogenous NHERF-1. High-resolution confocal images at the apical surface showed that the WT molecule localized in a punctate pattern at the cell surface, which highly co-localized with NHERF-1 (Pearson’s coefficient=0.9). The mutant ΔEC chPodocalyxin was also localized at the cell surface, and also highly colocalized with NHERF-1 (Pearson’s coefficient 0.8). Scale=5µm.  247  248  see bottom panels). These findings suggest that the extracellular domain is absolutely required for microvillus formation, but not apical podocalyxin targeting.  6.4 The role of actin and regulators of the actin cytoskeleton in podocalyxin localization and microvillus formation The data presented above indicates that NHERF-1 and ezrin, which normally link membrane proteins to the actin cytoskeleton were completely dispensable for podocalyxin-induced microvillus formation. Despite this finding, high-resolution TEM indicated that these microvilli clearly contained parallel bundles of actin filaments coursing longitudinally through their core (Nielsen et al., 2007). Therefore, I next asked if the actin cytoskeleton itself was required for podocalyxindependent microvillus/apical membrane projection formation.  6.4.1  Podocalyxin-induced microvillus formation requires actin polymerization  Latrunculin A effectively disrupts microfilament organization by binding to free, monomeric G-actin and blocking actin polymerization. High-resolution confocal analysis of untreated (DMSO; vehicle control) and wild type podocalyxin overexpressing cells revealed a high degree of co-localization between f-actin and ectopic podocalyxin in a punctate pattern at the apical surface, which was indicative of microvillus formation. In contrast, in control cells, f-actin appeared diffuse beneath the apical membrane and it was most abundant at areas of cell-cell contact (Figure 6-13A; DMSO/vehicle panels). Short-term (15 minute) latrunculin A treatment led to a complete loss of intact actin filaments in both control and 249  Figure 6-13: Actin polymerization is required for podocalyxin induced microvillus formation Monolayers of MCF-7 control cells or cells stably overexpressing podocalyxin were treated with the actin disrupting agent Latrunculin A or DMSO (vehicle control) for 15 minutes, fixed and either processed for immunostaining or SEM.  A) Treated monolayers were immunolabelled for ectopic podocalyxin and f-actin. Confocal images at the apical surface showed that untreated podocalyxin overexpressing cells displayed co-localization between ectopic podocalyxin and factin in a punctate pattern at the cell surface, while control cells displayed f-actin concentrated at cell-cell borders. Latrunculin A treatment completely disrupted actin filaments in both cell types and disrupted the localization of podocalyxin along the membrane. Scale=5µm.  B) SEM images showed that Latrunculin A treatment disrupted the microvilli at the apical surface of podocalyxin overexpressing cells compared to DMSO control treated cells (Top panels; Scale=2µm.). Confocal images along the Z-axis showed that despite the disruption of microvilli, the majority of podocalyxin was retained at the apical membrane (Bottom panels; Scale=5µm).  250  251  podocalyxin overexpressing cells. In the podocalyxin overexpressors it also caused a striking re-distribution of podocalyxin into a more diffuse pattern along the cell surface (Figure 6-13A; LatA panels). SEM analysis of podocalyxin expressing cells confirmed that latrunculin A treatment caused a dramatic dissolution of cell surface membrane projections cells (Figure 6-13B). Also, confocal images along the Z-axis of the podocalyxin overexpressors confirmed that, the large majority of podocalyxin protein was retained at the apical surface (Figure 6-13B; bottom right). These data demonstrate that podocalyxin-induced microvillus formation, but not apical podocalyxin localization, requires actin polymerization.  6.4.2  Balanced RhoGTPase and CDC42 activation are required for podocalyxin induced microvillus formation  How actin is regulated during microvilli formation remains unclear, but it likely involves the Rho family of GTPases, which are well known to regulate actin dynamics and induce the formation of plasma membrane protrusions that include lamellipodia and filopodia (Nobes and Hall, 1995); (For a reviews see, (Jaffe and Hall, 2005; Ridley, 2006). One hypothesis suggests that microvilli may initially form as modified filopodia (Zelhof and Hardy, 2004), which are finger-like membrane projections that also contain parallel bundles of actin filaments. The Rho GTPase family members CDC42 and RhoA have both been shown to promote this type of linear actin elongation (Ridley, 2006). Therefore, I chose to assess the requirement of CDC42 and RhoA for podocalyxin-dependent microvilli formation.  252  Confluent monolayers of MCF-7 cells overexpressing the wild type podocalyxin were transiently transfected with dominant negative CDC42 and RhoA mutants (termed DN-Rho and DN-CDC42; see Materials and Methods section 2.4.4) and were then subjected to high-resolution confocal imaging. In cells expressing either DN-CDC42 (Figure 6-14; top panel) or DN-Rho (Figure 6-14; bottom panels) there was a clear disruption of the densely packed podocalyxin-containing puncta that are indicative of apical surface microvillus induction. Interestingly, the DN-GTPase mutant-mediated loss of microvilli was not associated with a loss of apicallytargeted podocalyxin (Figure 6-14; Z-axis images).  6.5 The ability of podocalyxin mutants to induce microvillus formation and disrupt 3D tumor cell spheroids are positively correlated I have previously shown that podocalyxin overexpression perturbs 3D tumor spheroid formation, but it is not yet clear if podocalyxin’s ability to induce microvillus formation and actin plays a role in this histo-architectural disruption of the spheroids. To address this, MCF-7 cells stably expressing the various mutant forms of podocalyxin were placed on a gelled layer of basement membrane ECM (Matrigel) for 5 days. Under these conditions vector control cells formed compact 3D spheroids, while overexpression of the WT form, the ΔDTHL, and ΔTail mutants, all of which induce microvillus formation in monolayers, caused a significant disruption to the 3D spheroid architecture (Figure 6-15). In contrast, the cells expressing the ΔEC mutant, which does not induce microvillus formation, formed small compact spheroids that resembled those generated by control MCF-7  253  Figure 6-14: Expression of dominant negative forms of Rho and CDC42 GTPase disrupt microvilli, but do not affect apical localization of podocalyxin Monolayers of MCF-7 cells stably overexpressing podocalyxin were transiently transfected with myc-tagged dominant negative (DN) mutants of CDC42 (N17CDC42; top panel) or Rho (N19-Rho; bottom panel), fixed and immunolabelled for ectopic podocalyxin, endogenous NHERF-1 and myc to detect the presence of the GTPase mutant. Expression of either DN-CDC42 or DN-Rho resulted in a breakdown of podocalyxin-induced microvilli at the cell surface, but did not overtly disrupt podocalyxin and NHERF-1 localization at the membrane surface. Scale=5µm.  254  255  cells. While this analysis is only correlative, it nevertheless suggests that the morphological disruption of 3D spheroids, which is likely an important contributor to the increased tumorigenicity that is caused by podocalyxin overexpression, may depend on the ability of the molecule to rearrange the actin cytoskeleton and alter the structure of the free surface plasma membrane (ie. as occurs during microvillus formation).  256  Figure 6-15: Podocalyxin’s ability to form cell surface microvilli correlates with 3D tumor spheroid architectural disruption MCF-7 control cells and MCF-7 cells stably overexpressing various forms of the chicken podocalyxin gene were cultured for 5 days on a reconstituted basement membrane (Matrigel) to form 3D tumor cell spheroids. All forms of podocalyxin that contained the extracellular domain (WT, ΔDTHL, ΔTAIL) all formed irregular 3D aggregates compared to the compact spheroids formed by control cells. Strikingly, the ΔEC mutant overexpressing cells also formed compact spheroids similar to control cells. Representative phase micrographs are shown. Scale=50µm.  257  258  6.6 Summary and Conclusions Mechanisms that regulate the targeting of apical transmembrane proteins are complex, and can involve diverse types of sorting signals. For example, apical sorting signals can require both intracellular and extracellular protein domains as well as the protein’s ability to partition into specific membrane domains (Nelson and Yeaman, 2001). Also, once targeted to the membrane, in the case of many resident apical membrane proteins, their localization and function can depend on selective retention at the cell surface, which can be modulated by interaction with intracellular scaffolds. I hypothesized that, for podocalyxin, its ability to function at the membrane surface may be critical for its role in tumor progression. Thus, in this chapter I have attempted to examine aspects of how podocalyxin is regulated at the plasma membrane, with particular attention to the role of its intracellular scaffold, the NHERF-1/ezrin/actin complex, and how it may effect podocalyxin’s membrane localization and function.  Surprisingly, I found that podocalyxin was consistently targeted to the free apical membrane, and this did not critically depend on its association with NHERF-1 or ezrin. For example, podocalyxin was properly localized to the luminal face of mammary epithelial cells lacking NHERF-1. In MCF-7 cells, expression of a dominant negative form of ezrin failed to result in any significant mis-sorting of podocalyxin to the free surface. In fact, a truncated form of podocalyxin lacking the entire cytoplasmic domain, a mutant incapable of binding both NHERF-1 and ezrin,  259  was faithfully delivered to the apical membrane when expressed in MCF-7 cells. Together this implies that intracellular scaffolding by NHERF-1, or by any other potential direct binding partner on the cytoplasmic face, is not critical for the targeting and retention of the podocalyxin molecule at the apical membrane. Similarly, deletion of the entire extracellular domain did not disrupt podocalyxin targeting to the apical surface. This implies that apical sorting signals may potentially reside either in the transmembrane region, or in both the extracellular domain and the cytoplasmic region. A concurrent study also demonstrated this finding for the canine homologue of podocalyxin, gp135. Interestingly, these authors found that deletion of what they determined to be two putative apical sorting signals, including key residues in the extracellular domain and the entire cytoplasmic domain, resulted in only a partial mis-sorting of this ‘double mutant’ (Yu et al., 2007a). Importantly, despite the proportion of apical mis-sorting, the double mutant still localized to the plasma membrane, albeit with some aberrant lateral targeting between cells. This suggests that in polarized epithelia the targeting of podocalyxin to the plasma membrane is highly regulated and difficult to derail.  