UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Deciphering the role of beta-catenin in prostate cancer Mulholland, David J. 2006

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2006-130253.pdf [ 36.03MB ]
Metadata
JSON: 831-1.0092861.json
JSON-LD: 831-1.0092861-ld.json
RDF/XML (Pretty): 831-1.0092861-rdf.xml
RDF/JSON: 831-1.0092861-rdf.json
Turtle: 831-1.0092861-turtle.txt
N-Triples: 831-1.0092861-rdf-ntriples.txt
Original Record: 831-1.0092861-source.json
Full Text
831-1.0092861-fulltext.txt
Citation
831-1.0092861.ris

Full Text

DECIPHERING THE ROLE OF BETA-CATENIN IN PROSTATE CANCER By DAVID J MULHOLLAND M . S c , University of British Columbia, 1999 B.Sc , University of British Columbia, 1996 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY In THE F A C U L T Y OF G R A D U A T E STUDIES (Experimental Medicine) THE UNIVERSITY OF BRITISH COLUMBIA December 2005 © David J Mulholland, 2005 ) 11 Abstract Prostate cancer (PrCa) is the second leading cause of cancer related mortality in North American men. Treatment of advanced human PrCa occurs by way of hormone ablation therapy promoting an initial regression of tumour size and serum PSA levels. However, following androgen ablation is the inevitable occurrence of androgen independence (AI) and the lethal phenotype of metastasis. While the oncogenic Wnt/p-catenin/Tcf signaling axis is well correlated with colon cancer, increasing evidence suggests that P-catenin is also a contributor to progression of PrCa. p-catenin binds A R , in a ligand sensitive manner, to alter A R ligand specificity thereby promoting its response to non-gonadal ligands including androstenedione and estradiol. This thesis puts forth new knowledge regarding the functional interactions between AR, Wnt/p-catenin/Tcf and P B K / A k t signaling. Specifically, Wnt3a promotes ligand independent AR/p-catenin interactions, A R transactivation, PSA production and enhanced PrCa cell proliferation (Chapter 2). Upon exposure to androgens, A R and P-catenin undergo ligand dependent, nuclear co-translocation and formation of AR/p-catenin nuclear complexing. In the presence of androgens, p-catenin and A R complex at regulatory sites for A R transcription, thereby conferring the ability of P-catenin to function as an A R co-activator (Chapter 3). Upon androgen activation, a limited pool of P-catenin yields augmented A R signaling but diminished P-catenin/Tcf signaling. Functionally, this is achieved by way of a competitive balance of nuclear P-catenin between A R and Tcf transcriptional sites (Chapter 4). Accordingly, cells that are highly p-catenin/Tcf dependent demonstrate reduced cyclin DI activity, altered cell cycle and reduced cell proliferation. PTEN expression serves as a critical check-point in PrCa advancement, as a potent regulator both of p-catenin/Tcf function and, therefore, A R function. In vitro, PTEN regulates these effectors via GSK3P and ILK, two kinases with critical regulatory functions in PBK/Akt dependent cancers (Chapter 5). In vivo, gain-of-function (stable, inducible PTEN PrCa cells) and loss-of-function (PTEN -/- mouse model) analysis support that PTEN serves to regulate nuclear localization of p-catenin. Human tissue microarray analysis indicates that increased P-catenin expression is not only associated with increasing Gleason grade assignment and mitotic index but also with benign prostate pathologies (Chapter 6). These data indicate a dynamic and functional role for P-catenin in PrCa systems. Ul Table of Contents Abstract ii Table of Contents iii List of Figures viii List of Tables : xiv List of Abbreviations and Symbols xv Acknowledgements xviii Chapter One: Introduction 1 1.1 Prostate cancer 2 1.2 The prostatic epithelium 3 1.3 Genetic changes during progressive PrCa 5 1.4 Androgens, the hormone axis and treatment of PrCa 5 1.5 The AR: structure and function 7 1.6 Evidence that A R is functionally important in PrCa progression 8 1.7 Pro-survival signaling pathways and progression of PrCa 10 1.8 Wnt signaling: activation and inactivation 11 1.9 /?-catenin: structure & function 13 1.10 Nuclear import of /?-catenin 13 1.11 Transcriptional regulation by P-catenin 14 1.12 P-catenin in tumourigenesis 14 1.13 Alterations in E-cadherin/p-catenin mediated adhesion during PrCa progression 15 1.14'AR and PI3K/Akt provide essential signaling events for acquisition of androgen independent PrCa 16 1.15 Intact PBK/Akt and A R signaling: a mutualistic relationship 18 1.16 Phosphatase and Tensin Homologue Deleted (PTEN): structure, function and significance in PrCa 1.17 P-catenin and A R interactions: A statement of other published and related works 1.17.1 P-catenin Binding and Co-activation of A R 1.17.2 Glycogen synthase kinase (GSK3P) Modulation of AR/p-catenin Interactions... 1.17.3 P-catenin Co-trafficking and Transrepression of Tcf-mediated transcription by AR 1.17.4 PTEN as a Potent Regulator of A R and P-catenin Function 1:17.5 The Potential Clinical Relevence of P-catenin in Prostate Cancer 1.18 Thesis rationale and statements of hypotheses Chapter Two: Wnt/p-catenin Activation of Androgen Receptor 2.1 Introduction 2.2 Methods 2.2.1 Tissue culture : 2.2.2 Dextran-charcoal stripped fetal bovine serum (CSS) 2.2.3 Luciferase reporter assays 2.2.4 Generation of LNCaP-WT p-catenin and LNCaP-A(27-45)p-catenin cells 2.2.5 Generation of lentivirus 2.2.6 Colony forming assays 2.2.7 MTS assays 2.2.8 PSA ELISA 2.3 Results 2.3.1 Relative expression of Wnt effectors and activity in PrCa cell lines 2.3.2 Generation of inducible lentiviral expression constructs for WT and GSK3P-resistant P-catenin 2.3.3 The Wnt/p-catenin/Tcf axis is functional in PrCa cells 2.3.4 Wnt3a recapitulates the effects of stabilized P-catenin and promotes ligand independent activation of A R 2.3.5 Wnt3a enhances production of PSA 2.3.6 Wnt3a enhances ligand sensitive AR/p-catenin interactions 2.3.7 Wnt3a promotes ligand-independent proliferation of LNCap cells and provides resistance to LY294002 treatment in PC3 cells 2.3.8 Wnt3a enhances LNCaP cell colony formation 62 2.4 Discussion 81 Chapter Three: The Androgen Receptor Mediates Nuclear Translocation of p-catenin....87 3.1 Introduction 88 3.2 Methods 90 3.2.1 Cell culture '. 90 3.2.2 Dextran-charcoal stripped fetal bovine serum (CSS) 90 3.3.3 Immunocytochemistry 90 3.3.4 Western blotting 91 3.3.5 Co-immunoprecipitations 91 3.3.6 Cell fractionation and time course study 92 3.3.7 Glutathione synthase transferase (GST)-pull downs 92 3.3.8 Acrydite capture of DNA-binding complexes (ACDC) 92 3.3 Results 93 3.3.1 Co-localization of P-Catenin and A R 93 3.3.2 The A R - L B D interacts directly with the P-catenin arm repeats 94 3.3.3 Ligand-bound AR promotes nuclear translocation of cytosolic P-catenin 94 3.3.4 The A R - D B D / L B D is necessary and sufficient for nuclear translocation of P-catenin in PC3 cells 96 3.3.5 AR-mediated translocation of P-catenin is distinct from APC complexing 97 3.3.6 P-catenin and the A R complex directly on an androgen responsive promoter 97 3.3.7 p-catenin shows increased nuclear expression in castrastred mice reconstituted with exogenous androgen 98 3.4 Discussion 118 Chapter Four: AR Repression of Wnt/p-catenin/Tcf Signaling 123 4.1 Introduction 124 4.2 Methods 125 4.2.1 Polymerase chain reaction (PCR) 125 4.2.2 Plasmids 126 4.2.3 Luciferase reporter assays 127 vi 4.2.4 In vitro translation and recombinant proteins 127 4.2.5 FACS analysis 128 4.2.6 In vitro growth assays 128 4.2.7 Cell proliferation assay 128 4.2.8 Northern blot analysis 129 4.2.9 Western blotting 129 ' 4.3 Results 130 4.3.1 Repression of Tcf is both A R and androgen, dose dependent but can be relieved by Casodex 130 4.3.2 AR/DHT-mediated repression can overcome P-catenin activating mutants 131 4.3.3 Androgen-dependent localization of HcRed-Tcf with AR-EGFP in LNCaP cells.... 131 4.3.4 Confirmation of expression and transcriptional activity of AR-EGFP and HcRed-Tcfl32 4.3.5 In vivo and in vitro binding assays are suggestive that A R and Tcf compete for nuclear P-catenin 133 4.4 Discussion 154 Chapter Five: PTEN is a Potent Regulator of p-catenin/Tcf AR Signaling via GSK3p Axis 160 5.1 Introduction 161 5.2 Methods : 163 5.2.1 Cell culture & transfections 163 5.2.2 Western blotting 163 5.2.3 Immunocytochemistry '. 164 5.2.4 Luciferase reporter assays 164 5.2.5 Nuclear extracts and electrophoretic mobility shift assays 164 5.2.6 Nuclear extracts and co-immunoprecipitation assay 165 5.2.7 Pulse-chase analysis 166 5.2.8 Flow cytometric analysis 166 5.2.9 GSK3 kinase assay 166 5.2.10 Development of PTEN stable clones: clonal selection 167 5.2.11 Development of PTEN stable clones: Tentiviral system (Invitrogen) 167 5.2.12 RNA preparation 168 vii 5.2.13 Northern blotting 169 5.2.14 Cell proliferation (MTS) assays 169 5.2.15 AR-gfp, PTEN and DAPI cell counting 169 5.2.16 Colony forming assays 170 5.3 Results 170 5.3.1 Levels of nuclear P-catenin are reduced by expression of exongenous PTEN in PTEN-null PrCa cells 170 5.3.2 PTEN enhances degradation and phosphosphorylation of P-catenin 171 5.3.3 PTEN expression inhibits P-catenin/Tcf complexing and binding of a P-catenin/Tcf complex to a Tcf consensus oligonucleotide 171 5.3.4 Generation of stable "Knock-In" PTEN expressing cells by two different methods. 172 5.3.5 Maintaining non-"leaky" PTEN expression using the Invitrogen lentiviral tet-repressor and a VP-16 activator fusion 173 5.3.6 Validation of functional LNCaP-WT PTEN and C124S-PTEN clones 173 5.3.7 The PI3K chemical inhibitor, LY294002, causes a dose dependent inhibition of A R transcriptional reporter activity in LNCaP-PTEN cells (p45) 174 5.3.8 PTEN regulates P-catenin/Tcf signaling via PI3K/GSK3P signaling 174 5.3.9 PTEN can regulate AR by the PTEN/GSK3/p-catenin signaling axis 176 5.3.10 Androgens protect from PTEN mediated apoptosis 176 5.3.11 PTEN regulates the cellular distribution of AR 177 5.3.12 WT PTEN expression reduces nuclear AR and P-catenin, ligand-sensitive AR/p-catenin complexing but increases total A R expression 178 5.3.13 PTEN and androgens: opposing regulators of cell cycle and cell viability 179 5.3.14 PTEN as a regulator of cell proliferation, colony formation and the P-catenin/Tcf target gene, cyclin DI 181 5.4 Discussion 220 Chapter Six: In Vivo Analysis of P-catenin Expression 229 6.1 Introduction 230 6.2 Methods...: 232 6.2.1 LNCaP & LNCaP-PTEN xenograft models 232 6.2.2 Immunohistochemistry 233 Vlll 6.2.3 Prostate conditional PTEN knock-out 234 6.2.4 Neoadjuvant hormone therapy (NHT) tissue micro array (TMA) 234 6.2.5 Gleason grade tissue micro array (TMA) 235 6.2.6 Statistical analysis 235 6.3 Results 235 6.3.1 Profile of P-catenin transcript in the LNCaP xenograft tumour series 236 6.3.2 Analysis of p-catenin expression by LNCaP xenograft tissue array 236 6.3.3 Induction of PTEN expression in a LNCaP-WT PTEN xenograft 237 6.3.4 Induced PTEN may promote a more differentiated LNCaP xenograft 238 6.3.5 PTEN Regulates Distribution of A R and P-catenin in LNCaP xenografts 238 6.3.6 Conditional (prostate specific) deletion of PTEN promotes increased nuclear p-catenin : 239 6.3.7 P-catenin expression increases during histologically advanced PrCa 239 6.3.8 Increased P-catenin expression may be associated with benign prostate pathology..241 6.3.9 p-catenin expression decreases in response to anti-androgen therapy but increases in metastatic lesions 241 6.4 Discussion 271 Chapter Seven: Summary and Future Directions 279 7.1 Summary and future directions 280 References 288 Appendix 306 Supplementary Tables 306 Review: "Interaction of Nuclear Receptor with the Wnt/P-catenin/Tcf signaling Axis: Wnt you like to know?" 312 Review: "PTEN and GSK3p: Key Regulators in Progression to Androgen Independent Prostate Cancer." ; 330 ix List of Figures Chapter One: Introduction Figure 1.1 Histological and clinical grading system for progressive adenocarcinoma of the prostate 30 Figure 1.2 Progression of Human 32 PrCa Figure 1.3 Androgen epithelial entry and activation of A R gene targets 34 Figure 1.4 Rationale for total androgen blockade in PrCa therapy 36 Figure 1.5 Structural features of nuclear receptors and the androgen 38 receptor Figure 1.6 The Wnt degradative cascade depicted in an active or inactive 40 conformation Figure 1.7 Structural domains, binding partners, and armadillo repeats of 42 P-catenin Figure 1.8 The PI3K signaling axis in PrCa cells. 44 Figure 1.9 Structure and function of PTEN. 46 Chapter Two: Wnt and P-catenin activation of the Androgen Receptor Figure 2.1 The Wnt/P-catenin/Tcf axis is functional in PrCa cells 63 Figure 2.2 Generation of an activated 3-catenin lentiviral expression 65 system Figure 2.3 Use of Wnt3a overexpressing cells to generate Wnt3a conditioned media 67 Figure 2.4 GSK3P resistant P-catenin promotes enhanced A R transactivation in the presence of estrogen (E 2) and low concentrations of androgen 69 Figure 2.5 Wnt3a promotes enhanced A R transactivation in the presence of estrogen (E2) Figure 2.6 and low androgen concentrations Wnt3a/CSS media enhances PSA protein production in the absence and low levels of androgens x 71 73 Figure 2.7 Wnt3a, dose-dependently enhances complexing between A R and P-catenin but is antagonized by the presence of 75 Casodex Figure 2.8 Wnt3/CSS C M promotes proliferation of LNCaP cells both with low concentrations of androgens (0.1 n M R1881) and androgen 77 independently Figure 2.9 The "Dave-Grid" technique for counting soft agar colony formation 79 Chapter Three: Ligand dependent binding and nuclear co-translocation of the Androgen receptor & P-catenin Figure 3.1 Localization of the A R and P-catenin in methanol-fixed LNCaP cells treated with 10 n M R1881 for 60 min viewed with confocal laser microscopy 99 Figure 3.2 Mapping of in vitro interactions between recombinant A R deletion truncations) and in vitro translated P-catenin/S35 deletion truncations 101 Figure 3.3 LNCaP cell nuclear translocation of P-catenin in the presence of androgens over 60 min 103 Figure 3.4 Evaluation of P-catenin nuclear co-translocation with non-AR nuclear receptors and with truncating mutations for A R 106 Figure 3.5 Assessment of immunocomplexing between A R and P-catenin, GSK3, and APC in LNCaP cells treated with (+) and without (-) R1881 for 24 hours 110 Figure 3.6 Specific retention of proteins to androgen-regulated promoter regions using the acrydite capture of D N A complexes (ACDC) method 112 Figure 3.7 Assessment of p-catenin localization in mouse prostate after 7 days in animals 114 Figure 3.8 Proposed model showing that A R may facilitate translocation of cytoplasmic P-catenin to the nucleus and, thereby, increase A R transcription and possibly regulate Wnt 116 xi Chapter Four: The Androgen Receptor can Repress the P-catenin/Tcf Signaling Figure 4.1 A R represses P-catenin/function both dose and ligand dependency PC3 cells transiently transfected with A R and treated with 10 n M DHT for 24 hours 136 Figure 4.2 A R and androgens provide potent negative regulation in cells with hyperactive P-catenin/Tcf signaling 138 Figure 4.3 Distribution of HcRed-Tcf and AR-EGFP foci in transfected LNCaP cells treated with 10 n M DHT viewed in a single xy plane without and with deconvolution (DECON) 140 Figure 4.4 Colocalization of Tcf-HcRed with phosphorylated (Ser 2) R N A Pol II in the presence of DHT (SW480 cells) 142 Figure 4.5 Confirmation of expression and transcriptional activity of AR-EGFP and HcRed-Tcf fusion expression vectors in LNCaP cells and SW480 cells 144 Figure 4.6 Evidence from binding assays that A R and Tcf compete for P-catenin 146 Figure 4.7 Effect of AR-mediated repression of Tcf signaling on cell cycle and viability 148 Figure 4.8 Soft agar colony forming assay showing the effects of AR, androgen and A R antagonist treatment on the number of SW480 colonies formed 150 Figure 4.9 Mechanisms by which androgens may modulate cellular localization and expression of E-cadherin/P-catenin and Tcf activation 152 Chapter 5. PTEN is a negative regulator of P-catenin/Tcf and AR Signaling Figure 5.1 Levels of nuclear P-catenin are reduced by expression of exogenous PTEN in PTEN null PrCa cells 183 Figure 5.2 Increasing amounts of transfected PTEN-GFP in PC3 cells promotes decreased nuclear expression of P-catenin 185 Figure 5.3 PTEN enhances p-catenin degradation 187 Figure 5.4 PTEN inhibits D N A binding activities of P-catenin/Tcf complexes 189 Figure 5.5 Generation of stable "knock-in" PTEN cells by two different methods 191 xii Figure 5.6 VP 16 fusion and lentiviral tet-repressors maintain minimal non-induced ("leaky") expression of the PTEN transgene 194 Figure 5.7 Validation of PTEN expression and distribution in Dox induced stable LNCaP-PTEN/C124S clones , 196 Figure 5.8 PTEN inhibits A R and P-catenin/Tcf function in a phosphatase dependent manner in androgen dependent (LNCaP) PrCa cells 198 Figure 5.9 PC3-WT PTEN cells treated with 2 ng/ml Dox show reduced Topflash activity in the presence of endogenous and exogenous WT P-catenin 200 Figure 5.10 Induced PTEN expression in LNCaP cells (14M clone) reduces A R transactivation (ARR3-luc) in cells transfected with either WT P-catenin or 203 activated P-catenin Figure 5.11 Androgens oppose apoptotic effects conferred by WT PTEN expression in LNCaP PrCa cells 205 Figure 5.12 PTEN expression is correlated with decreased expression of AR-EGFP in LNCaP-PTEN cells (14M clone) 208 Figure 5.13 WT PTEN expression reduces nuclear A R and p-catenin, ligand-sensitive AR/P-catenin complexing but increases total A R expression 212 Figure 5.14 WT PTEN and androgens: opposing regulators of cell cycle and cell viability 214 Figure 5.15 PC3-PTEN cells are more resistant (than LNCaPs) to the anti-proliferative effects conferred by PTEN 216 Figure 5.16 Androgens and serum protects against PTEN inhibition of LNCaP-PTEN and PC3-PTEN soft agar colony formation 218 Chapter 6. In vivo analysis of P-catenin localization and function Figure 6.1 p-catenin transcript levels increase during acquisition of AI in the LNCaP xenograft series 243 Figure 6.2 Expression profiling of P-catenin in a LNCaP xenograft tissue microarray 245 Figure 6.3 Development of an inducible LNCaP-PTEN xenograft 247 Figure 6.4 Co-expression of PTEN and A R in LNCaP-PTEN cells 249 Figure 6.5 Co-expression of PTEN and P-catenin in LNCaP-PTEN cells 251 Xlll Figure 6.6 Prostate specific PTEN knock-out mouse model 254 Figure 6.7 A R and P-catenin immunoreactivity in the dorsal lateral lobe from 20 week old PTEN fl/fl Cre recombinase negative and positive mice 256 Figure 6.8 Expression of the proliferative antigen, Ki-67, and P-catenin increase with increasing Gleason grade 259 Figure 6.9 Ki-67 and P-catenin immunoreactivity in representative tissue cores that are Gleason 4 and 5 in grade 261 Figure 6.10 P-catenin cellular distribution and staining intensity in a Gleason grade tissue array that includes histology of benign hyperplasia (BPH) 264 Figure 6.11 P-catenin distribution and staining intensity in a human Neoadjuvant Hormone Treatment (NHT) tissue microarray 267 Figure 6.12 High immunoreactivity of P-catenin in PrCa metastatic lesions viewed at low, medium and high magnification 269 Chapter 7: Summary & Future Directions Figure 7.1 Alternative expression of PTEN and GSK3P in AI PrCa facilitates cross-regulation of phosphatidylinositol-3-kinase/Akt and Wnt/P-catenin signaling xiv List of Tables Table I. Summary of Related Works to this Thesis 48 Table II. PCR Primers 306 Table III. Expression Plasmids donated to or purchased by The Prostate Centre 306 Table IV. Plasmids cloned at The Prostate Centre 307 Table V. List of Antibodies Used 308 Table VI. List of Secondary Detection Reagents 309 Table VII. List of Ligands 309 Table VIII. List of Antibiotics 310 Table IX. List of Bacteria 310 Table X. Cell Line Origins and Attributes 311 List of Abbreviations and Symbols P-catenin/Tcf4 P-catenin/T-Cell Factor p-catenin ' Beta-catenin A C T H Androcorticotrophic hormone A D Androgen dependence/dependent ADF2 Adipophilin AF-2 activation function-2 AI Androgen independence/independent APC Adenomatous polyposis coli A R Androgen receptor AR-CTD Androgen receptor carboxy terminal domain A R - L B D Androgen receptor ligand binding domain AR-NTD or AR-Nt Androgen receptor amino terminal domain A R E Androgen response element APC Adenomatous polyposis coli B P H Benign prostatic hyperplasia C/EBP CCAAT/enhancer-binding protein Casodex Bicalutamide CBP Crebs binding protein CREB cAMP-response element binding protein CTD C-terminal Domain DBD D N A binding domain D E X Dexamethasone DHT Di h ydrotestosterone D R E Digital rectal exam E 2 Estradiol EGF-R Epidermal growth factor receptor ER Estrogen receptor FABP2 Fatty acid binding protein2 FRAT Frequently rearranged in advanced T-cell lymph GBP GSK3 binding protein GCs Glucocorticoids GnRH Gonadotrophin releasing hormone GR Glucocorticoid receptor; GRE Glucocorticoid response element GSK3(3 Glycogen synthase kinase 3 Beta Heat shock proteins Hsps H M G High mobility group IGF-1R Insulin-like growth factor-1 receptor LP Immunoprecipitation IL Interleukin ILK Integrin linked kinase J A K Janus kinus K G F Keratinocyte growth factor LAPC-4 Los Angeles prostate cancer line LNCaP Human adenocarcinoma prostate cancer cell line L B D ligand binding domain Lef Lymphoid enhancer factor L H Leutenizing hormone L H R H Leutenizing hormone release hormone M A P K Mitogen activating protein kinase M R Mineral corticoid receptor NR Nuclear receptor NTD, Nt Amino terminal domain PARP Poly(ADP-ribose)polymerase PC3 Prostate Cancer line derived from bone metastasis PI3K Phosphatidyl inositol 3-kinase PIN Prostatic intraepithelial neoplasia PDK1 PI3K-dependent kinase 1 PDK2 PI3K-dependent kinase 2 PPAR Peroxisome proliferator activated receptor PR Progesterone receptor PrCa Prostate cancer xvii PSA Prostate specific antigen PTEN Phosphatase and tensin homologue deleted on chromosome ten R A Retinoic acid Rb Retinoblastoma protein R A R Retinoic acid receptor R A R E Retinoic acid receptor response element Rb Retinoblastoma protein SHBG Sex hormone binding globulin SRE Steroid response element STAT3 Signal transducer activation transcription factor T Testosterone T 3 Triidothyronine Tcf T-cell factor TGF Transforming growth factor T N M Tumour, lymph node, metastasis TR Thyroid receptor TrCP Transducing receptor containing protein T R U G Transrectal ultrasound-guided biopsy TZD Thiazolidinediones VDR Vitamin D receptor V D R E Vitamin D receptor response element WLF1 Wnt inhibitory factor a Alpha u Micro 5 Delta A Delta P Beta xvm Acknowledgements I acknowledge my supervisor, Dr. Colleen Nelson, for putting forth endless hours of effort to elevate The Prostate Centre at Vancouver General Hospital to a recognized centre of excellence for cancer research. I also thank Colleen for allowing me freedom to develop and explore my many good (and not so good) ideas in the laboratory. I have found this mode of mentorship to have helped me develop into a highly independent and motivated investigator. I pay tribute to those who I have had many conversations regarding my research including Dr. Paul Rennie, Dr. Martin Gleave, Dr. Chris Ong, Dr. David Huntsman, Dr. Shoukat Dedhar, and Dr. Bi l l Salh. In particular, I'd like to thank Dr. Michael Cox for being a great mentor and collaborator. I thank other researchers with whom I have theorized with, discussed with, complained with and collaborated with at The Prostate Centre. They have included Dr. Rich Sobel, Helen Cheng, Dr. Jason Read, Latif Wafa, Dr. Dieter Fink, Bob Shukin, and Steven Hendy. I thank members of The Prostate Centre who have provided indispensable services including Terry and Nathan for washing my dishes for 5 years, Mary Bowden for maintaining a well kept tissue culture facility, Ladan Fazli for assistance in pathology, Bob Shukin for assistance in cloning and D N A sequencing, Steve Hendy and Helen Cheng for maintaining radiation safety and Brenda Prieur for keeping the administration in order. I thank Nadine Thompson and Robert Bell for their excellent assistance in bioinformatics. In addition, I also would like to thank Carol Fredrikson, who has happily supported me throughout my Ph.D and continues to always be by my side and help me when needed. Lastly, I dedicate this completed thesis to my bench mate, Wesley Sydor, whose companionship we recently lost. Chapter One: In t roduc t i on 2 1.1 Prostate cancer Prostate cancer (PrCa) is the most commonly diagnosed neoplasm in the western world and a leading cause of cancer related death in males (Kamo and Sobue 2004). PrCa is considered to be a slow progressing pathology that usually has clinical implications for men over the age of 60 (Greenlee, Murray et al. 2000). Patients who are diagnosed with PrCa generally have clinical symptoms that include urinary difficulties, dysuria, hematuria and elevated prostate specific antigen (PSA) levels (Yount, Cella et al. 2003). Those patients who have serum PSA levels greater than 10 ng/ml often undergo histological examination of biopsy tissue and transrectal ultrasound-guided biopsy (TRUG). The combined use of clinical and histological assessment is necessary to determine the progressiveness of a patient's disease and, therefore, the best course of treatment. Means for treating PrCa rely on the well described androgen dependency of the prostate gland, as originally described by Huggins and Hodges, who showed that removal of androgens causes regression of the prostate (Huggins C 1941). Therapies have changed little since these original observations, with current modalities of androgen ablation occurring by orchidectomy or administering of gonadotrophin releasing hormone analogues. Androgen receptor (AR) antagonists both bind and reduce the transcriptional activity of A R and, therefore, are thought to inhibit A R target genes that may be associated with androgen driven cell proliferation (Singh, Gauthier et al. 2000). As a result, chemotherapy is usually only employed if hormone blockade therapy has failed and is generally reserved for patients who are late stage and hormone refractory (Gulley and Dahut 2004). Furthermore, the use of androgen blockade therapy remains controversial and, in some instances, may provoke cell proliferation during androgen independence (AI) (Patterson, Balducci et al. 2002). Regardless, PrCa is a heterogeneous pathology requiring combination therapy and, in some cases, patient specific treatments are necessary for optimal clinical outcome (Moul and Chodak 2004). Despite advances in oncology, including clinical trials in hormone blockade therapy, chemotherapy and novel pharmacological inhibitors, development of treatments for AI PrCa have lagged in comparison to most other cancers. This is in part due to the lack of biologically relevant PrCa cells lines and animal models. It is also due to the heterogeneous nature of PrCa specimens and the associated difficulties in genetic analysis. Thus, treatments for progressive PrCa have changed little over the decades. 3 An important change that has occurred is related to early detection of PrCa. With the advent of PSA screening and increased advocation for digital rectal examinations (DREs), the rate of detection of PrCa has increased tremendously and, as a result, PrCa related mortalities have decreased. Nevertheless, approximately 25% of diagnosed men will die this year largely because early detection failed and the resulting cancer has evolved to a hormone refractory state (McDavid, Lee et al. 2004). Most patients who have late stage PrCa will die of metastatic lesions with little hope for a cure. In attempts to reduce deaths associated with advanced disease, intense investigation is focused on the mechanisms that allow PrCa cells to proliferate despite the removal of androgens. It also suggests that, A R likely serves an important functional role in progression from androgen dependent (AD) to independent PrCa. For example, the inability of hormone blockade therapy to succeed during androgen independence (AI) suggests that A R function is being altered and or bypassed. Thus, co-factor or alternative prosurvival signaling pathways that may alter the response of A R in low androgen environments is of particular interest. 1.2 The prostatic epithelium The prostatic epithelium consists of 3 major cell types including secretory luminal cells, basal cells and neuroendocrine cells, each which can be identified by immunological markers. For example, secretory luminal cells stain positive for cytokeratin 8, 18 and CD57; basal cells are positive for cytokeratin 5, 14 and CD44 (Weckermann, Muller et al. 2001) while neuroendocrine cells are identified by specific neuronal markers including CgA and NSE (Hvamstad, Jordal et al. 2003). Histologically, PrCa often appears fo have a mixture of benign glandular structures and neoplastic foci of varying severity. Thus, PrCa is considered a heterogeneous and multifocal disease in which neoplastic transformation leads to a progressive loss of histological differentiation (Helpap, Oehler et al. 1990). Prostatic intraepithelial neoplasm (PIN) lesions are thought to be the main precursors of invasive carcinoma and can be recognized by formation of tufting, micropapillary and cribiform histological features. PIN may precede carcinoma for 10 or more years and, thus, is thought as a continuum between low and high grade forms of PrCa. Similar to invasive carcinoma, PIN focal sites also display loss of basal cells and reduction of 4 differentiation markers such as E-cadherin (Montironi, Bartels et al. 1995; Battels, Montironi et al. 1998), though the basement membrane generally remains intact (Bostwick, Leske et al. 1995). Clinically, important differences can be made between PLN and non-malignant pathologies of the prostate including benign prostatic hyperplasia (BPH). For example, while carcinoma is found in the peripheral zone of the prostate, B P H is found in the transitional zone. Further, most PrCas (70%) occur in the peripheral zone of the prostate and almost all (95%) are glandular in origin (adenocarcinomas) (Helpap, Oehler et al. 1990). The heterogeneous nature of PrCa has lead to the use of the Gleason scoring system as an indicator of histological differentiation (Figure 1.1). Prostatic histopathology grading is based on a grading system of 1-5 (from most to least differentiated) based on the combination of the dominant and secondary histological features. Scored out of ten, Gleason grading is a useful predictor of pathological and clinical outcome (Lattouf and Saad 2002). Cancers that are well differentiated have low Gleason grades (2-4), with moderately well differentiated cancers of grade 5 or 6 and poorly differentiated cancers of grades 8-10. Low grade tumours are, in general, very slow to progress while high grade tumours are aggressive and more likely to be involved in metastasis (Calvete, Srougi et al. 2003). Scores of <4 have a 25% chance of progression, scores 5-7 have a 50% chance of progression while those >7 have a 75% chance of progressing extracapsularly (Calvete, Srougi et al. 2003). Patients with disease that extends from the prostate and Gleason scores >7 have a significantly worse clinical outcome and, thus, may be subjected to immediate radical prostatectomy, surgical or. medical castration and radiation therapy (Khoddami, Shariat et al. 2004). While the heterogeneity of prostatic lesions is the basis for Gleason scoring, the multifocality of PrCa has also created a challenge for researchers in identifying sufficient homogenous tissue that can be used for genetic analysis. This is mainly a result of the multifocal and multiclonal nature of PrCa tissue and has, in part, lead to a relatively poor understanding of genes involved in the PrCa progression as compared to other cancers. Grading also occurs during clinical examination. The T N M scaling system is intended to evaluate tumour size (T), lymph node involvement (N) and the presence of metastasis (M) (Eichelberger, 2004) (Figure 1.1). The T N M scaling is, thus, a good indicator of prognosis in patients who have progressive PrCa. The prostatic capsule acts as a barrier to metastatic invasion of local tissues. Thus, PrCa initially spreads within the gland but subsequently invades the lymphatics, bladder and seminal vesicles once the capsule is breached. Lymph node and 5 vascular invasion frequently results in secondary disease whereby cancerous cells metastasize to the pelvis and lumbar vertebrae. Patients who have a score indicative of extracapsular invasion have a significantly poorer clinical prognosis (Yamashita, Noguchi et al. 1999) while those with disease confined to the prostate generally have a more favourable outcome and are treated with radiation, radical prostatectomy or by watchful waiting (McLeod 1999). 1.3 Genetic changes during progressive PrCa Genetic instability serves an important role in allowing tumourigenic cells to escape programmed cell death. For example, during PrCa, loss of heterozygosity has been shown to allow for potential disruption of normal tumour suppressor functions in prostatic epithelium (Isaacs and Kainu 2001). More specifically, PrCa progression is characterized by loss of several signature genes including Nkx3.1 (8p21), PTEN (lOq), Rb (13q) and p53 (17pl3) (Figure 1.2). Loss of expression of such tumour suppressors is correlated with progression of normal epithelium to PIN, invasive carcinoma and metastasis (MacGrogan and Bookstein 1997). In fact, recent mouse model evidence suggests that PTEN loss may promote cell cycle senescence, which then must be followed by loss of p53 in order for invasion and potentially metastasis to occur (Kim, Kim et al. 2005). While the frequency of loss of critical cell cycle regulators is not readily agreed upon, prosurvival pathways that are regulated by these suppressors have been intensively studied with respect to promoting the acquisition of an AI phenotype. 1.4 Androgens, the hormone axis and treatment of PrCa Androgens function to maintain a balance of mitotic index and differentiation of the prostatic epithelium (Kang, Tsai et al. 2003). As such, androgen production is vital in maintaining the phenotype and function of the prostatic epithelium. Maintenance of androgen regulation is achieved by the hypothalamus-pituitary axis, whereby gonadotrophin releasing hormone (GnRH) is secreted by the hypothalamus to stimulate L H (lutinizing hormone) release from the pituitary gland. L H acts on the Leydig cells of the testes to induce androgen production. Therefore, the GnRH-LH hormone axis is critical in mediating androgen blockade therapy and provides an effective negative feedback loop for inhibition of gonadal androgen production (Labrie 2004) (Figure 1.4). In males, the majority (95%) of testosterone (T) is 6 produced by the testes with a minority of androgens produced by the adrenal glands. Sex Hormone Binding Globulin (SHBG) occupies the majority of circulating T and facilitates its targeting to cells of the prostate gland (Kaaks, 2003) (Figure 1.3) where either by diffusion or active import by a putative SHBG receptor, T is imported into prostate epithelium (Kaaks, 2003). The enzyme, 5 (alpha) a-reductase converts T into a more potent form of androgen, 5a-dihydrotestosterone (DHT) (Hellwinkel, Muller et al. 2000). Subsequently, DHT acts upon cytosolic A R to induce both transcriptional and non-transcriptional A R alterations. DHT that does not occupy A R is rapidly converted to 3-a-17-P-androstenediol (Nakhla and Rosner 1996). DHT binds A R to induce dissociation from heat shock proteins (Hsps) (Haendler, Schuttke et al. 2001), a change in structural conformation, dimerization and translocation to the nucleus. Subsequent to translocation, A R dimerizes and binds to hormone response elements, by way of a D N A binding domain, to activate A R responsive genes (Kokontis and Liao 1999) (Shaffer, Jivan et al. 2004). Consequently, 5a-reductase has been targeted for therapeutic inhibition with the goal of reducing available DHT for activation of A R (Andriole, Humphrey et al. 2004; Reddy 2004). Treatment for PrCa most often entails total hormone blockade, either by surgical or chemical hormone ablation therapy. The rationale of ablating androgens is to induce apoptotic regression of the prostate (Griffiths, Morton et al. 1997; Miyamoto, Yeh et al. 1998; Schneider 2003). This can be achieved by taking advantage of the feedback loop between the testes and the hypothalamus (Figure 1.4). While, administering of leutenizing hormone releasing hormone (LHRH) agonists (e.g. leuprolide, goserelin acetate or buserelin acetate) causes an initial secretion of gonadotropin and testosterone, within 2-3 weeks a reduction in serum testosterone levels takes place (Fretin 1990) (Figure 1.4). Though the mechanism is poorly understood, patients can be treated using L H R H agonists in conjunction with glucocorticoids in order to create a negative feedback inhibition of adrenocorticotrophic hormone (ACTH) secretion from the pituitary gland. Blockade can also occur at the level of A R using anti-androgens which are effective in competing with androgens for binding to A R (Berrevoets, Umar et al. 2002). Non-steroidal anti-androgens (e.g. flutamide, bicalutamide) bind to the A R without other endocrine effects. Differently, are steroidal anti-androgens (e.g. cyproterone acetate) which both bind AR, glucocorticoid receptor and progesterone receptor to inhibit production of T by interfering with the hypothalamus-pituitary axis (Berrevoets, Umar et al. 2002). Direct inhibition of androgen 7 production may be accomplished using ketoconazole (an anti-fungal agent) and aminoglutethimide, both which act directly on the Leydig cells (Bok and Small 1999). Specifically, Ketoconazole inhibits cytochrome P450, a required step for production in steroid synthesis, while aminoglutethimide inhibits 20-22-desmolase, thus, blocking conversion of cholesterol to pregenolone, a precurosor of T (Bonneterre 1986). More recently, evaluated preventative therapies, including finasteride and epristeride, have included the use of 5a-reductase inhibitors and have undergone large scale clinical testing. Such inhibitors targeting both type 1 and type 2 reductases have been shown to be effective in delaying BPH, though the effectiveness against prostate carcinogenesis is conflicting (Debruyne, Barkin et al. 2004). Of intense evaluation is the use of taxanes, including paclitaxel and taxotere, which are highly active anticancer drugs currently being investigated for effectiveness against human PrCa. Taxanes are effective in inducing apoptosis by promoting mitochondrial release of cytochrome c (Beedassy and Cardi 1999) and when combined with microtubule inhibitors have proven to be an effective anti-cancer combination (Pellegrini and Budman 2005). 