UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Analysis of mouse prostate development using serial analysis of gene expression (SAGE) Zhang, Tian-Jiao 2005

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

Item Metadata

Download

Media
831-ubc_2005-0725.pdf [ 13.74MB ]
Metadata
JSON: 831-1.0092285.json
JSON-LD: 831-1.0092285-ld.json
RDF/XML (Pretty): 831-1.0092285-rdf.xml
RDF/JSON: 831-1.0092285-rdf.json
Turtle: 831-1.0092285-turtle.txt
N-Triples: 831-1.0092285-rdf-ntriples.txt
Original Record: 831-1.0092285-source.json
Full Text
831-1.0092285-fulltext.txt
Citation
831-1.0092285.ris

Full Text

Analysis of Mouse Prostate Development Using Serial Analysis of Gene Expression (SAGE)  By TIAN-JIAO Z H A N G B . M . , HeBei Medical University, 1995  A thesis submitted in partial fulfillment of the requirements for the degree of M A S T E R OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Experimental Medicine)  T H E UNIVERSITY OF BRITISH C O L U M B I A June 2005  © Tian-Jiao Zhang, 2005  ABSTRACT  Mouse prostate organogenesis is initiated with the outgrowth of the urogenital sinus epithelium (UGE) into the  surrounding urogenital sinus mesenchyme  (UGM).  Mesenchymal-epithelial (M/E) interactions are important for prostate development and molecules involved in this process remain to be identified. To enhance our understanding of the molecular mechanisms regulating mouse prostate development, we generated a serial analysis of gene expression (SAGE) library from each of the following six tissues: El6.5 male urogenital sinus (UGS), E16.5 female UGS, E16.5 male U G M , E16.5 male UGE, postnatal day 0 (PO) prostate and adult dorsal-lateral prostate (DLP) libraries. Bioinformatic analyses revealed differential expression of several members of the WNT signaling pathway and expression of some members was confirmed by quantitative Real Time PCR (RT-qPCR). Of particular interest, a gene encoding an antagonist of the WNT pathway known as Secreted Frizzled-Related Protein 2 (SFRP2) was highly expressed in the El6.5 male UGS library and down-regulated in the adult D L P . Transcripts of WNT4, which interact with SFRP2, and P-catenin, a downstream molecule of the canonical W N T pathway, were also differentially expressed in the developing prostate as well as in male and female UGS. Expression patterns of Sfrp2 were examined using both RT-qPCR and Digital Northern analysis. Further, localization of Sfrp2 mRNA and its protein was carried out using in situ hybridization and immunofluorescence, respectively. Our studies show that WNT pathway members are expressed in the developing prostate, suggesting that the WNT pathway may play a role in prostate development.  n  Development of cancer often involves inappropriate reactivation of pathways that are active during normal development. Therefore, expression of Sfrp2, Wnt4 and fi-catenin during prostate cancer (PCa) initiation and progression was evaluated using the TRansgenic Adenocarcinoma of the Mouse Prostate (TRAMP) model. Sfrp2 and Wnt4 mRNA levels in the T R A M P D L P were upregulated at 4 weeks (PCa initiation) and 8 weeks (PCa progression) of age respectively, compared to their wild type littermates. Taken together, these studies provide the first evidence that W N T pathway members are differentially expressed in the developing prostate and in the D L P of the T R A M P model of PCa. Functional analyses are now required to establish the biological significance of these observations.  