Despite the overwhelming tendency for podocalyxin to be properly targeted to the apical membrane, its ability to regulate the ultrastructure of the membrane surface was entirely dependent on its extracellular domain. Although the mechanisms are unclear, podocalyxin-dependent microvilli formation also required actin polymerization at the cell cortex, which presumably occurs downstream of local Rho and/or CDC42 GTPase activation. Importantly, this process did not rely on its own linkage to the actin cytoskeleton via NHERF-1 and ezrin. This was surprising 260  since NHERF-1 and ezrin are known to form a complex in microvilli (Berryman et al., 1995; Reczek et al., 1997), and ezrin is well known to anchor actin filaments to the plasma membrane (Algrain et al., 1993). As the structure of microvilli seems to depend on anchoring actin bundles present in the microvillar core to the plasma membrane (Chhabra and Higgs, 2007), it is not unexpected that in other scenarios where NHERF-1/ezrin complexes are deregulated it can result in microvilli defects (Morales et al., 2004). However, the data presented in this chapter suggests that podocalyxin-dependent microvilli formation follows an alternative mechanism that does not directly rely on its ability to interact with NHERF-1 and ezrin, yet still promotes active actin rearrangement at the plasma membrane. Interestingly, it was recently shown that Podocalyxin can recruit RhoA to the apical plasma membrane in the absence of direct interaction with NHERF-1 (Schmieder et al., 2004), which supports a potential podocalyxin-RhoA dependent pathway for microvillus formation independent of NHERF-1.  It is well established that microvilli formation can occur by modulating intracellular microvilli-associated cytoplasmic proteins, such as the ERM family members and numerous actin bundling proteins such as villin, and espin. However, podocalyxin’s ability to induce microvilli suggests that integral constituents of the plasma membrane may also play an active role. Thus, it is intriguing to consider that there may be alternative general mechanisms to allow membrane proteins to dictate ultrastructural changes to the cell surface. In support of this idea, recent evidence has implicated other transmembrane molecules in the formation of cell surface membrane protrusions like microvilli. For example, overexpression of the drosophila cadherin, Cad99C, is critical for microvilli length in drosophila ovarian 261  follicle cells (D'Alterio et al., 2005; Schlichting et al., 2006). Similar to podocalyxin, deletion of the majority of Cad99C’s cytoplasmic domain including a PDZrecognition motif, still promoted microvilli formation (D'Alterio et al., 2005). This further supports, along with podocalyxin, that there may be other cell surface transmembrane proteins that promote extracellular-dependent ultrastructural changes to free membrane surfaces.  How podocalyxin is regulating plasma membrane structure is not clear. Early in vivo studies in the kidney have long postulated that podocalyxin regulates the intricate structure of kidney podocytes by virtue of the physicochemical properties of its highly negatively charged extracellular domain (Dekan et al., 1991; Kerjaschki et al., 1984; Kerjaschki et al., 1985). Thus it may not be surprising that the extracellular domain was absolutely required for microvilli formation in MCF-7 cells. It has been shown that neutralization of the sialic acid residues present on the surface of podocytes results in a severe disruption of the interdigitating structure of foot processes (Seiler et al., 1977). However, neuraminidase treatment of MCF-7 cells overexpressing podocalyxin failed to significantly reduce its ability to induce microvilli formation. This suggests the interesting possibility that there may be important mechanistic differences in podocalyxin’s contribution to the generation of the extensive foot process structures in podocytes (where negative charge clearly is important for this process), and its ability to induce smaller structural changes to the membrane surface, like microvilli formation.  The extracellular domain of podocalyxin is highly glycosylated and physically bulky. Thus, one explanation for the extensive microvilli that form as a result of its 262  overexpression could be biophysical. This possibility suggests that podocalyxin plays a rather indirect role in microvilli formation (Figure 6-16). In this model, high levels of bulky, cell surface podocalyxin could cause an expansion of the apical membrane, and in order to cope, the membrane reorganizes to generate membrane projections to more easily distribute podocalyxin along the cell surface. For this model, the mechanism for microvilli formation likely relies on the extensive glycosylation of podocalyxin’s extracellular mucin domain. Alternatively, podocalyxin may play a more direct role in microvilli formation. Podocalyxin’s extracellular domain could directly bind with a transmembrane protein present at the apical surface (Figure 6-16). Formation of this complex at the membrane could directly or indirectly transduce the appropriate signals to initiate actin polymerization and microvilli formation. Potential for protein-protein interaction within the extracellular domain could also depend on the appropriate glycosylation, or could involve the membrane proximal stalk or globular region. For either model, it will be important to investigate the specific region of podocalyxin’s extracellular domain required for microvilli formation.  Intriguingly, I found that the ability of podocalyxin to perturb the architecture of 3D tumor spheroids was also dependent on its extracellular domain, since MCF-7 cells overexpressing the mutant form of podocalyxin lacking this domain formed more compact spheroids, which were similar to control cells. Importantly, this suggests the possibility that there may be a mechanistic link between podocalyxin’s ability to induce structural changes to the free surface membrane, which in this case for MCF7 cells faces the cell-ECM interface, and its ability to perturb spheroid formation. However it remains to be tested whether the properties of the extracellular domain 263  Figure 6-16: Proposed models for podocalyxin-dependent microvillus formation Cells expressing low or no podocalyxin exhibited a flat apical surface with welldefined cell junctions and firmly adhered to the underlying substratum. Overexpression of podocalyxin expanded the apical membrane domain, led to a recruitment of NHERF-1/ezrin and f-actin to the apical surface and caused robust microvillus formation. The microvillus formation was dependent on the extracellular domain of podocalyxin, required actin polymerization and could be blocked by Rho and CDC42 GTPase inhibition. Surprisingly the ability of podocalyxin to form microvilli did not require direct or indirect interaction with either NHERF-1 or ezrin. This led to two potential mechanisms proposed for podocalyxin-dependent microvillus formation, the indirect biophysical model or a direct model involving interaction with an as of yet unidentified membrane associated protein (see text for details).  264  265  required for microvilli formation and its role in perturbing homotypic cell-cell aggregation (ie: anti-adhesion) are mutually exclusive. Therefore, further analysis is required to determine how the functions of podocalyxin at the free surface are integrated, particularly in a 3D context.  266  CHAPTER 7 : CONCLUDING REMARKS 7.1 Summary and Discussion Maintaining the organization of epithelial cells within a tissue is critical for its functional homeostasis. Any alteration to the normal polarized architecture, as a consequence, can lead to tumor formation. The dissolution of cadherin-mediated cell-cell junctions is detrimental to apicobasal membrane domain integrity and overall epithelial structure. Thus, E-cadherin has been the focus of much study in carcinoma development and metastatic progression (Birchmeier and Birchmeier, 1995). In contrast to many other epithelial-derived tumors, the majority of primary breast tumors that arise from the ductal epithelium often contain cohesive tumor cells that have retained E-cadherin at sites of cell-cell contact (Cleton-Jansen, 2002). These tumors do, however, exhibit a significant breakdown in normal polarized tissue organization. Although loss of cell polarity was once considered to be a consequence of oncogenic transformation and abnormal cell accumulation, recent evidence supports the idea that the disruption of cell-polarity mechanisms may play a causal role in tumor initiation and/or progression (Lee and Vasioukhin, 2008). Therefore, alterations in polarity determinants that perturb the regulated balance of apical and basolateral membrane domains, without targeting cell-cell junctions per se, could be potential candidates to promote breast tumor progression.  Podocalyxin has been best characterized as the anti-adhesive, transmembrane sialomucin covering the apical surface of kidney podocytes that is critical for  267  maintaining their elaborate structure (Doyonnas et al., 2001). However, its physiological functions in other tissues outside the kidney are not well established. It was recently demonstrated that podocalyxin might function as a membrane landmark to define and help establish the apical membrane domain during early stages of epithelial polarization (Meder et al., 2005). Using tissue microarray technology, we found that podocalyxin is overexpressed in a subset of invasive breast tumors and this served as a highly significant independent predictor of metastatic tumor progression and poor patient outcome (Somasiri et al., 2004). The goal of this project was to determine whether overexpression of this “apical membrane determinant” is functionally important in breast tumor progression. To this end, I force expressed podocalyxin in normal mammary epithelial cells and non-metastatic breast tumor cells.  I found that the normal mouse mammary epithelial EpH4 cell line is suitable to model mammary morphogenesis in vitro as they form organized 3D spheroids with polarized apical tight junctions that cage an interior lumen in response to α6 and β1 integrin dependent interactions with a laminin-rich ECM (Chapter 3). Based on previous gain of function analyses in MDCK cells (Takeda et al., 2000), I initially hypothesized that podocalyxin would directly target the ability of mammary epithelial cells to form apical tight junctions, and that this would effect overall 3D spheroid architecture and lead to defects in functional differentiation. This prediction was not supported by the data presented in this thesis. Specifically, ectopic podocalyxin in EpH4 cells was targeted exclusively to the membrane domain located apical to cell-cell junctions and did not perturb mammary morphogenesis. Instead, ectopic podocalyxin may have helped facilitate lumen 268  formation as these cells frequently formed spheroids with luminal cavities that were often larger than those observed in control cells. This was the first indication that podocalyxin overexpression may not be sufficient to overtly cause a dissolution of cell-cell junction complexes in mammary cells.  