1.5 The AR: structure and function Nuclear receptors (NRs) are a class of transcription factors that are regulated by small lipophilic ligands including steroids, thyroid hormone, retinoids (vitamin A metabolites) and vitamin D 3 . At least 100 nuclear receptors have been identified that can be divided into two general subfamilies. Type 1 receptors include: AR, estrogen receptor (ER), mineral corticoid receptor (Ohta, Elnemr et al.) and progesterone receptor (PR) while type 2 receptors include: the thyroid receptor (TR), Vitamin D receptor (VDR), retinoic acid receptor (RAR), retinoid x receptor (RXR) and peroxisome proliferator activated receptor (PPAR) (McKenna, Lanz et al. 1999; McKenna and O'Malley 2001; McKenna and O'Malley 2002). NRs mediate their effects not only by ligand binding and gene activation but also by way of phosphorylation dependent interactions with signaling transduction pathways including M A P K , PI3K and Wnt/p-catenin/Tcf (Rochette-Egly 2003). Importantly, novel ligands for NRs are continuously being developed and include numerous synthetic steroid agonists and antagonists, ligands that alter fatty acid synthesis, analogues of Vitamin A + D, ligands with anti-diabetic qualities, as well as ligands with anti-cancer attributes (Grun and Blumberg 2003) 8 (Yoshihara, Nguyen et al. 2003). Therapeutic intervention is perhaps best exemplified by retinoic acid which functions in a chemotherapeutic manner (Okuno, Kojima et al. 2004). Thus, NRs and their cognate ligands serve as potent regulators of development, cell differentiation and normal physiology with important implications for pathologies such as cancer (Weatherman, Fletterick et al. 1999). Given the fundamental role that NRs serve in maintaining normal cellular milieu it is not surprising that receptors and their cognate ligands can functionally interact with other key signaling systems. The A R is a transcription factor that binds D N A response elements in a ligand dependent manner. Upon binding, A R dissociates from cytosolic heat shock proteins (Hsps), changes conformation and translocates to the nucleus to activate target genes. Various structural features, that are both common to all nuclear receptors and those that are unique to AR, allow for a dynamic cellular response to hormone activation (Figure 1.5). The human A R gene has 9 exons and codes for a protein approximately 919 amino acids in size. A R consists of an amino terminal domain (NTD), D N A binding domain (DBD), hinge region and a ligand binding domain (LBD) (Culig, Klocker et al. 2002). The AR-NTD is the least conserved region and contains the AF1 region (aa 141-338) capable of ligand independent transcriptional activation. Several polymorphisms also exist in the NTD including a polyglutamine tract ranging from 17 to 29 repeats in addition to a polyglycine tract (Khorasanizadeh and Rastinejad 2001). This is validated and exemplified by the fact that GR and dexamethasone can mediate PSA expression in LNCaP PrCa cells but cannot mediate their cell proliferation (Cleutjens, Steketee et al. 1997). Distal to A R - D B D is the hinge region (separating the D B D and LBD) which codes a nuclear localization signal, thus allowing ligand dependent nuclear translocation of A R (Culig, Klocker et al. 2002). The A R - L B D is relatively conserved between NRs and facilitates differential response to ligands with either agonist or antagonist qualities. The transcriptional activity of A R involves interactions of the NTD and CTD interactions in the presence of agonist activation but not in the presence of most antagonists, including bicalutatmide (Culig, Klocker et al. 2002). 1.6 Evidence that AR is functionally important in PrCa progression Use of steroidal and non-steroidal antagonists are effective in early management of PrCa but are not effective in advanced forms (Richie, 1999). These well-established observations 9 strongly suggest that A R function is altered either directly or indirectly. Perhaps most importantly, despite total androgen blockade, the expression of A R and PSA continues during metastatic cancer and hormone resistant PrCa, indicating that other events can activate AR signaling. While expression of A R is static in normal prostate and BPH, various changes have been reported to occur in AR expression. These have included A R gene amplification, messenger ribose nucleic acid (mRNA) overexpression and alteration of A R protein expression and function (Suzuki, Ueda et al. 2003). For example, LNCaP sublines, such as LNCaP-abl, generated by continual growth in androgen depleted media show increased levels of A R message and protein (Culig, Hoffmann et al. 1999). CWR22 xenografts show increased expression of A R during AI growth (Tan, Sharief et al. 1997). Thus, during late stage PrCa, loss of sensitivity of A R to androgens has generally lead to four main ways by which PrCa could escape the need for androgens and apoptosis: 1) selection for oncogenic clones, 2) alteration of A R activity, 3) adaptation to low androgen environments and 4) alternative modes of signal transduction (Chodak 1998) (Walczak and Carducci 2002). Transcriptional co-regulators are thought to facilitate A R targeting to basal machinery and function, likely by competitive binding, to either activate or repress A R function (Shang, Myers et al. 2002). Co-regulator alteration of A R function usually occurs by way of histone acetylase and deacetylase activity. A fundamental hypothesis of co-regulator research is that activator overexpression during PrCa AI may allow A R function to escape the need for androgens (Rahman, Miyamoto et al. 2004). SRC1 is one of the first identified co-factors to augment A R activity both in the presence and absence of androgens by way of altered histone acetyl transferase function (Gregory, He et al. 2001). CBP/p300 is frequently overexpressed in PrCa and, in addition, to being a potent activator of A R transcription (Fronsdal, Engedal et al. 1998). The first AR specific co-regulator to be described was ARA70, enhancing A R target genes both in the presence of androgens and possibly non-androgen ligands including estradiol (Yeh and Chang 1996) (Miyamoto, Yeh et al. 1998). While select co regulators are known to be specific for interactions with AR, there are others which are highly promiscuous in their capacity to interact with nuclear receptors. Among these are members of the canonical Wnt signaling cascade (see Appendix Review: Interaction of the Wnt/P-catenin/Tcf Signaling Axis with Nuclear Receptors: Wnt you like to know?). 10 1.7 Pro-survival signaling pathways and progression of PrCa The continual expression of PSA and increasing tumour load during chronic androgen blockade suggests that alternative cell signaling mechanisms may facilitate AI activation of AR. Growth factors have been intensely studied for their abilities to activate receptor tyrosine kinases and resulting intracellular signaling cascades. Factors including IGF-1, EGF, K G F , IL6, and LL4 have all been shown to activate AR and instigate growth of PrCa cell lines by activation of major prosurvival signaling pathways including PI3K, M A P K and the JAK/STAT3 signaling cascades (Culig, Hobisch et al. 1994). Growth factors have been well documented to alter key signal transduction pathways, in PrCa systems. For example, Her-2/neu, an EGF receptor tyrosine kinase, shows gene amplification and overexpression at the protein level in patients with benign hyperplasia and PrCa. Over expression of Her2/neu can also lead to promotion of tumour growth and PSA production in LAPC-4 PrCa xenografts (Craft, Shostak et al. 1999), events likely mediated by M A P K signaling (Franco, Onishi et al. 2003). Reports have also implicated GTPase signaling pathways, as acting through R T K activation of Ras and recruitment of Raf kinase (Weber and Gioeli 2004). These events allow for phosphorylation of M A P K / E R K kinases (MEKs) and their downstream counterparts, ERK1 and ERK2 (Bell, Myers et al. 2003). Using forskolin to increase intracellular cAMP, the second messenger P K A has been implicated in ligand independent activation of the amino terminus (NTD) of A R (Sadar 1999). Curiously, LL-6, a pleiotropic cytokine, has been shown to have both stimulatory and inhibitory effects on PrCa cell lines (Chung, Y u et al. 1999). However, IL-6 can activate the AR-NTD through parallel signaling pathways consisting of STAT3 and M A P K (Yamamoto, Sato et al. 2003; Yang, Wang et al. 2003). Specifically, IL-6 binds IL-6Ra to phosphorylate STAT3 thereby promoting its dimerization, nuclear translocation and binding to AR. LL-6R occupation may also recruit Ras to increase M A P K phosphorylation and activation of AR-NTD (Michalaki, Syrigos et al. 2004). Collectively, these observations indicate that growth factor activation of receptor tyrosine kinases may lead to parallel initiation of proliferative pathways. Dissecting which pathway(s), confer signaling mechanisms sufficient to promote AI cell proliferation is an area of intense 11 investigation. However, of the major survival pathways including Ras, M A P K , JAK/STAT3 and PI3K/Akt, most in vitro and in vivo PrCa systems are predominately PI3K/Akt dependent. That is, while pharmacological abrogation of the former pathways can be tolerated (without an obvious apoptotic response), chemical inhibition of the IGF1-R/PI3K/Akt axis leads to cell arrest and a potent apoptotic response (Lin, Adam et al. 1999). 1.8 Wnt signaling: activation and inactivation The canonical Wnt pathway is highly regulated in maintenance of cell-cell adhesion and cell cycle control. Perturbations in Wnt regulators, either by mutation or by cross-regulation with other pathways, can allow for altered signaling with extra cellular matrix components and downstream targets potentially leading to epithelial pathogenesis (Kikuchi 2003). The Wnt signaling pathway consists of secreted ligands, a membrane receptor (Frizzled), the cytoplasmic receptor Dishevelled (Dsh), P-catenin, cytoplasmic regulators and nuclear transcriptional complexes (Polakis 2000) (Greaves 2003). Wnts are glycoprotein growth factors that function as ligands for the transmembrane, cysteine-rich family of Frizzled receptors regulating signaling both in an extracellular and intracellular manner (Bhanot, Brink et al. 1996) (Yang-Snyder, Miller et al. 1996). Although not all Wnt ligands have not been conclusively associated with the development of cancer, reports indicate that Wnt 1 and Wnt3a are tumourigenic (Lejeune, Huguet et al. 1995; Dale, Weber-Hall et al. 1996). Binding of secreted Wnt ligands to Frizzled results in receptor kinase action and phosphorylation, by poorly understood mechanisms, of the cytoplasmic mediator, Dsh. Extracellular modulation of Wnt signaling may occur through competitive binding of Wnt ligands, including Secreted Frizzled Related Protein (sFRP) (Lin, Xia et al. 2000) (Uren, Reichsman et al. 2000) or Wnt inhibitory factor (WLF) (Hsieh, Kodjabachian et al. 1999). Wnt binding and activation inhibits glycogen synthase kinase (GSK3{3), a serine/theronine kinase (Davies, Jiang et al. 2001) (Hagen, Di Daniel et al. 2002; Hagen and Vidal-Puig 2002), ultimately allowing for accumulation of (3-catenin (Lee, Shvartsman et al. 2003) (Semba, Kusumi et al. 2000) (Lee, 2003; Semba, 2000) (Figure 1.6). 12 The ability of P-catenin to locate numerous cellular compartments is testament to its role as a promiscuous effector molecule of the Wnt signaling pathway. P-catenin is also found in multiple cellular pools including (i) the adherens junctions in association with the transmembrane receptor, E-cadherin; (ii) the cytoplasm and (iii) in the nucleus (Blaschuk and Rowlands 2002). Binding of P-catenin to partners, including E-cadherin and Tcf4, occurs mostly through the 12 Armadillo repeats, each of which contain 3 helices (Pai, Kirkpatrick et al. 1996). Soluble and non-soluble forms of p-catenin are highly regulated by a proteosome/ubiquitination system consisting of glycogen synthase kinase (GSK3P) (Hagen, Di Daniel et al. 2002; Hagen and Vidal-Puig 2002), axin (Kikuchi 2003) and the tumour suppressor, adenomatous polyposis coli (APC) (Rosin-Arbesfeld, Cliffe et al. 2003). Upon increased cellular levels (Figure 1.6a) and nuclear accumulation (Rosin-Arbesfeld, Cliffe et al. 2003), P-catenin binds the amino terminus of Tcf altering its conformation and promoting interactions in the targeting of D N A sequences (A/T A/T C A A A G ) (Oosterwegel, van de Wetering et al. 1991) (van de Wetering, Oosterwegel et al. 1991). P-catenin binding to Tcf also causes displacement of the repressor, Groucho (Roose, Huls et al. 1999) (Poy, Lepourcelet et al. 2001) and a concomitant recruitment of co-activators such as CBP/p300, Brgl and CARM1 (Koh, L i et al. 2002) resulting in de-repression of Tcf transcription (Cavallo, Cox et al. 1998; Roose, Molenaar et al. 1998). Activation of P-catenin/Tcf signaling pathway, either by disengagement of the APC/GSK3 components or by Wnt, results in increased expression of downstream target genes including cyclin DI (Shtutman, Zhurinsky et al. 1999; Tetsu and McCormick 1999), c-myc (He, Sparks et al. 1998; Brabletz, Jung et al. 2002), PPAR8 (He, Chan et al. 1999), Tcf-1 (Oosterwegel, van de Wetering et al. 1991; van de Wetering, Oosterwegel et al. 1991), matrilysin (Crawford, Fingleton et al. 1999) and CD44 (Mikami, Saegusa et al. 2001). Physiologically, induction of these genes results in dramatic effects on cell and tissue development and oncogenesis (Peifer and Polakis 2000) (Behrens, von Kries et al. 1996; Barker, Morin et al. 2000; van Es, Barker et al. 2003). Negative Wnt signaling (Figure 1.6b) occurs by GSK3 phosphorylation of APC allowing binding and phosphorylation of free p-catenin at Ser 33, 37, 45 and Thr 41 (Hagen, Di Daniel et al. 2002; Hagen and Vidal-Puig 2002). This change in phosphorylation targets P-catenin for recognition by the F box protein, pTrCP, and degradation by the ubiquitin/proteosome pathway 13 (Hart, Concordet et al. 1999). P-catenin levels are also regulated by Siah, a p53 inducible pathway. p53 regulates Siah, which interacts with the NTD region of APC resulting in recruitment of a ubiquitination complex to the P-catenin NTD and targeting it for degradation (Liu, Stevens et al. 2001; Matsuzawa and Reed 2001). Phosphorylation of the P-catenin NTD is required both for GSK3P and Siah means of degradation (Orford, Crockett et al. 1997; Hagen, Di Daniel et al. 2002; Hagen and Vidal-Puig 2002). These observations indicate that Wnt signaling maintains strict control over cellular levels of P-catenin. However, it is also equally understood that pathologies harbouring forms of P-catenin with mutation within the p-catenin NTD may result in cytosolic accumulation of p-catenin and, therefore, its nuclear accumulation and activation of target genes. 1.9 P-catenin: structure & function P-catenin, an adhesion molecule with oncogenic potential, consists of three main domains including the NTD, Armadillo repeats and transactivating, CTD (Figure 1.7). The amino domain contains regions of GSK3P phosphorylation and degradation at serine and threonine sites. The Armadillo domain consists of 12 imperfect repeats each of 42 amino acids. Each repeat has 3 a helices termed HI , H2 and H3, except repeat 7 which lacks an HI helix. The helices of adjacent repeats form an elongated superhelix with helices at repeats 9 + 1 0 bending by 50° producing a large positively charged groove (Huber, Nelson et al. 1997). The CTD of P-catenin is highly acidic and is essential for transcriptional co-activation with Tcf/Lef. Basal transcriptional machinery also bind to this region (Vleminckx, Kemler et al. 1999). 1.10 Nuclear import of P-catenin As the oncogenic function of P-catenin depends on its nuclear accumulation, its controlled entry into the nucleus is critical for its function. P-catenin lacks a classical nuclear localization signal (NLS), thus other modes for its entry must exist in importin independent 14 mechanisms, possibly involving binding to the nuclear pore components (Fagotto, Gluck et al. 1998). Translocating pools of P-catenin appear to compete with cytosolic importins for passage through nuclear pore complexes (Fagotto, Gluck et al. 1998) and also appear to be independent of Ran mediated import (Yokoya, Imamoto et al. 1999). Alternatively, as over expression of Tcf/Lef promotes nuclear accumulation of P-catenin, Lef could form a cytosolic complex with P-catenin that then translocates (via a Lef NLS) to the nucleus. However, as high levels of nuclear P-catenin can be observed in cells with low levels of Tcf/Lef, it is likely that a host of other mechanisms regulate P-catenin trafficking into the nucleus. P-catenin also binds with greater strength to Lef-1 than does its homologue, plakoglobin, providing further uniqueness to the means by which P-catenin locates to the nucleus (Huber, Korn et al. 1996). 1.11 Transcriptional regulation by P-catenin The main nuclear binding partners of P-catenin are the transcription factors Tcf/Lef. While P-catenin provides the transcriptional domain, Tcf provides D N A binding elements, thereby, allowing P-catenin/Tcf complex binding to D N A and initiation of gene transcription. Additionally, both the NTD and CTD of P-catenin are necessary for connecting of the P-catenin/Tcf complex to that of the basal transcriptional machinery (van de Wetering, Oosterwegel et al. 1991) and CREB binding proteins (Hecht and Kemler 2000; Takemaru and Moon 2000). Well studied targets of Tcf include c-myc, c-jun, fra-1, PPAR5, uPA, MMP-7, fibronectin and cyclins (Schwartz, Wu et al. 2003). Interactions of P-catenin with Tcf are negatively regulated by the Tcf binding proteins N L K (NEMO-like kinase) (Ishitani, Ninomiya-Tsuji et al. 1999) and members of the Groucho family (Ishitani, Ninomiya-Tsuji et al. 1999). Transcription factors likely alter the response of Tcf by competing away activating P-catenin as exemplified with Soxl , R A R and AR. 1.12 P-catenin in tumourigenesis Initial reports concerning the importance of P-catenin in cancer arose upon the identification of inactivating mutations in the tumour suppressor, APC, rendering it unable to degrade P-catenin. Thus, APC mutations are associated with accumulation of nuclear P-catenin 15 and activation of Tcf/lef (Henderson and Fagotto 2002). Oncogenic mutations in colon cancer are also associated with the NTD of P-catenin (van Noort, Meeldijk et al. 2002) (Giles, van Es et al. 2003) as serine and threonine mutations within this domain prevent GSK3p from phosphorylating and degrading P-catenin. The implication of mutational activation of P-catenin/Tcf is increased activity of Wnt target proteins that can regulate the cell cycle (cyclin DI) (Shtutman, Zhurinsky et al. 1999; Tetsu and McCormick 1999), (c-myc) (He, Sparks et al. 1998), E C M adhesion (fibronectin) (Gradl, Kuhl et al. 1999) and E C M invasion (matrylisin) (Crawford, Fingleton et al. 1999). 1.13 Alterations in E-cadherin/p-catenin mediated adhesion during PrCa progression Cadherins are transmembrane glycoproteins that facilitate cell-cell adhesion in a calcium dependent manner. Cancer cells can escape contact inhibition allowing uncontrolled cell growth and layering. Frequently accompanying contact independent cell cycling is loss of cell-cell junctional integrity. Thus, reduced E-cadherin expression, found predominantly at epithelial junctions, serves as an important indicator for transformation during focus formation, contact independent growth and xenograft (tumour) potential (Beavon 2000) (Hajra and Fearon 2002). E-cadherin signaling to cell cycle regulators such as p27 K i p l , and its interaction with cyclin dependent kinases (CDKs), has been implicated in allowing regulation of contact inhibition. Formation of p27/cdk complexes deactivates cdk function and, thus, reduces cell progression (St Croix, Sheehan et al. 1998). Therefore, loss of E-cadherin function, either by C a 2 + depletion or neutralizing antibodies can reduce nuclear accumulation of p27, thus removing one of the required "brakes" for cell cycle progression (Levenberg, Yarden et al. 1999) (St Croix, Sheehan et al. 1998). E-cadherin expression can be reduced in a number of ways including: anchorage independence and epithelial migration; abnormal cellular location of E-cadherin; loss of expression by inactivating mutations; proteolytic degradation; and altered expression of oc-catenin in addition to over-phosphorylation in P-catenin (Beavon 2000) (Yoshida, Kimura et al. 2001). Thus, while junctional P-catenin does not directly regulate cell cycle regulators, loss of E-cadherin during progressive PrCa concomitant with anchorage independence and metastasis may 16 contribute, indirectly, to increased nuclear levels of P-catenin and, therefore, enhancement of P-catenin/Tcf function (Giroldi and Schalken 1993; Koksal, Ozcan et al. 2002). 1.14 AR and PI3K/Akt provide essential signaling events for acquisition of androgen independent PrCa A particularly exciting area of cytosolic signaling, with respect to AI PrCa, is the functional interaction of A R with the phosphatidylinositol 3-kinase (PI3-K) signaling pathway (Figure 1.8). It is now appreciated that both A R and PI3K cascades are crucial for progression to hormone refractory PrCa and, therefore, are being intensely investigated for therapeutic intervention. Recent data strongly argues that these pathways are dynamic in their interactions and, thus, should be considered collectively when interpreting basic or clinical data concerning progression PrCa (See Appendix Review: PTEN & GSK3/3: Critical Regulators in Progression to Androgen Independent Prostate Cancer). Several lines of evidence suggest that A R and PI3K/Akt signaling are the main modes by which the prostatic epithelium escapes apoptotic responses associated with androgen blockade. Specifically, many PSA secreting (e.g. LNCaP, LAPC4, C4-2, CWR22) and non-PSA secreting (e.g. PC3, DU145) PrCa cell lines have been shown to harbour constitutive PI3K/Akt signaling. Immunohistochemical analysis of high Gleason grade PrCa indicates that phosphorylated Akt is both highly expressed (Malik, Brattain et al. 2002) and an excellent predictor of PSA failure (Kreisberg, Malik et al. 2004). Furthermore, stable expression of activated Akt results in substantial increases in LNCaP xenograft growth (Graff, Konicek et al. 2000). Inhibitors of PI3K (e.g. LY294002, Wortmanin) induce rapid apoptosis while those targeting downstream effectors such as mTOR kinase (e.g. Rapamycin, RAD001, CCI-779) or M A P K signaling (e.g. PD98059) do not elicit an obvious apoptotic response in LNCaP cells cultured in vitro (Lin, Adam et al. 1999). These data suggest that upstream PI3K signaling events are necessary for progression to AI PrCa. Compelling evidence also exists for the importance of A R in AI PrCa. While treatment with the anti-androgen, bicalutamide, invokes decreased mitotic activity and growth, it does not induce an acute apoptotic response (Murillo, Huang et al. 2001). Despite this, recent experiments 17 using shRNA indicate that AR is necessary for progression to a hormone refractory state (Chen, Welsbie et al. 2004). This has been shown by in vivo chromatin immunoprecipitation (LP) experiments using PrCa xenografts whereby A R was found to bind the PSA promoter enhancer region during recurrent PrCa (Zhang, Johnson et al. 2003). Therefore, while androgen deprivation can be tolerated in many PrCa models, loss of A R generates a distinct loss of growth advantage (Wright, Tsai et al. 2003) (Chen, Welsbie et al. 2004). Growth factors, including IGF-1, EGF, K G F (Culig, Hobisch et al. 1994), Her2/Neu (Wen, Hu et al. 2000), IL-6 (Xie, Lin et al. 2004) and IL-4 (Lee, Lou et al. 2003), have been demonstrated to activate both PBK/Akt signaling as well as the A R signaling axis. IGF-1 is the most potent in its ability to activate A R and can induce a 6.-7 fold increase in PSA without androgens (Culig, Hobisch et al. 1994). Her-2/neu activates Akt signaling to increase AR phosphorylation and AI cell survival (Wen, Hu et al. 2000). Thus, while interleukins regulate AR through the STAT3 and M A P K signaling pathways, they can also function to activate A R transactivation and PSA production through P B K / A k t signaling. Interleukins 4 and 6 sensitize A R to low levels of androgens in an Akt dependent manner (Lee, Lou et al. 2003; Yang, Wang et al. 2003). Furthermore, while it is clear that A R signaling has the capacity to repress P B K -mediated neuroendocrine differentiation, hormone ablation therapy may, in fact, promote neuroendocrine forms of PrCa (Wang, Horiatis et al. 2004). Therefore, interleukins, growth factors and mitogenic stimuli can activate A R in a PI3K dependent manner. Activated Akt is a crucial mediator of P B K signaling and promotes growth through the mTOR and p70 S6 kinase pathways (Vivanco and Sawyers 2002). Akt promotes survival by f"*IP KTP repressing the C D K inhibitors, p21 and p27 , resulting in increased cell proliferation (Shin, Yakes et al. 2002). Of considerable importance in escaping androgen dependency is the ability of P B K / A k t to repress apoptotic signals including those elicited by B A D and the forkhead family of transcription factors (Huang, Muddiman et al. 2004). Akt can also directly inhibit central regulators of cell death including mdm2, p53 and p21 (Vivanco and Sawyers, 2002). Phosphorylation serves a critical role in A R function including receptor nuclear translocation and transcriptional activation (Wong, Burghoorn et al. 2004). Recent reports have implicated several components of P B K signaling with A R phosphorylation. Akt can phosphorylate A R at Ser210 and Ser790with an apparent ability to either inhibit or activate A R function (Lin, Hu et al. 2003) (Lin, Yeh et al. 2001; Lin, Wang et al. 2002) (Wen, Hu et al. 2000). Studies concerning P B K 18 signaling in LNCaP cells have indicated a differential response to androgens. At low DHT concentrations (0.1 nM >) Akt phosphorylates and activates A R function (Wen, Hu et a l , 2000) while at high androgen concentrations (> 1 nM) Akt represses A R transcription, and therefore inhibits apoptotic gene induction (Griffiths, Morton et al. 1997). Thus, it is possible that subsequent to androgen withdrawal therapy Akt confers a greater role in A R activation, thus, contributing to AI growth. IGF-1 stimulation appears to be required for Akt mediated phosphorylation of A R and may account for the variation in reports of Akt dependent phosphorylation of A R (Lin, Hu et al. 2003). Importantly, constitutive activation of Akt can facilitate A R transcription in the absence of androgens, promoting cell proliferation and protection from apoptosis (Wen, Hu et al. 2000). Akt induces other post-translational modifications including complexing with mdm2, resulting in phosphorylation and ubiquitination of A R (Lin, Wang et al. 2002). The IGF-l/Akt adaptor protein, APPL, has been shown to enhance the repressive abilities that Akt has upon A R transcription (Yang, Lin et al. 2003). Activation of AR, in androgen depleted conditions, may also be facilitated by increased association with potent transcriptional transactivators of A R including ARA70, ARA54 and TLF-2, events thought to be enhanced by Akt function (Yang, Lin et al. 2003). 1.15 Intact PI3K/Akt and AR signaling: a mutualistic relationship PI3K signaling has been shown to be necessary for proper A R expression as abrogation of PI3K (He and Wilson 2002) signaling , either by LY294002 or PTEN expression, results in decreased A R activity (Manin, Baron et al. 2002). Similarly, androgens and A R have the capacity to activate PI3K and many of its downstream effectors, non-genomically (Baron, Manin et al. 2004). A R activation of PI3K signaling requires constitutive complexing both with Src and the p85, catalytic subunit of PI3K (Baron, Manin et al. 2004). This results in phosphorylation of Akt and downstream initiators of apoptosis, B A D and F K H R (Huang, Muddiman et al. 2004). Interestingly, A R has been shown to bind and inactivate FKHR in DU145 PrCa cells, independently of Akt phosphorylation (Li, Lee et al. 2003). However, in cells expressing endogenous AR, androgens appear to be necessary for androgen/PI3K 19 interactions as treatment with the anti-androgen, bicalutamide, inhibits androgen activation of Akt (Baron, Manin et al. 2004). These observations are also recapitulated in the MC3T3-E1 osteoblast cell line in which A R is observed to both induce phosphorylation and nuclear translocation of Akt (Kang, Cho et al. 2004). Androgens function to maintain prostate epithelium in a differentiated state. In accordance, PC3-AR stable cell lines treated with androgens show reduced mitotic activity and capacity for invasion by interfering with AR/EGF complexing (Bonaccorsi, Muratori et al. 2004) and EGF7PT3K signaling (Bonaccorsi, Carloni et al. 2004). GSK3P, a relatively new player in A R signaling , phosphorylates regions of the AR-L B D leading to reduced transcription and proliferation of PrCa cells (Salas, K im et al. 2004). GSK3(3 may also regulate sites on the AR-NTD, thus affecting the ligand independent (APT) transactivation region (Wang, Lin et al. 2004). By binding and phosphorylation, GSKP promotes reduced A R NTD/CTD interactions and, therefore, A R transcription (Salas, K im et al. 2004). Others have shown that GSK3(3 action is required for A R function in addition to activation of A R gene targets (e.g. PSA, MMP2) and exogenous gene targets (Liao, Thrasher et al. 2004). Thus, it is possible that a basal level is required for A R function while increased GSK3p levels are inhibitory. GSK3p has diverse functions including regulation of the P B K and Wnt signaling pathways (Ding, Chen et al. 2000). During PrCa progression, increased PBK/Akt signaling reduces GSK3|3 activity resulting in decreased GSK3(3 mediated, repression of AR. As P-catenin is also regulated by GSK3P (Ding, Chen et al. 2000) and is a potent transcriptional co-activator of A R (Truica, Byers et al. 2000; Mulholland, Cheng et al. 2002), loss of GSK3P function may lead to increased transcriptional activation of AR. Pharmacological activation of GSK3P is, therefore, an attractive prospect for inactivation of aberrant A R function during AI PrCa. In PrCa systems harbouring elevated P B K / A k t activity, incremental increases in either Wnt or P B K dependent pools of GSK3P (Ding, Chen et al. 2000) could have a dual effect in reducing activity of AR. 1.16 Phosphatase and Tensin Homologue Deleted Homologue (PTEN): structure, function and significance in PrCa PTEN (Phosphatase and Tensin homologue Deleted on chromosome 10) is a dual phosphatase that is mutated up to 40% of glioblastomas (Liu, James et al. 1997; Tohma, Gratas et al. 1998) and 50% of advanced PrCas (Cairns, Okami et al. 1997; Whang, Wu et al. 1998) 20 mutational silencing. As a relatively recent defined tumour suppressor (Li, Yen et al. 1997), PTEN has gained considerable importance in its abilities as a lipid and protein phosphatase (Myers, Stolarov et al. 1997). The PTEN cDNA codes two transcripts (2 and 5 kb) and a peptide of 403 amino acid (47 kD). PTEN contains several functional domains including a tensin/auxillin homologous domain, a lipid/protein phosphatase domain and a PDZ binding motif (Figure 1.9). The NTD of PTEN has a high degree of homology with tensin and auxillin, two proteins involved in assembly of focal adhesion plaques (Wehrle-Haller and Imhof 2002). The phosphatase domain is structurally related to a family of dual specific phosphatases, CDC 25, that function to regulate cyclin-dependent kinases. The phosphatase domain is, thus, crucial for the tumour suppressor activities of PTEN since point mutations in this region are associated with loss of PTEN enzymatic activity and increased tumourigenesis (Myers, Stolarov et al. 1997). The implication, is therefore, that loss of PTEN phosphatase activity is highly associated with progression of many cancers. PTEN protein and lipid phosphatase activity have the capacity to regulate three, well defined pathways including that associated with (1) cell migration and invasion, (2) gene transcription and (3) apoptosis and proliferation. PTEN has been shown to decrease integrin mediated spreading of cells and achieves this effect by inhibiting tyrosine phosphorylation of Focal Adhesion Kinase (FAK) and its downstream target, p l 3 0 C A K (Crk associated substrate). This results in modulation of the actin skeleton, reduced cell migration and, therefore, invasion (Raftopoulou, Etienne-Manneville et al. 2004). The ability of WT PTEN to antagonize integrin signaling is solely dependent upon its protein phosphatase function. This is supported by the fact that PTEN lipid phosphatase mutants show an identical ability to PTEN in inhibiting focal adhesion formation (Tamura, Gu et al. 1998; Tamura, Gu et al. 1999; Tamura, Gu et al. 1999). PTEN can also inhibit general cellular transcription by interacting directly with She and the adapter proteins, Grb2/Sos (Weng, Smith et al. 2001). Thus, PTEN reduces E R K / M A P K activation of Elk and its transcriptional targets. PTEN's role as a tumour suppressor is, perhaps, best understood by its capacity to regulate PI3K signaling. PI3K functions to generate phosphatidylinositol (3, 4, 5) triphosphate (PLP3), a lipid substrate responsible for recruitment and activation of downstream effector kinases (Figure 1.8). PTEN functions to specifically remove the 3' lipid phosphate from the inositol ring of PI(3, 4, 5)P3 and also has a minor phosphatase activity towards PI(3)P, PI(3,4)2 21 and PI(1,3,4,5)P4 (Vivanco and Sawyers 2002). The ability of PTEN to decrease intracellular PLP3 pools implies that PTEN can reduce proliferation, cause G l arrest and induce apoptotic signals. Specifically, decreased pools of PLP3 results in reduced membrane localization of lipid binding kinases as well as decreased Akt-Thr308 and Akt-Ser473 activation by PDK1 and PDK2, respectively (Vivanco and Sawyers 2002). Importantly, PTEN induced cell cycle arrest is inhibited in the presence of high (10%) serum conditions (Furnari, Huang et al. 1998), indicating that the balance of IGF/PI3K and PTEN signaling effects dictate physiological and phenotypic outcome. Several lines of evidence indicate that PTEN is a crucial factor in PrCa progression: (1) PTEN expression is absent in most PrCa cell lines and xenografts (Cairns, Okami et al. 1997), (2) primary tumours show allelic loss of PTEN in up to 50% of advanced PrCa cases (Cairns, Okami et al. 1997), (3) prostate specific deletion of PTEN results in formation of PrCa in mice (Wang, Gao et al. 2003). Evidence that PTEN serves a crucial role in progression of PrCa is accumulating rapidly largely through development of novel animal models. Using such models it is appreciated that the dose of PTEN inactivation regulates the rate of progression, latency and frequency of PrCa in mice (Trotman, Niki et al. 2003). PTEN loss in humans is also correlated with high Gleason grade and progressiveness of the disease (McMenamin, Soung et al. 1999). While PTEN loss clearly causes PrCa in mouse prostate, not all humans who undergo radical prostatectomy have single allele or bi-allelic loss of PTEN (Cairns, Okami et al., 1997). More specifically, as up to 50 % of advanced PrCas show loss of PTEN, at least 50% have functional PTEN, thus, suggesting that PrCa can still occur even in the presence of PTEN. Interestingly, single allele loss of PTEN can act in a synergistic manner with Nkx3.1 allele loss to form invasive adenocarcinoma in mice of at least 1 year (Kim, Cardiff et al. 2002) (Abate-Shen, Banach-Petrosky et al. 2003). A recent report indicating that prostate specific deletion of PTEN forms progressive and metastatic cancer; provides not only compelling evidence that PTEN can cause PrCa but also provides one of the best models for the study of PrCa (Wang, Gao et al. 2003). Thus, the clinical implications are that PTEN expression is associated with elevated PI3K signaling and protection from apoptosis due to androgen withdrawal therapy. 22 1.17 P-catenin and AR interactions: A statement of other published and related works In carrying out experimental data presented in this thesis, other works of similar nature have appeared in published format either before, concurrently or after journal publications that were based upon this thesis data. While these reports are summarized in many instances in individual chapters, discussions and in the appendix review papers, the following serves as a succinct declaration and summary of related, published works at the time of this thesis writing. These works can be summarized into four main topics including: 1) P-catenin Binding and Co-activation of AR; 2) Glycogen synthase kinase (GSK3P) Modulation of AR/P-catenin Interactions; 3) Transrepression of P-catenin/Tcf Signaling by AR, 4) PTEN as a Potent Regulator of A R and P-catenin Function and 5) Clinical and Therapeutic Implications of p-catenin/AR Interactions. 1.17.1 P-catenin Binding and Co-activation of AR Initial reports by Truica et al., identified that P-catenin can complex in a ligand (androgen) sensitive manner to AR. This report not only identified that p-catenin serves as a transcriptional co-activator of A R and has the potential to promote responsiveness of P-catenin to non-androgens (e.g. estradiol) and to the presence of anti-androgens (e.g. OH-flutamide) (Truica, Byers et al. 2000). These important, initial observations have since been both confirmed and expanded upon by several other investigators. Specifically, it has been shown that AR and P-catenin interact by direct binding and complexing, as ascertained by yeast two-hybrid analysis (Yang, L i et al. 2002) (Song, Herrell et al. 2003), GST pull downs (Yang, L i et al. 2002) (Mulholland, Cheng et al. 2002), coimmunoprecipitations (Truica, Byers et al. 2000) (Yang, L i et al. 2002) (Mulholland, Cheng et al. 2002), and transcriptional reporter assays (Truica, Byers et al. 2000) (Yang, L i et al. 2002) (Song, Herrell et al. 2003) (Mulholland, Read et al. 2003) (Pawlowski, Ertel et al. 2002). AR/p-catenin interactions are ligand sensitive, whereby complexing occurs in the presence of dihydrotestosterone (DHT), or R1881, less in the absence of ligand (Truica, Byers et al. 2000) (Mulholland, Read et al. 2003), and not in the presence of the pure A R antagonist, bicalutamide (Song, Herrell et al. 2003) (Mulholland, Read et al. 2003). Transient transfections of deletion mutant expression plasmids and yeast two-hybrid studies suggest that the ligand binding domain of A R (AR-LBD) is necessary and sufficient for AR/P-catenin interactions (Yang, L i et al. 2002) (Song, Herrell et al. 2003). Reduced AR/p-catenin 23 binding, in the presence of pure A R antagonists, may be explained by the formation of unfavorable stoichiometry by helix 12 of A R - L B D , with respect to its binding pocket, thereby preventing efficient coactivator binding (Darimont, Wagner et al. 1998) (Mclnerney, Rose et al. 1998). Despite the conservation in structure between NRs, it is clear that the A R - L B D has unique structural aspects that facilitate binding to P-catenin. L X X L L binding motifs (L = leu, X = any amino acid) contained within the A R - L B D serve a different function than most other NRs. In AR, L X X L L binding regions of the L B D serve to mediate NTD (containing F X X L F ) and L B D interactions (Shen, Buchanan et al. 2005) (He and Wilson 2002), whereas in most other NRs, L X X L L binding motifs serve primarily to recruit transcriptional coactivators (Chang and McDonnell 2002). Mutation of A R - L B D helices 3 and 12 results in disruption of AR/p-catenin binding, with alteration of helix 3 affecting binding of all three, P-catenin, the AR-NTD, and the transcriptional regulator, TLF2 (transcriptional initiation factor) (Song, Herrell et al. 2003). In fact, as P-catenin / A R - L B D interactions have been shown to depend upon L X X L L binding motifs found within the AR, it has been suggested that P-catenin is required for changes in structural conformation that are coincident with ligand binding and dissociation of heat shock proteins (Song, Herrell et al. 2003). Regions of p-catenin necessary for interaction with A R have also been well defined. P-catenin Arm repeats 1-6 are required because mutation of repeats 5 or 6 abolishes binding, coactivation, and nuclear cotranslocation interactions with A R (Yang, L i et al. 2002). Importantly, Arm repeats 5 or 6 also bind Tcf4 and E-cadherin (Pai, Kirkpatrick et al. 1996) (Hulsken, Birchmeier et al. 1994), suggesting the therapeutic possibility of simultaneously disrupting the coactivating effects that P-catenin may confer upon either A R or Tcf4. That overexpression of Tcf4 or E-cadherin blocks P-catenin interactions with the AF2 region of A R indicates that P-catenin binds these molecules in close proximity of the Arm repeats (Song, Herrell et al. 2003) (Mulholland, Read et al. 2003). P-catenin also contains L X X L L motifs, found on the second cc-helix of Arm repeats 1, 7, 10, and 12 (Pai, Kirkpatrick et al. 1996) (Huber, Nelson et al. 1997). However, deletion mutants of repeats 7, 10, and 12 suggest that these sites may not be necessary for AR/p-catenin binding (Yang, L i et al. 2002) (Song, Herrell et al. 2003). Structurally, this may be explained by the fact that leucine residues of the Arm repeats are buried within hydrophobic cores, possibly rendering them inaccessible to NR binding (Huber, Nelson et al. 1997). P-catenin does, however, transactivate A R on minimal 24 transcriptional reporters (Yang, L i et al. 2002) (Song, Herrell et al. 2003) (Mulholland, Cheng et al. 2002) (Masiello, Chen et al. 2004) as well as endogenous targets such as prostate-specific antigen (PSA) (Masiello, Chen et al. 2004) at a magnitude similar to CBP (Yang, L i et al. 2002) and steroid receptor coactivator 1 (Truica, Byers et al. 2000) (Yang, L i et al. 2002), thereby demonstrating the potency of (3-catenin as an A R regulator. Interestingly, P-catenin i s more effective as an AR coactivator in cell lines harboring endogenous A R (Chesire, Ewing et al. 2002) (Chesire and Isaacs 2002), suggesting that the ability of P-catenin to enhance A R coactivation is sensitive to the endogenous cellular milieu of coregulators. The affinity of P-catenin/AR interactions is likely attributable to the unique structural identity of the A R - L B D but is likely also accounted for by differences in the supporting network of coregulators between cell lines. 