in  TABLE OF CONTENTS  Abstract  ii  Table of Contents  iv  List of Tables  vi  List of Figures  vii  List of Abbreviations  ix  Acknowledgements  xii  C H A P T E R 1 INTRODUCTION 1.1  1.2 1.3  1.4  1.5  1  Mouse Prostate Development  1  1.1.1 1.1.2 1.1.3 1.1.4  1 2 3  Initiation of Development and Morphogenesis Cytodifferentiation of the U G E and U G M Roles of Hormones in Prostate Development Mesenchymal-Epithelial (M/E) Interactions in Prostate Development 1.1.5 Molecular Mechanisms Regulating Prostate Development The Adult Mouse Prostate: Function, Anatomy and Histology Prostate Cancer (PCa) 1.3.1 Brief Overview of PCa 1.3.2 Genetic and Epigenetic Factors in PCa 1.3.3 Disregulation of Oncogenes and Tumor Suppressor Genes 1.3.4 Signaling Pathways Involved in PCa 1.3.5 Changes in Stromal-Epithelial Interactions in PCa 1.3.6 Modelling PCa 1.3.7 Xenograft Models 1.3.8 Knockout (KO) and Tg Mouse Models of PCa 1.3.9 TRansgenic Adenocarcinoma of the Mouse Prostate (TRAMP) 1.3.10 Difference Between Mouse PCa Model and Human PCa Gene Expression Profiling 1.4.1 Summary of Techniques 1.4.2 Technical Improvement of SAGE 1.4.3 Application of S A G E 1.4.4 Bioinformatic Analysis of SAGE Data 1.4.5 Gene Expression Profiling of the Prostate Wingless-Type Mouse Mammary Tumor Virus Integration Site Genes {Wnt) and Secreted Frizzled-Related Proteins (Sfrps) 1.5.1 The W N T Signaling Pathway 1.5.2 The Secreted Frizzled-Related Proteins (SFRPs)  iv  5 6 11 12 12 13 13 15 16 16 17 18 21 22 23 23 27 30 32 34 36 36 39  1.6  Hypotheses and Objectives  42  C H A P T E R 2 M A T E R I A L S A N D METHODS  45  2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10  45 45 47 51 54 57 58 60 60 64  Mouse Maintenance and Mating Strategy Tissue Collection R N A Isolation SAGE Library Construction Bioinformatic Analysis of the SAGE Libraries Reverse Transcription (RT) Quantitative Real Time PCR (RT-qPCR) Tissue Preparation for Immunofluorescence and in situ Hybridization In situ Hybridization Immunofluorescence  CHAPTER 3 RESULTS 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11  65  Analysis of the Prostate SAGE Libraries 65 3.1.1 Rationale for Library Selection 65 Validation of the Prostate SAGE Libraries 72 Identification of Sfrp2 As A Candidate Gene 78 Expression of WNT Family Members during Prostate Development 80 Assessment of the Expression Pattern of Sfrp2 by RT-qPCR and Digital Northern 87 RT-qPCR Analysis of Wnt4 and P-catenin Expression Levels 91 In situ hybridization of Sfrp2 94 Immunofluorescence to Detect SFRP2 and P-catenin Protein During Prostate Development 97 RT-qPCR analysis of Sfrp2, P-catenin and Wnt4 in the T R A M P Mouse Model of PCa 100 Immunofluorescence to Detect SFRP2 and p-catenin Protein in the Prostate of T R A M P mice 102 Summary 102  CHAPTER 4 DISCUSSION  106  4.1 4.2 4.3  SAGE Library Analyses SFRP2 and the WNT Pathway in Prostate Development SFRP2 and the WNT Pathway in PCa  106 110 115  4.4  Future Directions  118  REFERENCES  122  APPENDICES  145  v  LIST OF TABLES Table 2.1  Summary of samples collected for SAGE library construction  50  Table 2.2  Primer sequences used to amplify the W N T family members and housekeeping genes selected for RT-qPCR  59  Table 3.1  SAGE library summary  67  Table 3.2  Tag distributions in the 6 libraries  68  Table 3.3  Pair-wise comparisons of the SAGE libraries  71  Table 3.4  Top 10 tags more highly expressed in the El6.5 male U G S than in the P0 and adult prostate libraries  73  Top 10 tags more highly expressed in the P0 library than in the the El6.5 male UGS and adult prostate libraries  73  Top 10 tags more highly expressed in the adult library than in the El6.5 male UGS and P0 prostate libraries  74  Known genes used for bioinformatic validation of the E16.5 male UGS, P0 prostate and the adult prostate SAGE libraries  75  Expression of known mesenchymal or epithelial specific genes in the U G M and U G E SAGE libraries  77  Sequence and identity of tags expressed at higher levels in the developing prostate libraries  81  Expression of WNT pathway members in the 6 S A G E libraries analyzed  83-84  Digital Northern analysis of Sfrp2 expression in mouse S A G E libraries  88  Table 3.5  Table 3.6  Table 3.7  Table 3.8  Table 3.9  Table 3.10  Table 3.11  Table 3.12  ' Digital Northern analysis of Sfrp2 expression in human SAGE libraries  vi  89  LIST OF FIGURES Figure 