Non-metastatic MCF-7 breast tumor cells are highly epithelial and form polarized cell junctions in monolayer culture. Importantly, forced expression of podocalyxin in these cells also failed to disrupt cell-cell junction formation. This is an important aspect of this study, because it has been previously suggested that podocalyxin may alter the architecture of epithelial cells (as in podocyte development, for example) by breaking cell junction complexes in order to remodel the cell’s overall structure. I have found that this is not the case in mammary epithelial cells or breast tumor cells. An alternative possibility is that podocalyxin may dictate morphological changes not by direct junction disruption per se. Instead, high expression of this molecule at the free surface overwhelms and expands this membrane domain such that there is a relative descension of the apical junction complex towards the basal aspect of the cell monolayer. The recent evidence indicating that podocalyxin may participate in defining apical membrane domains supports this idea (Meder et al., 2005). Also, I observed that forced overexpression of podocalyxin re-defined free surface membrane domains by recruiting apical proteins to this domain, and altering the ultrastructure of the apical membrane surface in MCF-7 cells. Additionally, ectopic podocalyxin increased the exclusion of integrins from the apical membrane such that they were completely restricted to the basolateral membrane domain. (See Figure 7-1 for a model demonstrating the multiple consequences of podocalyxin overexpression in MCF-7 cells). 269  Figure 7-1: Overall model outlining the consequences of podocalyxin overexpression in MCF-7 breast tumor cells MCF-7 cells expressing low or no podocalyxin exhibited a flat apical surface, had well-defined cell junctions, and they adhered strongly to the underlying substratum. When they were maintained as single attached cells, β1 integrins were found both basally (ie. at sites of attachment) and apically (ie. on the free surface) of these low/no podocalyxin expressors. Ectopically expressed podocalyxin (PC overexpression) inhibited/delayed homotypic cell-cell aggregation in suspension, and it was apically targeted in single, attached cells where it excluded β1 integrins from that domain and decreased cell spreading. In monolayer culture where cellcell junctions between cell were present, PC overexpression expanded the apical membrane domain, induced microvillus formation, and it increased the recruitment of NHERF-1, f-actin and ezrin to the apical domain. This 'over-apicalization' was accompanied by a decrease in actin at the basal cell surface and it decreased the strength of cellular adhesion to the substratum, which likely contributed to the delamination of cells apically in the monolayer and the increase migratory potential. All of these phenotypes also likely contributed to the architectural disruption of 3-D MCF-7 cell spheroids. Additionally, I speculate (see text) that the relocalization of β1 integrins in PC overexpressing spheroids may alter proliferative potential and responses to growth factor stimulation via mechanisms of which are not yet clear (question marks) but may also involve alterations in the localization of the NHERF-1 and ezrin signal modifiers.  270  271  While podocalyxin overexpression did not disrupt cell junction complex assembly, it did alter the kinetics cell adhesion. When cells were maintained in suspension, ectopically expressed podocalyxin was targeted to the free surface membrane where it decreased the rate of homotypic cell-cell aggregation and inhibited initial tumor cell adhesion and spreading on ECM. These effects are likely dependent on steric hindrance and/or charge repulsion generated by the extracellular domain, but it remains to be determined which is required in breast tumor cells. It is worth noting that how the extracellular domain blocks adhesion may be dependent on the cell type in which it is force expressed, as this aggregation defect in MDCK cells was dependent on charge repulsion, while this is not likely involved in ovarian tumor cells (J. Cipollone, M. Graves and C.D Roskelley unpublished observations).  In MCF-7 cells overexpressing podocalyxin, attachment to a rigid substratum triggered an “over-apicalization” of non-adhesive free surface membranes, which appeared to facilitate the basolateral restriction of β1 integrins. At the apical surface, podocalyxin recruited NHERF-1, ezrin and f-actin to this domain and induced robust microvillus formation. Interestingly, actin recruitment to the apical surface appeared to be accompanied by a decrease in actin at the basal surface and a decrease in the strength of adhesion to the substratum. This likely contributed to the podocalyxin-driven delamination of cell monolayers that was often observed, and it may have helped facilitate the increased migratory potential for these cells. Taken together, the data suggests that podocalyxin overexpression causes an imbalance of membrane domain identity where the free, non-adhesive surfaces become “dominant” over adhesive surfaces. This change in membrane domain 272  definition caused both apical and basal resident proteins to be relocalized. An intriguing possibility is that podocalyxin’s ability to shift the organization of the plasma membrane, could link podocalyxin overexpression to multiple downstream effects. For example, given that derailing polarity regulation of breast tumor cells is frequently linked to hyperproliferation (Liu et al., 2005), it is possible that podocalyxin’s ability to alter membrane domain organization could contribute to the increased growth potential of these cells. This notion was supported by my finding that the increased proliferation of MCF-7 breast tumor cells was only observed in 3D tumor spheroids where alterations in membrane domain identity/cell polarity were most evident.  Both NHERF-1 and ezrin interact with a wide variety of molecules and function in cytoskeletally-associated signaling scaffolds that are important for cell morphology, proliferation and migration. Although the data presented in Chapter 6 precludes a requirement for NHERF-1 and ezrin for podocalyxin-induced microvillus formation, it is possible that either or both play a role in other podocalyxin-induced phenotypes. Thus, it will be of interest to further interrogate the mutant forms of podocalyxin and their effects on tumor cell migration, proliferation in 3D tumor spheroids, and tumor growth in vivo. For example, the ΔEC mutant, which lacks the extracellular region but is still targeted to the apical membrane and capable of recruiting NHERF-1, might identify important consequences of NHERF-1 recruitment that are independent of podocalyxin’s role as an anti-adhesin and microvilli-inducing morphogen.  273  Podocalyxin overexpression caused the formation of disorganized and noncohesive 3D MCF-7 tumor spheroids in which podocalyxin preferentially localized to the exterior surface. If podocalyxin, NHERF-1 and ezrin co-exist in a complex at this exterior cell-ECM interface, it is intriguing to consider that they could cooperate to promote collective migration and 3D invasion at this in vitro version of the “tumor front”, a process which does not require the dissolution of cell junctions. In fact, recent evidence has implicated another small transmembrane mucin, podoplanin, in mediating this type of collective tumor cell invasion that occurs in an EMT independent manner. Like podocalyxin, podoplanin interacts with ezrin, and functionally reorganizes the actin cytoskeleton (Nielsen et al., 2007; Schmieder et al., 2004; Wicki et al., 2006). Thus, future experiments could utilize the MCF-7 cells expressing the various mutant forms of podocalyxin to test their ability to invade ECM, with and without ezrin and NHERF-1 interaction. In addition, these experiments would be further benefited by the use of a stiffer, invasion-permissive matrix environment that would more closely resemble the properties of breast tumor associated stroma (Paszek et al., 2005).  It has been recognized for many years that multiple cell surface sialomucins are often upregulated in human tumors, but their functional roles in tumor progression are not well understood (Carraway et al., 1992). Given the functional similarities reported for podoplanin and the functions described here for podocalyxin, this raises the intriguing question whether there exists a generalized role for cell surface sialomucins to promote an EMT-independent mode of tumor progression in epithelially derived tumors. Therefore, it will be important to establish whether there is functional co-operativity among mucins that are co-overexpressed at the 274  cell surface of tumor cells.  7.2 Potential clinical importance and future directions Overall, the biological properties of podocalyxin as an anti-adhesive protein that is capable of altering cellular morphology suggest that it may function at a critical stage when primary tumors become disorganized and the cells acquire the capacity to break away from the primary tumor. Thus, locally invasive breast tumors that exhibit significant podocalyxin overexpression may represent a discrete group of tumors that are at high risk of metastatic progression. As such, podocalyxin might be a useful clinical biomarker to identify aggressive tumors particularly for early stage diagnosis, and to possibly predict those patients that might be at risk for tumor relapse. As described below, exciting new clinical correlations support this possibility.  At the time of diagnosis, the presence of axillary lymph node metastasis is one of the most important prognostic indicators of aggressive disease in breast cancer where the risk of recurrence is high and the overall survival is low (Vargo-Gogola and Rosen, 2007). In contrast, it is much more difficult to determine, with certainty, which individual patients diagnosed with lymph node-negative breast cancer are at a low or high risk of later tumor recurrence. Thus, identifying new biomarkers that improve the predictive certainty of node-negative patients, which constitutes the great majority of women diagnosed with breast cancer, would facilitate the identification of the patient sub-group that would most benefit from aggressive treatment. Just as importantly, biomarker-assisted improvements in predictive 275  certainty would also lead to the increased sparing of low risk, node-negative patients from unnecessary toxic treatments.  