1.17.2 Glycogen synthase kinase (GSK3P) Modulation of AR/p-catenin Interactions GSK3P resides at the junction of the PI3K/Akt and Wnt/p-catenin/Tcf survival pathways, thereby serving critical roles in cellular metabolism, growth, and proliferation (Frame and Cohen 2001) (Kim and Kimmel 2000). Under nonstimulated conditions, GSK3P pools are constitutively active, but they are phosphoinhibited upon PI3K/Akt or Wnt activation (Doble and Woodgett 2003) . Although the substrates for GSK3P are generally specific and allow for disparate signals between PI3K and Wnt signaling (Doble and Woodgett 2003) (Frame and Cohen 2001) (Ding, Chen et al. 2000), it appears that GSK3P may not only directly regulate NRs but also indirectly regulate A R function by modulation of PI3K/Akt and Wnt/p-catenin/Tcf pathways. Specifically, GSK3P both regulates cell cycle events in endocrine cancers and posttranslationally regulates NR. For example, GSK3p binds and phosphorylates both A R (Salas, K im et al. 2004) (Wang, Lin et al. 2004) at sites within the hinge and ligand binding regions resulting in decreased gene activation and cell proliferation. In accordance, overexpression of the GSK3P inhibitor, Akt, or treatment with L i C l results in increased A R activity (Salas, K im et al. 2004). This implies that endocrine pathologies, with mutational silencing of PTEN and Akt-inactivation of GSK3P may allow for more promiscuous NR function. Short interfering RNA-mediated silencing of GSK3P results in suppression of androgen-stimulated gene expression in PSA-secreting cells (Liao, Thrasher et al. 2004). This suggests that although a minimal level of GSK3P is required for proper A R expression, higher levels, achieved in lesser PI3/Akt-sensitive cells, decrease A R 25 activity. Interestingly, introduction of Akt-resistant forms of GSK3|3 (deleted amino acids 1-9) has been reported to promote nuclear colocalization of GSK3P and A R (Salas, K im et al. 2004); however, the physiological significance of this remains to be determined. The dynamics of GSK3P/NR interactions cannot be properly interpreted without consideration of Akt, an upstream regulator of GSK3P and key dictator in the determination of how GSK3(3 influences A R function. Loss of PTEN expression and the resulting constitutive activation of PI3K/Akt signaling (Li, Yen et al. 1997) (Steck, Pershouse et al. 1997) results in repressed GSK3P function. In these systems, GSK3P direct regulation of NRs is likely limited; however, in systems with reduced PI3K/Akt signaling, A R likely undergoes simultaneous inactivation by Akt and GSK3p dependent phosphorylation, with the net outcome likely dependent upon the activity of PI3K signaling and the local concentration of androgens. GSK3P has also been demonstrated to mediate the coactivating effects that P-catenin confers upon AR, thereby reflecting the functional integration of Wnt and PI3K signaling in PrCa cells (Persad, Troussard et al. 2001) (Sharma, Chuang et al. 2002) (Culig, Hobisch et al. 1994). Both Wnt3a (Verras, Brown et al. 2004) (Cronauer, Schulz et al. 2005) and IGF-I (Culig, Hobisch et al. 1994) enhance A R coactivation and production of endogenous PSA. Therefore, an attractive hypothesis is that Wnt and PI3K/Akt growth factors phosphoinactivate GSK3P promoting stabilization, nuclear localization of P-catenin (Mulholland, Read et al. 2003) (Pawlowski, Ertel et al. 2002), enhanced P-catenin/AR interactions (Verras and Sun 2005), and enhanced proliferation (Verras, Brown et al. 2004) (Cronauer, Schulz et al. 2005). Ultimately, however, it appears that the status of PTEN dictates these events and that, upon its loss, creates an environment that is highly amenable for NR gene activation. As such, it has been demonstrated that PTEN can effectively antagonize the functions both A R (Li, Nicosia et al. 2001) (Nan, Snabboon et al. 2003) and p-catenin (Persad et al., (Sharma et al). 1.17.3 P-catenin Co-trafficking and Transrepression of Tcf-mediated transcription by AR The coactivating effects that P-catenin exerts on A R have been intensively studied; however, it is apparent that A R can invoke equally dynamic changes in Wnt/p-catenin function 26 that include nuclear cotrafficking and transrepression. AR/p-catenin complexing and cotrafficking have been observed in PrCa cells containing both endogenous and exogenous A R (Mulholland, Cheng et al. 2002). AR/p-catenin nuclear cotrafficking has also been reported in nonprostate cell lines such as CV-1 cells (Yang, L i et al. 2002) and in pituitary cells (Pawlowski, Ertel et al. 2002). Furthermore, castrated mice reconstituted with androgen pellets show an obvious redistribution of both A R and p-catenin to the nucleus in normal prostatic epithelium (Chesire, Ewing et al. 2002). Because P-catenin lacks a nuclear localization signal, A R provides a vehicle to recruit P-catenin to nuclear foci, sites likely consisting of active transcription (Mulholland, Cheng et al. 2002). Consistent with binding requirements, P-catenin Arm repeats 1-6 have been shown to be necessary for efficient nuclear cotrafficking and resulting coactivation of A R in CV-1 cells (Yang, L i et al. 2002) (Pawlowski, Ertel et al. 2002). However, functional A R - N T D / L B D interactions or pi60 coactivator binding are not required for A R mediated nuclear entry of P-catenin (Pawlowski, Ertel et al. 2002). Furthermore, GSK3P or PI3K inhibitors fail to disrupt AR-mediated nuclear accumulation of P-catenin (Pawlowski, Ertel et al. 2002). Together, these data suggest that there are minimal structural requirements for AR-mediated import of P-catenin and that this import is not dependent on the P-catenin shuttling protein, APC (Mulholland, Cheng et al. 2002), or the PI3K pathway (Sharma, Chuang et al. 2002). Recent studies have indicated that ligand-activated A R potently inhibits Wnt signaling in colon cancer cell lines and to a moderate extent in PrCa cells (Pawlowski, Ertel et al. 2002) (Chesire and Isaacs 2002) (Song, Herrell et al. 2003) (Mulholland, Read et al. 2003) (Cullen, Killick et al. 2004) (Amir, Barua et al. 2003). Specifically, PrCa (CWR22-Rvl, L A P C -4, DU145) and nonprostate cell lines (TSU, HEK-293, SW480, HCT-116) that express AR demonstrate repression of both TOPFLASH and cyclin DI protein levels upon treatment with DHT (Chesire, Ewing et al. 2002). In accordance with a mechanism of P-catenin reciprocity, increased Tcf expression and antiandrogen treatment can rescue the repressive effects that androgens confer upon Wnt (Chesire, Ewing et al. 2002) (Mulholland, Read et al. 2003). These studies strongly suggest that a limited pool of P-catenin can associate with either Tcf or A R in an androgen-dependent manner. A R repression of Wnt signaling is not without clinical relevance because pathological expansion of the A R polyglutamine tract (increases of 20 to 51 glutamine repeats) results in diminished, inhibitory effects on T O P F L A S H reporter activity. These observations provide a potential connection between Wnt signaling and Kennedy's disease, a disease characterized by expanded glutamine repeats (Cullen, Kill ick et al. 2004). 27 1.17.4 PTEN as a Potent Regulator of AR and P-catenin Function The ability of P-catenin to serve as an A R activator likely also depends on variations in PI3K/Akt signaling status between cell lines. Constitutively activated PI3K signaling, found frequently in many PTEN -/- PrCa cell lines, promotes decreased GSK3P function and, consequently, high levels of P-catenin (Persad, Troussard et al. 2001) (Sharma, Chuang et al. 2002) (Verras and Sun 2005). Therefore, A R coactivation may be less apparent in cells containing high levels of total P-catenin with relatively little opposition by the Wnt degradative system. System differences in P-catenin coactivation effects could also reflect the availability of A R pools to bind and import P-catenin to the nucleus (Yang, L i et al. 2002) (Mulholland, Cheng et al. 2002) (Pawlowski, Ertel et al. 2002). Genetic silencing of PTEN is a frequent genetic event in advanced PrCa and has been clearly associated with accumulation of nuclear P-catenin (Persad, Troussard et al. 2001) (Sharma, Chuang et al. 2002) and cyclin DI (Tetsu and McCormick 1999). Despite the loss of PTEN and accumulation of p-catenin, activation of p-catenin/Tcf gene targets is low in most advanced PrCa systems suggesting that the pro-proliferative effects of P-catenin are mediated via AR, as opposed to P-catenin's cognate receptor, Tcf. 1.17.5 The Potential Clinical Relevence of P-catenin in Prostate Cancer Advanced PrCa is often treated by total androgen ablation therapy; however, the ultimate phenotype is one of androgen independence (AI) and death (Kojima, Suzuki et al. 2004). Transcriptional coregulators are hypothesized to serve a critical role in promoting a more aggressive A R during androgen-independent PrCa. Altered ligand responsiveness of A R has been postulated as a major mechanism by which PrCa continues to proliferate in low androgen environments (Chen, Welsbie et al. 2004). As such, several lines of evidence indicate that P-catenin may promote the oncogenicity of AR. In light of this, it is significant that P-catenin increases AR-mediated gene activation not only in the presence of DHT, but also in the presence of the weaker adrenal androgen, androstenedione (Truica, Byers et al. 2000), a steroid remaining present in chemically castrated patients. Importantly, in addition to the AR-specific coactivators ARA70 and ARA50, P-catenin is one of the few coactivators to enhance transcription in LNCaP cells upon treatment with 17P-estradiol (Truica, Byers et al. 2000). P-catenin also coactivates mutant forms of A R that are clinically relevant, including AR-W741C and AR-T877A, 28 mutations found in PrCa cell lines isolated from hormone refractory PrCa patients that have been treated with bicalutamide (AR W741C) (Chesire, Ewing et al. 2002) (Taplin, Rajeshkumar et al. 2003) and in lymph node metastatic lesions, respectively. These findings indicate that P-catenin acts as a coactivator both of wild-type (WT) and mutant AR. Therefore, p-catenin is not only altering the specificity of A R toward certain ligands but is also acting as a pure coactivator. The status of Wnt signaling in many PrCa cell lines (PC3, LNCaP) and bladder cancer (TSU) cell lines is low, though functionally, likely due to relatively low levels of endogenous Tcf (Mulholland, Read et al. 2003) (Chesire, Ewing et al. 2002) (Chesire and Isaacs 2002). Despite this, several lines of evidence indicate that P-catenin is important for progression of PrCa. Cre-mediated excision of the P-catenin (exon3) regulatory domain develops hyperplasia and transdifferentiation in mice at 18 wk of age but without metastatic behavior (Bierie, Nozawa et al. 2003) . In a similar model, stabilized P-catenin appears to be important for the initiation of prostatic neoplastic lesions (Gounari, Signoretti et al. 2002), a phenotype comparable to intestinal polyps, the precursor of invasive carcinoma and colon cancer. Also, gain-of-function, truncated forms of P-catenin occurring in metastatic prostate and breast specimens have been shown to preferentially locate to the nucleus, possibly serving as an additional "pool" of P-catenin to promote cell proliferation during the androgen-independent phenotype (Rios-Doria et al., 2004). AR/P-catenin interactions have distinct clinical relevance and therapeutic options. For example, P-catenin recruitment to A R is readily detected in a PrCa subline containing an AR-W741C mutation (isolated from a hormone refractory PrCa specimen) (Masiello, Chen et al. 2004) , suggesting that P-catenin/AR complexing may be increased in hormone refractory PrCa. Furthermore AR/p-catenin interactions are reduced or abolished in the presence of A R antagonists, suggesting that loss of P-catenin function may be associated with the beneficial effects of antiandrogen therapy for PrCa patients (Masiello, Chen et al. 2004). P-Catenin/AR-L B D interactions have been demonstrated to be dependent upon a single A R lysine (K720) for binding (Yang, L i et al. 2002), a site also necessary for proper A R N T D and TLF2 interactions (He, Kemppainen et al. 1999). 29 1.18 Thesis rationale and statements of hypotheses The aim of this thesis is to decipher a potential functional role(s) for P-catenin in PrCa systems. A central rationale for the following studies is precipitated from the capacity for P-catenin to function both as promoter of cell proliferation and oncoprotein in many cancers. The following represent statements of hypothesis for major, experimental themes. 1 ) Wnt/P-catenin/Tcf signaling is functional in A R and non-AR expressing PrCa cells. Activation of the P-catenin/Tcf axis, by either Wnt ligand occupation or stabilizing P-catenin mutations, serves to potentiate A R transactivation, cell proliferation, and tumourigenesis of PrCa cells. 2) p-catenin physically interacts with A R and can undergo androgen-sensitive nuclear translocation to complex and transactivate A R on hormone response containing elements. The androgen sensitive complexing of A R and P-catenin to AREs promotes sequestering of P-catenin from Tcf sites leading to diminished P-catenin/Tcf signaling by a mechanism of competition for a limited pool of P-catenin, between A R and Tcf transcription factors. 3) PTEN is a negative regulator of the P-catenin/Tcf signaling axis and, therefore, also functions to antagonize A R in vitro. In vivo, genetic silencing or overexpression of PTEN is, therefore, associated with abrogated or diminished expression of p-catenin and/or AR. If P-catenin expression is associated with advancing PrCa, then high Gleason grade PrCa is associated with increased total and/or nuclear expression of P-catenin in conjunction with increased cell proliferation index. 30 Figure 1.1 Histological and clinical grading system for progressive adenocarcinoma of the prostate, a) Stages range from most differentiated (grade 1) to least differentiated (grade 5). The combined grades of the primary and secondary histological features are equal to the Gleason score, a value that is used in conjunction with clinical evaluation to predict a patient's clinical prognosis (Gleason, 1974). b) Tumour, Lymph Node and Metastasis (TNM) grading is a strong clinical predictor for clinical outcome for PrCa patients. Importantly, patients with cancer confined to the prostate (<T3a) have a significantly more favourable clinical prognosis than patients with extra capsular cancer. Patients with T N M grading of T4 (Mia) or greater generally have less than one year to live (Eichelberger, 2004). a) Gleason Scoring/Grading PROSTATIC ADENOCARCINOMA Well differentiated Small uniform glands Well differentiated Medium sized acini still separated by stroma Cells closer together Most common PrCa histology Moderately differentiated Variation in glandular size Infiltration of stroma Poorly differentiated Atypical ceils Extensive infiltration Poorly differentiated Sheets of undifferentiated cells b) Clinical (TNM) Grading TO No evidence of primary tumour T1a Cancer present in less than 5% tissue during prostatectomy n i b Cancer present in greater than 5% during prostatectomy T1c Cancer detected in needle biopsy alone (usually as a result of elevated PSA) T2a Cancer still confined within one lobe T2b Cancer is confined two both lobes T3a Cancer has spread to nearby tissues but NOT the seminal vesicles T3b Cancer has spread to nearby tissues including the seminal vesicles T4 (NO) Cancer adherent to extracapsular structures but no lymph node metastasis T4 (N1-3) Regional lymph node metastasis T4 (MO) No distant metastasis T4(M1a) Distant metastasis in lymph beyond the pelvic region T4(M1b) Distant metastasis in the bone T4(M1c) Distant metastasis in other regions 32 Figure 1 . 2 Progression of human prostate cancer. PrCa proceeds through defined pathological stages including normal, neoplasia and dysplasia (PIN), invasive carcinoma and metastatic carcinoma. Loss of basal cells and basement membrane are hallmark features of progression including invasion and the potentially lethal phenotype of metastasis to the bone. Genetic changes are also critical in promoting PrCa progression and frequently include chromosomal deletions of tumour suppressors most frequently including Nkx 3.1, Retinoblastoma Protein (Rb), PTEN and p53. 34 Figure 1.3 Androgen epithelial entry and activation of A R gene targets. Androgens enter epithelia either by diffusion from the blood or chaperoning by way of a putative SFfJBG receptor. Cytosolic testosterone is then converted to dihydrotestosterone (DHT), by 5ce-reductase (type 2), binds AR, thereby, facilitating dissociation from Hsp90 and nuclear nuclear transport. A R binding to androgen response elements (AREs) promotes transcriptional activation of target genes. While, transcription is best understood to occur androgen dependency (®), it may also occur in a ligand independent manner (©). Blood Vessel Androgen dependent AR activation • • • • • • • Androgen independent AR activation ) M M m m SHBG/T M complexes (D ® AR dimerization 36 Figure 1.4 Rationale for total androgen blockade in PrCa therapy. Complex negative feedback loops regulate gonadotrophin releasing hormone (GnRH) secretion from the hypothalamus and its stimulation of the pituitary gland. GnRH secretion and adrenalcorticotrophic hormone (ACTH) promotes testosterone secretion both from the Leydig cells of the testes and adrenal gland. While, L H R H agonists stimulate the hypothalamus to create an initial surge (4-5 days) in testosterone, it is the eventual negative repression (~2 weeks) of gonadal androgen production that is therapeutically important. Non-steroidal (bicalutamide & flutamide) and steroidal anti-androgens (cyproterone acetate) act directly on A R to reduce its transcriptional activity. 37 38 Figure 1.5 Structural features of nuclear receptors (general) and the androgen receptor (AR). a) The general structure of nuclear receptors includes a variable NTD (A); a conserved D N A binding region (B); hinge region (C); ligand binding region (D); and a distal, CTD (E). Most nuclear receptors contain regions permitting either ligand independent (AF1) or ligand dependent activation (AF2). b) The A R contains polymorphic regions in its NTD including a polyglutamine (Gln)n and polyglycine tract (Rochette-Egly)n. Nuclear translocation of A R is mediated by a nuclear localization signal (NLS) found in the hinge region while D N A binding is facilitated by two Zinc fingers, c) AR transactivation is regulated by the proximity of upstream enhancers and repressors with promoter AREs. Binding of A R co-regulators and basal transcriptional machinery in conjunction with changes in chromatin architecture dictate whether A R undergoes transcriptional enhancement or repression. 39 a) A D E Y~S AF1 AF2 b) 1 141 338 (Gln)n AF-1 ZnZn (Gly)n WW Hint. AF-2 919 N-terminal Domain DNA Binding Domain Ligand Binding Domain c) 40 Figure 1.6 The Wnt degradative cascade depicted in an a) active or b) inactive conformation, a) Wnt ligands bind the receptor, Frizzled, to activate Dishevelled (Dsh), thus promoting phosphorylating and inhibition of GSK3(3, dissociation of GSK|3 from Axin but binding to FRAT. Cytosolic p-catenin accumulates and translocates to the nucleus to activate Tcf/Lef-1 responsive genes including those associated with cell cycle progression, b) Absence of Wnt stimulation results in the formation of the Axin/GSK3/APC complex, phosphorylation of P~ catenin by casein kinase 1 (CK1) and GSK3P followed by TrCp mediated proteosome degradation. Wnt target genes are, therefore, not activated. 41 a) Tcf/Lef CBP/p300 P-TrCP f V ^ GSK3B a-Cat Axin,APC, AR E-cad, Brg-1 PTEN /—*—» COOH Regulatory Domain Transactivation (Ser 33, 35,41, Thr 37) P-catenin Domain (CTD) (NTDt Armadillo Repeats h ] Repeat 7 +*l Pockets with lacking a-helix Neg. charge H 1 h e | i x J 1 / 2 \ 3 4 5 6 7 / 8 9 10 11 12 mwimmi Arm Repeat P-catenin Armadillo Repeats 42 Figure 1.7 Structural domains, binding partners, and armadillo repeats of P-catenin. a) P-catenin consists of three main structural domains including an NTD that is required for TrCP and G S K 3 P mediated phosphorylation and degradation of P-catenin; the armadillo repeats consisting of 12 imperfect repeats and a CTD transactivation domain. Each domain is associated with P-catenin binding to partners that permit diverse cellular functions, b) Each of 12 Armadillo repeats consists of 3 alpha helices creating negatively charged pockets facilitating interaction with a multitude of binding partners. a) Activated Canonic Wnt Signaling b) In-activated Canonic Wnt Signaling 44 Figure 1.8 The P B K signaling axis in PrCa cells. Loss of PTEN expression results in elevated levels of phosphatidylinositol (3,4,5) phosphate (PLP3) and increased activation of Akt. Growth factors (IGF-1, EGF, KGF, LL-6) driving P B K signaling promotes AI cell proliferation. P B K activation results in simultaneous repression of proteins with tumour suppressor activity (GSK3p\ B A D , Tuberin, FKHR) and augmenting those with the capacity to promote growth (mTOR, p70 S6K) and cell cycle progression (e.g. cyclin Dland c-myc). Growth Factors (IGF-1, EGF, KGF, IL-6) 46 Figure 1.9 Structure and function of PTEN. a) The structural components of PTEN including the phosphatase and C-terminal, PDZ domains. PTEN also contains a region that is highly homologous to the actin binding protein, tensin. b) PTEN serves to regulate 3 major pathways including (1) cell invasion via Focal Adhesion Kinase (FAK) signaling; (2) proliferation and apoptosis by regulating pools of PLP3 and the cell cycle regulators p27/Kipl and p21/CLP and (3) regulation of M A P K induced transcription. Phosphatase domain Tensin homology domain C2 domain PDZ binding domain NTD I 186 3 CTD 352 400-403 PTEN I Cytoskeleton l> Migration 4 Proliferation I Invasion t Apoptosis i p 1 3 0 C a s . 4 B A D t p27/Kip1 T p21/CIP Shc/MAPK i Transcription ink 48 Table I. Summary of works closely related to the contents of this thesis including effector molecules of the Wnt pathway, the nature of interactions with nuclear receptors, criteria for the interactions and according reference. 49 T A B L E I Effector Wnt-3a Wnt3a Wnt-11 Frizzled (FRP 5/6) Summary of Interactions between the Wnt/p-catenin/Tcf Axis and AR p-catenin P-catenin P-catenin P-catenin Tcf-4 Tcf-4 AR Interaction Activation Activation Activation/inhibition Inhibition Co-activator, Binding Co-activtor, Binding Co-activator, Binding Co-activator, Binding Co-activator Direct Binding, Complexing Activation Transcription Transcription DHT, R1881 Criteria Promotes ligand, independent AR activation (LNCaP cells) t Nuclear complexing of AR/P-catenin (LNCaP cells) | Cell proliferation, colony formation (LNCaP cells) T Cell proliferation in 22Rvl and LNCaP cells i activity in AD PrCa cells (LNCaP) No inhibition in A l PrCa cells (LNCaP-r) | Wntl I expression in AI cells (LNCaP-r) and xenografts I osteoblastic activity of PC3 cells DHT sensitive, MMTV-luc, Prob-luc PrCa (LNCaP, PC3) and bladder cancer cells (TSU) TSU cell in the presence of E2 Presence of androstenedione, DHEA 2xARE-luc, PSA-luc AR LBD and P-catenin Arm repeats \-6 sufficient for binding CV-1, LNCaP cells E-cadherin modulates P-catenin coactivation of AR AR Coactivator, binding 3xARE-luc LNCaP, PC3 cells AR-positive cells, PSA-luc, HK-2-luc Gene activation of endogenous targets Gene activation not in AR-negative cells Ligand-independent coactivation (TSU cells) Mutant P-catenin does not increase LNCaP proliferation DHT, hydroxyflutamide causes P-catenin recruitment L-39 and cyproterone acetate does not recruit P-catenin in LNCaPs Bicalutamide recruits P-catenin and activates AR in LNCaPW741C P-catenin/NTD GSK3P sites regulates in vivo PSA levels AR helices 3, 4, 5 + 12 required for P-catenin binding P-catenin L X X L L motifs not required for AR/P-catenin binding Mutant P-catenin activates AR in 22Rv-1, not LNCaP cells Weak binding in vivo + in vitro to AR-DBD region ARR3-Luc, weak binding in vivo I P-catenin/Tcf gene activation Rerference (Verras et al.,2004) (Cronauer et al., 2005) (Zhu et al., 2004) (Truica et al.„ 2000) (Yang et al., 2002) (Mulholland et al., 2002) (Chesire et al., 2002) (Masiello et al.„ 2004) (Song et al.„ 2003) (Cronauer et al.„ 2005) (Amiretal., 2003) (Mulholland et al., 2002) (Song et al., 2003) (Mulholland et al., 2003) (Pawlowski et al., 2002) (Shah et al., 2003) AR AR GSK3P GSK3P GSK3P R1881 R1881 AR AR AR Inhibits cyclin DI (colon not PrCa cells) Inhibits colon cancer cell proliferation No change in P-catenin levels in prostate cells In vitro, cotrafficking of AR and P-catenin (neuronal cell line) In vitro, cotrafficking of AR and P-catenin (LNCaP, PC3 PrCa cells) In vitro, cotrafficking of AR and P-catenin (CV-1 cells) Arm repeats 1-6 are necessary for AR/p-catenin binding and trafficking Corepression Binding/phosphorylation to AR hinge region Corepression Phosphorylates AR-NTD, interrupts AR-NTD/CTD interactions Regulator Required for androgen-stimulated gene expression (Mulholland et al., 2003) (Mulholland et al., 2003) (Chesire et al., 2002) (Pawlowski et al., 2002) (Mulholland et al., 2003) (Yang et al., 2002) (Salas et al., 2004) (Wang et al., 2004) (Liao et al.,2004) T A B L E 1. Summary of Interactions between the Wnt/p-catenin/Tcf Axis and AR (con't) Effector Interaction Criteria Rerference AR DHT, R1881 J. P-catenin Tcf gene activation Inhibits cyclin DI (colon not PrCa cells) Inhibits colon cancer cell proliferation (Song et al., 2003) (Mulholland et al., 2003) (Pawlowski et al 2002) (Shah et al., 2003) (Mulholland et al., 2003) AR AR R1881 R1881 No change in P-catenin levels in prostate cells In vitro, cotrafficking of AR and P-catenin (neuronal cell line) In vitro, cotrafficking of AR and P-catenin (LNCaP, PC3 PrCa cells) In vitro, cotrafficking of AR and P-catenin (CV-1 cells), necessary (Chesire et al., 2002) (Pawlowski et al., 2002) (Mulholland et al., 2003) (Yangetal., 2002) 51 Chapter Two: W n t / p - C a t e n i n A c t i v a t i o n o f the A n d r o g e n Recep to r 52 2.1 (Tamura, Gu et al. 1998)Introduction Wnts are secreted glycoproteins that have the capacity to bind the transmembrane receptor Frizzled and activate the (1) Canonical Wnt Pathway, (2) the Wnt Planar Pathway and/or (3) the Wnt C a 2 + dependent pathway (Moon 2005). While all three forms of Wnt signaling can induce dramatic spatial and temporal changes in cell and tissue development, Wnt canonical signaling is most associated with oncogenesis (Barker and Clevers 2000). Though approximately 20 Wnts have been characterized to date, their functional roles are poorly characterized in relation to the understanding of other Wnt/P-catenin signaling components (Moon 2005). This has been, in part, due to the lack of availability of Wnt antibodies and inability to generate functional, recombinant forms of Wnt. Recently, however, new technology of Wnt purification has increased the available tools by which to study this pathway. The ability to regulate Wnt signaling, either through activation or inhibition, is an important factor for evaluation of cross regulation with other signaling pathways. Although few techniques provide absolute certainly of strict Wnt activation, use of recombinant Wnts and overexpressing Wnt expressing L cells are preferable because they initiate the Wnt cascade from the Frizzled transmembrane receptor. Inhibition of negative regulators, including GSK3P by way of L i C l , or Axin and APC by R N A i , can also promote rapid accumulation of P-catenin. As a result of recent "Wnt advances", a rapid and acute increase in our understanding of how Wnts may be associated with cancer is occurring. A comprehensive and up to date list of reagents can be found at the website of Dr. Roel Nusse of Standford University (http://www.stanford.edu/~rnusse/wntwindow.html). In this study, we make use of a mouse fibroblast cell line with stable expression of functional active, human Wnt3a (ATCC, MBA-176) with respect to PrCa. Upon examination of PrCa cell lines, xenograft models and tissue obtained from radical prostatectomy's, there is a low rate (-5%) of activating mutations (Voeller, Truica et al. 1998) (Chesire, Ewing et al. 2000). While none of the examined PrCa cell lines (DU145, PC3, LNCaP) have revealed mutations in APC and P-catenin, Trcp (transducin repeat-containing protein), mutations have been detected in TSU cells (Voeller, Truica et al. 1998) (Chesire, Ewing et al. 2000). Despite the lack of Wnt mutations detectable in frequently studied PrCa cells, their presence in primary prostate tumours suggests an importance in during intiating tumour events (Voeller, Truica et al. 1998) (Chesire, Ewing et al. 2000). This is further underscored by the fact 53 that PrCa is a heterogeneous and multifocal disease implying that select epithelial "hotspots" may have activating mutations that therefore could be under reported. Thus, although P-catenin mutations are not found in PrCa cell lines with metastatic disease, the presence of activating Wnt mutations may be significant for initiation and/or progression of tumours. As a caveat, the frequency of P-catenin mutations in primary prostate tumours (-5%) has been reported to be more frequent than those found in some colon cancers (2%) (as opposed to frequent APC mutations) (Kitaeva, Grogan et al. 1997). Further studies facilitating techniques, such as laser capture dissection will likely be required to correlate the occurrence of activating Wnt mutations with progressive oncogenic phenotypes associated with PrCa. While the frequencies of total Wnt mutations are somewhat low in PrCa, the frequency of nuclear P-catenin is somewhat higher (-20-25%), suggesting that P-catenin may serve an important role in regulating transcription and oncogenesis (Chesire, Ewing et al. 2000). Therefore, it is possible that P-catenin can serve as an oncogene in select epithelial hotspots even when its regulators (APC, CK1, GSK3, Tcf4) do not have obvious defects, p-catenin is well associated with increases in proliferative markers such as Ki67 (Semba, Kusumi et al. 2000), a correlation which could be achieved through Tcf/Lef or other proproliferative signaling pathways that are predominant in PrCa. Importantly, the disparity of Wnt mutations between PrCa samples suggests that these genetic alterations are important only in a subset of PrCa cells. This is supported by the observation that within an individual human prostate tumour much variation in nuclear P-catenin staining can occur (Mulholland et al., 2004, Chapter 6, unpublished observations,). Nevertheless, the frequent rate of detected nuclear staining for P-catenin strongly suggests a functional role that is independent of activating mutations. PSA secreting PrCa cells are regulated by A R and its cognate ligand, DHT. Transcriptional co-activators, with the ability to activate A R responsive promoters in the presence of little or low androgens, may allow for promiscuous gene activation during androgen ablation therapy (Heinlein and Chang 2002). Co-factors may also facilitate altered responses to non-androgen ligands including estrogens, progesterone and anti-androgens (Heinlein and Chang 2002). This is well exemplified by SRC and CBP, which have been shown to increase the activity of several nuclear receptors. Thus, in addition to gene amplification and activating mutations, the ability of transcriptional co-activators to specifically increase the activity of AR, 54 in the presence of low androgens, could facilitate any pathogenic roles that A R may assume during AI PrCa. Since P-catenin is expressed at moderate levels in the nucleus in many PrCas, it could serve to modify the function of AR, thus, potentially facilitating aberrant A R activity. In this chapter, data is put forth demonstrating that P-catenin/Tcf signaling is both functional and responsive to the introduction of exogenous Wnt3a, WT (3-catenin and activated (A aa 27-47) P-catenin. While these data have recently been reported by others (Verras, Brown et al. 2004), we co-report the ability of Wnt signaling to dramatically enhance the transcriptional activity of A R both in the presence of altered ligands (estrogens) and AI. Analysis of A R endogenous targets indicates that Wnt3a and p-catenin (A aa 27-47) are both capable of increasing levels of PSA transcripts as compared to control conditioned media. We also illustrate that treatment of A R expressing PrCa cells with Wnt3a and/or lentiviral infection with P-catenin (A 27-47) can enhance cell viability and colony formation. 2.2 Methods 2.2.1 Tissue culture Generating Wnt3a Conditioned Media: Conditioned media was generated using mouse L cells stably expressing the human Wnt3a gene (ATCC, Cat # CRL-2647). Control conditioned media was made using L cells containing an empty vector (ATCC, Cat # CRL-2648). Media was collected from L cells cultured in 5% FBS/charcoal stripped serum (CSS) after 3 days, and 6 days of growth. Conditioned media was filtered using a 0.45 (im filter (Sarstedt, Cat 831826) and then applied to LNCaP or PC3 cells for 48 hours. To assay the content of Wnt3a both cell lysates and conditioned media were assayed for the presence of Wnt3a. 60 p,g of total cell lysate and 40 ul of conditioned media, concentrated using an Amicon Ultra-15 centrifugal filter device (Millipore), was separated by SDS-PAGE and probed for the presence of Wnt3a. As a further control, Wnt3a conditioned media was confirmed to be non-cross reactive with Wntl , a closely related Wnt (data not shown). Cells treated with conditioned media were harvested and assessed for activation of P-catenin/Tcf or A R signaling by Western Blotting, Northern Blotting, luciferase reporter activity, MTS cell viability assay and soft agar colony formation. 55 2.2.2 Dextran-charcoal stripped fetal bovine serum (CSS) Charcoal (1%) was coated with dextran (0.1% Dextran T70) for 16-20 hours at 4 °C. Subsequently, dextran-coated charcoal was pelleted by centrifugation (10 min x 3, 000 rpm at 4 °C) and the supernatant discarded. Fetal Bovine Serum (FBS) was mixed with dextran-coated charcoal for 3 x 30 min with centrifugation between each incubation. Stripped serum was passed through a 0.2 \im filter in a sterile environment and stored at -20 °C until use (Referece???). 2.2.3 Luciferase reporter assays Luciferase transcriptional assays were carried out using the ARR3-luc thymidine kinase driven luciferase reporter containing three flanking sequences of a non-consensus A R E (GGATCAgggAGTCTC); the PSA-luc reporter construct containing approximately 6 kb of upstream sequence (Snoek, Bruchovsky et al. 1998); the Topflash reporter construct (Upstate Biot, Cat #21-170) containing three tandem Tcf4 binding sites and its corresponding mutant (Upstate Biot, Cat #21-169). Reporter assays were performed using LNCaP which were first virally infected with pDEST4/TO/Lenti-WT (3-catenin and pDEST4/TO/Lenti-A (aa 27-45)-p-catenin constructs for 24 hours followed by 48 hours of recovery. Parental and 3-catenin infected cells were then transfected with luciferase reporter constructs for 12-16 hours followed by culturing in 5% FBS/RPM3/CSS media with or without appropriate ligand treatment. For PC3 cells, lipofectin was used for delivery of pDEST4/TO/Lenti-WT P-catenin and pDEST4/TO/Lenti-A (27-47)-P-catenin constructs (each containing a minimal C M V promoter). For luciferase transfections 2xl0 5 PC3 cells were seeded per well of a 6 well plate, while 5xl0 4 cells were plated/well of a 12 well plate. Cells were transfected for 12-16 hours using lipofectin as per manufacturer's guidelines and replaced with either full media (10% RPMS or 10% D M E M ) or androgen depleted media (CSS) with or without low (0.1 n M DHT) or high androgen (InM DHT). Firefly luciferase values were normalized to the results of Renilla Luciferase with standard deviations being generated for each treatment. For LNCaP cells Renilla Luciferase control plasmids proved to be androgen responsive and, thus, were not used to control for transfections (see Mulholland et al, The Prostate, 2004). To rule out any that any experimental differences could be due to variation in transfection efficiency, cells were also transfected with a GFP mammalian expression plasmid. Cells were inspected with at least 100 cells counted or 56 used for Western blotting for the presence of GFP. In general, transfection efficiencies within a particular cell line were highly consistant. To further add confidence to transfection efficiency, each experimental condition was carried out in replicates of 12 (i.e. 1 x 12 well plate for each condition). 2.2.4 Generation of LNCaP-WT P-catenin and LNCaP-A(27-45)|3-catenin cells Development of the pDEST4/TO/Lenti-WT P-catenin and pDEST4/TO/Lenti- A (27-45)-P-catenin was by two main steps. Both WT and A (27-45)-p-catenin forms were PCR amplified using the forward and reverse primers (GGG G A C A A G TTT G T A C A A A A A A G C A G G CTT A atg get act caa get gat ttt; G G G G A C C A C TTT GTA C A A G A A A G C T G G GTC tga ttt aca ggt cag tat caa), respectively, from approximately 50 ng of cDNA expression plasmid. The -2.3 kb PCR product was separated by gel electrophoresis and gel purified. Transfer of the attBl-P-catenin-attB2 and attBl-A (27-47)-P-catenin-attB2 products into the pDON201 vector was carried out by BP reaction (attB => attP) using BP Clonase II according to the manufacturer's directions (Invitrogen, Cat # 11789-020). 