2.1  Tissues dissected for S A G E library construction  48-49  Figure 2.2  Schematic diagram of SAGE library construction  55-56  Figure 2.3  Schematic diagram of Sfrp2 probe synthesis for in situ hybridization by in vitro transcription  62  Figure 3.1  Tag distributions in the 6 libraries  68  Figure 3.2 Figure 3.3  Pair-wise comparisons of the S A G E libraries Identification of Sfrp2 through bioinformatic comparisons of the SAGE libraries  70 79  Figure 3.4  RT-qPCR analysis of WNT family member expression during prostate development  85  RT-qPCR analysis of WNT family member expression in E16.5 male and female UGS  86  Figure 3.5  Figure 3.6  Comparison of Sfrp2 expression levels using SAGE (right panels) and RT-qPCR (left panels)  90  Figure 3.7  Wnt3a, 8a and 8b are not expressed in the P0 prostate  92  Figure 3.8  Comparison of Wnt4 expression levels using SAGE (right panels) and RT-qPCR (left panels) Comparison of /3-catenin expression levels using S A G E (right panels) and RT-qPCR (left panels)  Figure 3.9  93 95  Figure 3.10  In situ hybridization of Sfrp2 during prostate development  Figure 3.11  Immunofluorescent detection of SFRP2 in the developing and adult prostate  98  Immunofluorescent detection of P-catenin in the developing and adult prostate  99  RT-qPCR analysis of Wnt4, P-catenin and Sfrp2 expression in the DLP of T R A M P mice during PCa development  101  Immunofluorescent detection of SFRP2 protein in T R A M P mice and WT littermates  103  Figure 3.12  Figure 3.13  Figure 3.14  vii  96  Figure 3.15  Immunofluorescent detection of P-catenin protein in T R A M P mice and WT littermates  i  viii  104  LIST OF ABBREVIATIONS Ag AP Ape* APC* AR B2m BLAST Bmp4 BMP4 BPH CaMK CG CRD DHT DIG Dkk DKK DLP DP Dvl DVL E16.5 E17.5 Eff Egf EGF ER EST Ezh2 EZH2 FBS Fgf FGF Fzd FZD Gapdh GH GO HBSS HGPIN IGF lgfb 2 IGFBP2 P  Antigen Anterior prostate Adenomatous polyposis coli Adenomatous polyposis coli Androgen receptor P-2-microglobulin Basic Local Alignment and Search Tool Bone morphogenetic protein 4 Bone morphogenetic protein 4 Benign prostatic hyperplasia Calmodulin-dependent kinase Coagulating gland Cysteine-rich domain Dihydrotestosterone Digoxigenin Dickkopf Dickkopf Dorsal-lateral prostate Dorsal prostate Dishevelled Dishevelled Embryonic day 16.5 Embryonic day 17.5 Efficiency Epidermal growth factor Epidermal growth factor Estrogen receptor Expressed sequence tags Enhancer of zeste homologue 2 Enhancer of zeste homologue 2 Fetal Bovine Serum Fibroblast growth factor Fibroblast growth factor Frizzled Frizzled Glyceraldehyde-3-phosphate dehydrogenase Growth Hormone Gene Ontology Hank's Balanced Salt Solution High grade prostatic intraepithelial neoplasia Insulin-like growth factor Insulin-like growth factor Insulin-like growth factor binding protein 2 Insulin-like growth factor binding protein 2  ix  JNK KO LB Lef LEF LGPIN LP Lrp5 LRP5 Lrp6 LRP6 M/E MGC MMP NE NTR PO PBS PCa PCP PCR PFA PIN Prl PRL PSA Ptc PTC Pten PTEN Rb RB RefSeq RT-qPCR SAGE SCID Sfrp SFRP Shh SHH SV SV40 Tcf TCF Tfm Tg  Jun kinase Kockout Luria-Bertani lymphoid enhancer binding factor lymphoid enhancer binding factor Low grade prostatic intraepithelial neoplasia Lateral prostate Low density lipoprotein-receptor-related protein 5 Low density lipoprotein-receptor-related protein 5 Low density lipoprotein-receptor-related protein 6 Low density lipoprotein-receptor-related protein 6 Mesenchymal-epithelial Mammalian Gene Collection Matrix metalloproteinase Neuroendocrine Netrin domain Postnatal day 0 Phosphate Buffered Saline Prostate cancer Planar cell polarity Polymerase chain reaction Paraformaldehyde Prostatic intraepithelial neoplasia Prolactin Prolactin Prostate specific antigen Patched Patched Phosphatase and tensin homolog Phosphatase and tensin homolog Retinoblastoma Retinoblastoma Reference Sequences Quantitative Real Time PCR Serial analysis of gene expression Severe Combined