One clinical marker that has gained considerable attention as a prognostic indicator in lymph node negative patients is the histopathological detection of tumor cells inside blood or lymphatic vessels. The presence of lymphvascular invasion (LVI) has been shown to predict later lymph node metastases, and thus it is a useful prognostic indicator to identify an increased risk of tumor recurrence (Lauria et al., 1995; Lee et al., 2006; Marinho et al., 2008). Dr. Irene Andrulis’ group (our collaborators at the Samuel Lunnenfeld Research Institute, Toronto, ON) has been involved in following a cohort of women with node-negative breast cancer prospectively. They have found that LVI, along with tumor grade, is an important prognostic indicator within their cohort (Trudeau et al., 2005). Importantly, in a comparative gene array analysis, they found that podocalyxin was, by far, the most overexpressed gene among tumors that were positive for LVI (I. Andrulis, C. Forse, personal communications). We had previously found that overexpression of podocalyxin predicted poor prognosis independently of lymph-node involvement (Somasiri et al., 2004). Together, this provides strong clinical evidence that podocalyxin overexpression in locally invasive, node-negative breast tumors may be an important new biomarker of aggressive disease, and may be useful to predict potential tumor relapse.  Given the functional properties of podocalyxin, it is intriguing to consider that podocalyxin may contribute to primary tumor cell invasion into the local vasculature. I demonstrated that overexpression of podocalyxin altered tumor cell 276  morphology and adhesion, which likely contributed to an increase in their growth factor-induced migratory potential, all without altering cell-cell junctions or overtly inducing an EMT. Interestingly, LVI is more commonly detected in ductal breast tumors, the subtype that least often undergoes a prominent EMT when it becomes invasive. Also, strong E-cadherin expression has been positively correlated with the presence of LVI, and clusters of tumor cells observed within lymphatic vessels are characteristically E-cadherin positive (Gupta et al., 2003). It is tempting to speculate that podocalyxin may be actively participating in the collective movement or homing of E-cadherin positive ductal tumor cells towards tumor vasculature. Additionally, one could even imagine that the surface of tumor cells coated in podocalyxin could act as an anti-adhesive ‘Molecular Teflon’ that contributes to the squeezing of tumor cell clusters into the pericellular spaces between lymphatic endothelial cells.  In order to test this hypothesis, it is important to develop an appropriate in vivo breast tumor model overexpressing podocalyxin. In the subcutaneous xenograft experiments presented in Chapter 5 of this thesis, I found that there was little vessel infiltration within either the control or podocalyxin tumors as determined by H&E histological analysis and immunolabeling for CD31 and LYVE-1, which are markers of the hematogenous and lymphatic endothelium respectively. Thus, it was difficult to test the hypothesis that podocalyxin overexpression may functionally contribute to LVI in vivo. To address this, I plan to use an orthotopic breast tumor model by injecting either the control or podocalyxin overexpressing MCF-7 breast tumor cells suspended in a Matrigel solution into the mammary fat pads of immunocompromised mice. The fat pad offers the appropriate microenvironment 277  that may confirm a podocalyxin-dependent increase in tumor growth, and provide the right model to test for the increased presence of tumor cells within the vasculature (and lymph node) present in the mammary gland.  An alternative model will involve a transgenic approach. Dr. Kelly McNagny’s group is currently developing a transgenic mouse model that will allow us to either tissue-specifically ablate or overexpress podocalyxin in the mammary epithelium. Using these mice we will first determine the consequences of aberrant podocalyxin expression on mammary gland development in vivo. Next we will cross podocalyxin-modified animals with mice expressing the SV40 polyoma middle T antigen or HER2/neu in the mammary epithelium to assess changes in tumor progression. Polyoma middle T- expressing mice develop highly aggressive metastatic tumors rapidly with great frequency. Thus, if podocalyxin contributes to progression, knocking it out should decrease the number and frequency of metastatic lesions in polyoma middle T-mice. Clinically, HER2/neu overexpression is also an important prognostic indicator of recurrence in lymph node-negative breast tumor patients (Andrulis et al., 1998). Therefore, podocalyxin and HER2/neu may functionally co-operate to increase tumor aggressiveness in the crossed transgenic animals.  HER2/neu has proven to be a clinically relevant therapeutic target with the development of Herceptin. Like HER2/neu, podocalyxin is a transmembrane protein that is expressed at the cell surface. In fact, in vitro analysis has demonstrated that podocalyxin’s localization rarely deviates from the cell surface (Chapter 6). The propensity of podocalyxin to be membrane-localized makes it a 278  possible candidate to develop a targeted therapeutic strategy. I have demonstrated that the extracellular domain of podocalyxin is critical for its ability to function as an anti-adhesin and inducer of microvilli. Thus, if these processes are important for its role in tumor progression, then targeted inhibition of podocalyxin, perhaps via a functional blocking antibody, may likely be a worthwhile strategy to investigate.  I have shown that the localization of ectopic podocalyxin dominated the exterior cell surface of 3D tumor spheroids, and dramatically displaced the localization of β1 integrin (and could likely alter other cell surface proteins). Targeted immunotherapy against β1 integrin has been shown to be effective in reducing tumor burden in vivo, thus these antibodies have potential therapeutic value for treating breast cancer (Park et al., 2006). However, the dominant presence of podocalyxin at the cell surface may severely hamper access to the β1 integrin antigen. Thus it will be interesting to test whether podocalyxin overexpression causes resistance to certain treatments like β1 integrin inhibitory antibodies. In the same manner, however, podocalyxin-induced architectural changes to membrane domains may create an increased sensitivity to other possible treatments. These possibilities would be important to test, and could be easily tested first in vitro using the MCF-7 model cultured as 3D tumor spheroids described in this thesis, and later in the in vivo models described above.  In summary, podocalyxin overexpression in primary breast tumor cells appears to functionally contribute to a non-EMT-mediated architectural disruption of the tumor epithelium by localizing to, and expanding the non-adhesive free surface of  279  tumor cells, and interfering with cell-ECM interactions. This suggests that podocalyxin may be overexpressed before tumor cells have become locally invasive which may be particularly important for identifying potentially aggressive tumors amongst patients who are lymph node negative. Podocalyxin overexpression in breast tumor cells may serve to not only increase primary tumor growth, but also to “tip the balance” towards a non-EMT-mediated migration of cohesive tumor cells during the stromal infiltration and invasion of the local vasculature that marks the first step along the metastatic pathway. Thus, it will be of the utmost importance to better define podocalyxin as a clinically relevant biomarker to facilitate the development of potentially improved classification parameters for diagnosing and treating women with aggressive breast cancer.  280  CHAPTER 8 : REFERENCES Aggeler, J., Ward, J., Blackie, L.M., Barcellos-Hoff, M.H., Streuli, C.H., and Bissell, M.J. (1991). Cytodifferentiation of mouse mammary epithelial cells cultured on a reconstituted basement membrane reveals striking similarities to development in vivo. J Cell Sci 99 ( Pt 2), 407-417. Algrain, M., Turunen, O., Vaheri, A., Louvard, D., and Arpin, M. (1993). Ezrin contains cytoskeleton and membrane binding domains accounting for its proposed role as a membrane-cytoskeletal linker. J Cell Biol 120, 129-139. Andrulis, I.L., Bull, S.B., Blackstein, M.E., Sutherland, D., Mak, C., Sidlofsky, S., Pritzker, K.P., Hartwick, R.W., Hanna, W., Lickley, L., et al. (1998). neu/erbB-2 amplification identifies a poor-prognosis group of women with node-negative breast cancer. Toronto Breast Cancer Study Group. J Clin Oncol 16, 1340-1349. Aranda, V., Haire, T., Nolan, M.E., Calarco, J.P., Rosenberg, A.Z., Fawcett, J.P., Pawson, T., and Muthuswamy, S.K. (2006). Par6-aPKC uncouples ErbB2 induced disruption of polarized epithelial organization from proliferation control. Nat Cell Biol 8, 1235-1245. Armstrong, J.F., Pritchard-Jones, K., Bickmore, W.A., Hastie, N.D., and Bard, J.B. (1993). The expression of the Wilms' tumour gene, WT1, in the developing mammalian embryo. Mech Dev 40, 85-97. Asakura, T., Nakanishi, H., Sakisaka, T., Takahashi, K., Mandai, K., Nishimura, M., Sasaki, T., and Takai, Y. (1999). Similar and differential behaviour between the nectin-afadin-ponsin and cadherin-catenin systems during the formation and disruption of the polarized junctional alignment in epithelial cells. Genes Cells 4, 573-581. Baas, A.F., Kuipers, J., van der Wel, N.N., Batlle, E., Koerten, H.K., Peters, P.J., and Clevers, H.C. (2004). Complete polarization of single intestinal epithelial cells upon activation of LKB1 by STRAD. Cell 116, 457-466. Ball, R.K., Friis, R.R., Schoenenberger, C.A., Doppler, W., and Groner, B. (1988). Prolactin regulation of beta-casein gene expression and of a cytosolic 120-kd protein in a cloned mouse mammary epithelial cell line. EMBO J 7, 2089-2095. Barash, I. (2006). Stat5 in the mammary gland: controlling normal development and cancer. J Cell Physiol 209, 305-313. Barcellos-Hoff, M.H., Aggeler, J., Ram, T.G., and Bissell, M.J. (1989). Functional differentiation and alveolar morphogenesis of primary mammary cultures on reconstituted basement membrane. Development 105, 223-235. 