1 | i l of the BP product was transformed into Stbl 3 chemical competent bacteria (Invitrogen, Cat # C7373-03) and plated on L B agarose plate with 50 jag/ml Kan selection. At least 10 colonies were picked, D N A isolated and verified by restriction digest and sequence analysis. Resulting plasmid D N A was then used for an LR reaction (attL => attR) according to the manufacturer's protocol. 1 | i l the LR product was transformed into Stbl 3 chemical competent bacteria and plated on L B agarose plate with 50 Lig /ml Amp selection. At least 10 colonies were picked, D N A isolated and verified by restriction digest and sequence analysis. 2.2.5 Generation of lentivirus Lentiviral packaging plasmids (LP1, LP2 + VSVG) were transfected with the pDEST4/TO/Lenti-WT p-catenin and pDEST4/TO/Lenti-A (27-45)-P-catenin plasmids into 293FT cells(Invitrogen, Cat # R700-07) (Figure 2.2). Transfection media was removed and replaced with 5% F B S / D M E M and collected after 48 hours of virus production. The conditioned 57 media was filtered (0.45 p,m) and applied to parental LNCaP. LNCaP cells were found to infect at rate of -95% as determined by cell counting of a GFP-lenti positive control vector, and as a result cells could be treated as "semi-clonal" or modified populations of cells. PC3 cells, however, infected at a considerably lower rate, an observation that has been previously reported (Bastide et al., 2003). Unlike lentivirus clones generated for the PTEN transgene, cells infected with pDEST4/TO/Lenti-WT p-catenin and pDEST4/TO/Lenti-A(27-45)-|3-catenin were, in some instances, not co-infected with the pLenti6/TR and, thus, did not require doxycycline treatment for induction. Rather, the transgenes were constitutively expressed by a C M V promoter. 2.2.6 Colony forming assays Soft agar assays were carried out to evaluate the ability of Wnt3a, WT P-catenin and pDEST4/TO/Lenti-A (27-45)-p-catenin and combined treatment of Wnt3a and P-catenin to promote increased colony formation. First, a base layer of 1.5 ml of 0.4% agarose in treatment media (bottom layer) was applied after solidifying a 1.5 ml solution (middle layer) of 0.2% agarose in treatment media containing either LNCaP or PC-3 cells. LNCaP cells were lentivirally infected, as described above, with either WT P-catenin or A (27-45)-p-catenin and treated with varying amounts of androgen (DHT) over 12-14 days. Colony formation assays were also carried out using Wnt3a/CSS-conditioned media (Wnt3a CM) and CSS-control conditioned media (CM). Briefly, CSS-CM was collected and applied to 0.5 x 104 or 2.0 x 104/well with or without androgen. PC3 cells were also assessed for altered growth in the presence of either Wnt3a C M or control C M but without regard for androgens. Colonies were stained with 2 ml of 0.005% crystal violet and counting using the "Dave-Grid" technique. 10 squares per 35 mm well were counted and averaged over 6 wells of identical treatment to generate a standard deviation (Figure 2.9). 2.2.7 MTS assays LNCaP and PC3 were plated in 24 well plates in 200 pi of media at cell densities between O.lx 1x10s and 1x10s cells per well in either 5% FBS/RPMI or 5% F B S / D M E M . 58 LNCaP cells were either infected with WT P-catenin or A(27-47)-p-catenin and treated with or without Wnt3a C M for up to 7 days. Cell treatments for LNCaP cells included Wnt3a-FJ3S-CM (5% FBS/RPMI), Wnt3a-CSS-CM (dextran charcoal stripped serum) + InM R1881, Wnt3a-CSS-CM + 0.1 nM R1881 only and Wnt3a-CSS-CM only. The same androgen concentrations were also evaluated in the control-CSS-CM. Conditions for PC3 cells included either Wnt3a-F B S - C M (5% FBS/DMEM), or control-FBS-CM either 0, 2, 4 or 6 days followed by addition of MTS/PMS solution (2.5 ml of MTS solution + 125 ul of PMS/12 well plate). 40 u l of MTS/PMS solution was added into each well of the 12 well assay plates followed by incubation for 1-4 hours at 37 °C in a tissue culture incubator. To measure levels of cellular formazan, the assay plate was measured at 490 nM using a standard ELISA plate reader. 2.2.8 PSA ELISA A PSA ELISA (ClinPro International Co.) was used to assay culture media for changes in secreted PSA. 50 ul of Zero Buffer and 50 pi of culture media samples (25 pi media + 25 ul dH20) were added to each well of the ELISA plate, mixed for 30 sec and incubate at RT for 60 min. The plate was then washed 5x with dH20 followed by addition of 100 ul of Enzyme Conjugate Reagent into each well and incubated for 60 min at RT. The plate was then washed 5x with dH20 followed by addition of 100 u,l of T M B reagent for incubation at RT for 20 min. The reaction was stopped by addition of 100 u,l of Stop Solution to each well followed by mixing and optical density reading at 450 nM. Absolute PSA concentrations in culture media were determined by comparison to standards that range from 0 to 120 ng/ml. 2.3 Results 2.3.1 Relative expression of Wnt effectors and activity in PrCa cell lines. As a prerequisite to evaluating the cross regulation between the Wnt/p-catenin/Tcf axes and AR, we considered basal levels of Wnt3a, Wntl , Tcf4, P-catenin and A R in PrCa and colon cancer cell lines (Figure 2.1a). Importantly, we determined that all cell lines examined were 59 positive in expression for the Wnt transcriptional effector molecule, Tcf. PC3, PrCa cells and SW480 colon cancer cell lines contained the highest levels of Tcf, an observation that is consistent with constitutive (3-catenin/Tcf signaling detected in these cell lines. In cell lines that were positive for A R expression levels were observed, from highest to lowest, to be C4-2 > LNCaP > Wnt3a-CL > SW480, C L with PC3 and DU145 cells not expressing AR. 2.3.2 Generation of inducible lentiviral expression constructs for WT and GSK3|3-resistant P-catenin Up to 5% of lethal PrCas contain activating mutations (Ser 33, 37, 45, Thr 41) found in the NTD of P-catenin (Figure 2.2ai). Though alterations at the p-catenin regulatory region usually occur as single mutations; we chose to evaluate potential effects of P-catenin with complete inactivation; that is alternation to all GSK3P phosphorylation sites (Figure 2.2aii). Thus, by PCR amplification, a full length (WT P-catenin) and truncated version of P-catenin (A27-45) P-catenin was inserted into a lentiviral vector (Invitrogen), thereby allowing the comparison of accumulation of a degradable and non-degradable (GSK3P resistant) P-catenin. 293FT cells were transiently transfected with appropriate packaging vectors along with either the WT or A(27-45)-P-catenin transgene (Figure 2.2aiii). After infection and recovery, cells were transfected with the Topflash luciferase reporter construct and evaluated by way of Western blotting. Upon infection, LNCaP and C4-2 cells generally revealed a high level of viral uptake (>90%), while PC3 cells were less efficient during infection (-15%). Thus, while PC3-P-catenin stable expressing cells required co-selection using blasticidin (Tet-R) and zeocin (transgene), LNCaP-P-catenin cells did not. 2.3.3 The Wnt/p-catenin/Tcf axis is functional in PrCa cells To assay the effects of WT versus stabilized P-catenin on both Tcf activation and A R transactivation, cells were infected with either empty pDEST/Lenti vector (Invitrogen), WT P-catenin lenti or A27-45 P-catenin. Consistent with removal of GSK3P phosphorylation sites, augmented Topflash was observed upon lentiviral delivery of P-catenin (Figure 2.4a). 60 Considerable transactivation was observed of the minimal, A R binding (ARR3-luc) reporter upon introduction of P-catenin in to LNCaP PrCa cells that were treated with either R1881 or E2 (Figure 2.4b). Importantly, we observed that P-catenin not only dramatically enhances ARR3-luc activity in the presence of androgens but also confers ligand response to non-androgens, such as estradiol. Furthermore, lentiviral delivery of GSK3p resistant P-catenin into LNCaP cells treated with 0.1 nM R1881, resulted in ARR3-luc values that were comparable to levels observed in LNCaP cells harbouring exogenous WT P-catenin and 1 n M R1881. Thus, increased P-catenin results in both enhanced sensitivity to androgens, but also results in A R responsiveness to a non-androgen (Figure 2.4b). 2.3.4 Wnt3a recapitulates the effects of stabilized P-catenin and promotes ligand independent activation of AR Having observed that P-catenin considerably enhances A R transactivation, we chose to evaluate whether upstream activation of Wnt/P-catenin/Tcf signaling would result in similar effects. To evaluate this hypothesis, we made use of mouse L cells with stable overexpression of human Wnt3a and control L cells. Figure 2.3 clearly demonstrates that Wnt3-CL cells produce significant amounts of Wnt3a as detected by immunocytochemistry and by western blotting (Figure 2.3a, b). Importantly, secreted Wnt3a was also detected in the condition media of Wnt3a L cells but not in control cells. Thus, Wnt3a-CL conditioned media serves as an additional means by which to evaluate the effects of Wnt signaling upon A R and, therefore, in PrCa function (Figure 2.3c). Using increasing concentrations of DHT, we considered the effects both of A27-45 P-catenin and Wnt3a on A R transactivation (Figure 2.4). Importantly, culturing of LNCaP cells in Wnt3a/CSS media resulted in significant sensitization of A R to the activating effects of DHT. Enhanced sensitization was greatest at lower concentrations of DHT including 0.01 and 0.05 n M DHT (Figure 2.5b). Given the capacity of Wnt3a to significantly augment A R transactivation at low concentrations of DHT, we subsequently considered whether Wnt3a could promote ligand independent activation. We considered the effects of Wnt3a either alone or in combination with WT P-catenin or A27-45 P-catenin. Significantly, the combined effects of stabilized P-catenin with Wnt3a/CSS conditioned media resulted in a synergistic response (2.5x 61 that of Wnt3a alone) of the ARR3-luc reporter (Figure 2.5c,d). In either the absence or presence of androgens, this trend was not observed with the control conditioned media (CL). 2.3.5 Wnt3a enhances production of PSA An important corollary to determining whether Wnt3a can activate A R function is the assessment of a specific, endogenous target of AR. To do this we considered levels of PSA that was secreted into conditioned media, PSA protein and transcript. Consistent with ARR3-luc reporter data, we observed significant enhancement of PSA in the absence and at low levels of androgen, as seen in two independent Western blots (40 |ig and 20 pg total protein) (Figure 2.6a). Using a PSA specific ELISA to detect altered levels of PSA in tissue culture media of LNCaP cells, we observed that Wnt3a also enhanced secreted PSA in androgen ablative conditions and at low (0.1 nM DHT) androgen conditions (Figure 2.6b). Analysis of PSA transcripts revealed only a minor trend, relative to protein and ELISA data, in which treatment of LNCaP cells in androgen ablative conditions demonstrated moderately enhanced (2.5 fold) PSA transcript levels (relative to G3PDH loading controls) in comparison to C L treated LNCaP cells. Curiously, little difference in PSA message was observed with Wnt3a and control treated LNCaP cells when treated with 0.1 nM DHT (Figure 2.6c). 2.3.6 Wnt3a enhances ligand sensitive AR/p-catenin interactions To ascertain whether Wnt3a enhanced A R function by way of the Wnt/P-catenin signaling axis, we considered the known status of AR/P-catenin complexing. Using LNCaP cells treated with Wnt3a and control conditioned media, we immunoprecipitated A R and probed for the presence of P-catenin (Figure 2.7a). In the presence of Wnt3a, approximately 3x as much P-catenin was detected in A R precipitates in the absence of exogenous androgen. Upon treatment with exogenous androgen and increasing concentrations of Wnt3a/CSS (50% v/v and 100% v/v) we observed considerable enhancement of AR/P-catenin complexing (Figure 2.7b). To further validate that Wnt3a augments the known ligand sensitive interactions between A R and 3-catenin the A R antagonist, Casodex, was added to LNCaP cells while culturing in 62 Wnt3a/CSS condition media. Addition of 10 u,M Casodex reduced Wnt3a stimulated complexing between A R and P-catenin significantly (Figure 2.7c). These data recapitulate previously published observations and those presented here, (Chapter 4) that endogenous or stimulated interactions (as stimulated by Wnt3a or activated P-catenin expression) between A R and p-catenin can be antagonized by anti-androgen. 2.3.7 Wnt3a promotes ligand-independent proliferation of LNCaP cells and provides resistance to LY294002 treatment in PC3 cells Having observed a potent response of A R transactivation, Tcf activation and AR/P-catenin complexing, we evaluated the capacity of Wnt3a to promote cell viability of LNCaP cells cultured in CSS and in CSS with androgen. Importantly, after 6 days culture LNCaP cells, treated with Wnt3a, demonstrated a 5.5 increase in cell viability (compared to androgen ablative conditions) and a 1.5 fold increase, upon treatment with Wnt3a and androgen, in comparison to androgen alone (Figure 2.8a). While the PC3 PrCa line is inherently AI, we evaluated the capacity for Wnt3a to protect from treatment with low (10 uJVI) and high (40 pM) concentrations of LY294002, a chemical inhibitor of PI3K/Akt function. While treatment with Wnt3a/CSS (100% v/v) did not protect against 40 uJVI amounts of LY294002, it did provide significant protection (6-fold) against co-treatment with lower concentrations of LY294002 (Figure 2.8c). 2.3.8 Wnt3a enhances LNCaP cell colony formation To consider the potential significance of Wnt3a in PrCa tumourigenesis, we evaluated the capacity for Wnt3a to augment soft agar colony formation. While moderate in increase, Wnt3a did promote approximately a 1.5 increase in colony formation as compared with controls, as indexed by the "Dave Grid" counting technique (Figure 2.9a). Wnt3a also appeared to potentiate the ability of low androgen concentrations (0.1 nM) to promote tumourigenesis (Figure 2.9b). 63 Figure 2.1 Comparative expression of Wnt effector molecules including Wnt3a, Wntl , Tcf4, (3-catenin, and A R in PSA secreting PrCa cells (LNCaP, C4-2), non-PSA expressing PrCa cells (PC3, DU145) and non-PrCa cells (Wnt3a-L, C L , SW480 ft *> " s * V O 3 £ ~* £ W n t 3 a [ W n t l T c f 4 P - C a t | « » » - ^ " AR Actin 65 Figure 2.2 Generation of an activated (3-catenin lentiviral expression system, (ai) Up to 5% of lethal PrCas contain activating mutations (Ser 33, 37, 45, Thr 41) found in the NTD regulatory region of P-catenin. (aii) The NTD regulatory region is present in WT p-catenin but deleted from the GSK3P degradative resistant, A(27-45)-P-catenin mutant. By PCR amplification a full length (WT P-catenin) and truncated version of mutant p-catenin (A27-45 P-catenin) was inserted into a lentiviral vector (Invitrogen), thus allowing the comparison of accumulation of a degradable and non-degradable (GSK3P resistant) P-catenin. (aiii) Lentiviral based WT P-catenin and A(27-45)-p-catenin vectors were generated by way of Invitrogen technology. 293FT cells were transiently transfected with appropriate packaging vectors along with either the WT or A(27-45)-P-catenin transgene. After infection and recovery, cells were evaluated by way of Western blotting for appropriate induction of P-catenin. 66 Thr41 Mutant p-Catenin cDNA 2343 bp aii) P-catenin Regulatory Domain aiii) Lentiviral {Invitrogen) Production of WT fj-cat & A ( 2 7 ^ * 5 ) p-cat Stable PrCa (LNCaP) Cell Lines 293FT Transfection pDEST4/TC7V5 WT p-cat pDEST4/TO/V5 A(27-45) p-cat 0 0 0 HIV Rev Gag-Pol VSVG Viral Infection of PrCa cells Test p-cat infected cells by WB Dox: + - + WB: p-cat I" ~ " 67 Figure 2.3 Use of Wnt3a overexpressing L fibroblast cells to generate Wnt3a conditioned media, a) Wnt3a-L cells contain overexpressing mouse Wnt3a transgene resulting in dramatic accumulation of Wnt3a both in the cellular extracts and in tissue culture media, as detected by (a) immunocytochemistry of fixed Wnt3a-L cells and control L cells and (b) western blotting of Wnt3a and control L cell culture media. Importantly, no Wnt3a is detectable in the control-L cell line by either assay, b) Wnt3a conditioned media was applied to PrCa cells that are evaluated for changes in A R transactivation, growth and colony formation. 68 Wnt3a/Control Mouse L Cells ICC: Wnt3a C) Wnt3a CSS Control CSS WB: Wnt3a Conditioned media collected at day 3 & 7 and applied to LNCaP and P C 3 cells 4 Evaluate transcription, growth, colony formation in PrCa cells 69 Figure 2.4 GSK3{3 resistant P-catenin promotes enhanced A R transactivation in the presence of estrogen (E2) and low concentrations of androgen, a) P-catenin/Tcf signaling is functional in LNCaP and PC3 PrCa cells as demonstrated by viral infection with degradable (WT p-catenin) and non-degradable A(27-45)-P-catenin. The GSK3P resistant p-catenin mutant showed significantly higher levels of absolute Topflash reporter activity than WT p-catenin. Infection usg a lenti-GFP pasmid (field of view) and western blotting for GFP indicate that LNCaP cells are transfected a a very rate (-95%) while PC3 cells are less efficiently infected (-15-25%). bi) Both WT and A(27-45)-p-catenin can enhance transactivation of the ARR3-luc reporter in the presence of low concentrations of androgen (0.1 nM) and in the presence of 10 nM Estradiol (E2), with A(27-45)-P-catenin having capacity to promote minor ligand independent activation of the ARR3-luc reporter, bii) Importantly, treatment of LNCaP cells with 10 n M E2, in the presence of the empty pDEST4/TO/V5 control vector, produced activation of the ARR3-luc indicating that LNCaP cells are responsive to E2 without exogenous P-catenin. c) Expression of pDEST4/TO/V5-A(27-45)-p-catenin, pDEST4/TO/V5 WT P-catenin and the empty pDEST4/TO/V5 control vector was assayed in LNCaP cells by co-transfection with the TR (tet repressor), followed by culture with or without 2 pg /ml doxycycline. o w M * 71 ARR3-LUC (ALU) ro ^ Vi CO o CD S (Q | j? > a i TOPFLASH (ALU) 3 8 8 6 8 8 3 8 8 •o j i ) _ 71 Figure 2.5 Wnt3a promotes enhanced A R transactivation in the presence of low androgen concentrations, a) WT, A(27-45)-(3-catenin and mock lenti plasmids were introduced to LNCaP PrCa cells by way of lentiviral infection followed by overnight transfection with the ARR3-luc reporter plasmid. After full serum recovery culture media was replaced with CSS/phenol red free media containing increasing concentrations (0, 0.01, 0.05, 0.1, 1, and 10) of DHT. After 48 hours of culture treatment cells were assayed for luciferase reporter activity, (b) Wnt3a/CSS conditioned media promotes A R transactivation by up to 11-fold (1 n M DHT) compared to CT/CSS media. LNCaP cells were transfected with the ARR3-luc reporter plasmid followed by treatment with either Wnt3a/CSS or CT/CSS conditioned media and increasing concentrations (0, 0.01, 0.05, 0.1, 1, and 10) of DHT. (c, d) LNCaP cells infected with either WT p-catenin or A(27-45)-P-catenin and cultured with Wnt3a media demonstrate synergy in ARR3-luc activation both in the absence and in the presence of androgen (0.1 n M DHT). This synergistic effect is not observed when using the control C M . Lenti-GFP was used to confirm equal infection rates between the WT P-catenin and A(27-45)-P-catenin lenti viral constructs. to 73 Figure 2.6 a) Wnt3a/CSS media enhances PSA protein production in the absence or presence of low levels of androgens (0.1 n M DHT) as assayed by a) Western blotting and b) PSA ELISA and c) Norther analysis. Northern blots for PSA, A R and P-catenin transcripts after 48 hours of either Wnt3a/CSS or CT/CSS phenol-red free conditioned media indicates that Wnt3a augments PSA mRNA in a manner similar to protein expression. Curiously, A R and p-catenin mRNA levels were observed to decrease in the presence of Wnt3a/CSS + 0.1 n M DHT. a) Wnt3a CONT DHT (nM): WB: PSA — • WB: P S A — • 0 0.1 0 0.1 40 ug total protein 20 ug total protein Cytosolic b) 140 120 (A —. — E 80 W -= S c en Q. 40 20 0 0 N \ \ ° 0 0 N \ \ 0 DHT Wnt3 Media CONT Media c) Wnt3a CONT NB: PSA DHT NB: AR DHT NB: p -Cat DHT NB: G3PDH DHT 0 0.1 0 0.1 Wnt3a CONT DHT {nM) 0 0.1 0 0.1 0 0.1 0 0.1 Wnt3a CONT DHT (nM) 0 0.1 0.1 Wnt3a 0.1 CONT 0.1 0.1 DHT (nM) 0 0.1 DHT (nM) 0 0.1 75 Figure 2.7 (a, b) Wnt3a, dose-dependently enhances complexing between A R and (3-catenin but is c) antagonized by the presence of Casodex. 76 IP: AR IP: IgG input b) WB: p-Cat IP: AR 10% Input c) rv* WB: p-Cat IP: AR 10% Input 77 Figure 2.8 a) Wnt3/CSS C M promotes proliferation of LNCaP cells both with low concentrations of androgens (0.1 nM R1881) and androgen independently, b) Wnt3a/CSS C M can increase cell proliferation in AR-null PC3 PrCa cells but also provides resistance against low levels (10 uM) of the P B K inhibitor, LY294002, over 5 days, c) Lentiviral delivery of WT (3-catenin was poorly effective in enhancing cell proliferation in LNCaP (3-catenin while A(Ser33, 37, 45, Thr41)-[3-catenin promotes a moderate increase in cell proliferation at low androgen concentrations over 6 days. 78 LNCaP Cell Viability Wnt3a/CSS + 0.1 nM R1881 CT/CSS + 0.1 nM R1881 Wnt3a/CSS CT/CSS 6 days PC3 Cell Viability Wnt3a/CSS CT/CSS Wnt3a/CSS + 40 uM LY CT/CSS + 10 uMLY Wnt3a/CSS + 10 uM LY CT/CSS + 40 uM LY 1.5 5 days 3.5 3 cn " * 2 o 1 0.5 0 LNCaP Cell Viability P l ^ — * s 1 A(27-45) B-cat + 0.1 nM R1881 WT p-cat + 0.1 nM R1881 A(27-4S) B<at + CSS WT p-cat + CSS 0 2 4 6 days 79 Figure 2.9 a) The "Dave-Grid Technique" for counting soft agar colony formation. Each well consists of a 35 squares of which 10 from each well are used to calculate colony formation (colonies counted >0.2 mm2) size. The averages between wells are used to generate a standard deviation, b) Wnt3a/CSS conditioned media can promote significant colony formation in LNCaP PrCa cells with and without the presence of androgens. To validate statistical significance, a paired t-test (p > 0.05) used comparing (i) Wnt3a/CSS to CT/CSS treatments and Wnt3a/CSS + 0.1 R1881 to CT/CSS + 0.1 R1881 treatment. 80 81 2.4 Discussion While the continued production of PSA strongly suggests that A R remains functional during AI PrCa, the events that allow AR, and potential target genes, to remain active during androgen ablation therapy are elusive. Though the study of transcriptional coactivators has been intense through the previous 5-10 years, relatively little evidence exists supporting in vivo functional significance. This may be, in part, attributable to compensatory mechanisms by transcription factors that may prevent manifestation of an obvious functional change or phenotype, as studied in knock-out studies. Despite this, in vitro data has provided an abundance of information as to how bona fide transcriptional co-regulators interact with nuclear receptors to either enhance or diminish their activity (See Chapter 7 Reviews). For example, conserved motifs found in co-regulators contain an L X X L L motif (L= leu, X = any amino acid) facilitating binding to regions of receptor activation (AF1 and AF2 in AR). Similar to other well known co-activators, such as SRC, P-catenin has five L X X L L motifs that may be functionally involved in binding and activation of A R (Heinlein and Chang 2002). We have demonstrated that extracellular (Wnt3a) and intracellular (P-catenin) components of the Wnt/P-catenin/Tcf signaling cascade can activate AR-dependent transcription in the presence of androgens. We also draw clinical relevancy by showing that Wnt signaling has the capacity to facilitate a more promiscuous A R by altering specificity to ligands (agonists + antagonists). We further show that Wnt3a and activated forms of P-catenin can enhance cell proliferation of A R and non-AR expressing PrCa cells, possibly by alternative signaling mechanisms. One of the key regulators for Wnt signaling is the relative expression of Tcf4. While reports have indicated that in benign tissue Tcf4 expression is limited to gastrointestinal and mammary tissue (Barker, Huls et al. 1999), data presented here, and by others, clearly indicate that PrCa specimens contain low to moderate levels of Tcf4 (Chesire, Ewing et al. 2002; Chesire and Isaacs 2002) (Mulholland, Read et al. 2003). In many cell types, expression patterns of Tcf4 are directly related to oncogenic p-catenin/Tcf signaling. Thus, it is possible that increased expression of Tcf4 occurs upon epithelium transformation. Moreover, the ability of 82 PrCa cells to respond to Wnt/p-catenin/Tcf stimuli could be, important when considering factors involved in acquisition of an AI phenotype. It is well known that many cell lines of colon carcinoma origin have highly active Tcf/Lef activity as assayed, in most cases, using the Topflash/Fopflash luciferase reporter pair. In general, the presence of a chronic Wnt activating event in such systems is sufficient to promote a functional and/or constitutive Tcf/Lef signaling. For example, SW480 colon cancer cells contain a APC-/- defect (Morin, Sparks et al. 1997), thus displaying a high level of nuclear P-catenin and very high Topflash reporter counts. However, despite a large amount of Wnt3a expression, stable Wnt3a-L cells fail to show the same degree of Tcf/Lef activation. Thus, it is clear that in specific instances other factors may influence the capacity for effective Wnt/P-catenin activation. Furthermore, while nuclear p-catenin is frequently associated with increased Wnt/P-catenin activity, it is not a requirement for highly active Tcf/Lef transcription (Morin, Sparks et al. 1997). Data presented here indicate that PC3 and LNCaP PrCa cells contain a functional Wnt/P-catenin/Tcf signaling pathway. Evidence to support this includes detectable Tcf4, increased Topflash activity upon Wnt or P-catenin activation and enhanced growth cell proliferation and colony formation upon Wnt or P-catenin stimulation. Functional Wnt signalling in A R expressing PrCa cell lines has also been reported elsewhere (Chesire and Isaacs 2002) (Persad, Troussard et al. 2001). However, unlike colon cancer cells and many frequently used breast cancer cells (e.g. M D A , MCF7), the PrCa cells lines studied here demonstrate a moderate to low level of Tcf/Lef activity. Interestingly, though SW480 cells contain a similar abundance of Tcf4 protein as PC3 PrCa cells, the former line displays Topflash reporter values that are known to be hyperelevated (Bienz and Clevers 2000). It is of interest, and a caveat, that the A R signaling axis has been reported to inhibit Wnt signaling (Chesire and Isaacs 2002; Mulholland, Read et al. 2003), suggesting the possibility of selective pressure against high Tcf4 expression. It is important to appreciate, however, that nuclear P-catenin is not necessarily a prerequisite for P-catenin/Tcf activity; as some colon cancer cells (HCT 116) show little nuclear P-catenin staining but still have highly functional P-catenin/Tcf activity (Polakis 2000). Given that the PrCa cell lines studied here all have loss of PTEN expression, it is possible that such cells have increased availability of active (non-phoshorylated) P-catenin. Specifically, hyperelevated P B K / A k t signalling in PrCa cells combined with any potential for Wnt growth 83 factor activation of the Wnt/(3-catenin cascade may yield repression of multiple pools of GSK3(3 mediated degradation of P-catenin. Intuitively, this scenario suggests that non-Tcf bound P-catenin could be available to associate with transcription factors, including AR, among many others. Using lentiviral delivery of both WT-p-catenin and A27-47 P-catenin, we were able to infect LNCaP PrCa cells at a high rate (-95%) allowing more accurate evaluation of how growth and cell proliferation are effected by Wnt/p-catenin/Tcf activation. Infection in PC3 cells is for, uncertain reasons, considerably lower (-15%), as has been reported elsewhere (Bastide et al.„ 2003). . Thus, measurements of changes in cell proliferation are likely underreported As a result, caution should be used in comparing the efficacies of conditioned media and viral delivery of P-catenin in these cells. However, in both LNCaP and PC3 cells, it was clear that over expression of WT and A27-47 P-catenin were capable of enhancing Topflash reporter activity. Thus, Tcf/Lef activity was clearly demonstrated to be functional in LNCaP and PC3 PrCa cells. Importantly, LNCaP-P-catenin stable cells illustrated potent transactivation of the ARR3-luc reporter both in the presence of low androgens (0.1 n M R1881) and in the presence of 17P-Estradiol (E2). These results are significant for two reasons: First, mechanisms promoting A R transactivation at low androgen concentrations may facilitate A R regulated gene expression and mitotic activity during androgen ablation therapy. Second, few other transcriptional co-activators have been shown to activate A R in the presence of 17P-Estradiol. To date, co-regulators with the ability to transactivate A R in the presence of E2 include ARA55, ARA70 and P-catenin (Heinlein and Chang 2002). These data are clinically significant in light of the fact that while hormone ablation therapy significantly reduces intraepithelial androgen levels, circulating adrenal androgens and estrogens in conjunction with co-activators may promote A R functional promiscuity. While activation of Wnt/P-catenin/Tcf signaling is most frequently associated with P-catenin stabilization, the advent of better reagents has recently allowed assessment of activation initiated by Wnt ligands. Thus, using Wnt3a overexpressing L cells and parental control L cells, we evaluated changes in A R transcription, PrCa cell growth and tumourigenesis. Using LNCaP cells transfected with the ARR3-luc reporter and treated with Wnt3a/CSS conditioned media we ascertained that Wnt3a could also efficiently promote reporter activation as compared to control 84 conditioned media in the presence of increasing concentrations of DHT. In this titration, it is also evident that Wnt3a can promote activation in the absence of androgens. This is a highly significant observation as there are few bona fide co-regulators with the capacity for ligand independent activation of AR. These data have been recently reported by Verras et al., citing very similar and corroborative observations to those reported here (Verras, Brown et al. 2004). Given the significance of ligand independent induction by Wnt3a/CSS on A R reporter activation, we assayed the effect of Wnt3a/CSS both on intracellular (protein and mRNA) and secreted forms of PSA. The ability of Wnt3a to activate PSA in the absence of androgen is intriguing, but certainly not unprecedented. For example, protein kinase A (PKA) has been shown to increase intracellular cAMP levels thereby allowing ligand independent binding of A R to response elements (Navarro et al., 2002). Growth factors including IGF-1, K G F , and EGF can activate receptor tyrosine kinase pathways, thereby stimulating A R in ligand independent and low androgen environments (Culig, Hobisch et al. 1994). Her2/Neu is overexpressed in certain cases of PrCa and can also promote ligand independent activation of A R (Mellinghoff, Vivanco et al. 2004). IL-6 has been shown to be detected in higher levels in patients with advanced PrCa and can promote growth and increased A R gene expression in the absence of androgens (Michalaki, Syrigos et al. 2004). These growth factors all have potent mitogenic effects that may result from activation of PI3K, M A P K and/or JAK/STAT signaling pathways. The means by which Wnt3a could enhance PrCa growth and colony formation in soft agar conditions likely include increased co-activator effects between P-catenin and AR. This is a distinct possibility, as co-immunoprecipitation data, while only semi-quantitative, clearly indicate increased association between A R and P-catenin upon Wnt3a treatment. Thus, it is feasible while Wnt3a can augment Tcf/Lef activation, as assayed by Topflash activation, A R target genes may respond more readily. This effect could be attributed to the relatively low levels of Tcf4, the transcriptional mediator of Wnt activation, but high levels of AR. In other cancers, most notably colorectal, elevated activity of canonical Wnt signaling acts via increases in Tcf/Lef signaling . However, our data and those of others (Truica, Byers et al. 2000; Masiello, Cheng et al. 2002; Yang, L i et al. 2002; Masiello, Chen et al. 2004; Cronauer, Schulz et al. 2005), argue that the predominate transcriptional role of P-catenin, in PrCa cells, is of a transcriptional co-activator of AR. Given that both Tcf4 and A R can bind directly to P-catenin, perhaps the best explanation for this observation is the presence of relatively high 85 amounts of A R compared to that of Tcf4. Thus, an apparent balancing act may exist between both factors with A R predominating in PSA secreting PrCa cells such as LNCaP and C4-2 cells, and Tcf/Lef signaling in A R null PrCa cells such as PC3 and DU145 cells. The mechanism by which Wnt3a can enhance cell proliferation in A R expressing PrCa cells is also intriguing and equally perplexing given that cells express low levels of Tcf. The ability of Wnt3a, but not P-catenin, to ligand independently activate PSA protein expression suggests that Wnt3a may promote activation of other A R responsive signaling pathways. This possibility was recently entertained by Ohigashi et al., 2005 who showed that overexpression of WTF-1 (Wnt Inhibitory Factor) can decrease P-GSK3P (Ser9) levels and also decrease P-Akt (Ser473) levels, a positive marker for P B K activation (Ohigashi, Mizuno et al. 2005). Further, WIF-1 appears not to induce any increase in chemosensitization in PTEN +/+ positive DU145 cells, which maintain a much lower basal level of P-Akt (Ser473). WLF-1 is also known to be frequently downregulated in PrCa (Ohigashi, Mizuno et al. 2005). Thus, Wnt activators may have the capacity for cross-talk with other major prosurvival signaling in PrCa cells, an interpretation that would accommodate the lack of significant ligand independent induction of A R upon WT P-catenin or A27-47 P-catenin overexpression. Given the potential cross-talk with P B K / A k t signaling it would be interesting to evaluate the ability of Wnt3a ligands to activate effectors of growth such as mTOR and/or S6 kinase. Wnts may also facilitate increased PrCa cell proliferation by way of inhibiting mediators of apoptosis, an event that has been previous documented (Chen, Guttridge et al. 2001; You, Saims et al. 2002). Not all Wnts have potential for canonical Wnt activation and, therefore, would not be expected to activate A R (van Es, Barker et al. 2003). For example, while Wntl and Wnt3a have been implicated in promoting human cancer, there is considerably less evidence that non-canonical Wnts can activate Tcf/Lef mediated oncogenesis. Interestingly, it has been recently demonstrated that W n t l l can inhibit A R activation in PrCa cells (Zhu, Mazor et al. 2004). However, many Wnts have been shown to have potent morphogenic effects as studied in Xenopus Laevis and C. Elegans. While it is apparent that A R expressing PrCa cells may be highly responsive to Wnt3a, most PrCa cells express very little Wnt3a. Therefore, it is possible that PrCa cells show increased expression of Frizzled receptors 1 or 2, as compared to benign cells, thus allowing for increased sensitivity to Wnt3a. A recently published comprehensive analysis of Wnt transcript expression in PrCa cells indicates little or no expression of Wntl or 86 Wnt3a in PrCa cells, two Wnts that are thought to be very similar in function (Zhu et al., 2004). Given this, PrCa cells appear to be low expressors of many Wnts but highly responsive to Wnt3a exposure. Identification of Frizzled expression patterns in PrCa and alternative routes for intracellular signaling activation by Wnts is an area requiring future attention. Chapter Three: T h e A n d r o g e n Recep to r M e d i a t e s N u c l e a r T r a n s l o c a t i o n o f P-catenin 88 3.1 Introduction There is strong documentation of steroid receptor shuttling upon exposure to cognate ligand. Such studies have pertained to A R (Georget, Lobaccaro et al. 1997; Tyagi, Lavrovsky et al. 2000; Tomura, Goto et al. 2001), the glucocorticoid receptor (GR) (Hache, Tse et al. 1999), the estrogen receptor (ER) (Stenoien, Mancini et al. 2000), the mineralocorticoid receptor (Ohta, Elnemr et al.) (Fejes-Toth, Pearce et al. 1998) and the thyroid receptor (TR) (Zhu, Hanover et al. 1998). These receptors show a certain degree of trafficking either to or from the nucleus, but also in a subnuclear fashion. Those that show a strong migration to the nucleus upon exposure to ligand are termed "translocating receptors" and can be contrasted with receptors that are constitutively nuclear (Hager, Lim et al. 2000). The ER shows expression that is mainly nuclear in the absence of ligand (Htun, Holth et al. 1999), whereas the A R and GR shows a distribution that is both cytoplasmic and nuclear (Ogawa, Inouye et al. 1995; Htun, Holth et al. 1999). Curiously, there are varying reports as to relative abundance of A R in the cytoplasm (Simental, Sar et al. 1991) and in the nucleus in many cell types (Jenster, van der Korput et al. 1991) Upon receptor stimulation by DHT, or potent analogues of DHT, including R1881, A R will dissociate from heat-shock proteins, translocate to the nucleus, and form transcriptionally active DNA-protein complexes (Hager, Lim et al. 2000). In general, this two-step model for steroid hormone receptor action can be applied to A R and GR whereby unliganded receptor is localized in the cytoplasm and upon ligand binding undergoes conformational change that permits translocation to the nucleus. This homodimerization leads to initiation of target gene regulation (Tyagi, Lavrovsky et al. 2000). Within the nucleus most members of the receptor superfamily form focal accumulations in the presence of ligand (Hager, L im et al. 2000). Proteins that are carried or shuttled with steroid receptors into the nucleus have not been thoroughly explored. Therefore, with the ability of the A R to translocate to the nucleus, we hypothesize that A R could co-traffic other molecules that are known to bind, though examples of this phenomenon are few in number. Functionally, cytoplasmic to nuclear and subnuclear trafficking could allow for formation of multiprotein-DNA complexes and enhanced AR transcriptional activation (Tyagi, Lavrovsky et al. 2000). 89 Given the documented ligand-dependent association of A R and P-catenin (Truica, Byers et al. 2000), we hypothesize that P-catenin could be part of a complex that translocates to the nucleus as a pre-requisite to forming transcriptionally active, nuclear complexes. When P-catenin accumulates, it is phosphorylated by GSK3P and targeted for ubiquitination by the proteasome complex (Hart, de los Santos et al. 1998). Stimulation of the Wnt/Wingless pathway inhibits GSK3p and allows for accumulation of hypo-phosphorylated P-catenin in the cytoplasm. Stabilized P-catenin can then translocate to the nucleus, with Lef, and interact with the Lef/Tcf family to stimulate gene expression of cell cycle-associated proteins. Targeted genes, that are thought to be regulated by a P-catenin-Tcf complex, include c-myc, tcf-1, and cyclin DI (Hecht and Kemler 2000). An intriguing feature of Wnt signaling is the manner in which P-catenin is actively translocated to the nucleus. Although P-catenin does not have a nuclear localization signal (Henderson 2000; Henderson and Fagotto 2002), APC (adenomatous polyposis coli) can act as a nuclear-cytoplasmic shuttling protein (Henderson 2000). Such studies have shown that alteration of the amino nuclear export sequence on APC could accumulate nuclear P-catenin and have concluded that APC can shuttle between the nucleus and cytoplasm while directing P-catenin to functionally important locations. APC also contains two nuclear localization signals that are necessary for optimal nuclear APC activity (Zhang, Otevrel et al. 2001) and likely for its tumour suppressor function. Recently, studies have shown that nuclear export of P-catenin can occur independent of the CRM1 export protein. This suggests that there could be alternative pathways associated with P-catenin transport (Eleftheriou, Yoshida et al. 2001). Previous studies have also shown that P-catenin can localize to the nucleus independent of the shuttling protein Ran (Yokoya, Imamoto et al. 1999). Beyond these mechanisms, little is understood about the means by which P-catenin is shuttled to the nucleus. In this chapter we: 1) used confocal microscopy and a novel D N A binding assay to provide evidence that P-catenin binds in a ligand-dependent manner, via AR, to an androgen-regulated promoter; 2) defined the domains of A R and P-catenin as they related to in vitro binding interactions; 3) demonstrated that A R can translocate P-catenin to the nucleus in an A R ligand dependent fashion without association with APC; 4) identified the structural components 90 of A R and 3-catenin that are necessary and sufficient for co-translocation; 5) demonstrated the potential for androgen dependent translocation of P-catenin, in vivo. 3.2 Methods 3.2.1 Cell culture PC3 and HeLa F L A G - A R cells were cultured in Dulbecco's modified Eagle's medium containing 5% FBS and LNCaP cells in RPMI media containing 5% FBS. Cells were transiently transfected in serum-free media for 6-20 hours with the following deletions mutant, expression cDNAs: WT b-catenin HA-tagged, WT P-catenin myc-tagged, P-catenin-Nt, P-catenin-Arm repeats HA-tagged, P-catenin- Nt/Ct HA-tagged, rat AR-WT, AR-Nt , AR-Nt /LBD, AR-DBD, A R - D B D / L B D , GR-WT, TR-WT and RAR-WT where Nt = amino terminal coding cDNA, Ct = carboxy terminal coding cDNA, WT = full length coding cDNA. Cells were treated with 10 nM steroid receptor ligand in 5% dextran/charcoal-stripped serum. Stable A R FLAG-tagged HeLa cells were kindly donated by Dr. Michael Carey (UCLA School of Medicine). 3.2.2 Dextran-charcoal stripped fetal bovine serum (CSS) Charcoal stripped serum was produced as described in Chapter 2 Methods. 3.3.3 Immunocytochemistry LNCaP cells were grown on glass coverslips, transfected with 3 u,g of plasmid per well and treated for 48 h with or without ligand. Subsequently, cells were fixed with cold methanol for 3 min, air dried for 10 min and reconstituted in blocking buffer (0.1% Tween PBS/BSA/4% NGS) for 20 min. Following blocking, cells were treated with primary and fluorescent secondary antibodies, each for 1-2 hours at room temperature with 3x5 min washes (0.1% 91 Tween PBS/BSA/1% NGS) after each incubation. Antibodies used for staining included phosphoserine R N A Pol II (Babco). 3.3.4 Western blotting Subconfluent cell cultures to be used for Western blotting were washed with PBS and lysed with RIPA lysis buffer (0.1% SDS, 150mm NaCl, 50mm Tris, pH 8.0) on ice for 30 min with the addition of a protease inhibitor cocktail (Roche Scientific). Samples were standardized for total protein and separated by SDS-PAGE. Alternatively, cells used for immunoprecipitation (IP) were lysed with NP-40 lysis buffer (1% NP-40, 150mm NaCl, 50mm Tris, pH 8.0), standardized for total protein content and immunoprecipitated with antibodies (16 h at 4°C) and protein A / G beads (Santa Cruz). Following SDS-PAGE and blotting to PVDF (Amersham Life Sciences), membranes were blocked with 5% nonfat milk in TPBS (0.05% Tween-20 in PBS) for 1 hour and then incubated for 1-2 hours at room temperature with primary antibodies. Antibodies included p-catenin (Transduction Lab., Santa Cruz Biotechnology), A R (ABR, Pharmingen, Santa Cruz Biotechnology), Tcf4 (Santa Cruz, Upstate Biotechnology), FUS-tag (New England BioLabs), myc-tag (Santa Cruz), GFP (Sigma), actin (Sigma) or cyclin DI (Santa Cruz Biotechnology). Detection was carried out with either anti-mouse-FLRP (Santa Cruz) or anti-rabbit-HRP (Santa Cruz) and E C L Western blotting detection reagents (Amersham Life Science). 3.3.5 Co-immunoprecipitations Cells were grown to approximately 70% in 5-10% FBS/RPMI and then replaced with CSS with or without ligand (10 n M R1881 treatment for 24-48 hours. After treatment cells were collected with 5 m M EDTA/PBS and lysed in co-IP buffer with (150 m M NaCl, 50 m M Tris-Cl pH 7.5, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS) or without SDS (150 m M NaCl, 50 m M Tris-Cl pH 7.5, 1% NP-40, 0.5% sodium deoxycholate). The cell lysate was centrifuged for 10 min at 14,000 rpm with lmg of total cell lysate being used per IP input. 1-2 "Xg of treatment antibody was used for (co)-IPs with an equal mass of control non-immune IgG. 92 Immunocomplexes were precipitated using 50 ui Protein A / G plus agarose beads (Santa Cruz Biotechnology). 3.3.6 Cell fractionation and time course study Subsequent to transfection, growth in full serum conditions and at least 12 hours growth in charcoal-stripped serum, LNCaP, PC3, and HeLa F L A G - A R cells were treated with 10 nM ligand for up to 60 min. Prior to harvest, cells were washed once in cold PBS and separated into cytoplasmic and nuclear fractions using the Nuclear and Cytoplasmic Extraction Reagent (Pierce, Cat # 78833) at 10-min intervals. Fractions were assayed for total protein using the B C A protein assay (Pierce, Cat # 23223) with nuclear and cytosolic fractions being tested for purity by probing for Histones and E-cadherin, respectively. 3.3.7 Glutathione synthase transferase (GST)-pull downs Recombinant proteins were labeled with [S ] methionine using either the SP6 or T7 Quick Coupled T7 TNT in vitro transcription/translation kit (Promega) in a volume of 50 Lil. The pGEX vector was used to generate GST-AR in BL21 E.Coli with purification proceeding as follows: after 16-20 hours of incubation of radiolabeled proteins with GST-AR (plus protease inhibitors) complexes were isolated by 1 hour of incubation with glutathione-agarose beads. Complexes were washed 4x with binding buffer (20 m M HEPES, pH 7.6, 150 m M KC1, 5 mM MgCl2, 1 m M EDTA, 0.05% Nonidet P-40, and protease inhibitors). Finally, complexes were isolated from agarose beads by boiling in Laemmli sample buffer and then separated by 10% SDS-PAGE. Detection gels were treated with enhancing solution N A M P 100 (Amersham), dried and exposed to film. 3.3.8 Acrydite capture of DNA-binding complexes (ACDC) 93 The A C D C assay developed in our laboratory (Cavanaugh et al., 2002) was carried out by incubating recombinant His-tag androgen receptor D N A binding domain (His-tag ARDBD): (1) without D N A , (2) with NF-1 acrydite binding sites or (3) with A R E acrydite binding sites (ARE-ac). No D N A and NF-1 D N A served as negative controls for the binding assay. LNCaP cells were treated with or without 10 n M R1881 and nuclear extracts isolated by with hypotonic/hypertonic extraction. Extracts were incubated with acrydite-DNA probes for 12-16 hours with D N A bound complexes being polymerized in the well of a 5% polyacrylamide acrydite gel overlaying a 15% SDS-PAGE. After the first electrophoresis, protein that was specifically bound to DNA-acrydite was extracted from the acrydite gel and separated again by SDS-PAGE (8.5%). D N A was transferred to PVDF and probed for the presence of A R and f3-catenin. A R E binding sites were obtained from the Probasin promoter (P-ac) (-286 to +28) or the pBluescript acrydite multiple cloning site (MCS-ac) created by PCR amplification using specific primers with a 5'-acrydite moiety attached. 3.3 Results 3.3.1 Co-localization of P-Catenin and AR The localization of A R (Figure 3.1, ai) and P-catenin (aii) was viewed using immunocytochernistry and confocal microscopy with antibodies against the D N A binding domain of A R and the Arm repeats/CTD portion of P-catenin in fixed LNCaP cells, that were treated with and without R1881 for 60 min (Figure 3.1). Images captured by confocal microscopy were compiled in image analysis software and suggest that in the absence of androgen (bi), A R is diffusely localized throughout the cell, whereas P-catenin (bii) was found to be localized predominately at cell borders, diffusely in the cytoplasm, and within the nucleus, (ai) In the presence of ligand a punctate nuclear staining pattern of A R was observed, (aii) p-catenin also showed nuclear localization upon addition of ligand and co-localized in many instances with A R (arrowheads), though this pattern was not exclusive (arrows). Levels of P-catenin at cell borders were not observed to change in response to A R ligand, whereas cytoplasmic levels appeared to decrease moderately. Nuclei were stained using DAPI (Figure 3.1, aiv and biv). 94 3.3.2 The AR-LBD interacts directly with the P-catenin arm repeats As a prerequisite to assessing the trafficking properties of the AR, we determined the relative affinity of binding between the A R and P-catenin. To do this we used recombinant A R fragments, including AR-Nt, A R - D B D / L B D , AR-DBD, and AR-Nt/DBD (Figure 3.2a). Additionally, we employed P-catenin S35-labeled deletion mutants included Nt-(aa 1-140), Arm repeats-(aa 140-664), and Ct-(aa 664-781) domains (Figure 3.2b) translated in vitro. Although relatively weak interactions were detected between the P-catenin (Nt), or P-catenin (Ct), with any portion of the A R (only slightly greater than the GST-negative control), strong interactions were detected between the p-catenin Arm repeats and A R (Figure 3.2c). The Arm repeats showed moderate interactions with the AR-DBD but increased affinity with the A R - D B D / L B D , suggesting a role both for the DBD and L B D in binding A R and P-catenin. Despite the differences detected in our physical mapping, using our recombinant truncations, we were unable to detect a significant difference (10%) between AR-GST/p-catenin binding affinity in the presence or absence of ligand. We have attributed this to potential masking of interaction sites by chaperone proteins in the cell in absence of ligand, which is not present in the recombinant system. In general, these data support transcriptional assays, IP studies, and shuttling time courses. 3.3.3 Ligand-bound AR promotes nuclear translocation of cytosolic P-catenin To determine if P-catenin was able to translocate to the nucleus, in an androgen-dependent manner, we performed a series of cell fractionations: using LNCaP cells and HeLa-F L A G A R cells over a time course of androgen exposure (Figure 3.3). Cytoplasmic and nuclear compartments were efficiently separated, as determined by probing nuclear fractions for pan-cadherin and probing cytoplasmic fractions for total histone (Figure 3.3A). Low levels of pan-cadherin in nuclear fractions and histone in the cytosol, suggested an efficient separation of cellular compartments. We further demonstrated equal total protein loads through a time course by assessing total cytosolic actin and nuclear histone (Figure 3.3B). Non-transfected LNCaP cells that were treated with R1881 and were fractionated into cytoplasmic and nuclear compartments at 10-min intervals over 60 min (0, 10, 20, 30, 40, 50, and 60 min) showed a 95 moderate accumulation of endogenous, nuclear (3-catenin (Figure 3.3B i , ii). We also observed a similar increase in nuclear AR. Densitometry evaluation of nuclear accumulation of A R and [3-catenin suggested a similar rate of increase while loading controls (actin) remained constant (Figure 3.3Bii). A less noticeable decrease in cytosolic P-catenin was observed when compared with changes in nuclear (3-catenin levels. This was likely accounted for by the large amount of P-catenin that remained non-migratory; that is, it does not migrate upon addition of A R ligand. Interestingly, the shift in P-catenin found in cytosolic and nuclear compartments was considerably more noticeable in LNCaP cells, when transiently transfected with tagged expression constructs for P-catenin (Figure 3.3Ci, ii). More specifically, by transfecting LNCaP cells with either HA-tagged and myc tagged P-catenin constructs, a greater amount of nuclear accumulation of the de novo synthesized P-catenin was observed as compared with non-transfected cells. Densitometry analysis (Figure 3.3Cii) indicated a similar trend in nuclear accumulation of P-catenin with a plateau at 50-min post-ligand addition suggesting a decreased rate of nuclear import. We further chose to consider whether AR-dependent movement of p-catenin could occur using other A R ligands, including the physiological androgen, DHT. DHT, like its analogue, R1881, promoted nuclear accumulation of both the A R and P-catenin, although to a slightly lesser extent than with R1881 (data not shown). We next investigated whether other non-PrCa cell lines, including stably AR-FLAG-transfected HeLa cells could mediate AR-mediated P-catenin nuclear translocation. To do this we isolated cell fractions at 0 and 60 min post-ligand addition and immunoprecipitated AR, F L A G , and P-catenin (Figure 3.3D). In the absence of androgen, there were greater amounts of A R in the cytosol as judged by anti-FLAG and anti-AR antibodies; whereas there was greater amounts in the nuclear fractions of cells when treated with ligand. We found detectable P-Catenin with A R and F L A G - A R immunoprecipitates in the absence of androgen confirming the presence of a constitutive interaction, as observed with LNCaP IPs (Figure 3.3D). Interestingly, AR/P-catenin complexing was also detected in HeLa F L A G - A R cells cultured without androgen but substantially higher in its presence (Figure 3.3D). 96 3.3.4 The AR-DBD/LBD is necessary and sufficient for nuclear translocation of P-catenin in PC3 cells Having found evidence that the A R can translocate p-catenin to the nucleus in LNCaP and HeLa cells, we chose to ascertain whether other nuclear receptors have this capability in PC3 PrCa cells, a cell line that does not express endogeneous AR. When PC3 cells were transiently transfected with R A R (Figure. 3.4a), ER (Figure. 3.4b), GR (Figure 3.4c), TR (Figure 3.4c), or P-Catenin (HA-tag) we observed an inability to move P-catenin to the nucleus with ligand cognate exposure. Although GR shows an ability to translocate to the nucleus upon exposure to ligand, co-trafficking with P-catenin was not detected in this cell line despite minor fluctuations in GR cellular levels. With P-catenin translocation being detected, it was most pronounced with AR. We also chose to assess P-catenin movement using various A R deletion mutants (Figure 3.4, E-Hi) . In PC3 cells the empty, control expression vector (PRC/CMV) and that of the amino terminus (Nt) of A R demonstrated little ability to move P-catenin to the nucleus (Figure 3.4F). However, constructs expressing the AR-Nt/DBD showed a more prominent cellular expression, but did not show ligand-dependent changes in nuclear distribution of itself or p-catenin (Figure 3.4G). When cells were transfected with constructs expressing both the A R - L B D and the AR-DBD, ligand-dependent translocation of P-catenin was readily apparent (Figure 3.4Hi). This indicated to us the necessity of both the D N A binding domain and the ligand binding domain of the A R for efficient p-catenin nuclear translocation. Densitometry analysis indicated a coincident movement between the A R and P-catenin (Figure 3.4 Hii). Additionally, accumulation in PC3 cells was similar to that in LNCaPs, since both appeared to plateau at 50-60 min post-ligand addition. We further demonstrated that the Arm repeats were both necessary and sufficient for AR-dependent nuclear translocation of P-catenin. Constructs expressing the Nt and Ct components of P-catenin showed little, if any, fluctuation between compartments whereas the Arm repeats showed accumulation to the nucleus over 60 min (Figure 3.41). 97 3.3.5 AR-mediated translocation of P-catenin is distinct from APC complexing To determine if AR-mediated import of P-catenin was a distinct pathway, we assessed A R complexing with APC and GSK3 in LNCaPs (whole cell lysates) treated with and without R1881. As controls we probed A R IPs for A R (Figure 3.5a) and 3-catenin (Figure 3.5b). While we detected little fluctuation in A R levels as a function of hormone treatment, ligand sensitive interactions were observed between A R and P-catenin, through still detectable in the absence of androgen. Immunoprecipitated GSK3 showed a distinct band at about 45 kDa with only minor detection in A R immunoprecipitates. Although APC immunoprecipitates showed some degradation at -180 and 200 kDa, major species were detected at 300 kDa. These immunoprecipitates also did not contain detectable A R either with or without ligand, thus indicating the absence of A R / A P C complexes. 3.3.6 P-catenin and the AR complex directly on an androgen responsive promoter Acrydite experiments operate on the premise that protein specifically bound to a D N A sequence during a binding reaction will remain intact after being exposed to an electrophoretic current. Such species are then denatured when separated by SDS-PAGE and immunodetected. Our results showed that A R from nuclear receptor extracts, treated with R1881, bound specifically to ARE-acrydite (ARE-ac) PCR products (Figure 3.6). This was verified by the presence of the HIS-tag component of the A R - D B D in the elution from ARE-ac binding reactions (Figure 3.6a). A R antibodies also reacted strongly with androgen response elements and with only a small amount of nonspecific binding with negative controls, including a binding reaction for nuclear factor-1 (NF-1) and binding reactions without D N A (Figure 3.6bi). Figure 3.6bii indicates the extent of recovery of the A R in nuclear extract binding reactions, and was measured by comparing the reactivity with antibodies toward the D N A binding region of the A R which is specifically bound to the ARE-ac D N A sequence. Figure 3.6c suggests that, when binding reactions were probed for P-cateninusing a promoter sequence known to contain four cooperative AREs (-286/ +28), high amounts of 3-catenin were detected. Only low levels of P-catenin were detected in negative control binding reactions, including the multiple cloning site-acrydite PCR 98 reaction (MCS-ac) or binding reactions without DNA. These data strongly support the presence of a ligand-dependent transcriptionally active AR/P-catenin complex directly associated with the probasin promoter. Treating LNCaP nuclear extracts with 2 M NaCl and separating the non-extracted nuclear components by SDS-PAGE indicated a greater amount of both A R and P-catenin retained in the presence of ligand (Figure 3.6d). This suggests that both are incorporated into the nuclear matrix to a greater extent in the presence of ligand. 3.3.7 P-catenin shows increased nuclear expression in castrated mice reconstituted with exogenous androgen Mice were castrated, followed 14 days later with surgical implantation of exogenous DFfT (pellet) for 48 hours. Prostates were dissected and assessed for relative distribution of A R and P-catenin (Figure 3.7). Prostate tissue from castrated animals showed involution and regression, while that obtained from animals reconstituted with androgen appeared normal. Localization of A R in involuted tissue was predominantly cytosolic, but surprisingly demonstrated a high amount of basal nuclear A R expression (data not shown). Importantly, in repressed tissue p-catenin was found to be predominantly localized to the cell membrane and cytosolic compartments. However, upon reconstitution with the androgen, p-catenin showed dramatic relocalization to the nuclear compartment, with detection at the plasma membrane still being readily observed. 99 Figure. 3.1 Localization of (ai) A R and (aii) P-catenin in methanol-fixed LNCaP cells treated with 10 n M R1881 for 60 min viewed with confocal laser microscopy. In the absence of R1881, (bi) A R staining appears diffuse and throughout the cytoplasm whereas that for (bii) P-catenin is localized at the cell membrane, in the cytosol, and in the nucleus. When serial images collected by confocal microscopy were digitally compiled, the (ai) distribution of the A R appeared punctate and nuclear. Similarly, (aii) P-catenin showed a punctuate pattern that co-localized in some instances (arrowheads) whereas in others arrows) did not (overlay aiii, biii; DAPI aiv, biv) (bar 5 jUm). AR p-Cat Overlay DAPI 1 nM R1881 MOCK 101 Figure 3.2 Mapping of in vitro interactions between recombinant A R deletion truncations (WT, Nt, D B D / L B D , D B D , Nt/DBD) and in vitro translated P-catenin/S35 deletion truncations (WT, Arm repeats, Nt, and Ct regions). Complexes of GST-P-catenin and A R - S 3 5 were washed, decanted and resuspended in sample buffer. Samples were separated by SDS-PAGE, dried and exposed to developer, fixative and autoradiography, a) Relative to the total input the highest affinity of interaction was detected between the P-catenin/S35 Arm repeats and the AR-D B D / L B D . Interestingly, there was little difference (-10%) in the efficiency of binding reactions with or without the presence of A R ligand (data not shown), b) Graphical representation of the relative densitometry of in vitro AR/P-catenin interactions. Relative Binding Affinity v _ i <D Q> < DO 3 Q. 3 (Q > 3! (J-Cal B ! S 3 I I I 3 I I I I 1 I I I I § i • i i i > I a cn D D D 5 § J > 73 • CD CO -as • o M i CD 3 CD 3 a M l 3 ( Q o 103 Figure 3.3 LNCaP cell nuclear translocation of P-catenin in the presence of androgens over 60 min (10-, 20-, 30-, 40-, 50-, and 60-min intervals). Cells were fractionated into cytoplasmic and nuclear components and assessed for relative amounts of A R and P-catenin. Efficient separation was assessed by probing cytoplasmic fractions for histone and nuclear fractions for cadherins. (A) Total protein loads were assessed by the abundance of actin and histone in cytoplasmic and nuclear fractions, respectively. (Bi) LNCaP nuclear translocation of endogenous P-catenin and accompanying densitometry (Bii). (Q')Translocation of H A - and myc-tagged P-catenin in LNCaP cells and (CU) accompanying densitometry. (D) Stable AR-FLAG-tagged HeLa cells show a moderate movement of A R and P-catenin from a cytoplasmic fraction (0 min) when compared with R1881 treated fractions (60 min) as assessed by F L A G IPs of cellular fractions. Cytoplasmic Nucleus 0 10 20 30 40 50 60 0 10 20 30 40 50 60 ] WB: histone WB: pan-cadherin WB: Actin WB: Histone 104 Bl. Endogenous P-catenin B i i . Cytoplasmic Nucleus Min. 0 10 20 30 40 50 60 0 10 20 30 40 50 60 1m9 mm* n ^ - H mm -WB: AR Cytoplasmic Nucleus 0 10 20 30 40 50 60 0 10 20 30 40 50 60 — m - — WB: P-Cat Cytoplasmic Nucleus 0 10 20 30 40 50 60 1 ^ m m m w m m w mmw 0 10 20 30 40 50 60 WB: actin R1881 (Min): 0 10 20 30 40 50 60 WB: Histone (H3) C i . LNCaP (tagged p-catenin) Cytoplasmic Nucleus Min. 0 10 20 30 40 50 60 0 10 20 30 40 50 60 C i i . WB: AR Cytoplasmic Nucleus Min. 0 10 20 30 40 50 60 0 10 20 30 40 50 60 — WB: HA Cytoplasmic Nucleus Min. 0 10 20 30 40 50 60 0 10 20 30 40 50 60 R1881 (Min): 0 10 20 30 40 50 60 WB: actin WB: Histone (H3) D. HeLa (FLAG-tag) WB: p-cat WB: histone 106 Figure 3 .4 Evaluation of P-catenin nuclear co-translocation with non-AR nuclear receptors (A) RAR, (B) ER, (C), GR and (D) TR and with truncating mutations for AR. Evaluation of full-length p-catenin (HA-tagged) with (A) RAR, (B) ER, (C) GR, or (D) TR. Little shift in cellular compartmentalization of P-catenin is observed when other non-AR nuclear receptors are introduced into PC3 cells and treated with cognate ligand. PC3 cells were transiently transfected with (E) no A R (PRC/CMV vector), (F) AR-Nt, (G) AR-Nt/DBD, (Hi) A R - D B D / L B D and (Chuu, Hiipakka et al.)accompanying densitometry. (I)Time course examination of LNCaP cells transfected with AR-WT and P-catenin truncations, including that with the amino and carboxyl components and that with only the Arm repeats. 107 ^ Transfect: RAR + 10 nM trans retinoic acid Cytoplasmic Nuclear Min. 0 10 20 30 40 50 60 0 10 20 30 40 50 60 RAR p-Cat B. Transfect: ER + 10 nM Estradiol Cytoplasmic Nuclear Min. 0 10 20 30 40 50 60 0 10 20 30 40 SO 60 ^ - m m m ~ - -Q e Transfect: GR + 10 nM Dexamethasone Cytoplasmic Nuclear Min. 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Qm Transfect: TR + 10 nM T3 Cytoplasmic Nuclear Min. 0 10 20 30 40 50 60 0 10 20 30 40 50 60 108 E. Transfect: PRC/CMV Empty Vector + 10 nM R1881 Cytoplasmic Nuclear 0 10 20 30 40 50 60 0 10 20 30 40 50 60 F. Transfect: AR-Nt + 10 nM R1881 Cytoplasmic 0 10 20 30 40 50 60 AR P-Cat Nuclear 0 10 20 30 40 50 60 * 4 • » « • * * «t «f $ M G. Transfect: AR-Nt/DBD + 10 nM R1881 Cytoplasmic Nuclear 0 10 20 30 40 50 60 0 10 20 30 40 50 60 AR p-Cat mmm-' WmW ^PP W ^ 1 • Hi.Transfect: AR LBD/Ct + 10 nM R1881 Cytoplasmic Nuclear 0 10 20 30 40 50 60 0 10 20 30 40 50 60 110 Figure 3.5 Assessment of immunocomplexing between A R and a) P-catenin, b) GSK3, and APC c) in LNCaP cells treated with (+) and without (-) R1881 for 24 hours. Although A R interacted with P-catenin in a ligand-sensitive manner,1 low detection occurred with GSK3P and no detectable complexing occurring with APC. Control IPs for A R (a) and A P C (c) showed species at 120 and 300 kDa, respectively.(arrow heads = specific interactions; line = IgG band) WB: GSK3p GSK3p APC APC 112 Figure 3.6 Specific retention of proteins to androgen-regulated promoter regions using the acrydite capture of D N A complexes (ACDC) method, a) Retention of the His-tag A R - D B D was observed only for the specific ARE-ac binding site, not for the non-cognate NFl-ac binding or in the absence of DNA. bi) The A C D C method was used to capture the full-length A R and a proteolytic cleavage product of the A R that binds DNA. c) The A C D C assay was used to capture P-catenin from LNCaP nuclear extracts treated with 10 n M R1881. Nuclear extracts were incubated without D N A or with acrydite probasin (PB-ac) promoter (-286 to +28) or the pBluescript acrydite multiple cloning site (MCS-ac). Retention of p-catenin for the AR-containing LNCaP Nuclear Extract was observed only for PB-ac, which contains four known AREs. No retention was observed with the non-specific MCS-ac or in the absence of DNA. d) Nuclear bound AR/B-catenin complexes are present in high salt (2M NaCl) nuclear extracts. a) WB: His-tag AR-DBD No DNA NF-1 ARE-ac b i ) . WB: AR No DNA NF-1 ARE-ac b i ) WB: AR ARE-ac LNCaP NE Input WB: p-catenin c) No DNA MCS-ac PB-ac d) 2M NaCl 2M NaCl R 1 8 8 1 : + WB: p-cat WB: AR 114 Figure 3.7 Assessment of P-catenin localization in mouse prostate after 7 days in animals that were either a) castrated or b) castrated and supplemented with exogenous androgen pellet for 48 hours. 115 116 Figure 3.8 Proposed model showing that AR may facilitate translocation of cytoplasmic p-catenin to the nucleus and, thereby, increase A R transcription and possibly regulate Tcf signaling. Upon ligand occupation, heat-shock proteins (Hsp) dissociate from the A R allowing for augmentation of P-catenin binding. Hsp dissociation also exposes the A R nuclear localization site, thereby promoting translocation of AR-P-catenin complexes to the nucleus and binding to one or more A R promoters. (SRC = steroid receptor co-activator; CBP = Creb's Binding Protein) 118 3.4 Discussion A R trafficking of P-catenin is a particularly attractive hypothesis for several reasons. First, P-catenin does not have an identified nuclear localization signal, making it dependent upon other chaperone molecules for nuclear import. Second, the nuclear interactions between the A R and P-catenin are ligand sensitive. Third, AR-mediated transport of P-catenin appears to be distinct from APC transport of P-catenin. Fourth, transcriptional activity of A R promoters is augmented with increased levels of transfected P-catenin when in the presence of ligand. In this study, we provided evidence for a novel and distinct mechanism by which p-catenin can enter the nucleus of A R expressing cancer cell lines. We showed that upon exposure to androgen, the A R is capable of shuttling P-catenin to the nucleus, thus providing a means by which P-catenin can augment the transcriptional activity of AR reporters. This mode of import appears to be functionally independent of the p-catenin transporter protein, APC. Immunofluorescence studies strongly suggest co-localization of endogenous A R and P-catenin at punctate complexes in the nucleus of LNCaP PrCa cells in the presence of R1881. Co-IP studies suggest a cytosolic AR/P-catenin interaction but not with APC or GSK3. Further evidence for an AR/P-catenin complex that can directly bind to an AR-regulated promoter was provided using a novel D N A binding assay. Finally, we provided evidence that the P-catenin Armadillo repeats, and the AR-D B D / L B D domains are required for ligand-dependent nuclear translocation and transcriptional activation. A R and P-catenin can co-localize at transcriptionally active, multiprotein nuclear complexes Using confocal laser microscopy we demonstrated that P-catenin can co-localize with A R using antibodies specific for the D N A binding region of the receptor in the presence of androgen. The localization of other steroid receptors in the nucleus has been described, and, as with the GR (Hache, Tse et al. 1999), ER (Htun, Barsony et al. 1996), and TR (Bunn, Neidig et al. 2001), the A R shows a distinct punctate appearance upon exposure to ligand. Although the R A R and TR do not show substantial migration to the nucleus upon addition of ligand, it is reasonable to assume 119 that there is dynamic subnuclear trafficking of these receptors. Focal accumulations of nuclear proteins in androgen-treated cells likely consist of transcriptionally active protein complexes with AR, P-catenin, and other associated co-factors, including steroid receptor coactivator (SRC), and Crebs Binding protein (CBP), and T-cell factor-4 (Tcf4) being present (see Chapter 4 of this thesis). The co-localization of A R and P-catenin was not exclusive, as seen by confocal microscopy data which indicted that there were many A R complexes unoccupied by P-catenin and, similarly, P-catenin complexes that are unoccupied by AR. The co-localization of P-catenin with the A R also implies that P-catenin could be involved with the transcriptional machinery of LNCaP cells. This prospect was explored with the use of a novel D N A binding assay. Using an acrydite polymer, we were able to show specific AR-dependent binding of P-catenin to the probasin promoter. This suggests that P-catenin could have the capacity to modulate the cell cycle in PrCa cells, in an androgen-dependent fashion; likely by altering levels of known AR-modulated transcription factors such as c-myc (Bieche, Parfait et al. 2001) and the cyclins (Reutens, Fu et al. 2001). Nuclear translocation of P-catenin by A R . Detection of P-catenin migration from the cytosolic compartment to the nucleus was readily apparent using both light level and biochemical experiments. In both endogenously expressing A R cells (LNCaP) and transfected AR-expressing cells (PC3), time series cell fractionation studies showed a modest, ligand-dependent, nuclear translocation of P-catenin. This data determined that the majority of migrating A R located to the nucleus subsequent to 60 min of exposure to androgen. Time-course experiments extended beyond this time point (data not shown) did not yield significant amounts of A R or P-catenin movement between compartments. The movement of P-catenin was, in general, much less obvious in the cytosolic fraction. We attribute this to the fact that the percentage of P-catenin associated with the A R and the ability to move to the nucleus is relatively small compared with the total cytosolic pool of P-catenin. Additionally, it is likely that the amount of "free," non-bound, P-catenin varies considerably between different cell lines. The issue of different cytoplasmic pools of P-catenin has been raised in previous studies and could be functionally important in A R transcription. Previous reports (Stewart and Nelson 1997) have identified distinct pools of P-catenins, 120 suggesting the possibility of distinct APC and A R associating pools of P-catenin. It would be interesting to evaluate how other cytoplasmic signaling pathways and adhesion complexes are altered as a function of the removal of androgen-associated P-catenin during its androgen mediated nuclear translation. Nuclear compartments of P-catenin are likely in a constant state of flux, thereby allowing nuclear receptors and Tcf4-related complexes to "trade-off or compete for available P-catenin. Time series experiments showed relatively little, if any, movement of ER between the cytosol and the nucleus. This finding supports previous reports showing that the ER is found predominantly in the nucleus with or without the presence of ligand (Htun, Holth et al. 1999). The absence of nuclear co-migration with P-catenin also holds true with TR and RAR. However, given that P-catenin is a known co-activation of TR, ER and RAR, suggests the possibility of co-trafficking occurring in a strictly nuclear manner. Although we were not able to detect trafficking of P-catenin with these receptors, we also cannot rule out the possibility that during the initial translation and modification of these receptors some P-catenin is brought to the nucleus. Such a form of catenin trafficking would not be detectable with the techniques used in this study. The finding that GR does not facilitate P-catenin translocation in the same manner as A R does could be a function of the cell lines used in the present studies; that is, cells that express high levels of GR likely show a greater potential to translocate P-catenin as compared with PrCa cell lines. However, other corroborative studies (Pawloski et al) have also failed to detect GR/P-catenin interactions. Future studies, using such cell lines, will help clarify if a functional relationship exists between GR and P-catenin. We sought to determine the minimally required components for translocation of P~ catenin to the nucleus. By transiently transfecting a series of deletion mutants to PC3 cells we are able to determine that p-catenin requires the A R - D B D region for binding, but requires the AR-D B D / L B D for significant ligand-dependent nuclear translocation, p-catenin showed a small degree of nuclear movement in cells that were AR-Nt/DBD transfected but not in a prominent, ligand-dependent manner as observed when cells were transfected with A R - D B D / L B D . By using GST fusion protein interactions we determined that the A R - D B D domain is capable of a strong receptor/P-catenin interaction without the A R - L B D , indicating that minor amounts of P-catenin, although not detected by our methods, may be carried into the nucleus by the AR-DBD 121 independent of the L B D or more likely in association with other chaperones responsive to androgen. When the relative affinities between the A R and P-catenin were examined with GST binding assays, we did not observe any difference as a function of ligand exposure, likely due to the absence of chaperone proteins that would otherwise mask binding. LNCaP and PC3 PrCa cells are in many respects good systems to demonstrate the hypothesis of AR-mediated nuclear translocation of P-catenin. This is mainly due to constitutively low activity of GSK3 and highly active Akt (Kreisberg, Malik et al. 2004), a regulator of GSK3. Irregular GSK3 and Akt levels can be attributed to the mutated tumour suppressor PTEN, which both cells are null expression for. Furthermore, the relative amounts of unphosphorylated P-catenin are likely higher than other cell types where GSK3 function may not be inhibited by way of a hyper activated P B K / A k t pathway. This implies that any additional cytosolic P-catenin could have the potential for other binding partners such as cytosolic steroid receptors. Although AR-dependent movement of P-catenin appears to be functionally independent of APC, the continual movement of P-catenin from the cytoplasm to the nucleus and back to the cytoplasm in cancer cells suggests that these pathways can be functionally independent but can also draw from similar pools of p-catenin. Our proposed model of A R mediated transport of P-catenin (Figure 3.8) shows that upon ligand occupation, Hsps dissociate from A R allowing for augmentation of P-catenin binding. Hsp dissociation also exposes the A R nuclear localization site, which could promote translocation of an AR-P-catenin complex into the nucleus and binding to one or more A R promoters. Our finding of an AR-mediated trafficking pathway leads to the logical implication for a role in oncogenesis, whereby stimulation of the Wnt pathway could promote greater cytosolic P-catenin, and therefore greater A R transcriptional activity. Increased nuclear p-catenin could result in altered cell cycle or increased cell proliferation. Future studies will likely include investigations of how the A R and other nuclear receptors interact with the Wnt pathway. With structural conservation between the members of the nuclear receptor family, it is likely that there are many commonalities between the manner by which AR, GR, and ER interact with Wnt/P-catenin/Tcf signaling . Similarly, the ability of P-catenin to augment transcriptional activity of other nuclear receptors is a significant finding and 122 will likely lead to further studies in specific tissues such as the thyroid-targeted tissues, breast and liver. While there are likely many similarities in how p-catenin interacts with various nuclear receptors, it is likely there are many caveats to how P-catenin could influence individual receptor pathways and the endocrine functions which they control. Ultimately, such interactions may lead to more invasive studies, whereby repression of Wnt members could attenuate steroid receptor associated transcriptional activation. A key to elucidating how the Wnt pathway interacts with nuclear receptors lies in understanding the likely dynamic interplay between Tcf components, nuclear receptor promoters, and P-catenin, but also shared co-factors such as CBP and many, yet, undefined ones with the potential to facilitate both pathways. Chapter Four: A n d r o g e n Recep to r R e p r e s s i o n o f W n t / p - c a t e n i n / T c f S i g n a l i n g 124 4.1 Introduction Nuclear hormone receptors have been shown to have capacity for dual transcriptional function, allowing either activation or repression of target promoters (Damm, Thompson et al. 1989) (Sap, Munoz et al. 1989) (Schulman, Juguilon et al. 1996). Of recent testament to this is the functional and reciprocal interactions of NRs with the 3-catenin pathway (see Appendix Review: Interactions of Nuclear Receptors in Wnt//3-catenin/Tcf Signaling Wnt you like to know). For example, it has been shown that 3-catenin increases the activity of the retinoic acid receptor (RAR) on RAR-responsive promoters with a reciprocating down regulation of the 3-catenin/Tcf signaling pathway by the R A R (Easwaran, Pishvaian et al. 1999). Other reports have shown Wnt signaling to be functional in human thyroid cells (Helmbrecht, Kispert et al. 2001) with triiodothyronine (T3), also having the capacity to silence the 3-catenin/Tcf pathway (Miller, Park et al. 2001). With the structural conservation between members of the steroid nuclear receptor family, it is reasonable to hypothesize that similar interactions with the Wnt pathway may exist with other nuclear steroid receptor members. Reports have shown a role for the 3-catenin/Tcf pathway in PrCa including nuclear localization and 3-catenin mutations in primary prostate tumour samples (Chesire, Ewing et al. 2002) (Chesire and Isaacs 2002). 