281  Bauman, D.E., Mather, I.H., Wall, R.J., and Lock, A.L. (2006). Major advances associated with the biosynthesis of milk. J Dairy Sci 89, 1235-1243. Belkin, A.M., and Stepp, M.A. (2000). Integrins as receptors for laminins. Microsc Res Tech 51, 280-301. Berryman, M., Gary, R., and Bretscher, A. (1995). Ezrin oligomers are major cytoskeletal components of placental microvilli: a proposal for their involvement in cortical morphogenesis. J Cell Biol 131, 1231-1242. Berx, G., Cleton-Jansen, A.M., Nollet, F., de Leeuw, W.J., van de Vijver, M., Cornelisse, C., and van Roy, F. (1995). E-cadherin is a tumour/invasion suppressor gene mutated in human lobular breast cancers. Embo J 14, 6107-6115. Bilder, D., Schober, M., and Perrimon, N. (2003). Integrated activity of PDZ protein complexes regulates epithelial polarity. Nat Cell Biol 5, 53-58. Birchmeier, W., and Birchmeier, C. (1995). Epithelial-mesenchymal transitions in development and tumor progression. Exs 74, 1-15. Bissell, M.J., and Radisky, D. (2001). Putting tumours in context. Nat Rev Cancer 1, 46-54. Bissell, M.J., Weaver, V.M., Lelievre, S.A., Wang, F., Petersen, O.W., and Schmeichel, K.L. (1999). Tissue structure, nuclear organization, and gene expression in normal and malignant breast. Cancer Res 59, 1757-1763s; discussion 1763s-1764s. Bonilha, V.L., Finnemann, S.C., and Rodriguez-Boulan, E. (1999). Ezrin promotes morphogenesis of apical microvilli and basal infoldings in retinal pigment epithelium. J Cell Biol 147, 1533-1548. Boudreau, N., Sympson, C.J., Werb, Z., and Bissell, M.J. (1995). Suppression of ICE and apoptosis in mammary epithelial cells by extracellular matrix. Science 267, 891893. Bretscher, A. (1983). Purification of an 80,000-dalton protein that is a component of the isolated microvillus cytoskeleton, and its localization in nonmuscle cells. J Cell Biol 97, 425-432. Bretscher, A., Chambers, D., Nguyen, R., and Reczek, D. (2000). ERM-Merlin and EBP50 protein families in plasma membrane organization and function. Annu Rev Cell Dev Biol 16, 113-143. Butta, N., Larrucea, S., Alonso, S., Rodriguez, R.B., Arias-Salgado, E.G., Ayuso, M.S., Gonzalez-Manchon, C., and Parrilla, R. (2006). Role of transcription factor Sp1 and CpG methylation on the regulation of the human podocalyxin gene promoter. BMC Mol Biol 7, 17.  282  Cao, T.T., Deacon, H.W., Reczek, D., Bretscher, A., and von Zastrow, M. (1999). A kinase-regulated PDZ-domain interaction controls endocytic sorting of the beta2adrenergic receptor. Nature 401, 286-290. Carraway, K.L., Fregien, N., Carraway, K.L., 3rd, and Carraway, C.A. (1992). Tumor sialomucin complexes as tumor antigens and modulators of cellular interactions and proliferation. J Cell Sci 103 ( Pt 2), 299-307. Casey, G., Neville, P.J., Liu, X., Plummer, S.J., Cicek, M.S., Krumroy, L.M., Curran, A.P., McGreevy, M.R., Catalona, W.J., Klein, E.A., et al. (2006). Podocalyxin Variants and Risk of Prostate Cancer and Tumor Aggressiveness. Hum Mol Genet. Cheng, H.Y., Lin, Y.Y., Yu, C.Y., Chen, J.Y., Shen, K.F., Lin, W.L., Liao, H.K., Chen, Y.J., Liu, C.H., Pang, V.F., et al. (2005). Molecular identification of canine podocalyxin-like protein 1 as a renal tubulogenic regulator. J Am Soc Nephrol 16, 1612-1622. Chhabra, E.S., and Higgs, H.N. (2007). The many faces of actin: matching assembly factors with cellular structures. Nat Cell Biol 9, 1110-1121. Chrenek, M.A., Wong, P., and Weaver, V.M. (2001). Tumour-stromal interactions. Integrins and cell adhesions as modulators of mammary cell survival and transformation. Breast Cancer Res 3, 224-229. Cipollone, J.A., Gilks, C.B., McCoy, M.L., Zhou, L., McNagny, K.M., Roskelley, C.D. (2006). The apical membrane protein podocalyxin is highly expressed in epithelial ovarian carcinoma. 97th Annual Meeting for The American Association for Cancer Research. Washington, DC. 47a: 5017. Cleton-Jansen, A.M. (2002). E-cadherin and loss of heterozygosity at chromosome 16 in breast carcinogenesis: different genetic pathways in ductal and lobular breast cancer? Breast Cancer Res 4, 5-8. Couse, J.F., Curtis, S.W., Washburn, T.F., Eddy, E.M., Schomberg, D.W., and Korach, K.S. (1995). Disruption of the mouse oestrogen receptor gene: resulting phenotypes and experimental findings. Biochem Soc Trans 23, 929-935. Crepaldi, T., Gautreau, A., Comoglio, P.M., Louvard, D., and Arpin, M. (1997). Ezrin is an effector of hepatocyte growth factor-mediated migration and morphogenesis in epithelial cells. J Cell Biol 138, 423-434. D'Alterio, C., Tran, D.D., Yeung, M.W., Hwang, M.S., Li, M.A., Arana, C.J., Mulligan, V.K., Kubesh, M., Sharma, P., Chase, M., et al. (2005). Drosophila melanogaster Cad99C, the orthologue of human Usher cadherin PCDH15, regulates the length of microvilli. J Cell Biol 171, 549-558. Debnath, J., and Brugge, J.S. (2005). Modelling glandular epithelial cancers in threedimensional cultures. Nat Rev Cancer 5, 675-688. 283  Debnath, J., Mills, K.R., Collins, N.L., Reginato, M.J., Muthuswamy, S.K., and Brugge, J.S. (2002). The role of apoptosis in creating and maintaining luminal space within normal and oncogene-expressing mammary acini. Cell 111, 29-40. Debnath, J., Muthuswamy, S.K., and Brugge, J.S. (2003). Morphogenesis and oncogenesis of MCF-10A mammary epithelial acini grown in three-dimensional basement membrane cultures. Methods 30, 256-268. Dekan, G., Gabel, C., and Farquhar, M.G. (1991). Sulfate contributes to the negative charge of podocalyxin, the major sialoglycoprotein of the glomerular filtration slits. Proc Natl Acad Sci U S A 88, 5398-5402. Dennis, S.L., Manji, S.S., Carrington, D.P., Scarcella, D.L., Ashley, D.M., Smith, P.J., and Algar, E.M. (2002). Expression and mutation analysis of the Wilms' tumor 1 gene in human neural tumors. Int J Cancer 97, 713-715. Doyonnas, R., Kershaw, D.B., Duhme, C., Merkens, H., Chelliah, S., Graf, T., and McNagny, K.M. (2001). Anuria, omphalocele, and perinatal lethality in mice lacking the CD34-related protein podocalyxin. J Exp Med 194, 13-27. Doyonnas, R., Nielsen, J.S., Chelliah, S., Drew, E., Hara, T., Miyajima, A., and McNagny, K.M. (2005). Podocalyxin is a CD34-related marker of murine hematopoietic stem cells and embryonic erythroid cells. Blood 105, 4170-4178. Ebnet, K., Suzuki, A., Horikoshi, Y., Hirose, T., Meyer Zu Brickwedde, M.K., Ohno, S., and Vestweber, D. (2001). The cell polarity protein ASIP/PAR-3 directly associates with junctional adhesion molecule (JAM). EMBO J 20, 3738-3748. Economou, C.G., Kitsiou, P.V., Tzinia, A.K., Panagopoulou, E., Marinos, E., Kershaw, D.B., Kerjaschki, D., and Tsilibary, E.C. (2004). Enhanced podocalyxin expression alters the structure of podocyte basal surface. J Cell Sci 117, 3281-3294. Ediger, T.R., Kraus, W.L., Weinman, E.J., and Katzenellenbogen, B.S. (1999). Estrogen receptor regulation of the Na+/H+ exchange regulatory factor. Endocrinology 140, 2976-2982. Ediger, T.R., Park, S.E., and Katzenellenbogen, B.S. (2002). Estrogen receptor inducibility of the human Na+/H+ exchanger regulatory factor/ezrin-radixinmoesin binding protein 50 (NHE-RF/EBP50) gene involving multiple half-estrogen response elements. Mol Endocrinol 16, 1828-1839. Elliott, B.E., Qiao, H., Louvard, D., and Arpin, M. (2004). Co-operative effect of c-Src and ezrin in deregulation of cell-cell contacts and scattering of mammary carcinoma cells. J Cell Biochem 92, 16-28. 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. 284  Faraldo, M.M., Deugnier, M.A., Tlouzeau, S., Thiery, J.P., and Glukhova, M.A. (2002). Perturbation of beta1-integrin function in involuting mammary gland results in premature dedifferentiation of secretory epithelial cells. Mol Biol Cell 13, 35213531. Fata, J.E., Werb, Z., and Bissell, M.J. (2004). Regulation of mammary gland branching morphogenesis by the extracellular matrix and its remodeling enzymes. Breast Cancer Res 6, 1-11. Fialka, I., Schwarz, H., Reichmann, E., Oft, M., Busslinger, M., and Beug, H. (1996). The estrogen-dependent c-JunER protein causes a reversible loss of mammary epithelial cell polarity involving a destabilization of adherens junctions. J Cell Biol 132, 1115-1132. Fogg, V.C., Liu, C.J., and Margolis, B. (2005). Multiple regions of Crumbs3 are required for tight junction formation in MCF10A cells. J Cell Sci 118, 2859-2869. Folkman, J., and Moscona, A. (1978). Role of cell shape in growth control. Nature 273, 345-349. Friedl, P., and Wolf, K. (2003). Tumour-cell invasion and migration: diversity and escape mechanisms. Nat Rev Cancer 3, 362-374. Furuse, M., and Tsukita, S. (2006). Claudins in occluding junctions of humans and flies. Trends Cell Biol 16, 181-188. Gamallo, C., Palacios, J., Suarez, A., Pizarro, A., Navarro, P., Quintanilla, M., and Cano, A. (1993). Correlation of E-cadherin expression with differentiation grade and histological type in breast carcinoma. Am J Pathol 142, 987-993. Gates, J., and Peifer, M. (2005). Can 1000 reviews be wrong? Actin, alpha-Catenin, and adherens junctions. Cell 123, 769-772. Ginger, M.R., Gonzalez-Rimbau, M.F., Gay, J.P., and Rosen, J.M. (2001). Persistent changes in gene expression induced by estrogen and progesterone in the rat mammary gland. Mol Endocrinol 15, 1993-2009. Goldstein, B., and Macara, I.G. (2007). The PAR proteins: fundamental players in animal cell polarization. Dev Cell 13, 609-622. Grindstaff, K.K., Yeaman, C., Anandasabapathy, N., Hsu, S.C., Rodriguez-Boulan, E., Scheller, R.H., and Nelson, W.J. (1998). Sec6/8 complex is recruited to cell-cell contacts and specifies transport vesicle delivery to the basal-lateral membrane in epithelial cells. Cell 93, 731-740. Gumbiner, B., and Simons, K. (1986). A functional assay for proteins involved in establishing an epithelial occluding barrier: identification of a uvomorulin-like polypeptide. J Cell Biol 102, 457-468.  285  Gupta, A., Deshpande, C.G., and Badve, S. (2003). Role of E-cadherins in development of lymphatic tumor emboli. Cancer 97, 2341-2347. Hara, T., Nakano, Y., Tanaka, M., Tamura, K., Sekiguchi, T., Minehata, K., Copeland, N.G., Jenkins, N.A., Okabe, M., Kogo, H., et al. (1999). Identification of podocalyxin-like protein 1 as a novel cell surface marker for hemangioblasts in the murine aorta-gonad-mesonephros region. Immunity 11, 567-578. Hennighausen, L., and Robinson, G.W. (1998). Think globally, act locally: the making of a mouse mammary gland. Genes Dev 12, 449-455. Hernando, N., Deliot, N., Gisler, S.M., Lederer, E., Weinman, E.J., Biber, J., and Murer, H. (2002). PDZ-domain interactions and apical expression of type IIa Na/P(i) cotransporters. Proc Natl Acad Sci U S A 99, 11957-11962. Herrera, L., Ottolenghi, C., Garcia-Ortiz, J.E., Pellegrini, M., Manini, F., Ko, M.S., Nagaraja, R., Forabosco, A., and Schlessinger, D. (2005). Mouse ovary developmental RNA and protein markers from gene expression profiling. Dev Biol 279, 271-290. Heukamp, L.C., Fischer, H.P., Schirmacher, P., Chen, X., Breuhahn, K., Nicolay, C., Buttner, R., and Gutgemann, I. (2006). Podocalyxin-like protein 1 expression in primary hepatic tumours and tumour-like lesions. Histopathology 49, 242-247. Hirose, T., Izumi, Y., Nagashima, Y., Tamai-Nagai, Y., Kurihara, H., Sakai, T., Suzuki, Y., Yamanaka, T., Suzuki, A., Mizuno, K., et al. (2002). Involvement of ASIP/PAR-3 in the promotion of epithelial tight junction formation. J Cell Sci 115, 2485-2495. Holbro, T., and Hynes, N.E. (2004). ErbB receptors: directing key signaling networks throughout life. Annu Rev Pharmacol Toxicol 44, 195-217. Hood, J.D., and Cheresh, D.A. (2002). Role of integrins in cell invasion and migration. Nat Rev Cancer 2, 91-100. Hoover, K.B., Liao, S.Y., and Bryant, P.J. (1998). Loss of the tight junction MAGUK ZO-1 in breast cancer: relationship to glandular differentiation and loss of heterozygosity. Am J Pathol 153, 1767-1773. Horvat, R., Hovorka, A., Dekan, G., Poczewski, H., and Kerjaschki, D. (1986). Endothelial cell membranes contain podocalyxin--the major sialoprotein of visceral glomerular epithelial cells. J Cell Biol 102, 484-491. Hovey, R.C., and Trott, J.F. (2004). Morphogenesis of mammary gland development. Adv Exp Med Biol 554, 219-228. Howlin, J., McBryan, J., and Martin, F. (2006). Pubertal mammary gland development: insights from mouse models. J Mammary Gland Biol Neoplasia 11, 283-297. 