3-catenin has also been described as a ligand dependent coactivator of A R (Truica, Byers et al. 2000) (Mulholland, Cheng et al. 2002) (Pawlowski, Ertel et al. 2002) (Yang, L i et al. 2002). Previously, we have shown that translocating A R can provide a means of nuclear entry and accumulation of 3-catenin (Chapter 3, Thesis) (Mulholland, Cheng et al. 2002). This mode of 3-catenin trafficking has also been shown to hold true in neuronal cells (Pawlowski, Ertel et al. 2002). Co trafficking of A R and 3-catenin to the nucleus likely has important implications both for A R and Wnt signaling. However, while nuclear accumulation of 3-catenin has been shown to correlate with increased A R transcriptional activity, the effects on Tcf signaling have only begun to be explored. Recently, it has been shown that A R has the ability to inhibit the 3-catenin/Tcf signaling pathway, ligand-dependently (Chesire and Isaacs 2002). By way of transcriptional reporter assay, this report showed reduced luciferase activity for the Tcf reporter in a ligand-dependent manner in several prostate and colon cancer cell lines. This study also observed reduced transcriptional activity with the use of an A R deletion mutant (for 3-catenin binding), suggesting the possibility of a reciprocal balance of nuclear 3-catenin between the A R and Tcf. 125 The data presented in this chapter corroborates these results but provides a mechanism to support the hypothesis that repression of the (3-catenin/Tcf signaling is mediated by ligand-occupied A R that is in competition with Tcf for nuclear p-catenin. Specifically, using transcriptional reporter assays, we show that overexpression of WT Tcf, reduces the activity of an A R responsive reporter (ARR3-Luc), while overexpression of a ANTD-Tcf mutant does not achieve this effect. We employ the use of fluorescent fusion proteins and image analysis software to quantify a reciprocal balance of P-catenin-EGFP between HcRed-Tcf and HcRed-AR. We further show that the A R antagonist, Casodex has the potential to diminish AR-mediated depletion of Tcf transcription and the ability to diminish androgen-sensitive complexing of endogenous A R and P-catenin. Our data are also novel in showing that AR-mediated reduction of the Tcf signaling pathway can translate into altered cell cycle and reduced cell growth in cell lines with hyperactive p-catenin/Tcf function. 4.2 Methods 4.2.1 Polymerase chain reaction (PCR) D N A amplification was generally carried out using 50-100 ng template D N A with polymerase with (Pfx) or without proofreading (Taq) functions with respect to the manufacturer's protocol. Using a 250 ul PCR tube and a D N A cycler (Engine Peltier Thermal Cycler, M J Research) D N A amplification was performed as follows: Cycle 1 = (Denaturation: 90 sec at 95 °C; Annealing: 60 sec at 55 °C; Elongation: 180 sec at 72 °C), Cycle 2 = (Denaturation: 90 sec at 95 °C; Annealing: 60 sec at 60-68 °C; elongation: 180 sec at 72 °C). Cycle 2 was repeated 25 times with a final incubation of 10 min at 72 °C. PCR products were analyzed on 1-2% agarose gels containing Ethidium Bromide. 126 4.2.2 Plasmids A R & AR-deletions: Full-length hAR was PCR amplified using both forward and reverse primers containing BamHl restriction sites and pSV40-AR0 as a template (Brinkman, University of Rotterdam). Forward and reverse primer sequences were 5' G G A T C C G 3' and 5' GGATCTT 3', respectively. The PCR product was digested with BamHl and ligated into BamHl sites of pBlueScript plasmid (Stratagene). A R cDNA was excised and subcloned into the BamHl sites of pcDNA3.1. The ligand binding domain deletion (A-RNt/DBD) construct, amino-terminal A R deletion (AR-DBD/LBD) construct and the ARR3-Luc reporter were cloned and characterized as previously described (Snoek et al., 1996). AR-EGFP: The AR-EGFP plasmid was obtained from Dr A Roy (University of Texas Health Science Centre at San Antonio); WT and mutant (Al-130 bp) P-catenin cDNAs from Dr B Gumbiner (Univ of Virginia); human ANt-Tcf4 and Tcf construct (pHRhTCF4) from Dr B Vogelstein (Johns Hopkins, Baltimore, USA); and cyclin Dl-luc construct from Dr R Pestell (Albert Einstein College of Medicine, N Y , USA). The Tcf4 luciferase reporter (Topflash) and a mutated control reporter (Fopflash) were purchased from Upstate Biotechnology (Cat #21-170, 21-169). HcRed-hTcf4: HcRed (Clontech) is a recently developed far-red fluorescent protein derived from the coral Heteractis crispa and has been shown to be efficient for colabeling with green fluorescent proteins (Knop et al., 2002). The HcRed-hTcf4 fusion vector was generated by cloning the BamHl fragment of pHR-hTcf4 into the BamHl site of pBlueScript SK-(Stratagene), followed by excision of the reading frame fragment by an EcoRl digest and subsequent cloning into the EcoRl site of the pHcRedl-Cl vector. HcRed-AR: The HcRed-AR contains a PCR product digested with BamHl and ligated into a BamHl ligated into the pHcRed-Cl vector (Clontech, Cat # 63-631). The P-catenin-HcRed fusion vector was generated by cloning the BamHI-liberated P-catenin coding sequence from pGEX4Tl-P-catenin (A Hecht, Max-Planck-Institute of Immunobiology, Freiburg) into the BamHl site of the pHcRed-Cl vector. The 6xHis-tagged hTcf4 bacterial expression vector was generated by subcloning the hTcf4 coding sequence out of the pHR-hTcf4 mammalian expression vector. pHR-hTcf4. The pHR-hTcf4 vector was digested with BamHl, and a D N A fragment of 2.4 kb containing the hTcf4 coding sequence was gel extracted. This fragment was cloned into the BamHl site of the pBlueScript SK (-) plasmid and the resulting clones sequenced to identify clones in which the EcoRl site of pBlueScript was downstream of the hTcf4 coding sequence. An appropriate clone was isolated and digested with BamHl in the presence of ethidium bromide to prevent cutting of both BamHl sites. The linear plasmid was precipitated to 127 remove ethidium bromide and then digested with EcoRl. The 2.4 kb band corresponding to the hTcf4 coding sequence was gel extracted and cloned into the BamHT/EcoRI sites of the pENTR2B Gateway vector (Invitrogen). The resulting colonies were sequenced to verify the reading frame and then used in an LR recombination reaction as per the manufacturers' instructions to transfer the hTcf4 coding sequence to the pDEST17 bacterial 6xHIS-tag Gateway Destination vector (Invitrogen). Clones were screened for insert direction and reading frame integrity by sequence analysis. 4.2.3 Luciferase reporter assays After 24 hours of transfection, cells were cultured in 5% charcoal dextran-treated fetal bovine serum treated with DHT (0, 0.01, 0.025, 0.05, 0.075, 0.1, 1 or 10 nM) dissolved in dH20, 10 u M pure Casodex or vehicle. At 24 hours post-transfection, cells were harvested and luciferase activity was evaluated using the Dual Promega Luciferase (Promega). Luciferase assays for PC3 cells were performed in six-well plates while all other cell lines were analyzed in 12-well plates, both with a maximum of 10 p,g of transfected DNA/plate. For each transfection, the total mass of D N A (Steck, Pershouse et al.) was kept equal by the addition of empty vector pcDNA 3.1 (Invitrogen). To normalize for transfection efficiency, several means were used including cotransfection with pRL-TK Renilla vector (Promega) and cotransfection with C M V -EGFP-tagged empty vectors followed by Western blotting for EGFP. 4.2.4 In vitro translation and recombinant proteins Radio-labeled (S 3 5-Met) Tcf4-HIS and hAR protein was prepared using the Quick Coupled T7 TNT in vitro transcription/translation kit (Promega). Recombinant (3-catenin-HIS was expressed in the BL21 Escherichia coli strain and purified as described previously (Mulholland et al., 2002). In vitro binding reactions occurred in binding buffer (20mm HEPES pH 7.6, 150mm KC1, 5m m MgC12, 1mm EDTA, 0.05% NP-40) and were precipitated using N i -N T A beads (Qiagen). Precipitates were washed four times with binding buffer, separated by 128 SDS-PAGE, fixed, amplified with N A M P 100 (Amersham) and exposed to autoradiographic film. 4.2.5 FACS analysis SW480 cells were transfected in six-well dishes with and without 10 ug of WT A R treated with increasing concentrations of DHT (0, 0.01, 0.025, 0.05, 0.075, 0.1, 1 or 10 nM). Cells were collected using PBS/EDTA (5 mM) and fixed, dropwise, to a final concentration of 70% cold EtOH. Cells were pelleted and resuspended in staining buffer (20 mg/ml RNAse Abpropidium iodide) for 30 min at 37°C and then 41°C for 60 min. Samples were run in triplicate with a minimum of 10,000 cells/replicate. 4.2.6 In vitro growth assays The in vitro growth of SW480 cells was assessed by the in vitro 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (MTT assay) as previously described (Gleave et al., 1991). Briefly, 3x l0 J cells were seeded into each well of a 96-well microtiter plate. Cells were transfected with 0.1 mg of D N A per well and treated with ligand for 24 hours. Subsequently, cells were treated with 20 ui of MTT (Sigma) in PBS followed by incubation for 4 hours at 37°C. The absorbance was determined with a microculture plate reader (Becton Dickinson) at 540 nM. Absorbance values were normalized to the values obtained for the vehicle-treated cells to determine the percent of cell viability. Each assay was performed in triplicate. 4.2.7 Cell proliferation assay Cells were plated onto a 24 well plate at a density of l-2x 105 cells/well in RPMI containing 10% FBS. Treatment media after 72 hours, 100 (0,1 of PMS-MTS mixture was added to each well (Cell Titer 96™ Non-radioactive Cell Proliferation Assay A Q , Promega, Madison, Wl). lOOui of media were transferred to 96 well microtiter plates and the absorbance was 129 determined with a microculture plate reader (Becton Dickinson Labware, Lincoln Park, NJ) at 490 nM. Each assay was performed in triplicate with evidence for transfection being determined by western blotting for AR. 4.2.8 Northern blot analysis Total R N A was extracted from cultured cells or xenografts using the acid-guanidinium thiocyanate-phenol chloroform method. Frozen samples were treated with Trizol and phenol-chroloform to extract total RNA. After separation by centrifugation the aqueous layer R N A was precipitated using 90% ethanol and washed using 70% ethanol. For analysis of specific message levels a total of 20 Lig of RNA from each sample was subjected to gel electrophoresis (1.2% agarose/30% formaldehyde) and transferred to nylon membranes overnight according to standard procedures. (3-catenin, A R and Tcf-4 R N A probes of approximately 500 bp in size were generated by restriction enzyme digest of expression plasmids. D N A was denatured at 95 °C for 5 min followed by a 1 hour incubation with 50 uCi of [ 3 2P]-dCTP and Ready Prime labelling system. Membranes were probed with pre-hybridization fluid for 1 hour (42 °C) followed by overnight incubation with the labeled probe (40 °C). After 2 x 15 min in washing buffer 1 (0.1 % SDS, 2x SSC, lx phosphate buffered saline) and 2 x 15 min washing buffer 2 (0.05% SDS, I X SSC, 0.5X Phosphate Wash Buffer) blots were exposed to radiographic film for 1-20 hours. Analysis of message levels were normalized to levels of Glyceraldehyde-3-phosphate dehydrogenase levels (G3PDH) were used to normalize expression of experimental probes. 4.2.9 Western blotting For androgen withdrawal, experiments cell lysates were removed from tissue culture plates by 5 m M EDTA/PBS treatment, centrifuged and lysed in a non-SDS based lysis buffer (20 m M Tris-HCl, 1% Triton, Na Deoxycholate) for 30 min on ice followed by centrifugation for 10 at 10,000 rpm. Cells lysates were then separated by SDS-PAGE and probed for the presence of E-cadherin (Transduction Lab.), p-catenin (Santa Cruz, Transduction Lab.), A R (Santa Cruz) and PSA (Santa Cruz). 130 4 . 3 Results 4.3.1 Repression of Tcf is both AR and androgen dose dependent but can be relieved by Casodex To evaluate the repressive effects that A R and its physiological ligand confer upon Tcf signaling, several titrations were performed using PC3 cells. Specifically, we observed decreased Topflash activity with increased amounts of transfected A R with no detectable changes observed in either P-catenin or Tcf4 protein levels (Figure 4.1a). Basal levels of Topflash reporter in untreated PC3 cells were low, while in SW480 cells luciferase counts were 6-8-fold higher. In either cell line, control Fopflash values were about 10% of basal Topflash levels. To confirm the ligand dependency of Tcf inhibition, we used A R deletion mutants (Figure 4.1b), including those coding the N-terminal and DNA-binding domain regions (Nt/DBD), as well as the DNA-binding region plus ligand-binding region (DBD/LBD). While the AR-Nt/DBD mutant showed little ability to repress Topflash, the A R - D B D / L B D mutant was capable of a 3.5-4-fold repression in the presence of DHT. This suggests that in an A R overexpressed state, the AR-Nt is dispensable for Tcf repression. Further verification that the L B D is vital for repression is shown by increased relief of AR-mediated repression in cells treated with the pure A R antagonist, Casodex (1 and 10 pm), which was able to efficiently relieve AR (DBD/LBD)-mediated repression of Topflash. Having shown that repression is both A R and ligand dependent, we next evaluated the effect of increasing the concentration of DHT. First, we confirmed that our titration of DHT (0, 0.01, 0.025, 0.05, 0.075, 0.1, 1 and 10 nM) was an appropriate titration for the ARR3-Luc reporter. Both PC3s (•) and SW480 cells (•) showed similar trends of ARR3-Luc activation when transfected with A R and treated with increasing DHT concentrations (Figure 4.1c). When these conditions were applied using the Topflash reporter, we observed moderate diminishing of activity in PC3 cells (2.5-3-fold) and SW480 cells (4-4.5-fold) (Figure 4.Id). 131 4.3.2 AR/DHT-mediated repression can overcome P-catenin activating mutants Using SW480 cells transfected (2 ng/35 mm well) with either WT P-catenin or an activated P-catenin mutant (Al-130 bp) to remove the possible GSK3 phosphorylation site, we were able to show that A R plus 10 n M DHT has the ability to repress both forms of P-catenin-driven Tcf signaling with a slightly lesser efficiency for cells containing exogenous, activated P-catenin (Figure 4.2a). In addition, by using SW480 cells that are A P C -/-, we verified that DHT mediated repression is distinct from the main route of P-catenin ubiquitination. Experiments using HCT116 cells which harbor an activating P-catenin mutation were also prone to Wnt repression in the presence of DHT (data not shown). In SW480 cells, the Wnt target, cyclin DI , was inhibited to a similar degree as Topflash with and without the presence of exogenous P-catenin (Figure 4.2b). A control form of the cyclin DI reporter (pA3-Luc) showed little fluctuation with treatments (data not shown). Furthermore, by increasing DHT concentration, we were also able to repress cyclin DI protein levels but with little fluctuation in P-catenin protein levels (Figure 4.2c). A R showed increased stabilization and a slight increase in protein levels with increased DHT concentrations, while actin showed equal loading. 4.3.3 Androgen-dependent localization of HcRed-Tcf with AR-EGFP in LNCaP cells Subsequent to the addition of DHT and the nuclear translocation of the A R fusion, we observed partial colocalization between HcRed-Tcf and AR-EGFP (Figure 4.3a). Using deconvolution microscopy (DECON) of a single xy plane, to reduce fluorescence contribution from different focal planes, and Northern Eclipse Image analysis software (Empix Imaging, Inc.)/Adobe Photoshop 5.5, we calculated colocalization (as demarked with white arrowheads) of the A R and Tcf fusions to be 35%, with both proteins forming approximately 300 foci per nucleus. This value appeared to be consistent between cell lines (LNCaP, SW480 and PC3) subsequent to A R nuclear translocation. Colocalization (Figure 4.3bi, i i , arrowheads) was deemed as foci that contained yellow (Y, colocalization) rather than red (R) or green (G) (Figure 4.3biii), which were counted as no colocalization (Figure 4.3bi, arrows). In carrying out these experiments, LNCaP cells were cotransfected with either P-catenin-EGFP + HcRed-AR (Figure 3ci) or P-catenin-EGFP + HcRed-Tcf (Figure 4.3cii) and treated with or without 10 nM DHT for 132 48 h. Cell nuclei were evaluated for percentage fusion colors as a function of ligand treatment. Transfection with (3-catenin-EGFP + HcRed-AR and treating with DHT, 62.0% of fusion color appeared yellow, while in the absence of androgen this value dropped to 14.5%, with green and red values changing from 15.5 to 69.0% and 22.5 to 16.5%, respectively. These data validate the ability of 3-catenin-EGFP to become associated with HcRed-AR in an androgen-dependent manner. We observed reduced colocalization of (3-catenin-EGFP with HcRed-Tcf (15.3% Y, 40.2% G, 44.5% R) upon treatment with 10 nM DHT, but an augmentation to 63% Y in cells treated only with vehicle. This is suggestive that f3-catenin-EGFP has the capacity to reciprocate between A R and Tcf, androgen-dependently. To account for partial overlapping, focal colors and varying number of foci per nucleus, total values of yellow, red or green per nucleus were quantitated as a relative percentage of total focal area per nucleus. Furthermore, a minimum of 10 representative nuclei per cell treatment were evaluated. Standard deviations were included to show statistical variation within each treatment. Phosphorylated R N A Pol localization to HcRed-Tcf foci provided support that HcRed-Tcf foci could contain sites of transcriptional activity. Specifically, SW480 cells were transfected with HcRed-Tcf and stained for a phosphorylated form of R N A Pol II (phospho Ser 2) in the presence of 10 nM DHT (Figure 4.4). Subsequent to deconvolution, we observed an overlap of signals (30%) suggesting that HcRed-Tcf sites or complexes have the capacity to be phosphorylated. These data suggest that these sites are not simply a result of an overexpressed plasmid, but that they may be physiologically and transcriptionally active. Interestingly, we observed little difference in phospho R N A Pol U staining frequency with and without DHT (data not shown). 4.3.4 Confirmation of expression and transcriptional activity of AR-EGFP and HcRed-Tcf Since we chose to evaluate the distribution of both A R and Tcf4 by way of fluorescent fusion vectors, confirmation of correct molecular weight (Figure 4.5a) and transcriptional activity (Figure 4.5b) was necessary. Transfections of the AR-EGFP into LNCaP cells showed an expected doublet signifying the endogenous (arrow) and the fusion form of A R (arrowhead) with increased levels present following androgen exposure (10 n M R1881). SW480 cells also showed a band at approximately 135kDa (arrowhead, upper band), indicating a correctly 133 migrating AR-EGFP for an A R Western Blot (a, top panel). Furthermore, low but detectable levels of endogenous A R in both HCT116 (data not shown) and SW480 cells were apparent (arrow, lower band). Further, minor fluctuation of AR-EGFP expression occurred as a function of androgen exposure. Expression of endogenous Tcf in SW480 cells was recognized by the presence of a band at about 60 kDa and a tagged form approximately 25kDa larger upon probing for Tcf protein. The fusion also expressed well in LNCaPs, although considerably lower levels of endogenous Tcf were detected. We further verified the functional activity of HcRed-Tcf in SW480 cells (Figure 4.5b, right) and observed augmentation of the Topflash reporter, albeit to a slightly lesser extent than with the nontagged form, with only slight activation of the mutant control. Similarly, we were able to verify that the AR-EGFP vector could activate the ARR3-Luc reporter in the presence of androgen, again, with only slightly less activity than the untagged form (Figure 4.5b, left). To rule out the possibility of false-positive or aggregating fluorescent proteins, we transfected LNCaPs with CMV-HcRed (no insert) either alone or with AR-EGFP. The CMV-HcRed showed a distribution that was nonlocalized and without the formation of nuclear foci (Figure 4.5c). 4.3.5 In vivo and in vitro binding assays are suggestive that AR and Tcf compete for nuclear P-catenin Transcriptional and morphological data strongly suggest the possibility of competition between Tcf and A R for a nuclear pool of p-catenin. To evaluate this possibility further, we used transcriptional assays with a WT and a mutant P-catenin-binding Tcf expression plasmid (ANt/ Tcf), co-IPs in SW480 cells treated with and without androgen, and an in vitro binding and competition assay (Figure 4.6). Using the ARR3-Luc reporter, we found that increased levels of full-length Tcf resulted in decreased promoter activity, while a ANt Tcf deletion mutant (non-P-catenin binding) did not display this trend (Figure 4.6a). Verification of expression by the WT and mutant Tcf is seen from a Western blot for Tcf and myc tag, respectively (Figure 4.6a, arrows). While increased Tcf protein levels were observed with the titration, fluctuation in A R levels in the presence of androgen was not observed. As part of our evaluation of whether A R and Tcf compete for nuclear P-catenin, we assayed whether we could detect an AR/Tcf complex. As a prerequisite, we verified that we could, in fact, detect an AR/P-catenin complex upon transfection with A R (Figure 4.6b, arrow). Consistent with our previous studies (Chapter 3, 134 Thesis) (Mulholland et al., 2002), more AR/(3-catenin complexing from cells treated with DHT was detected confirming a ligand-sensitive interaction. We also detected an association with (3-catenin and Tcf-HIS, but did not detect this association in the ANt Tcf-myc deletion mutant. Importantly, using this assay, interactions between A R and Tcf were very weak as compared to our detected AR/(3-catenin and P-catenin/Tcf complexes both in prostate and colon cancer cells. Nonimmune control precipitations were only slightly less than those of the AR/Tcf complex, suggesting that only a small fraction of A R associates with Tcf. Next we considered whether Casodex, which abrogated both HcRed-Tcf focal accumulation (data not shown) and Topflash activity, could alter the binding of A R and (3-catenin. By treating cells with 5 \im Casodex (dissolved in EtOH) with and without 5 nM DHT, we observed reduced physical interactions between A R and P-catenin (arrowhead), while also relieving AR-mediated repression of the p-catenin/Tcf-HIS complex (Figure 4.6b, star and open arrow heads). To evaluate whether a decreased P-catenin/TcfHIS complex could be directly affected by the A R Tcf-HIS and A R S 3 5 , in vitro translated products were used in a series of binding assays with HIS-tagged recombinant P-catenin (Figure 4.6c). Using Ni -NTA beads to precipitate in vitro binding reactions, we observed that upon increased levels of AR-S (a-d) there was a corresponding decrease in Tcf-S 3 5 (a-e) precipitated with the HIS-tagged recombinant p-catenin. Although A R - S 3 5 levels up to 10 ml were added per binding reaction, we observed no further increase in detectable AR/P-catenin-HIS binding greater than 6 | i l (Figure 4.6 c, point d) or 60% of input by volume. Total binding reaction volumes were kept constant using nonprogrammed (NP) reticulolysate. A volume of 10 ml of each probe is shown in Figure 4.6c (right), indicating that each probe is efficiently labeled and with molecular weights easily separated by SDS-PAGE. Given the apparent capacity of A R to deplete Tcf sites of P-catenin binding and the reduction in cyclin DI transcription and protein levels, we assessed whether this translated into altered cell cycle in SW480 cells. To evaluate this, we used FACs analysis of A R transfected and nontransfected SW480 cells treated with increasing concentrations of DHT (Figure 4.7a). In nontransfected cells, we observed a basal G l level of 43.2%, which increased to 57% upon addition of 10 nM DHT for 24 h. In cells transfected with AR, resting G l was 53.1% which increased to 69.8% at maximum DHT concentrations. Therefore, cells containing AR+ DHT 135 showed a sub G l increase of approximately 26.4% over mock-transfected and untreated cells. Over the titration of DHT, sub-Gl levels remained relatively constant, suggesting that cell death or apoptosis was not occurring, while S phase population decreased as G l levels increased. To evaluate whether detected arrest translated into decreased cell proliferation SW480 cells was evaluated by in vitro mitogenic assay (Figure 4.7b). Assessing treatments of vehicle alone, 10 nM DHT, A R + 10 nM DHT, and A R + 10 n M DHT + 1 pJVI Casodex, we observed counts that were consistent with transcriptional and cell cycle data. Treatment with androgen alone promoted a 12% decrease in cell proliferation, while cells transfected with A R and treated with DHT showed a reduction of 37% over cells treated mock transfected and treated with vehicle alone. With clear evidence that increased levels of AR, accompanied by androgen treatment, could reduce cell proliferation we tested the influence of androgens on capacity of SW480 colony formations. While these experiments were performed using transient transfection of AR, the high mitotic index of SW480 cells allowed us to assess alterations in colony formation within 6 days (Figure 4.8). Western blot analysis confirmed that SW480 cells still expressed a moderate amount of transfection D N A even after 6 days (data not shown). SW480 cells treated with 10 n M DHT for 6 days showed a 2.5-fold reduction in total colony numbers as compared to vehicle treated cells. Cells transfected with full length hAR prior to plating on soft agar indicated a 5-fold reduction in colony number. Interestingly, cells transfected with full length A R with no androgen treatment elicited a similar suppression of colony formation to cells delivered with the A R A L B D mutant (1.4-fold). Given that we use a transient mode of A R delivery, it is conceivable that stable integration of A R would result in a much more potent inhibition. The reciprocal of P-catenin between A R and Tcf sites upon acute exposure of ablation of androgens is shown schematically in Figure 4.9. Upon exposure to androgens, AR/p-catenin undergo nuclear co-trafficking and associated with sites of A R activation. Upon acute androgen ablation, residual P-catenin is predominately associated with Tcf, thus accounting for decreased Tcf signalling (Figure 4.9). 136 Figure 4.1 A R represses P-catenin/ function both dose and ligand dependently a) PC3 cells transiently transfected with A R and treated with 10 n M DHT for 24 hours show inhibition of Topflash activity in a dose-dependent manner, but does not affect P-catenin or Tcf4 protein levels, b) The A R - D B D / L B D deletion mutant was sufficient for Tcf repression, while A R - L B D is required for repressive activity. Casodex can relieve repression in a dose-responsive manner (1 and 10 uM). c) Increasing amounts of DHT (0, 0.01, 0.025, 0.05, 0.075, 0.1, 1 and 10 nM) promoted increased ARR3-Luc activity; while d) the same titration of DHT decreased Topflash activation both in SW480 and PC3 cells, (b-d). 138 Figure 4.2 A R and androgens provide potent negative regulation in cells with hyperactive P-catenin/Tcf signaling, a) Transiently transfected SW480 cells with A R and either WT P-catenin or an activating P-catenin mutant (Al-130 bp) treated with 10 n M DHT showing Topflash repression, b) SW480 cells transfected with the cyclin DI reporter show activation by WT P-catenin, but inhibition upon introduction of AR. c) Cyclin DI protein levels decrease with increasing DHT concentration in AR-transfected SW480 cells, but no change in P-catenin levels was detected. A R levels showed slight stabilization with higher levels of DHT treatment. In all cases, 2 u,g of A R expression plasmid was transfected per six wells. 139 a) b) Cyclin 01-Luc AR DHT C) WB: AR WB: Cyclin 01 WB: p-Cat WB: Actin OnM [DHT] 10 nM 140 Figure 4.3 a) Distribution of HcRed-Tcf and AR-EGFP foci in transfected LNCaP cells treated with 10 n M DHT viewed in a single xy plane without and with deconvolution (DECON). b) HcRed-Tcf and AR-EGFP foci colocalized (yellow) in approximately 35% of the fusion signal. Sites that did not colocalize (i, arrows) or did codistribute (i, i i , arrowheads) were used to quantitatively evaluate the degree of focal colocalization using manual counting, Photoshop image software and Northern Eclipse Imaging software. Baseline values were set using non-colocalized (R, G) and colocalized (Y) foci (iii) to evaluate P-catenin-EGFP flux to either HcRed-AR or HcRed-Tcf with (+DHT) or without (-DHT) 10 n M androgen, c) LNCaP cells were transfected with either (Ci) (P-catenin-EGFP + HcRed-AR) or (Cii) (P-catenin-EGFP + HcRed-Tcf). Cells were treated with or without 10 nM DHT and evaluated for the degree of colocalization (Y) as compared to non-colocalized (G or R). Bar 5 p,m. 142 Figure 4.4 Colocalization of Tcf-HcRed with phosphorylated (Ser 2) R N A Pol II in the presence of DHT (SW480 cells) suggesting that approximately 30% foci could be sites of transcriptional activity, Bar 5 uM. Fusion Tcf immunostain HcRed-Tcf (DECON) • V r v • • P-RNAPol II (DECON) • > 1 * 144 Figure 4.5 a) Confirmation of expression and transcriptional activity of AR-EGFP and HcRed-Tcf fusion expression vectors in LNCaP cells and SW480 cells treated with and without 1 n M R1881. Bands indicate Western Blots for either A R (top panels) or Tcf (lower panels). In transfected cells, the A R and Tcf fusions appear as doublets, b) Transcriptional reporter assays show androgen-dependent activity of the A R fusion in LNCaP cells and little nonspecific (Fopflash) activation with the HcRed-Tcf plasmid in SW480 cells, c) SW480 cells transfected with AR-EGFP and, as a control, the null HcRed vector (CMV-HcRed) illustrating that Tcf foci do not occur in the presence or absence (not shown) of androgen. a) LNCaP R1881: + . + AR-EGFP SW480 R1881: AR-EGFP R1881: + - + . R 1 M 1 . + HcRed-Tcf 0 HcRed-Tcf 0 b) LNCaP AR T AR-EGFP RLU 1 • . I m I * ARRS-Lue: + + + + + + R1881: + - + + SW480 RLU TOPFLASH FOP FLASH LNCaP 146 Figure 4.6 Evidence from binding assays that A R and Tcf compete for P-catenin. a) Increased amounts of transfected WT Tcf result in decreased ARR3-Luc activity in SW480 while the non-P-catenin binding Tcf mutant (ANt Tcf) did not show this result. Expression, both for the Wt and mutant Tcf, is shown by Western blot, b) SW480 cells treated with DHT and immunoprecipitated for A R showed more associated p-catenin (arrow), while cells transfected and immunoprecipitated for HIS (Tcf-HIS) showed less associated P-catenin (star). Casodex (5 [im) was able to abrogate the ligand-dependent formation of an AR/P-catenin complex (solid arrowhead), while also reducing AR-mediated depletion of the Tcf/P-catenin complex (hollow arrowhead). Both ANt-Tcf-myc/P-catenin and AR/Tcf-HIS complexes were weakly detected, c) In vitro binding assays were immunoprecipitated for P-catenin-HIS with constant levels-of Tcf -S 3 5 , P-catenin-HIS and increasing levels of A R - S 3 5 . Higher levels of A R resulted in decreased Tcf-S 3 5 associated with P-catenin-HIS/Ni-NTA bead complexes, suggesting that, in a pool containing these three species, at a certain concentration A R has the ability to compete for P-catenin. Non-programmed (NP) reticulolysate was used to maintain a constant total pool of lysate. ARR3-LUC + + + + + AR + + • + + Tcf 0 1 2 4 8 « - B U r - H - T c f AN* Tcf 8 ^ I ANt Tcf DHT + + + + + + WB: Tcf4 WB: AR b) Input (10%) DHT IP: HIS (Tcf) WB: jScat (Casodex) •DHT -DHT V IP: Myc (ANt-Tcf) WB: Beat IP: AR WB: /Beat IP: HIS (Tcf) WB: AR Binding f * R , 5 S l T c f B S 1 2 4 6 8 10 9 8 6 4 2 0 10 10 10 10 10 10 ul 10uJ A R C S T c f « » IP: BCat-HIS 148 Figure 4.7 Effect of AR-mediated repression of Tcf signaling on cell cycle and viability ai) SW480 cells were transfected with A R and treated with increasing DHT concentrations (0, 0.01, 0.025, 0.05, 0.075, 0.1, 1 and 10 nM). Cells transfected with A R and treated with DHT showed an increase of 26.2% in G l population over non-AR transfected cells, aii) A R western blotting indicates that cells used for cell cycle analysis were efficiently transfected. FACs analysis of cells transfected with a GFP expression plamid indicates approximately 70% of cells were GFP positive (data not shown), b) In vitro cell viability as measured by M T T incorporation, showed a decrease of 37% in cells containing exogenous A R and treated with 10 n M DHT over mock-transfected cells treated with vehicle. Casodex (5 \xm) abrogated the repressive effects of AR/DHT on SW480 cell growth. ai) 80 70 | 00 £ 50 a 40 -I 30 149 aii) SW480 AR + DHT + DHT hAR: + 0 .01 .025 .05 .075 .1 1 10 WB: AR WB: Actin DHT(nM) b) •S3 D H Vehicle l- DHT - AR + DHT + Casodex AR + DHT o J31 025 05 XI73 0.1 1 10 100 DHT(nM) 150 Figure 4.8 Soft agar colony forming assay showing the effects of AR, androgen and A R antagonist treatment on the number of SW480 colonies formed. Cells were seeded at a concentration of 10,000/ml in 0.4% agarose/media over a base layer of 0.4% agarose/media. Cells were treated for 7 days with the number of colonies being counted using the "Dave-Grid" technique and averaged between 6 x 35 mm 2 wells. Colony formation evaluated for stastistical significance by paired t-test with WT/DHT conditions differing from WT/0 and A L B D + DHT treatments at p < 0.05 and 0/0 treatment differing from WT/DHT at the p < 0.02 level of confidence. SW480 Colony Forming Assay Control AR + DHT 152 Figure 4.9 Mechanisms by which androgens may modulate cellular localization and expression of E-cadherin/P-catenin and Tcf activation, (a') Subsequent to A R mediated nuclear entry of (3-catenin, competition between A R and Tcf could occur for nuclear P-catenin. In the absence of androgens, cells with activated Tcf signaling, could show activation of Tcf gene targets, (a") In the presence of androgens, the Tcf/P-catenin signaling axis would be lessened due to depletion of P-catenin, while AR-mediated targets would be activated. a') Acute androgen ablation a") Acute addition androgens 154 4.4 Discussion In this study, we determined that A R can repress P-catenin/Tcf transcriptional activation both in PrCa cells and in colon cancer cells. We further validated the necessity of the AR ligand-binding domain by way of deletion mutants and followed that Casodex can relieve this inhibition. By using red and green fluorescent fusions in conjunction with in vitro and in vivo binding studies, we provide strong mechanistic data supporting a reciprocal balance of nuclear P-catenin between A R and Tcf sites. Specifically, we showed a novel distribution of Tcf and illustrated its partial colocalization with ligand-occupied A R by way of high-resolution microscopy and image analysis software. By costaining with phosphorylated R N A Pol II, we showed that Tcf foci are likely dynamic with the capacity for transcriptional activation. Increased p-catenin-EGFP localization to HcRed-AR sites, but decreased association with HcRed-Tcf foci in the presence of DHT, is strongly suggestive of a limited pool of nuclear P-catenin that has the capacity of shuttling to the more predominating transcription factor. Our data are also novel by linking these transcriptional data to changes in the cell cycle and proliferation. By observing declined transcriptional levels of Tcf, and the immediate Tcf target cyclin DI , we were able to link A R and DHT to cell cycle quiescence in cells with hyperactive P-catenin/Tcf signaling . Functional localization of AR, Tcf4 and P-catenin Previous reports have shown P-catenin to form nuclear morphologies that have included both punctate foci (Simcha, Shtutman et al. 1998) and rods (Simcha, Shtutman et al. 1998; Giannini, Vivanco et al. 2000). It has also been shown that A R forms a mixture of both transcriptional and nontranscriptional foci in the presence of agonist, as seen with high-resolution three-dimensional microscopy (Tomura, Goto et al. 2001). Other examples supporting a physiological role for foci include reports that nuclear p-catenin has the capacity to interact with and affect transcriptional regulation of the promyelocytic leukemia (PML) gene (Shtutman et al., 2002). In fact, recently, P M L has also been shown to functionally associate with A R (Rivera, Song et al. 2003), a finding not surprising when considering what is known about the interaction of A R and P-catenin. It has been further suggested that cofactors such as CBP and 155 SRC are necessary components of A R transcriptional foci and that most other steroid receptors form foci as a prerequisite to transcription (Saitoh, Takayanagi et al. 2002). Proteins forming nuclear foci are thought to be localized to protein complexes in a highly dynamic manner, whereby exposure to ligands can alter their distributions (Saitoh, Takayanagi et al. 2002) (Shtutman et al., 2002). Therefore, given what is understood about the dynamic and transient nature of nuclear receptors and cofactors interaction, it is also not surprising that the A R and Tcf could engage in a reciprocal balance of P-catenin. While colocalization data implies interaction, it is in fact, only informative to the micrometer level and does not necessarily mean protein-protein interaction. While we observed a degree of colocalization between AR-EGFP and HcRed-Tcf, we did not detect strong physical interactions. Physical AR/p-catenin and Tcf/P-catenin complexes were, however, easily detected. Considering this, one possible interpretation for the close association of A R and Tcf is to facilitate sharing of common transcriptional coactivators and/or repressors including P-catenin. The combination of morphological data with in vitro and in vivo binding studies provides compelling evidence for nuclear competition for p-catenin. Despite the evidence for competative balance of P-catenin between A R and Tcf nuclear deposits, other published reports provide a potential an alternative explanation for the repressive effects that A R confers upon P-catenin/Tcf signalling. While the report by Balk and colleagues confirmed that A R promotes repression of P-catenin/Tcf signalling, this study also indicated that A R and Tcf can bind and interact at c-myc promoter sites in CV-1 and LNCaP cells (Amir, Barua et al. 2003). Amir et al indicated that the A R - D B D can directly bind Tcf, thereby suggesting the potential for a balance of AR/Tcf and 3-catenin/Tcf dictating the signalling outcomes of Tcf and A R signalling, though this mechanism was not overtly demonstrated. The partially oppossing results of this study to those of ours may be explained by the fact Amir et al made use in vitro binding techniques (GST binding assays) and chromatin immunoprecipitations while our study made use of in vivo, complexing assays and light level data. Given that other nuclear receptors have been shown to trans-repress the 3-catenin/Tcf axis, it would be interesting to evaluate their potential for interactions with Tcf. 156 Wnt in prostate versus colon cancer cells In this study, we used colon cancer cells to evaluate the potential repressive ability of A R on Tcf/Lef signaling. Colon cancer cells, perhaps, provide the ideal means of testing the potential role of A R to interact with the Wnt pathway, as SW480 cells have hyperelevated |3-catenin/Tcf signaling , but low levels of endogenous AR. Constitutive Tcf signaling in colon cancer cells is due to defects in the Wnt pathway including defects in the A P C protein and P-catenin deletions, resulting in a mainly nuclear P-catenin distribution (Easwaran, Pishvaian et al. 1999). When comparing the effects of A R + DHT on Tcf signaling, we observed similar degrees of repression in SW480 cells (~4.5-fold) and in PC3 cells (3.5 to 4-fold). The physiological significance of A R repression of the Wnt pathway was assessed using a cyclin DI transcriptional reporter, cell cycle analysis and cell growth assays. In colon cancer cells, the cyclin DI reporter showed an A R and DHT dose-dependent repression similar to Topflash counts. Our data are also corroborated by a recent report showing that the AR/p-catenin interaction is androgen, dose-dependent (Song, Herrell et al. 2003). Paradoxically, we were unable to observe a reduction in the cyclin DI reporter in PrCa cells. We interpret this apparent discrepancy between colon cancer cells and prostate cells in several ways. In colon cancer cells, a predominating Wnt signaling pathway drives cyclin DI . In PrCa cells, Wnt signaling is much less prominent and, therefore, cyclin DI activity is not constitutive. Furthermore, studies have shown cyclin DI not to be a reliable marker of A R (Tomura, Goto et al. 2001). This is unlike the situation in many breast cancer cell lines where cyclin DI has been shown to be a reliable marker for cancer prognosis and has been shown to be correlated with estrogen receptor expression (Barnes and Gillett, 1998). In fact, some studies have shown that elevated cyclin DI levels can in turn inhibit transcriptional activation of A R in a cell cycle-independent manner (Knudsen, Cavenee et al. 1999). An initial stimulation of cyclin DI by A R could result in a feedback mechanism whereby cyclin DI inhibits AR. Considering this, it is possible that an initial repression of cyclin DI occurs (via the Wnt pathway), but is followed by cyclin dependent inhibition of AR. While Tcf activity is not high in PrCa cells, the function of p-catenin as an A R transcriptional coactivator suggests an important role of the Wnt signaling in PrCa progression. PrCa cells show indirect dysregulation of Wnt via PTEN -/-resulting in elevated Akt levels, reduced GSK3P and increased P-catenin (Cairns, Okami et al. 157 1997) (Davies, Koul et al. 1999), which can reside both in the cytosol and in the nucleus (Chesire and Isaacs 2002). Therefore, lesser Tcf activity in prostate cells could be accounted for by distribution of p-catenin, rather than the total amount of cellular P-catenin. Considering this, molecules that are capable of changing P-catenin distribution, such as the A R (Mulholland, Cheng et al. 2002), would have the capacity to alter Wnt signaling . With this large pool of cytosolic P-catenin, there is increased opportunity to associate with other binding partners such as the A R which, in the presence of DHT, will cotraffic to the nucleus (Mulholland, Cheng et al. 2002). AR-driven nuclear P-catenin could then also be directed to AREs and potentially deplete Tcf-binding sites. It is also possible that extracellular stimulation by Wnts may cross regulate with A R signaling . Recent reports have indicated a mechanism of counter regulation between A R and Wntl 1, a Wnt that has been shown to be elevated in hormone-independent PrCa cells (Zhu et al., 2004). W n t l l , a non-canonical Wnt, increases in expression in hormone depleted media and inhibits A R transcriptional activation. Interestingly, W n t l l inhibits the ability of Wnt3a to activate canonical Wnt/P-catenin/Tcf signaling in Hek293 cells and represses basal P-catenin/Tcf activity in LNCaP cells (Zhu et al., 2004). The means by which W n t l l can antagonize A R signaling could be explained by the finding that non-canonical Wnts can antagonize canonical Wnts (Torres, Yang-Snyder et al. 1996; Ishitani, Kishida et al. 2003). Wnts are expressed both in normal and cancerous cells PrCa cells. As such, the function of Wnts may facilitate not only oncogenic regulation but also promotion of developmental growth. Similar to W n t l l , Wnt5a also inhibits canonical Wnt signaling and, thus, its overexpression could retard cell growth and proliferation. Use of ectopic (recombinant) Wnts, that are known to either promote or antagonize canonical Wnt signaling , may serve to discern the function of the observed genetic alterations. AR-mediated repression of Tcf is a nuclear event Several lines of evidence suggest that AR-mediated repression is a nuclear event. 