286  Hurd, T.W., Gao, L., Roh, M.H., Macara, I.G., and Margolis, B. (2003). Direct interaction of two polarity complexes implicated in epithelial tight junction assembly. Nat Cell Biol 5, 137-142. Hynes, R.O. (1992). Integrins: versatility, modulation, and signaling in cell adhesion. Cell 69, 11-25. Ikenouchi, J., Umeda, K., Tsukita, S., and Furuse, M. (2007). Requirement of ZO-1 for the formation of belt-like adherens junctions during epithelial cell polarization. J Cell Biol 176, 779-786. Ivanov, A.I., Nusrat, A., and Parkos, C.A. (2005). Endocytosis of the apical junctional complex: mechanisms and possible roles in regulation of epithelial barriers. Bioessays 27, 356-365. Jaffe, A.B., and Hall, A. (2005). Rho GTPases: biochemistry and biology. Annu Rev Cell Dev Biol 21, 247-269. Joberty, G., Petersen, C., Gao, L., and Macara, I.G. (2000). The cell-polarity protein Par6 links Par3 and atypical protein kinase C to Cdc42. Nat Cell Biol 2, 531-539. Kamei, M., Saunders, W.B., Bayless, K.J., Dye, L., Davis, G.E., and Weinstein, B.M. (2006). Endothelial tubes assemble from intracellular vacuoles in vivo. Nature 442, 453-456. Kanwar, Y.S., Liu, Z.Z., Kumar, A., Usman, M.I., Wada, J., and Wallner, E.I. (1996). D-glucose-induced dysmorphogenesis of embryonic kidney. J Clin Invest 98, 24782488. Kanwar, Y.S., Liu, Z.Z., and Wallner, E.I. (1997). Influence of glucose on murine metanephric development and proteoglycans: morphologic and biochemical studies. Lab Invest 76, 671-681. Kelley, T.W., Huntsman, D., McNagny, K.M., Roskelley, C.D., and Hsi, E.D. (2005). Podocalyxin: a marker of blasts in acute leukemia. Am J Clin Pathol 124, 134-142. Kerjaschki, D. (1978). Polycation-induced dislocation of slit diaphragms and formation of cell junctions in rat kidney glomeruli: the effects of low temperature, divalent cations, colchicine, and cytochalasin B. Lab Invest 39, 430-440. Kerjaschki, D., Sharkey, D.J., and Farquhar, M.G. (1984). Identification and characterization of podocalyxin--the major sialoprotein of the renal glomerular epithelial cell. J Cell Biol 98, 1591-1596. Kerjaschki, D., Vernillo, A.T., and Farquhar, M.G. (1985). Reduced sialylation of podocalyxin--the major sialoprotein of the rat kidney glomerulus--in aminonucleoside nephrosis. Am J Pathol 118, 343-349.  287  Kershaw, D.B., Beck, S.G., Wharram, B.L., Wiggins, J.E., Goyal, M., Thomas, P.E., and Wiggins, R.C. (1997a). Molecular cloning and characterization of human podocalyxin-like protein. Orthologous relationship to rabbit PCLP1 and rat podocalyxin. J Biol Chem 272, 15708-15714. Kershaw, D.B., Thomas, P.E., Wharram, B.L., Goyal, M., Wiggins, J.E., Whiteside, C.I., and Wiggins, R.C. (1995). Molecular cloning, expression, and characterization of podocalyxin-like protein 1 from rabbit as a transmembrane protein of glomerular podocytes and vascular endothelium. J Biol Chem 270, 29439-29446. Kershaw, D.B., Wiggins, J.E., Wharram, B.L., and Wiggins, R.C. (1997b). Assignment of the human podocalyxin-like protein (PODXL) gene to 7q32-q33. Genomics 45, 239-240. Kirshner, J., Chen, C.J., Liu, P., Huang, J., and Shively, J.E. (2003). CEACAM1-4S, a cell-cell adhesion molecule, mediates apoptosis and reverts mammary carcinoma cells to a normal morphogenic phenotype in a 3D culture. Proc Natl Acad Sci U S A 100, 521-526. Kominsky, S.L., Argani, P., Korz, D., Evron, E., Raman, V., Garrett, E., Rein, A., Sauter, G., Kallioniemi, O.P., and Sukumar, S. (2003). Loss of the tight junction protein claudin-7 correlates with histological grade in both ductal carcinoma in situ and invasive ductal carcinoma of the breast. Oncogene 22, 2021-2033. Korach, K.S., Couse, J.F., Curtis, S.W., Washburn, T.F., Lindzey, J., Kimbro, K.S., Eddy, E.M., Migliaccio, S., Snedeker, S.M., Lubahn, D.B., et al. (1996). Estrogen receptor gene disruption: molecular characterization and experimental and clinical phenotypes. Recent Prog Horm Res 51, 159-186; discussion 186-158. Kramer, F., White, K., Kubbies, M., Swisshelm, K., and Weber, B.H. (2000). Genomic organization of claudin-1 and its assessment in hereditary and sporadic breast cancer. Hum Genet 107, 249-256. Lau, A.G., and Hall, R.A. (2001). Oligomerization of NHERF-1 and NHERF-2 PDZ domains: differential regulation by association with receptor carboxyl-termini and by phosphorylation. Biochemistry 40, 8572-8580. Lauria, R., Perrone, F., Carlomagno, C., De Laurentiis, M., Morabito, A., Gallo, C., Varriale, E., Pettinato, G., Panico, L., Petrella, G., et al. (1995). The prognostic value of lymphatic and blood vessel invasion in operable breast cancer. Cancer 76, 17721778. Lazar, C.S., Cresson, C.M., Lauffenburger, D.A., and Gill, G.N. (2004). The Na+/H+ exchanger regulatory factor stabilizes epidermal growth factor receptors at the cell surface. Mol Biol Cell 15, 5470-5480. Lee, A.H., Pinder, S.E., Macmillan, R.D., Mitchell, M., Ellis, I.O., Elston, C.W., and Blamey, R.W. (2006). Prognostic value of lymphovascular invasion in women with lymph node negative invasive breast carcinoma. Eur J Cancer 42, 357-362. 288  Lee, M., and Vasioukhin, V. (2008). Cell polarity and cancer - cell and tissue polarity as a non-canonical tumor suppressor. J Cell Sci 121, 1141-1150. Li, N., Zhang, Y., Naylor, M.J., Schatzmann, F., Maurer, F., Wintermantel, T., Schuetz, G., Mueller, U., Streuli, C.H., and Hynes, N.E. (2005). Beta1 integrins regulate mammary gland proliferation and maintain the integrity of mammary alveoli. EMBO J 24, 1942-1953. Li, Y., Li, J., Straight, S.W., and Kershaw, D.B. (2002). PDZ domain-mediated interaction of rabbit podocalyxin and Na(+)/H(+) exchange regulatory factor-2. Am J Physiol Renal Physiol 282, F1129-1139. Lin, D., Edwards, A.S., Fawcett, J.P., Mbamalu, G., Scott, J.D., and Pawson, T. (2000). A mammalian PAR-3-PAR-6 complex implicated in Cdc42/Rac1 and aPKC signalling and cell polarity. Nat Cell Biol 2, 540-547. Litjens, S.H., de Pereda, J.M., and Sonnenberg, A. (2006). Current insights into the formation and breakdown of hemidesmosomes. Trends Cell Biol 16, 376-383. Liu, H., Radisky, D.C., and Bissell, M.J. (2005). Proliferation and polarity in breast cancer: untying the gordian knot. Cell Cycle 4, 646-649. Liu, H., Radisky, D.C., Wang, F., and Bissell, M.J. (2004). Polarity and proliferation are controlled by distinct signaling pathways downstream of PI3-kinase in breast epithelial tumor cells. J Cell Biol 164, 603-612. Lochter, A., Galosy, S., Muschler, J., Freedman, N., Werb, Z., and Bissell, M.J. (1997). Matrix metalloproteinase stromelysin-1 triggers a cascade of molecular alterations that leads to stable epithelial-to-mesenchymal conversion and a premalignant phenotype in mammary epithelial cells. J Cell Biol 139, 1861-1872. Loeb, D.M., Evron, E., Patel, C.B., Sharma, P.M., Niranjan, B., Buluwela, L., Weitzman, S.A., Korz, D., and Sukumar, S. (2001). Wilms' tumor suppressor gene (WT1) is expressed in primary breast tumors despite tumor-specific promoter methylation. Cancer Res 61, 921-925. Lydon, J.P., DeMayo, F.J., Funk, C.R., Mani, S.K., Hughes, A.R., Montgomery, C.A., Jr., Shyamala, G., Conneely, O.M., and O'Malley, B.W. (1995). Mice lacking progesterone receptor exhibit pleiotropic reproductive abnormalities. Genes Dev 9, 2266-2278. Madin, S.H., Andriese, P.C., and Darby, N.B. (1957). The in vitro cultivation of tissues of domestic and laboratory animals. Am J Vet Res 18, 932-941. Makretsov, N., Gilks, C.B., Coldman, A.J., Hayes, M., and Huntsman, D. (2003). Tissue microarray analysis of neuroendocrine differentiation and its prognostic significance in breast cancer. Hum Pathol 34, 1001-1008. 289  Marinho, V.F., Metze, K., Sanches, F.S., Rocha, G.F., and Gobbi, H. (2008). Lymph vascular invasion in invasive mammary carcinomas identified by the endothelial lymphatic marker D2-40 is associated with other indicators of poor prognosis. BMC Cancer 8, 64. Martin-Belmonte, F., Gassama, A., Datta, A., Yu, W., Rescher, U., Gerke, V., and Mostov, K. (2007). PTEN-mediated apical segregation of phosphoinositides controls epithelial morphogenesis through Cdc42. Cell 128, 383-397. Maudsley, S., Zamah, A.M., Rahman, N., Blitzer, J.T., Luttrell, L.M., Lefkowitz, R.J., and Hall, R.A. (2000). Platelet-derived growth factor receptor association with Na(+)/H(+) exchanger regulatory factor potentiates receptor activity. Mol Cell Biol 20, 8352-8363. Mauro, L., Sisci, D., Bartucci, M., Salerno, M., Kim, J., Tam, T., Guvakova, M.A., Ando, S., and Surmacz, E. (1999). SHC-alpha5beta1 integrin interactions regulate breast cancer cell adhesion and motility. Exp Cell Res 252, 439-448. McLachlan, E., Shao, Q., and Laird, D.W. (2007). Connexins and gap junctions in mammary gland development and breast cancer progression. J Membr Biol 218, 107-121. McNagny, K.M., Pettersson, I., Rossi, F., Flamme, I., Shevchenko, A., Mann, M., and Graf, T. (1997). Thrombomucin, a novel cell surface protein that defines thrombocytes and multipotent hematopoietic progenitors. J Cell Biol 138, 1395-1407. Meder, D., Shevchenko, A., Simons, K., and Fullekrug, J. (2005). Gp135/podocalyxin and NHERF-2 participate in the formation of a preapical domain during polarization of MDCK cells. J Cell Biol 168, 303-313. Mercurio, A.M. (1995). Laminin receptors: achieving specificity through cooperation. Trends Cell Biol 5, 419-423. Miettinen, A., Solin, M.L., Reivinen, J., Juvonen, E., Vaisanen, R., and Holthofer, H. (1999). Podocalyxin in rat platelets and megakaryocytes. Am J Pathol 154, 813-822. Morales, F.C., Takahashi, Y., Kreimann, E.L., and Georgescu, M.M. (2004). Ezrinradixin-moesin (ERM)-binding phosphoprotein 50 organizes ERM proteins at the apical membrane of polarized epithelia. Proc Natl Acad Sci U S A 101, 17705-17710. Mosmann, T. (1983). Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J Immunol Methods 65, 55-63. Mostov, K., and Martin-Belmonte, F. (2006). Developmental biology: the hole picture. Nature 442, 363-364. Mostov, K., Su, T., and ter Beest, M. (2003). Polarized epithelial membrane traffic: conservation and plasticity. Nat Cell Biol 5, 287-293. 290  Moyer, B.D., Duhaime, M., Shaw, C., Denton, J., Reynolds, D., Karlson, K.H., Pfeiffer, J., Wang, S., Mickle, J.E., Milewski, M., et al. (2000). The PDZ-interacting domain of cystic fibrosis transmembrane conductance regulator is required for functional expression in the apical plasma membrane. J Biol Chem 275, 27069-27074. Mueller, H., Loop, P., Liu, R., Wosikowski, K., Kueng, W., and Eppenberger, U. (1994). Differential signal transduction of epidermal-growth-factor receptors in hormone-dependent and hormone-independent human breast cancer cells. Eur J Biochem 221, 631-637. Muschler, J., Lochter, A., Roskelley, C.D., Yurchenco, P., and Bissell, M.J. (1999). Division of labor among the alpha6beta4 integrin, beta1 integrins, and an E3 laminin receptor to signal morphogenesis and beta-casein expression in mammary epithelial cells. Mol Biol Cell 10, 2817-2828. Muthuswamy, S.K., Li, D., Lelievre, S., Bissell, M.J., and Brugge, J.S. (2001). ErbB2, but not ErbB1, reinitiates proliferation and induces luminal repopulation in epithelial acini. Nat Cell Biol 3, 785-792. Nelson, C.M., and Bissell, M.J. (2006). Of extracellular matrix, scaffolds, and signaling: tissue architecture regulates development, homeostasis, and cancer. Annu Rev Cell Dev Biol 22, 287-309. Nelson, W.J. (2003). Adaptation of core mechanisms to generate cell polarity. Nature 422, 766-774. Nelson, W.J., and Yeaman, C. (2001). Protein trafficking in the exocytic pathway of polarized epithelial cells. Trends Cell Biol 11, 483-486. Ney, J.T., Zhou, H., Sipos, B., Buttner, R., Chen, X., Kloppel, G., and Gutgemann, I. (2007). Podocalyxin-like protein 1 expression is useful to differentiate pancreatic ductal adenocarcinomas from adenocarcinomas of the biliary and gastrointestinal tracts. Hum Pathol 38, 359-364. Ng, T., Parsons, M., Hughes, W.E., Monypenny, J., Zicha, D., Gautreau, A., Arpin, M., Gschmeissner, S., Verveer, P.J., Bastiaens, P.I., et al. (2001). Ezrin is a downstream effector of trafficking PKC-integrin complexes involved in the control of cell motility. Embo J 20, 2723-2741. Nielsen, J.S., Doyonnas, R., and McNagny, K.M. (2002). Avian models to study the transcriptional control of hematopoietic lineage commitment and to identify lineage-specific genes. Cells Tissues Organs 171, 44-63. Nielsen, J.S., Graves, M.L., Chelliah, S., Vogl, A.W., Roskelley, C.D., and McNagny, K.M. (2007). The CD34-related molecule podocalyxin is a potent inducer of microvillus formation. PLoS ONE 2, e237.  291  Nobes, C.D., and Hall, A. (1995). Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia. Cell 81, 53-62. O'Brien, L.E., Jou, T.S., Pollack, A.L., Zhang, Q., Hansen, S.H., Yurchenco, P., and Mostov, K.E. (2001). Rac1 orientates epithelial apical polarity through effects on basolateral laminin assembly. Nat Cell Biol 3, 831-838. O'Brien, L.E., Zegers, M.M., and Mostov, K.E. (2002). Opinion: Building epithelial architecture: insights from three-dimensional culture models. Nat Rev Mol Cell Biol 3, 531-537. Ojakian, G.K., and Schwimmer, R. (1988). The polarized distribution of an apical cell surface glycoprotein is maintained by interactions with the cytoskeleton of Madin-Darby canine kidney cells. J Cell Biol 107, 2377-2387. Ojakian, G.K., and Schwimmer, R. (1994). Regulation of epithelial cell surface polarity reversal by beta 1 integrins. J Cell Sci 107 ( Pt 3), 561-576. Oji, Y., Suzuki, T., Nakano, Y., Maruno, M., Nakatsuka, S., Jomgeow, T., Abeno, S., Tatsumi, N., Yokota, A., Aoyagi, S., et al. (2004). Overexpression of the Wilms' tumor gene W T1 in primary astrocytic tumors. Cancer Sci 95, 822-827. Oji, Y., Yamamoto, H., Nomura, M., Nakano, Y., Ikeba, A., Nakatsuka, S., Abeno, S., Kiyotoh, E., Jomgeow, T., Sekimoto, M., et al. (2003). Overexpression of the Wilms' tumor gene WT1 in colorectal adenocarcinoma. Cancer Sci 94, 712-717. Orlando, R.A., Takeda, T., Zak, B., Schmieder, S., Benoit, V.M., McQuistan, T., Furthmayr, H., and Farquhar, M.G. (2001). The glomerular epithelial cell antiadhesin podocalyxin associates with the actin cytoskeleton through interactions with ezrin. J Am Soc Nephrol 12, 1589-1598. Ormandy, C.J., Camus, A., Barra, J., Damotte, D., Lucas, B., Buteau, H., Edery, M., Brousse, N., Babinet, C., Binart, N., et al. (1997). Null mutation of the prolactin receptor gene produces multiple reproductive defects in the mouse. Genes Dev 11, 167-178. Palmer, R.E., Kotsianti, A., Cadman, B., Boyd, T., Gerald, W., and Haber, D.A. (2001). WT1 regulates the expression of the major glomerular podocyte membrane protein Podocalyxin. Curr Biol 11, 1805-1809. Park, C.C., Zhang, H., Pallavicini, M., Gray, J.W., Baehner, F., Park, C.J., and Bissell, M.J. (2006). Beta1 integrin inhibitory antibody induces apoptosis of breast cancer cells, inhibits growth, and distinguishes malignant from normal phenotype in three dimensional cultures and in vivo. Cancer Res 66, 1526-1535. Paszek, M.J., Zahir, N., Johnson, K.R., Lakins, J.N., Rozenberg, G.I., Gefen, A., Reinhart-King, C.A., Margulies, S.S., Dembo, M., Boettiger, D., et al. (2005). Tensional homeostasis and the malignant phenotype. Cancer Cell 8, 241-254. 292  Perez-Moreno, M., Jamora, C., and Fuchs, E. (2003). Sticky business: orchestrating cellular signals at adherens junctions. Cell 112, 535-548. Perou, C.M., Sorlie, T., Eisen, M.B., van de Rijn, M., Jeffrey, S.S., Rees, C.A., Pollack, J.R., Ross, D.T., Johnsen, H., Akslen, L.A., et al. (2000). Molecular portraits of human breast tumours. Nature 406, 747-752. Petersen, O.W., Ronnov-Jessen, L., Howlett, A.R., 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. Proc Natl Acad Sci U S A 89, 9064-9068. Pinkas, J., and Leder, P. (2002). MEK1 signaling mediates transformation and metastasis of EpH4 mammary epithelial cells independent of an epithelial to mesenchymal transition. Cancer Res 62, 4781-4790. Plant, P.J., Fawcett, J.P., Lin, D.C., Holdorf, A.D., Binns, K., Kulkarni, S., and Pawson, T. (2003). A polarity complex of mPar-6 and atypical PKC binds, phosphorylates and regulates mammalian Lgl. Nat Cell Biol 5, 301-308. Prag, S., Parsons, M., Keppler, M.D., Ameer-Beg, S.M., Barber, P., Hunt, J., Beavil, A.J., Calvert, R., Arpin, M., Vojnovic, B., et al. (2007). Activated ezrin promotes cell migration through recruitment of the GEF Dbl to lipid rafts and preferential downstream activation of Cdc42. Mol Biol Cell 18, 2935-2948. Reczek, D., Berryman, M., and Bretscher, A. (1997). Identification of EBP50: A PDZcontaining phosphoprotein that associates with members of the ezrin-radixinmoesin family. J Cell Biol 139, 169-179. Reichmann, E., Ball, R., Groner, B., and Friis, R.R. (1989). New mammary epithelial and fibroblastic cell clones in coculture form structures competent to differentiate functionally. J Cell Biol 108, 1127-1138. Reichmann, E., Schwarz, H., Deiner, E.M., Leitner, I., Eilers, M., Berger, J., Busslinger, M., and Beug, H. (1992). Activation of an inducible c-FosER fusion protein causes loss of epithelial polarity and triggers epithelial-fibroblastoid cell conversion. Cell 71, 1103-1116. Ren, S., Cai, H.R., Li, M., and Furth, P.A. (2002). Loss of Stat5a delays mammary cancer progression in a mouse model. Oncogene 21, 4335-4339. Renshaw, M.W., Toksoz, D., and Schwartz, M.A. (1996). Involvement of the small GTPase rho in integrin-mediated activation of mitogen-activated protein kinase. J Biol Chem 271, 21691-21694. Richert, M.M., Schwertfeger, K.L., Ryder, J.W., and Anderson, S.M. (2000). An atlas of mouse mammary gland development. J Mammary Gland Biol Neoplasia 5, 227241.  293  Ridley, A.J. (2006). Rho GTPases and actin dynamics in membrane protrusions and vesicle trafficking. Trends Cell Biol 16, 522-529. Ridley, A.J., and Hall, A. (1992). The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. Cell 70, 389-399. Roberts, S.G. (2005). Transcriptional regulation by WT1 in development. Curr Opin Genet Dev 15, 542-547. Robinson, G.W. (2007). Cooperation of signalling pathways in embryonic mammary gland development. Nat Rev Genet 8, 963-972. Rodriguez-Boulan, E., Paskiet, K.T., and Sabatini, D.D. (1983). Assembly of enveloped viruses in Madin-Darby canine kidney cells: polarized budding from single attached cells and from clusters of cells in suspension. J Cell Biol 96, 866-874. Roskelley, C.D., Desprez, P.Y., and Bissell, M.J. (1994). Extracellular matrixdependent tissue-specific gene expression in mammary epithelial cells requires both physical and biochemical signal transduction. Proc Natl Acad Sci U S A 91, 1237812382. Russo, J., Moral, R., Balogh, G.A., Mailo, D., and Russo, I.H. (2005). The protective role of pregnancy in breast cancer. Breast Cancer Res 7, 131-142. Sahai, E., and Marshall, C.J. (2003). Differing modes of tumour cell invasion have distinct requirements for Rho/ROCK signalling and extracellular proteolysis. Nat Cell Biol 5, 711-719. Sarrio, D., Rodriguez-Pinilla, S.M., Dotor, A., Calero, F., Hardisson, D., and Palacios, J. (2006). Abnormal ezrin localization is associated with clinicopathological features in invasive breast carcinomas. Breast Cancer Res Treat 98, 71-79. Sasaki, T., Fassler, R., and Hohenester, E. (2004). Laminin: the crux of basement membrane assembly. J Cell Biol 164, 959-963. Sathyanarayana, P., Menon, M.P., Bogacheva, O., Bogachev, O., Niss, K., Kapelle, W.S., Houde, E., Fang, J., and Wojchowski, D.M. (2007). Erythropoietin modulation of podocalyxin and a proposed erythroblast niche. Blood 110, 509-518. Sawada, H., Stukenbrok, H., Kerjaschki, D., and Farquhar, M.G. (1986). Epithelial polyanion (podocalyxin) is found on the sides but not the soles of the foot processes of the glomerular epithelium. Am J Pathol 125, 309-318. Schlichting, K., Wilsch-Brauninger, M., Demontis, F., and Dahmann, C. (2006). Cadherin Cad99C is required for normal microvilli morphology in Drosophila follicle cells. J Cell Sci 119, 1184-1195.  294  Schmidhauser, C., Casperson, G.F., Myers, C.A., Sanzo, K.T., Bolten, S., and Bissell, M.J. (1992). A novel transcriptional enhancer is involved in the prolactin- and extracellular matrix-dependent regulation of beta-casein gene expression. Mol Biol Cell 3, 699-709. Schmieder, S., Nagai, M., Orlando, R.A., Takeda, T., and Farquhar, M.G. (2004). Podocalyxin activates RhoA and induces actin reorganization through NHERF1 and Ezrin in MDCK cells. J Am Soc Nephrol 15, 2289-2298. Schnabel, E., Dekan, G., Miettinen, A., and Farquhar, M.G. (1989). Biogenesis of podocalyxin--the major glomerular sialoglycoprotein--in the newborn rat kidney. Eur J Cell Biol 48, 313-326. Schopperle, W.M., Kershaw, D.B., and DeWolf, W.C. (2003). Human embryonal carcinoma tumor antigen, Gp200/GCTM-2, is podocalyxin. Biochem Biophys Res Commun 300, 285-290. Schwimmer, R., and Ojakian, G.K. (1995). The alpha 2 beta 1 integrin regulates collagen-mediated MDCK epithelial membrane remodeling and tubule formation. J Cell Sci 108 ( Pt 6), 2487-2498. Seiler, M.W., Rennke, H.G., Venkatachalam, M.A., and Cotran, R.S. (1977). Pathogenesis of polycation-induced alterations ("fusion") of glomerular epithelium. Lab Invest 36, 48-61. Shackleton, M., Vaillant, F., Simpson, K.J., Stingl, J., Smyth, G.K., Asselin-Labat, M.L., Wu, L., Lindeman, G.J., and Visvader, J.E. (2006). Generation of a functional mammary gland from a single stem cell. Nature 439, 84-88. Shenolikar, S., Voltz, J.W., Minkoff, C.M., Wade, J.B., and Weinman, E.J. (2002). Targeted disruption of the mouse NHERF-1 gene promotes internalization of proximal tubule sodium-phosphate cotransporter type IIa and renal phosphate wasting. Proc Natl Acad Sci U S A 99, 11470-11475. Shenolikar, S., and Weinman, E.J. (2001). NHERF: targeting and trafficking membrane proteins. Am J Physiol Renal Physiol 280, F389-395. Shin, K., Fogg, V.C., and Margolis, B. (2006). Tight junctions and cell polarity. Annu Rev Cell Dev Biol 22, 207-235. Silberstein, G.B., Van Horn, K., Strickland, P., Roberts, C.T., Jr., and Daniel, C.W. (1997). Altered expression of the WT1 wilms tumor suppressor gene in human breast cancer. Proc Natl Acad Sci U S A 94, 8132-8137. Sizemore, S., Cicek, M., Sizemore, N., Ng, K.P., and Casey, G. (2007). Podocalyxin increases the aggressive phenotype of breast and prostate cancer cells in vitro through its interaction with ezrin. Cancer Res 67, 6183-6191.  