1) Cells transfected with an activating P-catenin mutant showed a fivefold increase in Topflash activity over basal levels both in prostate and colon cancer cells (data not shown). However, even under these circumstances, introduction of AR/DHT could efficiently reduce Tcf signaling. The fact 158 that Tcf repression occurs when A R accumulates in the nucleus, that is, when A R is exposed to DHT, is suggestive of a nuclear event for Tcf silencing rather than a degradative one. (2) By using a cell line that is APC-/- , we rule out a main mode of P-catenin degradation (ubiquitination). If the actions of androgens involve targeting cytoplasmic P-catenin for proteosome degradation, we would have observed changes in levels of cytoplasmic, nonphosphorylated P-catenin in cells containing functional APC. (3) Despite the fact that Casodex interferes with A R transcription, it has been well documented to promote nuclear localization of the A R (Waller, Sharrard et al. 2000). Importantly, in our study of Casodex, we observed a dose dependent relief of AR/DHT-mediated Tcf repression. (4) The reduction of A R transcription (in the presence of DHT) upon overexpression of Tcf provides evidence of competition. These data are consistent with the notion that modulation of the P-catenin-Tcf/Lef assembly may be the mechanism by which A R exerts its repression on Tcf/Lef signaling. To test this further, we evaluated the amount of P-catenin associated with the Tcf/Lef complex with and without AR/androgen, both in vivo and in vitro, results that corroborate morphological data and our overall hypothesis. It is interesting how two stimulators of cell proliferation (AR/DHT and Wnt) can interact in a "counter-repressive" manner. One possible interpretation is that androgens may not increase proliferation but, rather, promote growth and differentiation in prostate epithelia. PrCa cells have cell cycle deregulation compounded by a cell survival response to androgens. The contributions of Wnt signaling in PrCa is likely complex, although it is clear that nuclear P-catenin can serve as a potent A R coactivator. We suggest a scenario whereby P-catenin could be shuttling between Tcf/Lef-binding sites and A R elements. In the absence of androgen, A R resides mainly in the cytosol, while nuclear P-catenin associates with Tcf. In the presence of androgen, P-catenin could be shuttled by translocating nuclear receptors to both Tcf- and AR-associated response elements to promote coactivation (Figure 9a). Consequently, less P-catenin would be associated with Tcf and more with AR. This would simultaneously lower Tcf activity and augment A R transactivation. Such a model is also considerate of other reports of nuclear receptor repression of Tcf, including the R A R (Easwaran, Pishvaian et al. 1999) and the thyroid receptor (Miller, Park et al. 2001), two hormone receptors that are expressed constitutively in the nucleus (Figure 4.14). (See Appedix Review: "Interaction of Nuclear Receptors with the Wnt//3-catenin/Tcf Signaling Axis: Wnt you like to know"). 159 In conclusion, our data provide strong support that the A R can inhibit the P-catenin/Tcf signaling axis by competing away nuclear P-catenin from our Tcf sites. Our findings both corroborate initial reports (Chesire and Isaacs 2002) showing that A R can inhibit Tcf transcriptional activity and augment these observations with mechanistic data showing that Tcf and A R entertain a reciprocal balance of nuclear P-catenin. While Wnt signaling is clearly different, in many aspects, between prostate and colon cancer cells, it is possible that ligand-occupied A R could be important therapeutically in cancers with hyperactive Tcf/Lef transactivation or where this pathway contributes to cancer progression. Interestingly, some reports have even shown differences in A R expression levels between cancerous (less AR) and noncancerous (more AR) colon tissue (Catalano, Pfeffer et al. 2000). It is tempting to speculate that pathologies with low levels of RAR, thyroid receptor (Miller, Park et al. 2001) or A R could be a prosurvival adaptation for cancer cells to allow for increased Tcf signaling and, therefore, survival and cell proliferation. Chapter Five: P T E N is a Potent R e g u l a t o r o f P-catenin/Tcf a n d A R S i g n a l i n g via the G S K 3 P A x i s 161 5.1 Introduction Both androgens and the PBK/Akt signaling axis function to protect PrCa cells from apoptosis during androgen deprivation. This may be achieved, in part, by growth factor activation of downstream pro-survival effectors including P B K , Akt, integrin linked kinase (DiGiovanni, Kiguchi et al.), S6-Kinase, mTOR and cell cycle regulators, including cyclin DI . P B K signaling also functions to inhibit molecules with tumour suppressor functions including glycogen synthase kinase p (GSK3(3), B A D , caspases and the Forkhead (FKHR) transcription factor. Furthermore, P B K / A k t signaling inhibits the negative cell cycle regulators including KTP C IP p27 and p21 , thus preventing cell cycle arrest and apoptosis (Vivanco and Sawyers 2002). Activation of Akt also leads to inhibition of GSK3P, a key regulator both of the Wnt and P B K signaling pathways (See Appendix for Review Article: PTEN & GSK3/3: Key Regulators for progression to AI PrCa). Under stimulation by either Wnt or P B K / A k t activating growth factors, GSK3p is inhibited by Dishevelled or by direct phosphorylation by Akt (Ding, Chen et al. 2000). Thus, cells activated both by Wnt and P B K signaling experience a "double" selection against the activity of GSK3p\ Given that P-catenin is regulated both by Wnt and PBK/Akt signaling pathways, simultaneous activation of these pathways results in accumulation of P-catenin. The tumour suppressor, PTEN is a dual specific phosphoprotein/phospholipid that is mutated in primary PrCa and most PrCa cell lines (McMenamin, Soung et al. 1999). PTEN antagonizes P B K signaling by inhibiting Ser 4 7 3 and Ser 3 0 8 phosphorylation of Akt (Davies, Koul et al. 1999). It is also known that PTEN antagonizes A R transcription and androgen mediated cell growth of PrCa cells (Li, Nicosia et al. 2001) (Nan, Snabboon et al. 2003) (Lin, Hu et al. 2004). The use of PrCa mouse models will likely provide important clues as to whether in vivo loss of PTEN is a direct causal agent for promoting AR-specific gene activation. However, until such mouse genetics are accurately carried out, important observations gained by in vitro studies will continue to de.ne the relationship between PTEN and AR. In vitro observations by L i et al. (Li, Nicosia et al. 2001) suggest an antagonistic relationship, as PTEN can negatively regulate A R gene targets including expression of PSA (Li, Nicosia et al. 2001). Forced expression of PTEN can also reduce nuclear localization of A R and promote receptor degradation by way of caspase-3 or the proteosome system (Lin, Wang et al. 2002). As a caveat, such studies have 162 indicated that expression of PTEN, in the presence of androgens, is suf.cient to reduce cell proliferation but not to induce an apoptotic response (Li, Nicosia et al. 2001) (Nan, Snabboon et al. 2003) (Lin, Hu et al. 2004).. These data imply the presence of AR-regulated 'survival' genes that may function independently of PI3K/Akt signalling. Identication and therapeutic targeting of such genes could prove effective in sensitizing tumours to inhibitors, such as rapamycin, which target the PI3K/Akt pathway (Sharp and Bartke 2005). While it is probable that during PrCa progression A R and PI3K/Akt signaling form a necessary and, likely synergistic interaction (Xin, Ide et al. 2003) (L Xin & ON Witte, personal communication, David Geffen School of Medicine, U C L A ) , it is less clear how gain or loss of PTEN functions to alter A R activity. Given that PTEN regulates the integrity of the PI3K signaling pathway and the understanding that GSK3(3 pools function to regulate both Wnt and PI3K signaling , we propose that PTEN regulates the activity of AR, and therefore, P-catenin by modulation of GSK3p\ In this study, we demonstrate that upon re-expression of the wild-type PTEN (PTEN-WT), GSK3 activity is increased, leading to an increase in P-catenin phosphorylation, increased p-catenin degradation, and subsequent decrease in the expression of the protein in the nucleus. Transfection of PC3 PrCa cells with the dominant negative form of I L K (kinase-deficient ILK) or WT GSK3P also suppressed the elevated expression of nuclear p-catenin. Furthermore, PTEN, I L K - K D , and WT GSK3P induced comparable suppression of Tcf/P-catenin transcriptional activity in PC3 cells. In correlation with this, we noted that cyclin DI expression and promoter activity, which are constitutively elevated in PC3 cells, are suppressed upon re-expression of PTEN or expression of I L K - K D and WT GSK3P. These results delineate a novel role for PTEN in the regulation of cell growth and proliferation via its regulation of nuclear P-catenin and cyclin DI . To validate our hypothesis we have made use of both gain-of-function and loss-of-function PTEN systems. Taking advantage of the PTEN null background that exists in the commonly used PrCa cell lines, LNCaP and PC3, we have developed stable, inducible PTEN expressing PrCa cells. The development of these cells was technically difficult, likely as a result of the PI3K/Akt dependency of LNCaP and PC3 PrCa cells. However, we have provided a novel and useful tool by which to evaluate the critical role that PTEN expression may serve over longer (days-weeks) durations. The use of doxycycline inducible systems allows for a controlled 163 analysis of the effects of WT and catalytic mutant PTEN induction. Using a combination of clonal selection, lentiviral infection and a tet response element silencer, we have created relatively non "non-leaky" system in which to analyze both |3-catenin/Tcf and A R function. We have also taken advantage of a recently developed mouse model with prostate specific deletion of PTEN (see Chapter 6). Using these novel gain and loss of function PTEN systems, we will address our hypothesis that PTEN can negatively regulate the cellular functions of P-catenin/Tcf and AR. We also provide further data to support and validate that GSK3P is critical in regulating the means by which PTEN modulates the functions of P-catenin and AR. 5.2 Methods 5.2.1 Cell culture & transfections The PTEN null PrCa cells lines PC3 and LNCaP (ATCC) were used for studies in this chapter. PC3 cells were cultured in D M E M containing 10% FBS while LNCaP cells carried in RPMI-1640 media supplemented with 10% FBS. A l l cells were passaged in 5% C 0 2 at 37 °C. Transfection of PC3, LNCaP, LNCaP-PTEN/C124S and PC3-PTEN/C124S cells was carried out using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions but generally entailed the use of 2 (il of lipofectin for each |ig of delivered D N A . For transient transfections of PC3 or LNCaP cells, the efficiency was determined by flow cytometric analysis of cells by co-transfection with eGFP cDNA. Efficiency of PTEN-GFP transfection was determined without co-transfection with eGFP. Cytometric analysis was interpreted using WinMDO software. 5.2.2 Western blotting Equal protein concentrations were resolved by 10% SDS-PAGE with proteins being transferred to polyvinylidene difluoride membrane (PVDF). Western blotting analysis was carried out using enhanced chemiluminescence detection reagent as per manufacturer's instructions. Antibodies used in this study are listed in the Appendix of this thesis. 164 5.2.3 Immunocytochemistry PrCa cells were immunostained and cultured in Lap-Tek chamber slides (Nunc) 24 hours before transfection on 25 mm 2 glass coverslips. Cells were transfected for 12-16 hours with 3 u,g using lipofectin according to the manufacturer's guidelines. Cells were fixed and permeabilized 24-48 hours later using cold (-20 °C) methanol for 2-3 min followed by 15 min of air drying. Cells were treated with blocking buffer (4% NGS/0.1% BS A/0.005% tween/PBS) for 20 min followed by 1-2 hours incubation with primary antibodies diluted in 1%NGS/0.1%BSA70.005% tween/PBS as specified in the Appendix. Primary antibodies used in this chapter included (3-catenin, PTEN, N-cadherin, Lef-1, Tcf4, GFP, ILK and histone. 5.2.4 Luciferase reporter assays Luciferase transcriptional assays were carried out using the minimal A R responsive reporter (ARR3-Luc), the Topflash reporter construct (Upstate Biot, Cat # 21-170) containing three tandem Tcf4 binding sites, and the corresponding Fopflash mutant (Upstate Biot, Cat #21-169). For PC3 and PC3-PTEN cells the TK-Renilla Luciferase transfection control plasmid was co-transfected with the Firefly Luciferase expressing constructs. Firefly luciferase values were normalized to the results of Renilla Luciferase with standard deviations being generated for each treatment. For LNCaP and LNCaP-PTEN cells the Renilla Luciferase control plasmids proved to be androgen responsive and, thus, were not used to control for transfections (see Mulholland et al, The Prostate, 2004). Rather, replicates per experiments were doubled in order to increase our confidence of transfection efficiency. For transfections, 2x l0 5 PC3 cells were seeded per well of a 6 well plate while 5xl0 4 cells were plated/well of a 12 well plate. Cells were transfected for 12-16 hours using lipofectin as per manufacturer's guidelines and replaced with either full media (10% RPMS or 10% D M E M ) or androgen depleted media (CSS) with or without low (0.1 n M R1881) or high androgen (InM R1881). 5.2.5 Nuclear extracts and electrophoretic mobility shift assays Nuclear extracts were prepared by the miniextraction method as described previously (Andrews and Faller 1991). PC3 cells transfected with empty vector (control), ILK-Kinase Dead 165 (KD), or PTEN-WT were washed with ice cold PBS and harvested by being scraped in 1.5 ml PBS. Cells were then pelleted and resuspended in 400 | i l of 10 m M Hepes-potassium hydroxide, pH 7.9, 1.5 m M magnesium chloride, 10 m M potassium chloride, 0.5 mM dithiothreitol, 0.2 m M PMSF. After 10 min of incubation on ice, nuclei were pelleted and resuspended in 50 [i\ of 20 m M Hepes-potassium hydroxide, pH 7.9, 25% glycerol, 420 mM sodium chloride, 1.5 m M magnesium chloride, 0.2 m M EDTA, 0.5 m M dithiothreitol, 0.2 mM PMSF. Tubes were incubated for 20 min on ice and then centrifuged to clear cellular debris. Nuclear extracts were stored at -70 °C. Electrophoretic mobility shift assays were performed using 4 \xg of the nuclear extracts for 20 min at room temperature with a P 3 2 end-labelled D N A fragment containing the putative protein binding site, a Tcf binding element obtained from the cyclinDl promoter within the cyclin Dl-luc reporter construct (5'-TGC C G G GCT TTG ATC TTT GCT-3'). [y- 3 2P]ATP was from Amersham Pharmacia Biotech. Reaction products were analyzed on a nondenaturing 5% polyacrylamide gel (0.5% Tris-borate- EDTA, 3.5% glycerol). The specificity of the DNA-protein interaction was established by competition experiments using lOx cold Tcf oligonucleotide as the competitor. For the supershift assays, 1 p.g of mouse anti-P-catenin antibody or nonspecific IgG was added to the reaction mixture, subsequent to the addition of the P32-labeled oligonucleotide probe. This mixture was incubated for 45 min at room temperature and complexes resolved by electrophoresis as described for the mobility shift assays. 5.2.6 Nuclear extracts and co-immunoprecipitation assay PC3 cells were transfected with an empty vector (control), I L K - K D , or PTEN-WT cDNA and harvested at 48 h post-transfection. The cells were washed in cold PBS and nuclear extracts were prepared by the mini extraction method as, previously described (Andrews and Faller 1991). Immunoprecipitations for P-catenin were performed using 400 |ig of nuclear lysate and polyclonal anti-P-catenin antibody. Immunocomplexes were isolated with protein A/G-PLUS agarose (Santa Cruz Biotechnology, Inc.) and separated on 7.5% SDS-PAGE gels. The gels were then Western probed with either monoclonal Tcf4 or monoclonal Lef-1 antibodies to detect the presence of these transcription factors in the immunocomplexes. 166 5.2.7 Pulse-chase analysis PC3 cells were cultured to 70-80% confluence and transfected with empty vector (control) or PTEN-WT. For each chase time point, the cells were washed once with D M E without cysteine/methionine (starve media; Sigma-Aldrich) and incubated in the starve media for 1 hour at 37 °C. Cells were pulsed for 1 hour at 37 °C using [ SJPromix (Amersham Pharmacia Biotech), 100 Ci/ml. Cells were then chased with DME/10% FBS for the indicated time points, harvested at the various time points, and lysed in RIPA buffer. Immunocomplexes were isolated with protein G Sepharose (Santa Cruz Biotechnology, Inc.) and separated on 7.5% SDS-PAGE gels, stained with Coomassie blue, destained; and incubated with Amplify (Amersham Pharmacia Biotech) fluorographic reagent; dried and exposed to film. 5.2.8 Flow cytometric analysis Cells were plated on 150 mm dishes; when the cells were 70% confluent, and Doxycycline (2 Lig/ml) was added to the medium. Two days later, all cells were harvested using 5 m M EDTA in PBS, centrifuged at 2000 rpm for 5 min, and fixed with 70% ethanol. Cells were treated with buffer solution (4 m M citric acid in 0.2M NaiHPO^), vortexed for 5 min; incubated with 0.5 mg/ml of RNAse (Sigma) at 37 °C for 30 min; re-centrifuged and stained with 1 ml of 50 ug/ml propidium iodide (Sigma). The D N A profile was analyzed using a dual-laser flow cytometer (EPICS X L - M C L , Beckman Coulter, Miami, FL). 5.2.9 GSK3 kinase assay GSK3 kinase activity was determined in cell extracts from PC3 cells transiently transfected with either empty vector (control), ILK-WT, I L K - K D , PTEN-WT or dominant negative Akt (Akt-DN). Cells were lysed in 50 m M HEPES buffer, pH 7.5, containing NaCl (150 mM), NP-40 (1%), sodium deoxycholate (0.5%), leupeptin (10 Ug/ml), PMSF (1 mM), aprotinin (2.5 uiVml), sodium fluoride (5 mM) and sodium orthovanadate (ImM). Equivalent protein concentrations of cell lysates (determined by Bradford Assay) were precleared with nonspecific IgG and protein A Sepharose. After centrifugation the supernatants were immunoprecipitated with anti-GSK3|3 antibody bound to protein A Sepharose beads. Kinase 167 assays were carried out as described previously (Delcommenne, Tan et al. 1998) using GS-1 peptide as a substrate. Phosphorylated GS-1 peptide was electrophoresed on a tricine gel, then visualized and quantified by autoradiography or phosphoimage analysis. 5.2.10 Development of P T E N stable clones: clonal selection Stable expressing cells for PTEN (WT and C124S) were developed using two approaches. The first method included clonal selection of cells transfected with three constructs including (1) a tet-on VP16 fusion (G418 resistance), (2) PTEN WT or PTEN C124S (hygromycin resistance), and (3) a tet response element repressor (puromycin resistance) (Figure 5.7A). In brief, cells were transfected with linearized D N A coding the VP16-Tet On activator and selected under 0.1-0.3 pg /ml G418 at a concentration of 0.3 p:g/ml. Clones were tested for their ability to activate a luciferase reporter driven by tandem tetracycline response elements (TREs). Clones displaying high reporter induction in the presence of 2 Ug/ml doxycycline and low induction in its absence were selected for incorporation of the WT PTEN or catalytic (C124S) transgene. By a second round of transfection with linearized D N A and selection using 0.1-0.3 pg /ml hygromycin Dox-PTEN/C124S, inducible clones were developed either in LNCaP or PC3 cells. To minimize "leakiness" of transgene expression or non-induced expression, a Tet-repressor (TR) was incorporated using lentiviral delivery. In brief, using viral packaging components (GP3 = Gag-Pol; V S V G = vesicular stomatitus repressor glycoprotein G) along with a plasmid coding the TR (gift from Dr. Robert Kay, Terry Fox Laboratories, Vancouver, BC) under puromycin selection (2 ug /ml), clones with excellent induction of PTEN were developed. Clones were evaluated by western blotting, immunocytochemistry and FACs analysis. 5.2.11 Development of P T E N stable clones: lentiviral system (Invitrogen) (a) Cloning: Creation of the pDEST4/TO/Lenti-PTEN/C124S was achieved by two Steps. Using the forward (5' G G G G A C A A G T T T G T A C A A A A A A G C A G G C T T A atg aca gcc ate ate aaa gag 3') and reverse ( 5 ' G G G G A C C A C T T T G T A C A A G A A A G C T G G G T C gac tu tgt aat ttg tgt atg 3') primers, the 1.1 kB PCR product of WT PTEN and C124S PTEN mutant containing overhanging attL and attR sites 168 was amplified from pUHD-PTEN (gift from M E Cox, The Prostate Centre, VGH) and gel purified. By BP reaction (attB => attP), the PTEN PCR product was inserted into the pDONOR201 vector (Invitrogen) according to the manufacturer's directions. 1 JLXI of the BP product was transformed into TOP10 chemical competent bacteria (Invitrogen, Cat # C4040-10) and plated on L B agarose plate with 50 pg /ml Kan selection. At least 5 colonies were picked, with plasmid D N A being isolated and verified by restriction digest and sequence analysis. Resulting plasmid D N A was then used for a LR reaction (attL => attR) according to the manufacturer's instructions. 1 p.1 of the BP product was transformed into Stbl 3 chemical competent bacteria (Invitrogen, Cat # C7373-03) and plated on L B agarose plate with 50 Ug/ml Amp selection. At least 5 colonies were picked, D N A isolated and verified by restriction digest and sequence analysis. (b) Generation of lentivirus: Lentiviral packaging plasmids (LP1, LP2 + VSVG) were transfected with the pDEST4/TO/Lenti-PTEN/C124S plasmid into 293T cells. Transfection media was removed and replaced with 5% Optimum (Gibco) and collected after 48 hours of virus production. The conditioned media was filtered using a 0.40 pm and applied to parental LNCaP or PC3 cells. LNCaP and C4-2 cells were found to infect at rate of -90%, while PC3 cells infected at approximately 10-15%. Because of this, LNCaP and C4-2 infected cells could be treated as "semi-clonal", while PC3-PTEN clones involved selection and testing of numerous clones. PTEN-PrCa cells obtained by transfected-mediated clonal selection have been found to maintain their degree of inductiveness for at least 12 months. PTEN-PrCa cells generated by Lentivirus have been used for up to 6 weeks of passaging and subsequently replaced by newer clones. 5.2.12 RNA preparation R N A was extracted from each 15 cm plate of PTEN/C124S stable LNCaP cells with 4 ml of Trizol. 200 pi of Phenol was added per 1 ml of Trizol and then hand-shaken for 15 sec and incubated at RT for 2-3 min. Centrifuge tubes were then rotated at 14,000 rpm for 10 min with the aqueous component separated and mixed with 500 pi of isopropanol for overnight precipitation. The RNA/isopropanol solution was then centrifuged for 10 min at 14,000 rpm. The R N A pellet was then washed with 70% EtOH/depC H2O, pulsed, air-dried for 10 min and then washed with depC H 2 0 169 5.2.13 Northern blotting Total R N A was extracted from cultured cells or xenografts using the acid-guanidinium thiocyanate-phenol chloroform method. Frozen samples were treated with Trizol and phenol-chloroform to extract total RNA. After separation by centrifugation the aqueous layer was precipitated using 90% ethanol, and washed using 70% ethanol. For analysis of specific message levels, a total of 20 u,g of R N A from each sample was subjected to gel electrophoresis (1.2% agarose/30% formaldehyde) and transferred to nylon membranes overnight according to standard procedures. P-catenin, AR, PTEN, and Tcf-4 R N A probes of approximately 500 bp in size were generated by restriction enzyme digest of expression plasmids. D N A was denatured at 95 °C for 5 min followed by a 1 hour incubation with 50 u,Ci of [ 3 2P]-dCTP and Ready Prime labelling system. Membranes were probed with pre-hybridization fluid for 1 hour (42 °C), followed by overnight incubation with the labeled probe (40 °C). After 2 x 1 5 min washes in washing buffer 1 (0.1 % SDS, 2x SSC, lx phosphate buffered saline) and 2 x 15 min washes in washing buffer 2 (0.05% SDS, I X SSC, 0.5X Phosphate Wash Buffer), blots were exposed to radiographic film for 1-20 hours. Analysis of message levels were normalized to levels of Glyceraldehyde-3-phosphate dehydrogenase levels (G3PDH) and were used to normalize expression of experimental probes. 5.2.14 Cell proliferation (MTS) assays MTS assays were covered as per Chapter 2 Methods. 5.2.15 AR-gfp, PTEN and DAPI cell counting LNCaP-PTEN/C124S and PC3-PTEN cells were transfected with WT hAR-GFP and treated with InM R1881 for 48 hours. Cells expressing WT hAR-GFP were assessed for the relative levels of PTEN expression in the presence of Dox. 100 WT hAR-GFP positive cells were counted per slide and scored for the measure of WT PTEN or C124S induced expression as 170 either low, medium or high in expression. In addition to determining whether the observed association between levels of PTEN and WT hAR-GFP were due to the cell apoptosis, the apoptotic index was determined by scoring 200 WT hAR-GFP positive cells for nuclear condensation and were scored as being either condensed (apoptotic) or non-condensed (non-apoptotic). 5.2.16 Colony forming assays Colony forming assays were carried out in 35 mm wells each containing between 0.3 x lO 4 and 4.0 x lO 4 cells/well. Over a base layer of 1.5 ml of 0.4% agarose in treatment media (bottom layer), a 1.5 ml solution (middle layer) of 0.2% agarose in treatment media containing either PC3-PTEN or LNCaP-PTEN was applied to evaluate the ability of PTEN to modulate colony formation. Cells were treated with or without 2 Lig/ml doxycycline in either full serum, androgen depleted serum or serum-free conditions. Cells were allowed to grow for 10-15 days and subsequently stained for at least 1 hour with 0.005% Crystal Violet. Using an upright Zeiss light microscope, colonies of at least 0.15 mm in diameter were counted. To facilitate counting of colonies, the "Dave-Grid" technique was employed as per Chapter 2 of this thesis. 10 squares from each grid (35 mm) was used for counting of colonies with 6 wells for calculating averages and standard deviations (S.D.). An average of at least 6 wells was considered along with SDs. 5.3 Results 5.3.1 Levels of nuclear (3-catenin are reduced by expression of exogenous PTEN in PTEN-null PrCa cells. In the presence of full serum conditions (10% FBS/DMEM), the PC3 PTEN-negative PrCa cells, maintain a moderate amount of nuclear P-catenin in approximately 70% of the cell population (Figure 5.1a). High levels of nuclear p-catenin were also observed in nuclear extracts from LNCaP cells (Figure 5.1b) and PC3 cells (Figure 5.2). Immunocytochemistry analysis indicates that cells transfected with PTEN-GFP show p-catenin to be mostly extra nuclear in localization with -16% expressing nuclear P-catenin (Figure 5.2). No change was observed in 171 cells transfected with increasing amounts of eGFP cDNA (Figure 5.2) while transfection efficiency in both PC3 and LNCaP cells was -80% as calculated by FACs (data not shown). Thus, transient transfection of increasing amounts of WT PTEN into PC3 cells (Figure 5.1a, c and Figure 5.2) and LNCaP cells (Figure 5.1b) leads to inhibition of nuclear expression of P-catenin both at the light level and protein level. Of interest was the observation that total cellular P-catenin was noticeably altered upon increasing amounts of transfected PTEN. Also of interest was the observation that PC3 cells, which do not express E-cadherin, under basal conditions contain moderate levels of N-cadherin and show dramatic induction of E-cadherin upon re-introduction of PTEN (Figure 5.2). 5.3.2 PTEN enhances degradation and phosphosphorylation of P-catenin To examine the stability of P-catenin as a function of PTEN expression, pulse chase analysis was used in PC3 cells that were transiently transfected with PTEN. Introduction of increasing amounts of exogenous PTEN resulted in a dramatic increase in the rate of degradation of p-catenin (Figure 5.3a). P-catenin degradation is associated with increased phosphorylation of P-catenin at serines 33/37 and threonine 41. Using an anti-phospho-P-catenin (Ser 33/37,Thr 41) antibody, it was clear that these sites were increased in phosphorylation upon increased expression of WT PTEN (Figure 5.3b). 5.3.3 PTEN expression inhibits P-catenin/Tcf complexing and binding of a P-catenin/Tcf complex to a Tcf consensus oligonucleotide. P-catenin is known to bind to Tcf transcription factors, thereby, activating Wnt/P-catenin/Tcf target genes. Thus, the potential for P-catenin to form complexes with a Tcf oligonuceotide was evaluated in the presence of WT PTEN or empty vector. To do this, PC3 cells were transfected with WT PTEN, with nuclear extracts of this preparation being assessed for nuclear localization of P-catenin. Using a gel mobility shift assay, it was revealed that a significant amount of protein-DNA complex was present in the mock transfected PC3 cells. However, the level of protein-DNA complexes in PC3 cells transfected with WT PTEN was significantly reduced (Figure 5.4a). To validate that these DNA-protein complexes contained 172 nuclear (3-catenin, a super-shift assay was performed in order to verify immunodetection of (3-catenin. Figure 5.4a, lane 3, indicates a shift upon the addition of 1 |ig of (3-catenin antibody which is depleted upon transfection with WT PTEN (lane 4). To assess in vivo complexing, (3-catenin immunoprecipitates were probed for the presence of Tcf4 and Lef-1 (Figure 5.4b). Under mock conditions, an abundance of both P-catenin binding proteins was detected. However, upon introduction of PTEN or a parallel PI3K regulator, I L K (KD= kinase deficient), considerably less Tcf4 or Lef-1 associated with precipitated p-catenin. 5.3.4 Generation of stable "Knock-In" PTEN expressing cells by two different methods With the knowledge that transient expression of PTEN in PTEN-null backgrounds can alter P-catenin/Tcf signaling we developed novel, stable PTEN expressing cells to further consider the effects that PTEN may confer upon the P-catenin/Tcf/AR signaling axis. Creation of stable PTEN expressing PrCa cells was a challenging prospect; however, using two approaches including clonal selection and lentiviral infection we have developed clones with potent, inducible expression of PTEN that have been maintained in culture for extended periods of time (months). The ability to express WT, or catalytic mutant (C124S) PTEN, expression in a controlled manner has provided a valuable tool in which to study the functional expression of target molecules such as p-catenin and AR. Clonally selected LNCaP-PTEN and LNCaP-C124S sublines both potently express PTEN upon the addition of 2 |ig/ml doxycyclin (dox) and express minimal non-induced ("leaky") expression of PTEN in the absence of dox (Figure 5.5a). Characterization of these cells lines have been performed using immunocytochemistry, Western blotting, Northern blotting and FACs analysis. It has, however, been observed that not all cells show an equal intensity of PTEN expression, suggesting the possibility that these cells are "semi-clonal". Despite this, FACs analysis indicates that upon Dox treatment, a significant proportion (>90%) of the of the cells show robust PTEN expression (data not shown). The use of lentiviral delivery of the PTEN/C124S transgene has also yielded a high percentage of the parental LNCaP population to express PTEN (Figure 5.5b). While the population of PTEN infected cells using the lentiviral system was high (>95%), the induction of PTEN expression was relatively less than with PTEN expressing cells expressors obtained by clonal selection. 173 5.3.5. Maintaining non-"leaky" PTEN expression using the Invitrogen lentiviral tet-repressor and a VP-16 activator fusion An extremely important aspect of generating cell lines with inducible expression of tumour suppressors is maintaining a "non-leaky" gene expression system. To achieve this we employed both a (a) "Tet-on" and (b) "Tet-off' repressor system. The VP16 tet-on fusion allows displacement of a co-expressing tet-repressor, while the Tet-off system is constitutive bound except in the presence of Dox. This allows for titratible displacement of a bound repressor. 5.3.6 Validation of functional LNCaP-WT PTEN and C124S-PTEN clones The LNCaP-PTEN/14M and LNCaP-C124S/5-4 clones were tested for appropriate induction of PTEN expression in various culture conditions including full serum, androgen depletion and serum free conditions. At high magnification immunofluorescence, PTEN localizes to a submembraneous position (arrow, Figure 5.7a). Low magnification immunofluorescence shows PTEN expression to be expressed in all cells; however, with the amount of PTEN varying between cells (Figure 5.7a). Efficient induction is apparent by Northern blotting for PTEN transcripts at 2 and 5 kb (Figure 5.7b). Western blotting for PTEN demonstrated a 25 fold induction as compared to non-induced expressors; thereby suggesting that the repressor system was working efficiently (Figure 5.7c). To verify the phosphatase activity of the LNCaP-PTEN/14M clone we assayed levels of phosphorylated P-Akt (Ser4 7 3) in serum free, androgen depleted and full serum conditions and observed efficient down regulation of PTEN in all three conditions (Figure 5.7c). Figure 5.7di validates that LNCaP-PTEN clones can induce cleaved PARP expression both as a function of serum conditions and time of doxycycline exposure. Our LNCaP-PTEN/14M cells clones demonstrated maximal induction of PTEN at 2 |ig/ml and at approximately 24-48 hours of 2 u\g/ml Dox treatment. Importantly, the induction of WT PTEN in serum free media efficiently promoted the potent induction of cleaved PARP, in a time dependent manner that closely resembled PTEN induction (Figure 5.7di). At 24 hours, intense expression of cleaved PARP was observed while the C124S lipid phosphatase mutant elicited virtually no induction of cleaved PARP (Figure 5.7dii). 174 5.3.7 The PI3K chemical inhibitor, LY294002, causes a dose dependent inhibition of AR transcriptional reporter activity in LNCaP-PTEN cells (p45) Our previous data suggests that Wnt/P-catenin modulation of A R is dependent upon GSK3(3 function (Chapter 2). As p-catenin serves as a potent regulator of AR, we extend our hypothesis to suggest that modulation of PBK/Akt signaling (i.e. P B K pools of GSK3P) can also alter A R signaling . Thus, to evaluate this hypothesis we used both induction of PTEN and the P B K / A k t chemical inhibitor, LY294002, both of which are know to inhibit PBK/Akt by similar means (Figure 5.8a). Upon treating cells with 1 n M R1881 and concentrations ranging from 0 to 10 p M LY294002 we observed a dose dependent decline (5.7 fold) in ARR3-luc reporter activity after 24 hours treatment. Induction of PTEN expression also resulted in a dose dependent decline (4.4 fold) in activity of ARR3-luc which closely correlated the concentration of administered Dox (Figure 5.8b). It is known that antagonizing P B K results in increased activity of GSK3P, a direct negative regulator of P-catenin. We chose to evaluate the effects of titrated PTEN induction on the activity of the Topflash reporter construct which contains tandem Tcf4 binding elements (Figure 5.8c). Importantly, induced PTEN had little effect on the mutant Tcf4 control reporter (Fopflash) (Figure 5.8d). These data indicate that PTEN has the capacity to reduce binding of P-catenin to Tcf/DNA complexes, either by increased nuclear export, P-catenin degradation or both. These data also demonstrate the advantage of using an inducible system whereby the amount of PTEN induction can be regulated in the manner of a biological "rheostat". 5.3.8 PTEN regulates p-catenin/Tcf signaling via PI3K/GSK3P signaling With the implication that PTEN regulates PBK/P-catenin signaling, we chose to investigate this possibility further by modulating the activity of key regulators of this pathway. Thus, we evaluated Topflash activity as function of transient introduction of mock, WT p-catenin, or GSK3P degradation resistant A(Ser 27-45)P-catenin and induction of PTEN. PTEN represses Topflash activity in the presence of endogenous P-catenin and exogenous WT P-catenin (Figure 5.9a). However, transfection of A(Ser 27-45)P-catenin resulted in less 175 modulation (1.1 fold reduction) of Topflash activity upon PTEN expression, as opposed to at least 3.8 reduction with co-expression of exogenous, WT (3-catenin expression (Figure 5.9a). To affirm the pathway via which PTEN may be regulating P-catenin, we investigated the effect of the various members of the pathway downstream of PI3K/PTEN; namely, ILK, Akt, and GSK3, on nuclear P-catenin expression (Figure. 5.9b). GSK3 is known to be a member of the complex that regulates cellular levels of P-catenin in a phosphorylation-dependent manner (Morin, 1999; Behrens, 2000). ILK has been demonstrated to have a critical role in PTEN-dependent cell cycle regulation (Persad et al., 2000) and has also been demonstrated to inhibit GSK3 activity upon cell-extracellular matrix interaction (Troussard et al., 1999). Although transient transfection of ILK-WT has no effect on nuclear P-catenin expression, expression of dominant negative ILK (Persad et al., 2000) or GSK3-WT resulted in significant reductions in nuclear p-catenin expression (Figure 5.9b). Since dominant negative I L K has been shown to inhibit A k t S e r 4 7 3 phosphorylation in PC3 cells, we assayed whether I L K - K D and PTEN-induced inhibition of nuclear P-catenin involved Akt activity. Surprisingly, we determined that the regulatory effects of Akt upon P-catenin expression was lower compared with dominant negative ILK, PTEN-WT, or GSK3-WT. Total cellular levels of P-catenin remained unchanged by all plasmid transfections. Thus, in PC3 cells, PTEN suppresses nuclear accumulation of P-catenin by way of the downstream effectors, ILK and GSK3, and to a lesser extent, Akt. Further implication that GSK3 is a critical mediator of PTEN and P-catenin regulation was obtained by evaluating the activity of GSK3 in PC3 cells (Figure 5.9c). PC3 cells were transiently transfected with PI3K regulators including ILK, PTEN and Akt. Relative to controls, both the I L K kinase dead (-3.5 fold) and WT PTEN (-3 fold) induced significant kinase activity of GSK3 using GS-1 peptide as a substrate. The WT form of I L K did not alter the activity of GSK3, however, a dominant negative form of Akt (DN-Akt) had a minor induction of GSK3 kinase activity. These data indicate that while P-catenin may be considered the main Wnt substrate, other PI3K molecules, including ILK and Akt, have the capacity to indirectly regulate P-catenin, via the kinase activity of GSK3. Further, while I L K and Akt can phosphorylate GSK3, PC3 cells appear to show greater dependency upon PTEN and I L K - K D for the phosphorylation activities of GSK3 (Figure 5.9c). Given the potential importance of ILK as a critical regulator of GSK3 in PrCa cells, the ability of ILK to phosphorylate GSK3 in vitro was 176 considered. To test the potential involvement of ILK in the regulation of P-catenin in PC3 cells, LLK was co-immunoprecipitated from PC3 cell lysates and incubated with purified GSK3p or kinase dead GSK3P, with [y- 3 2P]ATP or cold ATP and kinase reaction buffer with or without the ILK inhibitor, KP-SD-1 (Figure 5.9d). GSK3 was successfully co-immunoprecipitated with ELK with increased phosphorylation detected either by [y-32P] emission or Western probing for phospho-GSK3-Ser9. This suggests that ILK can specifically phosphorylate, and therefore, inhibit GSK3. While it is possible that immunoprecipitated LLK is also associating with Akt, the use of a specific LLK inhibitor, KP-SD-1, likely rules this out (Figure 5.9d). 5.3.9 PTEN can regulate AR by the PTEN/GSK3/p-catenin signaling axis Given that P-catenin is a potent transcription co-activator of A R in the presence of androgens, we chose next to evaluate how PTEN and A(Ser 27-45)P-catenin, versus WT P-catenin, may alter A R transcriptional reporter activity (Figure 5.10a). As anticipated, induction of PTEN reduced reporter activity (ARR3-luc) by 2.2 and 6.6 fold in the presence of endogenous and WT P-catenin, respectively. Introduction of the mutant P-catenin construct only elicited a 1.3 fold reduction upon PTEN expression, over 48 hours, in the presence of 1 nM R1881 (Figure 5.10a). Further implications that GSK3P is a key regulator of the PI3K/P-catenin/AR signaling cascade is apparent upon the addition of a specific inhibitor of GSK3P, Lithium Chloride (20 p M LiCl) (Figure 10b). Importantly, increased ARR3-luc activity (1.5 fold) was observed in cells that were transfected with WT P-catenin in the presence of PTEN. Moreover, A(Sers 27-45)P-catenin was unresponsive to 48 hours of 20 p M L i C l treatment, suggesting that high steady state levels of P-catenin were already achieved (Figure 5.10b). Thus, PTE N is a critical regulator of P-catenin, and by extension, A R by way of GSK3P and LLK regulation. 5.3.10 Androgens protect from PTEN mediated apoptosis As PTEN is a phosphatase with tumour suppressor qualities, an important consideration is whether PTEN could be reducing A R transactivation and Wnt/P-catenin/Tcf transcription simply by a matter of cell death or by activation of an apoptotic signal. FACs analysis of 177 lentiviral infected LNCaP-WT PTEN cells cultured in FBS, CSS + 1 n M R1881, CSS and serum free conditions for 48 hours (+24 hour Dox pulse) revealed little change in sub Go in the presence of androgens, but a 10% and 15% increase in sub Go in cells cultured in CSS and serum free media, respectively. These data suggest in the presence of serum and androgens that PTEN can mediate signal transduction events independent of a detectable, acute apoptotic response. It also suggests that signaling events mediated by non-apoptotic events are still likely to occur in androgen ablative, containing conditions (Figure 5.11a). As expected, lentiviral infected LNCaP-C124S PTEN cells cultured under this treatment paradigm demonstrated little detectable modulation of sub Go accumulation upon expression of C124S PTEN (Figure 5.11b). 5.3.11 PTEN regulates the cellular distribution of AR Immunocytochemistry analysis of LNCaP-PTEN cells indicates global cytosolic induction of PTEN expression upon the addition of 2 p,g/ml Dox in CSS + InM R1881 culture conditions. Given our understanding that PTEN has the capacity to diminish levels of nuclear p-catenin and Topflash activity, and that A R and P-catenin can undergo co-trafficking, we chose to evaluate the distribution of AR-GFP in LNCaP-PTEN, stable expression cells. Transient transfection of AR-GFP into LNCaP-PTEN cells treated with 1 n M R1881 and 2 u-g/ml Dox indicates that PTEN negatively regulates expression and/or distribution of AR. Low magnification views showing co-distribution of PTEN and AR-GFP in LNCaP-PTEN cells treated with 2 u,g/ml Dox and InM R1881 for 48 hours (Figure 5.12a) indicate that cells containing AR-GFP show reduced levels of PTEN as compared to cells not containing exogenous AR. Of 15 AR-GFP positive cells (Figure 5.12a, arrows) in the field of view, 13 appear to follow this trend. DAPI staining for these cells indicates that while some of the nuclei in cells containing AR-GFP have condensed, the vast majority are viable cells and not undergoing an obvious apoptotic response (bar = 20 |im) (data not shown, Figures 5.12 bi, bii). Figure 5.12a clearly indicates that not all cells contain equal expression of either AR-GFP or PTEN. Interestingly, some AR-GFP positive cells show very little or no expression of PTEN (e.g. cells 5, 7, 10, 11, 12 and 13), while others show greater PTEN expression (e.g. cells 1, 2, 3, and 6). Higher magnification views of LNCaP-PTEN co-distribution with AR-GFP clearly show reduced levels of PTEN as compared to non-AR-GFP containing cells (Figure 5.12bi, bii). However, it is also evident that not all AR-GFP positive cells contain an equal amount of PTEN. 178 For example, cells a, b and c contain a moderate amount of PTEN, while cells "d" and "e" contain little or no PTEN expression (Figure 5.12b, arrows). To validate that reduced AR-GFP is not purely due to PTEN induced apoptosis or, alternatively, solely due to a general decrease in transcription, we performed two controls (Figure 5.12ci, cii, ciii). First, the apoptotic index of AR-GFP cells was determined by counting 200 AR-GFP positive cells and qualifying whether cells were low, medium or high expressors of PTEN. Nuclei were also determined to be either condensed (apoptotic) or non-condensed (non-apoptotic). Of 200 AR-GFP positive cells counted, 160 (80%) were low PTEN expressors, 15% were medium PTEN expressors and 10% were high PTEN expressors. Of the 160 low PTEN expressors 35/160 (21%) were apoptotic compared to 8/30 (27 %) of medium PTEN expressors and 7/20 (35%) of low PTEN expressors. Second, when LNCaP-WT PTEN cells are transfected with three different Renilla luciferase based vectors and treated with Dox for 48 hours in CSS + 1 n M R1881, a decrease in absolute numbers was observed in the presence of PTEN. However, these differences are considerably less than the fold change observed with the experimental ARR3-Luc reporter plasmids (Figure 5.12ciii). 5.3.12 WT PTEN expression reduces nuclear AR and P-catenin, ligand-sensitive AR/p-catenin complexing but increases total AR expression LNCaP WT-PTEN cells were assessed for the relative expression of AR, PTEN and AR-GFP in nuclear and cytosolic fractions. To do this, LNCaP-PTEN (14M) cells were pulsed with 2 pg/ml Dox for 24 hours and then cultured with 1 n M R1881 for 48 hours. Considerable induction of PTEN was observed in the presence of Dox and, interestingly, increased levels of PTEN were also detected in the nucleus. However, cellular compartmental leakage, due to incomplete nuclear/cytoplasmic fractionation, cannot be ruled out. Importantly, levels of transfected, nuclear AR-GFP were reduced in the presence of PTEN with a moderate reduction also being observed for P-catenin (Figure 5.13a). Knowing that A R and P-catenin complex in an androgen sensitive manner, we also evaluated whether PTEN expression could abrogate AR (endogenous)/P-catenin interactions. In a phosphatase manner, PTEN efficiently antagonized the amount of p-catenin detected in A R immunoprecipiations from nuclear cell fractions. These data further underscore that fact that WT PTEN can not only regulate localization of AR and P-179 catenin, but also the interaction of A R and binding transcriptional co-activators (Figure 5.13b). We subsequently chose to evaluate the effect of PTEN on A R and 3-catenin steady state levels with and without the presence of serum and androgens. Thus, we cultured LNCaP-WT PTEN cells in FBS, CSS + InM R1881, CSS and S F M culture conditions, pulsed cells for 24 hours with 2 u,g/ml Dox followed by either 24 and 48 hours of treatment. After 24 hours expression, androgens efficiently protected LNCaP cells from PTEN mediated reduction in phospho-A k tser473 ( p j g u r e 5 1 3 c ) However, by 48 hours a 4-fold reduction in P - A k t S e r 4 7 3 levels were observed (Figure 5.13d). Importantly, both at 24 and 48 hour time points little change in total (3-catenin levels were observed (Figure 5.13c, d). Curiously, however, we observed an efficient induction of total A R levels most notably in the LNCaP-PTEN cells cultured in FBS and CSS + InM R1881. Given this perplexing result, we also assayed levels of A R transcription by northern blotting. A clear induction of A R transcripts was observed (Figure 5.13ei) upon expression of PTEN; however, this trend was not readily observed in P-catenin transcripts. While minor deceases were observed in G3PDH levels (Figure 5.13eiii) total R N A input was relatively constant. 5.3.13 PTEN and androgens: opposing regulators of cell cycle and cell viability With the knowledge that WT PTEN can abrogate transcriptional events associated with Wnt/p-catenin/Tcf and A R signaling, and can potently reduce cell viability, we considered how PTEN might modulate critical cell cycle regulators. After 24 hours of dox pulsing and 48 hours of treatment, we assayed expression of direct targets of PTEN and cell cycle regulators including p27K I P, cyclin DI and Cdk2. As controls for the function of WT PTEN, we assayed levels of phospho-Akt S e r 4 7 3 / T h r 3 0 8 . Under these conditions, PTEN failed to induce significant levels of activated PARP, but in the presence of androgens, represses levels of P-Akt (Figure 5.14a, iii). Correlated with induction of PTEN, either with or without androgens, was the increased expression of the negative cell cycle regulator, p27 K I P (Figure 5.14a iv, v). Interestingly, a marked reduction in Cdk2 levels was not observed as a function of PT E N expression. Further of interest, though perplexing, was the A D induction of cyclin DI by PTE N (Figure 5.14a, vi). Thus, LNCaP cells cultured in the presence of androgens for 72 hours (24h pulse + 48 h treatment) with WT PTEN expression appear to survive but do not maintain a high proliferative index. With the implication that PTEN can negatively regulate both P-catenin/Tcf and A R 180 signaling, but that androgens can maintain survival of PrCa cells, we assayed cell viability over 7 days. To do this we used an MTS assay and lentiviral infected LNCaP cells cultured in full serum conditions (10% FBS), androgens (InM R1881), androgen depleted (CSS) conditions and serum free conditions (SFM). In the presence of 10% FBS, we ascertained that LNCaP cells harbouring PTEN have the capacity to reduce cell proliferation by a factor of 0.79 compared to the catalytic mutant treated with Dox (LNCaP-C124S + Dox) (Figure 5.14b). Cells virally infected with WT PTEN and C124S-PTEN treated with 1 n M R1881 demonstrated a greater degree of separation in cell viability at 7 days PTEN induction (WT PTEN at 0.59 to that of C124S PTEN cells). LNCaP-PTEN cells cultured in 1 nM R1881 conditions showed a minor drop in proliferation at 7 days (6.2) as compared with cells grown in 1 n M conditions (Figure 5.14b, ii). Cells grown in androgen depleted conditions showed a marked reduction in cell viability over 7 days as compared with the C124S PTEN mutant (-Dox). WT PTEN infected (+Dox) cells showed a proliferative rate of 0.56 in comparison to PTEN LNCaP (-Dox) cells, and 0.46 in comparison to C124S PTEN LNCaP (-Dox). Reduced proliferative activity in LNCaP WT PTEN cells (+Dox) was evident by day 3 when cultured in serum free conditions. By day 7, cells showed a proliferative rate that was 0.22 to that of LNCaP C124S PTEN (-Dox) cells. LNCaP cells carrying C124S PTEN (-Dox) showed little alteration in proliferative capacity (0.95 in comparison to day 0 control). Clearly, low levels of androgens are sufficient to maintain LNCaP cell viability when expressing catalytically active PTEN. It is possible that over longer durations, these levels of androgens would be insufficient to prevent a global apoptotic response. Cell viability of PC3 cells harbouring WT PTEN and C124S PTEN was also considered by MTS assay under high (10%), low (1%) and no serum culture conditions. While serum is clearly important for the ability of PC3 cells to tolerate the presence of WT PTEN (Figure 5.15a), PC3 cells demonstrated greater resistance to the presence of catalytically active PTEN in comparison to LNCaP cells. Since cyclin DI is both a cell cycle regulator and established target for P-catenin, our next aim was to examine the status of this cyclin in PC3 cells. Figure 15b demonstrates that the presence of cyclin DI protein is high in PC3 cells, consistent with the expression of Tcf and activation of the Topflash reporter. Consistent with the finding that LLK is inhibited by PTEN (Morimoto et al., 2000; Persad et al., 2000), transfection of dominant negative ILK results in substantial inhibition of cyclin DI expression (Figure 5.15b). 181 Transfection of ILK-WT did not alter the expression of this cyclin, demonstrating the specificity of the dominant negative effect of LLK-KD. In agreement with the fact that cellular P-catenin levels are regulated by GSK3, the expression of cyclin DI is also decreased dramatically by transfection of GSK3-WT (Figure 15b). A surprising observation was the fact that transfection with WT PTEN, I L K - K D , or GSK3-WT in PTEN null PC3 cells (Figure 5.15b) did not alter the expression level of the cyclin-dependent kinase (CDK) inhibitors p27 K I P and p21 c n > . These results are in contrast to the findings of Sun et al. (1999), who reported that PTEN appears to regulate the levels of C D K inhibitors in PTEN -/- ES cells. Collectively, Figure 5.15b results indicate that PTEN and LLK regulate cyclin DI expression via GSK3, a finding consistent with those recently reported for mammary carcinoma cells (D'Amico et al., 2000). Dysregulation of cyclin DI expression in PTEN mutant cells can be restored by either inhibiting LLK, replacing PTEN, or augmenting GSK3 expression. Furthermore, these data indicate that PTEN and LLK regulate the expression of cyclin DI in a parallel manner to their regulation of P-catenin (Figure 5.15b). 5.3.14 PTEN as a regulator of cell proliferation, colony formation and the P-catenin/Tcf target gene, cyclin DI With the development of stable inducible PTEN expressing LNCaP cells and the understanding that PTEN antagonizes p-catenin/Tcf/AR signaling potential, we evaluated the effects of PTEN on LNCaP tumourigenesis. To do this approximately 80,000 LNCaP-WT PTEN cells were plated in triplicate in the presence of 2 pg/ml Doxycycline. The soft agar/cell composite was overlayed in full serum for 24 hours and then replaced with treatment media including FBS, CSS, InM R1881, 0.1 nM R1881, 10 p M Casodex and serum free media (SFM). 40,000 PC3-PTEN cells/ml were plated with full serum for 24 hours followed by treatment with either 10% FBS, 1% FBS or SFM. In order to accurately count colony formation, an underlying acetate grid containing 36 squares was employed. 10 representative squares from each 35 mm 2 well were counted, totalled and averaged over 6 wells to generate a standard deviation (see "Dave Grid Technique" from Chapter 2). Though serum and androgens appeared efficient in maintaining viability in the presence of PTEN, considerable less apoptotic protection was observed with respect to the anti-colony formation effects of PTEN. Specifically, stable PTEN 182 expressing LNCaP cells (passage 45) grown in full serum (FBS) or CSS + androgens for 12 days showed at least a 60% reduction in colony formation. Cells cultured in androgen ablative conditions also showed little tumourigenic activity, and with the addition of PTEN, showed very little colony formation (Figure 5.16a). Cells cultured in serum free conditions showed a similar response to those cells cultured in the presence of the A R antagonist, Casodex. As a prelude to assaying the effects of PTEN on colony formation in PC3 cells, we measured cyclin DI promoter activity by transiently cotransfecting PC3 cells with a plasmid containing the cyclin DI promoter and a luciferase reporter gene with components of the PTEN/PI-3-kinase pathway including: ILK-WT, I L K - K D , PTEN-WT and GSK3-WT. As shown in Figure 5.16b, the cyclin DI reporter gene activity is constitutively elevated in the control PC3 cells. Cotransfection of ILK-WT did not further enhance the cyclin DI promoter activity in these cells. However, PTEN or dominant negative I L K imposes a substantial inhibition of the cyclin DI promoter activity. Interestingly, cotransfection of GSK3-WT resulted in inhibition of the cyclin DI promoter activity by a similar magnitude to that observed due to PTEN-WT or I L K - K D . Transfection efficiency of all the plasmids was similar, and the enhancement of the endogenous levels of proteins due to transfection of the respective plasmids is quite comparable (Figure 5.16b). Northern blot analysis of total R N A from serum-starved PC3 cells was used to determine the effects of dominant negative ILK, PTEN-WT, and GSK3-WT on cyclin DI transcription. As shown in Figure 16c, transfection of PC3 cells with I L K - K D , PTEN-WT, or GSK3-WT results in a dramatic decrease in mRNA expression of cyclin D L These results indicate that PTEN and ILK are linked in a common pathway that regulates cyclin DI gene transcriptional activity. The fact that GSK3 inhibits the cyclin DI promoter activity by a similar degree to that observed due to PTEN-WT and dominant negative ILK indicates that GSK3 is very likely the downstream effector of ILK by which this regulation of the cyclins is achieved. Consistent with cell proliferation analysis and analysis of cyclin DI regulation, stable PTEN expressing PC3 cells showed significantly impaired colony formation in serum free conditions and an attenuated response in the presence of low (1%) and high (10%) serum conditions. After 12 days of growth, the number of colonies (> 2 mm) found in 10 representative squares within the 36 square grid was averaged over 6 x 3 5 mm 2 wells (Figure 16d). Figure 16d clearly indicates that PC3 cells are capable of colony formation in the presence of PTEN and that the presence of serum appears to serve as a "protective agent" against the anti-proto tumourigenic effects of PTEN. 183 Figure 5.1. Levels of nuclear P-catenin are reduced by expression of exogenous PTEN in PTEN null PrCa cells, a) Immunofluorescence analysis of PC3 cells indicates that in full serum conditions a significant population (70%) of these cells express nuclear P-catenin (arrowheads) (bar = 5 uM). b) Re-expression of PTEN in PTEN null LNCaP cells cultured in 10% FBS serum conditions results in a dose-dependent reduction in the expression of nuclear P-catenin. Total cellular P-catenin remains constant as does the total nuclear input as measured by histone HI Levels, c) Introduction of PTEN-WT-GFP into PC3 PrCa cells decreases levels of nuclear P-catenin. Cells expressing exogenous PTEN-GFP (arrows) show reduced levels of nuclear P-catenin while those not expressing GFP tagged PTEN (arrow heads) show higher levels of nuclear p-catenin. Transfected GFP-cDNA alone did not alter levels of P-catenin. (bar = 5 uM). 184 a) PC3 ICC: p-catenin b) LNCaP o \— WT PTEN (ug) 2 A _ O r -O 2 4 6 8 WB: p-catenin Nuclear P-catenin Histone H1 WB: H1 Histone c) PC3 WB: p-catenin Total Cellular P-catenin • G F P ICC: p-catenin ICC: GFP (PTEN) 185 Figure 5.2 Increasing amounts of transfected PTEN-GFP in PC3 cells promotes decreased nuclear expression of p-catenin but also leads to a dramatic induction of E-cadherin. Levels of N-cadherin are not altered substantially upon the addition of PTEN-GFP. GFP cDNA (jig) / A v C 2 4 6 8 Nuclear p-catenin WB: p-catenin WT PTEN GFP cDNA (ng) < A N C 2 4 6 8 WB: p-catenin Nuclear p-catenin E-cadherin WB: E-cadherin WB: P-catenin WB: N-cadherin Total Cellular p-catenin N-cadherin Histone H1 WB: Histone H1 187 Figure 5.3 PTEN enhances P-catenin degradation, a) PC3 cells transfected with PTEN or empty vector (control) were starved with D M E without cysteine and methionine, pulsed with [35S] Promix for 1 h, and then chased for the indicated time-points with cold D M E containing cysteine/methionine and FBS. Cells were then harvested, lysed, and immunoprecipitated for P-catenin. Results were analyzed by densitometry and expressed as a percentage of the value at 0 hours, b) Increased phosphorylation of P-catenin within its regulatory domain (Ser33/37 and Thr41) was determined using a phospho-specific P-catenin antibody, upon increasing amounts (0, 2, 4, 6, and 8 Lig) of transfected PTEN. Total levels of P-catenin remain constant. — **1 CONTROL (stable protein) P-catenin (PTEN transfected) Chase Time: 0 2 4 6 8 (hrs) WT P T E N (ug) / A s 8 P-(Ser 33/37/Thr41) j|*j ^ WB: P-catenin WB: p-catenin 189 Figure 5.4 PTEN inhibits D N A binding activities of P-catenin/Tcf complexes, a) Electrophoretic mobility shift assays, using an oligonucleotide containing a Tcf4 binding site demonstrated an abundance of DNA-protein complex in empty vector-transfected PC3 cells (lane 1). The quantities of the protein-DNA complexes are reduced significantly in PC3 cells that have been transfected with PTEN (lane 2). Electrophoretic mobility shift assays performed in the presence of a supershifting antibody (p-catenin antibody) (lane 3) confirmed that the transcription factor complex binding to the oligonucleotide includes P-catenin. b) Co-IPs of Tcf-4 or Lef-1 with P-catenin using nuclear lysates from empty vector (control), J L K - K D (Kinase Dead), or WT PTEN-transfected PC3 cells show reduced complex formation in LLK-KD and PTEN-WT-transfected cells compared with control cells. 190 b) £> .O 4? & cV ±F x « ^ ° ^ ^ Tcf4 IP: p-catenin WB: Tcf4 Lef-1 IP: p-catenin WB: Lef-1 191 Figure 5.5 Generation of stable "knock-in" PTEN cells by two different methods, a) Linearized WT PTEN and C124S PTEN cDNAs (with hygromycin cassettes) and a VP16/tet-on encoding construct (with G418 cassettes) were introduced by lipofectin transfection into LNCaP and PC3 PrCa cells. Clones were selected using 0.1-0.3 pg/ml G418 and 0.1-0.3 pg/ml of Hygromycin, expanded and evaluated for induction of either a TRE-luc reporter construct or for the induction of PTEN. To minimize PTEN "leaky" expression, LNCaP-PTEN/C124S and PC3-PTEN/C124S clones were virally infected with a Tet-repressor (puromycin selection). LNCaP-PTEN/C124S + TR and PC3-PTEN/C124S + TR cells were then triple selected (Hyg-, G418, Puro-), expanded and evaluated for non-leaky induction of PTEN expression, b) LNCaP-PTEN and PC3-PTEN stable clones were also developed using the Gateway (Invitrogen) Lentiviral system. 293FT cells were used to produce lentivirus packaging the pDEST/Lenti-PTEN/C124S transgene (Zeocin selection) and the Tet-repressor (pLenti/TR) (Blasticidin selection). Subsequent to viral infection cells were allowed to recover and then were selected using 1-50 pg/ml Blasticidin and 10-300 pg/ml Zeocin. Due to a high rate of viral infection (>90%) LNCaP-PTEN/C124S cells were treated as single clones (i.e. "pseudo clonal"). Differently, were the PC3 cells which had a very low rate of infection, thus colonies were selected, expanded and evaluated by WB, ICC and FACs analysis. a) Clonal Selection of PTEN PrCa Stable Cell Lines (2) Transfect: using LNCaP or P C 3 cells (3) Select & expand cells: 0.1-0.3 ng/ml G418 0.1-0.3 ng/ml Hygromycin (5) Virus infection: for 24-48 hours OO * o TR Gag-Pol VSVG Tet-Repressor (Puromycin) GP3 (Gag-Pol) VSVG (virus envelope) WB PTEN: 2 ug/ml Dox: + - + + - + X X X y/ (7) Test PTEN-PrCa clones: inspect for good induction and low "leakage" b) Lenti Viral (Invitrogen) Production of PTEN-PrCa Stable Cell Lines (1) 293FT Transfection pLenti6/TR pDEST4/TG7V5 PTEN/C124S (2) Virus Infection of LNCaP, PC3 or C4-2 cells OO -o HIV Rev Gag-Pol VSVG I) (3) Select & Expand Cells 5-50 ug/ml Blasticidin 10-300 ug/ml Zeocin LNCaP (4) Test PTEN-PrCa Clones: WB and ICC 194 Figure 5.6 V P 16 fusion and lentiviral tet-repressors maintain minimal non-induced ("leaky") expression of the P T E N transgene. The VP16-Tet-on fusion binds Dox and binds directly to TPvEs to displace the constitutively bound Tet repressor, thus allowing gene expression. In the presence of doxycycline, binding occurs to the Tet repressor resulting in its removal from two tandem Tet response elements (TREs) found upstream of the minimal C M V promoter, b) In the Invitrogen "Tet-off" system, the addition of Dox promotes a conformational change and dissociation of the Tet repressor, (red triangles = doxycycline) 195 a) VP16 Fusion Tet "On" system © TRE TRE minCMV PTEN TRE TRE minCMV PTEN -Dox +Dox b) Invitrogen Tet "off" Repressor system TRE TRE minCMV ©0 PTEN • TRE TRE minCMV PTEN -Dox +Dox 196 Figure 5.7 Validation of PTEN expression and distribution in Dox induced stable LNCaP-PTEN/C124S clones, a) Immunocytochemistry (ICC) illustrating that at high magnification PTEN is observed to localize to a submembraneous position while at low magnification PTEN appears global in expression, b) Induction of PTEN message is identified by transcripts at 5 kb and 2 kb. c) 24 hours of 2 |ig/ml Dox treatment results in expression of PTEN in full serum, androgen ablation and serum free conditions at a ratio of ~ 25:1 compared to non-induced cells. Compared to the catalytic (C124S) mutant, PTEN cells diminish phospho-Akt (Ser4 7 3) levels by at least 6 fold, di) PTEN expression is both dose (0, 0.5, 1, 2, 4, 10 \xg/m\ Dox) and time (0, 8, 16, 24, 36, 48 hours Dox) dependent with the addition of Dox. dii) Induction of PTEN in serum free conditions results in dramatic induction of cleaved (activated) PARP. The PTEN catalytic (C124S) mutant fails, however, to generate such a response after 48 hours of 2 Lig/ml Dox induction. 197 High Mag b) LNCaP-PTEN Clones (14M) (5-4) Dox: + - + -PTEN (5 kb) PTEN (2 kb) NB: PTEN ICC: PTEN C) Phosphatase dependent reduction of P - A k ^ 4 7 3 LNCaP-WT PTEN Clone (14M) +Dox ODU = 25 -Dox ODU = 1 WB: PTEN WB: PTEN /fr <S? of & « +Dox C124S PTEN WB: P-AktSert73 WT PTEN WB: P-AktSert" di) Dose & time dependent induction of PTEN Dox: 0 0.5 1 2 4 10 (ng/ml) WB: PTEN Dox: 0 8 16 24 36 48 hours : " WB:PARP (cleaved) dii) Phosphatase dependent induction of apoptosis PTEN: C124S Dox: - A WT WB: PTEN WB: PARP (cleaved) 198 Figure 5.8 PTEN inhibits A R and P-catenin/Tcf function in a phosphatase dependent manner in A D (LNCaP) PrCa cells, a) Treatment of LNCaP cells (p45) with increasing concentrations (0, 1, 5 and 10 uM) of the P B K chemical inhibitor, LY294002, in the presence of 1 nM R1881 results in a dose dependent inhibition of A R transactivation, as measured with the ARR3-luc reporter, b) Increases in Dox treatments (0, 0.05, 0.1, 0.5, 1, and 2 p.g/ml) results in a similar repression of A R transactivation (ARR3-luc reporter) in LNCaP-PTEN cells, however, this trend was not observed in cells expressing the C124S PTEN, catalytic mutant (data not Shown), c) Increased Dox treatment (0, 0.05, 0.1, 0.5, 1, and 2 pg/ml) results in a similar repression of Tcf transactivation (Topflash) but not in the d) mutant reporter (Fopflash) construct. 200 Figure 5.9 a) PC3-WT PTEN cells treated with 2 pg/ml Dox show reduced Topflash activity in the presence of endogenous and exogenous WT P-catenin. Cells transiently infected with GSK3P resistant P-catenin (A 27-45) lentiviral vectors show greater Topflash activity in the presence of PTEN than cells containing endogenous or exogenous WT p-catenin. b) Nuclear P-catenin is mainly regulated by PTEN and LLK via GSK3. The expression of nuclear p-catenin is dramatically and comparably abrogated by expression of PTEN, I L K - K D and GSK3-WT in PC3 cells. Nuclear P-catenin expression is altered to a lesser extent by the dominant negative Akt-A A A and completely unaffected by expression of ILK-WT. The absence of any alterations in the total cellular P-catenin expression demonstrates the specific effect of the various components upon nuclear P-catenin. c) Bar graph representing quantification of GSK3 kinase activities by densitometric analysis (Odu/mm2) in PC3 cells transiently transfected with empty vector (control), ILK-WT, I L K - K D , Akt-DN or PTEN-WT. Top panel is a representative autoradiograph of GSK3 kinase activities in the various transfectants. To evaluate stimulation of GSK3 activity, transfected cells were serum starved for 18 h, refed with serum for 1 h, and then analyzed for GSK3 kinase activity by using GS-1 peptide as a substrate. Although PTEN and I L K - K D induced a dramatic increase in GSK3 kinase activity (3-4-fold) the effect of A K T - D N was more modest (1.6-fold). Western blotting for GSK3 shows equivalent amounts of GSK3 in each extract (bottom), d) Bar graph representing quantification of I L K kinase activity by densitometric analysis (Odu/mm2) in serum-starved (18 h) PC3 cells. ILK, purified by LP with ILK antibody, was coincubated with purified GSK3-KD and [y-32P] ATP in the presence or absence of the I L K inhibitor KP-SD-1. Bottom panel represents an in vitro LLK kinase assay, where recombinant ILK prepared in insect cells was coincubated with GSK3-KD and ATP in the presence or absence of an LLK inhibitor, KP-SD-1. Phosphorylated GSK3 was detected by Western blot analysis using an GSK3-Ser-9 antibody. Odu = optical density units. a ) PC3- WT PTEN Cells o 3 -I CO 3 3 PTEN (Dox): Topflash: Q r — _ _ 1:1.1 — A — j —t T X 1— 1:3.8 "V „ / • A \ J i , 1:4.6 r 1" j . , J — ' K g • • u + + - + -+ + + + + + WT P-Cat AP-Cat (ASer27-45) b) C * # ^ ^ cr o WB: P-catenin Nuclear P-catenin WB: Histone HI WB: p-catenin WB: V5 WB: GFP _ Total Cellular P-catenin - ILK-V5 -PTEN-GFP P K A - A A A GSK3-WT WB: H A i Activity ? 9 2 W I3UUUU M (0 c <o 100000 o • • • • 50000 0 Serum + - + . + . + . + Serum + . + - + . + . + -— ~ mMmm*m*^^~\ <*~GS-1 Peptide GSK3p Kinase Activity Total Gsk3p 203 Figure 5.10 a) Induced PTEN expression in LNCaP cells (14M clone) reduces A R transactivation (ARR3-luc) in cells transfected with either WT P-catenin or activated P-catenin. After 48 hours of 2 Ug/ml Dox treatment, LNCaP-PTEN cells transfected with GSK3P resistant P-catenin (ASer27-45) are less responsive to PTEN expression as measured by ARR3-luc reporter values, b) In the presence of 20 u,M L i C l , a GSK3P inhibitor, PTEN is less efficient in reducing A R transactivation of the ARR3-luc reporter in LNCaP-PTEN cells with either endogenous or exogenous WT P-catenin. In LNCaP-PTEN cells treated with 2 u,g/ml Dox and transfected with P-catenin ASer27-45, L i C l has little affect on ARR3-luc. "0 —I m ^ 73 o co co > 73 73 co CO D O x Absolute Luciferase Values o o w o o o o o o 1 o 000 000 000 000 ,000 000 O 205 Figure 5.11 Androgens oppose apoptotic effects conferred by WT PTEN expression in LNCaP PrCa cells, a) Cell cycle analysis (FACs) indicating that androgens (1 n M R1881) protects LNCaP-WT PTEN cells from PTEN, phosphatase dependent, mediated accumulation of a sub Go cell fraction. LNCaP-PTEN cells were pulsed with 2 [ig/m\ Doxycylin for 24 hours followed by treatment with 2 jxg/ml of Dox in FBS, CSS + 1 nM R1881, CSS or S F M for 48 hours. In androgen ablative conditions (CSS) PTEN mediates a 11% increase in sub Go cell fraction while in serum free conditions an increase of 14% is observed. No change in sub Go was observed in cells cultured in the presence of InM R1881. b) Using PTEN-C124S PTEN stable LNCaPs, little accumulation of sub Go fraction is observed in the presence of Dox. LNCaP-WT PTEN Dox none LNCaP-C124S PTEN 208 Figure 5.12 PTEN expression is correlated with decreased expression of AR-EGFP in LNCaP-PTEN cells (14M clone), a) Low magnification view showing co-distribution of PTEN and AR-GFP in LNCaP-PTEN cells (14M clone) treated with 2 ug/ml Dox and InM R1881 for 48 hours. Cells expressing high levels of PTEN are correlated with reduced nuclear expression of AR-GFP, however, appear to remain viable. Of 15 A R GFP positive cells (arrows) in the field of view, 13 appear to follow this trend (bar = 30 um) (bi, bii). Higher magnification views of LNCaP-PTEN cells (14M clone) showing PTEN co-distribution with AR-GFP. While some AR-GFP positive cells (bi, arrows; bii,a-e) clearly show reduced levels of PTEN as compared to non-AR GFP containing cells (bi, arrows; bii arrows d + e), other AR-GFP positive cells express moderate to low levels of PTEN (bii arrows a, b, + c). ci) 200 AR-GFP positive cells were counted and assessed for their relative expression (low, med, high) of PTEN intensity as measured by immunofluorescence. A R GFP positive cells were also stained using DAPI in order to identify cells that may be undergoing nuclear condensation or apoptosis. Cells counted as apoptotic were indicated in number as a function of PTEN expression, cii) Examples of a nuclei counted as apoptotic and non-apoptotic. ciii) Transient transfection of Renilla luciferase reporter vectors under varying promoters (CMV, SV40, TK) demonstrating the general effects that PTEN may have on global transcription. 209 LNCaP-WT PTEN 1 nM R1881 + 2 u.g/ml Dox (48 hours) AR gfp I 5 8 * • 10 13, 9 * "J. * • I 12 5 P T E N ' 2 4 1« < 5 J* , 8 10 13 12 30 uM LNCaP-PTEN 1 nM R1881 + 2 ng/ml Dox (48 hours) bi) Field of view 1 Field of view 2 PTEN AR gfp DAPI 211 (Cii) Apoptotic Non-apoptotic (Ciii) 800,000 700,000 600,000 500,000 J 400,000 fl C 300,000 200,000 a 100,000 \- \ 0 R1881: DOX: + + Tl + + + + CMV SV40 Ren Luc Ren Luc TK Ren Luc 212 Figure 5.13 WT PTEN expression reduces nuclear A R and (3-catenin ligand-sensitive AR/P-catenin complexing but increases total A R expression, a) The presence of PTEN promotes reduced levels of nuclear A R and P-catenin. LNCaP-WT PTEN cells were pulsed with 2 ng/ml doxycyclin (Dox) for 48 hours in the presence of 1 nM R1881 and assayed for nuclear A R and P-catenin relative to histone (nuclear) and cytoplasmic (E-cadherin) loading controls, b) Androgen dependent complexing in LNCaP cells (14M/5-4 clones) between AR-GFP and P-catenin is less efficient in the presence of WT PTEN (+Dox) but not with the phosphatase (C124S PTEN Mutant, c) After 24 hours of treatment, PTEN expression in LNCaP cells cultured in androgen ablative conditions or in serum free conditions showed marked reduction of P-A k t S e r 4 7 3 (ci) but relatively little apoptotic response. Little flux in either total A R (cii) or total p-catenin (ciii) levels was detected as a function of PTEN expression, d) After 48 hours of Dox treatment, phospho-Akt4 7 3 levels decreased (di), with cleaved PARP being detected in FBS, CSS and S F M conditions but not in the presence of 1 nM R1881 (dii). Curiously, PTEN appears to promote an increase in total levels of A R though this trend is not apparent with P-catenin (diii, div). e) Induction of PTEN in LNCaP-PTEN (14M) cells as assayed by Northern Blot (NB) Analysis. Cells were cultured in full serum (FBS) in the presence of 2 u,g/ml Dox for 24 hours and then replaced with treatment media including FBS, CSS + InM R1881, CSS and S F M for 48 hours. Samples were probed for the changes in PTEN, A R and P-catenin transcripts. 213 a) b) Dox: + - + R1881: + + + + WT PTEN C124S PTEN IP:AR | WB: PTEN WCL WB: GFP (AR-gfp) R1881: Dox: WB: p-cat loading — mm M i d * E-Cad H1 Histone . 4 mmmmW -0-Cat IgG WB: PTEN + + + + + + A N A A 4 Dox: + d) A ^ A A PTEN P-Akt4" PARP (cleaved) - m mm Dox: + PTEN P-Akt4" PARP I - r AR I *m ~ mm mmm mim mmm — -mm I | | P Cat | Wl» • » ^ » aj> " * " Actin (cleaved) AR P-Cat Actin 2 (ng/ml) for 24 hours j . — —mm*, mm1 « « . ^ | 2 (ug/ml) for 48 hours Dox (2 ug/ml): + NB: PTEN NB: AR N w H y 5kb 2k