295  Slade, M.J., Coope, R.C., Gomm, J.J., and Coombes, R.C. (1999). The human mammary gland basement membrane is integral to the polarity of luminal epithelial cells. Exp Cell Res 247, 267-278. Somasiri, A., Nielsen, J.S., Makretsov, N., McCoy, M.L., Prentice, L., Gilks, C.B., Chia, S.K., Gelmon, K.A., Kershaw, D.B., Huntsman, D.G., et al. (2004). Overexpression of the anti-adhesin podocalyxin is an independent predictor of breast cancer progression. Cancer Res 64, 5068-5073. Somasiri, A., Wu, C., Ellchuk, T., Turley, S., and Roskelley, C.D. (2000). Phosphatidylinositol 3-kinase is required for adherens junction-dependent mammary epithelial cell spheroid formation. Differentiation 66, 116-125. Soule, H.D., Vazguez, J., Long, A., Albert, S., and Brennan, M. (1973). A human cell line from a pleural effusion derived from a breast carcinoma. J Natl Cancer Inst 51, 1409-1416. Stanhope-Baker, P., Kessler, P.M., Li, W., Agarwal, M.L., and Williams, B.R. (2004). The Wilms tumor suppressor-1 target gene podocalyxin is transcriptionally repressed by p53. J Biol Chem 279, 33575-33585. Stemmer-Rachamimov, A.O., Wiederhold, T., Nielsen, G.P., James, M., PinneyMichalowski, D., Roy, J.E., Cohen, W.A., Ramesh, V., and Louis, D.N. (2001). NHERF, a merlin-interacting protein, is primarily expressed in luminal epithelia, proliferative endometrium, and estrogen receptor-positive breast carcinomas. Am J Pathol 158, 57-62. Stingl, J., Eirew, P., Ricketson, I., Shackleton, M., Vaillant, F., Choi, D., Li, H.I., and Eaves, C.J. (2006). Purification and unique properties of mammary epithelial stem cells. Nature 439, 993-997. Stokes, D.L. (2007). Desmosomes from a structural perspective. Curr Opin Cell Biol 19, 565-571. Streuli, C.H., Bailey, N., and Bissell, M.J. (1991). Control of mammary epithelial differentiation: basement membrane induces tissue-specific gene expression in the absence of cell-cell interaction and morphological polarity. J Cell Biol 115, 13831395. Streuli, C.H., and Bissell, M.J. (1990). Expression of extracellular matrix components is regulated by substratum. J Cell Biol 110, 1405-1415. Streuli, C.H., Schmidhauser, C., Bailey, N., Yurchenco, P., Skubitz, A.P., Roskelley, C., and Bissell, M.J. (1995). Laminin mediates tissue-specific gene expression in mammary epithelia. J Cell Biol 129, 591-603. Takeda, T., Go, W.Y., Orlando, R.A., and Farquhar, M.G. (2000). Expression of podocalyxin inhibits cell-cell adhesion and modifies junctional properties in MadinDarby canine kidney cells. Mol Biol Cell 11, 3219-3232. 296  Takeda, T., McQuistan, T., Orlando, R.A., and Farquhar, M.G. (2001). Loss of glomerular foot processes is associated with uncoupling of podocalyxin from the actin cytoskeleton. J Clin Invest 108, 289-301. Tan, P.C., Furness, S.G., Merkens, H., Lin, S., McCoy, M.L., Roskelley, C.D., Kast, J., and McNagny, K.M. (2006). NHERF-1 is an Hematopoietic Ligand for a Subset of the CD34 Family of Stem Cell Surface Proteins. Stem Cells. Tanentzapf, G., and Tepass, U. (2003). Interactions between the crumbs, lethal giant larvae and bazooka pathways in epithelial polarization. Nat Cell Biol 5, 46-52. Thiery, J.P. (2002). Epithelial-mesenchymal transitions in tumour progression. Nat Rev Cancer 2, 442-454. Trudeau, M.E., Pritchard, K.I., Chapman, J.A., Hanna, W.M., Kahn, H.J., Murray, D., Sawka, C.A., Mobbs, B.G., Andrulis, I., McCready, D.R., et al. (2005). Prognostic factors affecting the natural history of node-negative breast cancer. Breast Cancer Res Treat 89, 35-45. Underwood, J.M., Imbalzano, K.M., Weaver, V.M., Fischer, A.H., Imbalzano, A.N., and Nickerson, J.A. (2006). The ultrastructure of MCF-10A acini. J Cell Physiol 208, 141-148. Vargo-Gogola, T., and Rosen, J.M. (2007). Modelling breast cancer: one size does not fit all. Nat Rev Cancer 7, 659-672. Vega-Salas, D.E., Salas, P.J., Gundersen, D., and Rodriguez-Boulan, E. (1987). Formation of the apical pole of epithelial (Madin-Darby canine kidney) cells: polarity of an apical protein is independent of tight junctions while segregation of a basolateral marker requires cell-cell interactions. J Cell Biol 104, 905-916. Vega-Salas, D.E., Salas, P.J., and Rodriguez-Boulan, E. (1988). Exocytosis of vacuolar apical compartment (VAC): a cell-cell contact controlled mechanism for the establishment of the apical plasma membrane domain in epithelial cells. J Cell Biol 107, 1717-1728. Visvader, J.E., and Lindeman, G.J. (2006). Mammary stem cells and mammopoiesis. Cancer Res 66, 9798-9801. Vitureira, N., McNagny, K., Soriano, E., and Burgaya, F. (2005). Pattern of expression of the podocalyxin gene in the mouse brain during development. Gene Expr Patterns 5, 349-354. Walker, S.J., and Brugge, J.S. (2006). SePARating polarity and proliferation in ErbB2 oncogenesis. Nat Cell Biol 8, 1220-1222. Wang, A.Z., Ojakian, G.K., and Nelson, W.J. (1990a). Steps in the morphogenesis of a polarized epithelium. I. Uncoupling the roles of cell-cell and cell-substratum 297  contact in establishing plasma membrane polarity in multicellular epithelial (MDCK) cysts. J Cell Sci 95 ( Pt 1), 137-151. Wang, A.Z., Ojakian, G.K., and Nelson, W.J. (1990b). Steps in the morphogenesis of a polarized epithelium. II. Disassembly and assembly of plasma membrane domains during reversal of epithelial cell polarity in multicellular epithelial (MDCK) cysts. J Cell Sci 95 ( Pt 1), 153-165. Wang, A.Z., Wang, J.C., Ojakian, G.K., and Nelson, W.J. (1994). Determinants of apical membrane formation and distribution in multicellular epithelial MDCK cysts. Am J Physiol 267, C473-481. Wang, F., Hansen, R.K., Radisky, D., Yoneda, T., Barcellos-Hoff, M.H., Petersen, O.W., Turley, E.A., and Bissell, M.J. (2002). Phenotypic reversion or death of cancer cells by altering signaling pathways in three-dimensional contexts. J Natl Cancer Inst 94, 1494-1503. Wang, F., Weaver, V.M., Petersen, O.W., Larabell, C.A., Dedhar, S., Briand, P., Lupu, R., and Bissell, M.J. (1998). Reciprocal interactions between beta1-integrin and epidermal growth factor receptor in three-dimensional basement membrane breast cultures: a different perspective in epithelial biology. Proc Natl Acad Sci U S A 95, 14821-14826. Weaver, V.M., Lelievre, S., Lakins, J.N., Chrenek, M.A., Jones, J.C., Giancotti, F., Werb, Z., and Bissell, M.J. (2002). beta4 integrin-dependent formation of polarized three-dimensional architecture confers resistance to apoptosis in normal and malignant mammary epithelium. Cancer Cell 2, 205-216. Weaver, V.M., Petersen, O.W., Wang, F., Larabell, C.A., Briand, P., Damsky, C., and Bissell, M.J. (1997). Reversion of the malignant phenotype of human breast cells in three-dimensional culture and in vivo by integrin blocking antibodies. J Cell Biol 137, 231-245. Weinman, E.J., Hall, R.A., Friedman, P.A., Liu-Chen, L.Y., and Shenolikar, S. (2005). The Association of NHERF Adaptor Proteins with G Protein-Coupled Receptors and Receptor Tyrosine Kinases. Annu Rev Physiol. Weinman, E.J., Steplock, D., Corry, D., and Shenolikar, S. (1993). Identification of the human NHE-1 form of Na(+)-H+ exchanger in rabbit renal brush border membranes. J Clin Invest 91, 2097-2102. Weinman, E.J., Steplock, D., Donowitz, M., and Shenolikar, S. (2000). NHERF associations with sodium-hydrogen exchanger isoform 3 (NHE3) and ezrin are essential for cAMP-mediated phosphorylation and inhibition of NHE3. Biochemistry 39, 6123-6129. Weinman, E.J., Steplock, D., Wang, Y., and Shenolikar, S. (1995). Characterization of a protein cofactor that mediates protein kinase A regulation of the renal brush border membrane Na(+)-H+ exchanger. J Clin Invest 95, 2143-2149. 298  Weir, M.L., Oppizzi, M.L., Henry, M.D., Onishi, A., Campbell, K.P., Bissell, M.J., and Muschler, J.L. (2006). Dystroglycan loss disrupts polarity and beta-casein induction in mammary epithelial cells by perturbing laminin anchoring. J Cell Sci 119, 4047-4058. Weis, W.I., and Nelson, W.J. (2006). Re-solving the cadherin-catenin-actin conundrum. J Biol Chem 281, 35593-35597. Wicki, A., Lehembre, F., Wick, N., Hantusch, B., Kerjaschki, D., and Christofori, G. (2006). Tumor invasion in the absence of epithelial-mesenchymal transition: podoplanin-mediated remodeling of the actin cytoskeleton. Cancer Cell 9, 261-272. Wong, A.S., Maines-Bandiera, S.L., Rosen, B., Wheelock, M.J., Johnson, K.R., Leung, P.C., Roskelley, C.D., and Auersperg, N. (1999). Constitutive and conditional cadherin expression in cultured human ovarian surface epithelium: influence of family history of ovarian cancer. Int J Cancer 81, 180-188. Yamanaka, T., Horikoshi, Y., Izumi, N., Suzuki, A., Mizuno, K., and Ohno, S. (2006). Lgl mediates apical domain disassembly by suppressing the PAR-3-aPKC-PAR-6 complex to orient apical membrane polarity. J Cell Sci 119, 2107-2118. Yamanaka, T., Horikoshi, Y., Suzuki, A., Sugiyama, Y., Kitamura, K., Maniwa, R., Nagai, Y., Yamashita, A., Hirose, T., Ishikawa, H., et al. (2001). PAR-6 regulates aPKC activity in a novel way and mediates cell-cell contact-induced formation of the epithelial junctional complex. Genes Cells 6, 721-731. Yang, J., Mani, S.A., Donaher, J.L., Ramaswamy, S., Itzykson, R.A., Come, C., Savagner, P., Gitelman, I., Richardson, A., and Weinberg, R.A. (2004). Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell 117, 927-939. Yeaman, C., Grindstaff, K.K., and Nelson, W.J. (2004). Mechanism of recruiting Sec6/8 (exocyst) complex to the apical junctional complex during polarization of epithelial cells. J Cell Sci 117, 559-570. Yoder, B.J., Wilkinson, E.J., and Massoll, N.A. (2007). Molecular and morphologic distinctions between infiltrating ductal and lobular carcinoma of the breast. Breast J 13, 172-179. Yu, C.Y., Chen, J.Y., Lin, Y.Y., Shen, K.F., Lin, W.L., Chien, C.L., ter Beest, M.B., and Jou, T.S. (2007a). A bipartite signal regulates the faithful delivery of apical domain marker podocalyxin/Gp135. Mol Biol Cell 18, 1710-1722. Yu, W., Datta, A., Leroy, P., O'Brien, L.E., Mak, G., Jou, T.S., Matlin, K.S., Mostov, K.E., and Zegers, M.M. (2005). Beta1-integrin orients epithelial polarity via Rac1 and laminin. Mol Biol Cell 16, 433-445.  299  Yu, W., Fang, X., Ewald, A., Wong, K., Hunt, C.A., Werb, Z., Matthay, M.A., and Mostov, K. (2007b). Formation of cysts by alveolar type II cells in three-dimensional culture reveals a novel mechanism for epithelial morphogenesis. Mol Biol Cell 18, 1693-1700. Zelhof, A.C., and Hardy, R.W. (2004). WASp is required for the correct temporal morphogenesis of rhabdomere microvilli. J Cell Biol 164, 417-426.  300  

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.24.1-0066552/manifest

Comment

Related Items