Immunodeficient Secreted Frizzled-Related Protein Secreted Frizzled-Related Protein Sonic Hedgehog Sonic Hedgehog Seminal vesicle Simian virus 40 T-cell factor T-cell factor Testicular feminized Transgenic  X  Tgf TGF TIMP TRAMP TS TSG Ubc UGE UGM UGS uPA UTR VP Wif WIF Wnt WNT WT 2-DE  Transforming growth factor Transforming growth factor Tissue inhibitors of metalloprotease TRansgenic Adenocarcinoma of the Mouse Prostate Theiler stage Tumor suppressor genes Ubiquitin C Urogenital sinus epithelium Urogenital sinus mesenchyme Urogenital sinus Urokinase plasminogen activator Untranslated region Ventral prostate Wnt inhibitory factor Wnt inhibitory factor Wingless-Type Mouse Mammary Tumor Virus Integration Site Genes Wingless-Type Mouse Mammary Tumor Virus Integration Site Genes Wild type Two-dimensional electrophoresis  * Genes are indicated by italics with the first letter captalized; proteins are indicated by all letters captalized.  XI  ACKNOWLEDGEMENTS  I would like to express my appreciation to my supervisor, Dr. Cheryl D Helgason, for her supervision and inspiration, and for giving me the opportunity to be involved in such an interesting project. I would like to thank Dr. YuZhuo Wang and Dr. Pamela Hoodless for being on my supervisory committee and for their wonderful suggestions. Moreover, I would also like to acknowledge everyone from the Helgason lab for their great help.  Xll  CHAPTER 1. INTRODUCTION  1.1 Mouse Prostate Development 1.1.1  Initiation of Development and Morphogenesis  Mouse prostate organogenesis is initiated shortly after embryonic day 17.5 (E17.5) from the urogenital sinus (UGS) in males (1). The UGS is derived from the cloaca which is a caudal extension of the hindgut. The cloaca is subdivided by the urorectal septum into the UGS ventrally and the rectum and anal canal dorsally. The ventral UGS is further subdivided into the bladder and the definitive UGS from which the prostate buds develop (2). Both the UGS and its derivative prostate are lined with an endoderm-derived epithelial layer and a mesoderm-derived mesenchymal layer. Expression of testosterone by the fetal testes induces prostatic morphogenesis with the outgrowth of solid buds from the urogenital sinus epithelium (UGE) into the surrounding urogenital sinus mesenchyme (UGM) in a precise spatial pattern that specifies the lobular subdivisions of the prostate (3, 4). Postnatal day 0 (PO) represents the initiation of extensive epithelial branching. The nascent prostatic buds elongate and branch to form an extensive network while at the same time the solid buds canalize in a proximal to distal direction along the developing ducts. Each lobe has a characteristic branching pattern. Growth of the prostatic epithelium does not occur uniformly. The ductal tips proliferate more rapidly than proximal ductal areas (5). The first 15 days postnatal are a period of intense morphogenetic activity with about 70-80% of the branch points formed during this time. Ductal branching is completed by 60-90 days postnatal (5). The adult prostate on the other hand is relatively growth "quiescent" (6).  1  1.1.2  Cytodifferentiation of the UGE and U G M  The differentiation of the prostatic epithelium and mesenchyme occurs concurrently with morphogenesis (7). Epithelial differentiation is characterized by the expression of different cytokeratins. From its inception the U G E uniformly expresses cytokeratins 5, 7, 8, 14, 18 and 19. These cytokeratins are subsequently segregated to either the basal or luminal epithelial subtype with differentiation. Basal cells lose expression of cytokeratins 8 and 18, while luminal cells lose expression of cytokeratins 5, 7 and 14. Cytokeratin 19 is expressed in basal cells as well as a subset of luminal cells in the adult prostate (8). There is a clear temporal and spatial relationship between ductal canalization and basal and luminal cell segregation. Epithelial differentiation proceeds in a proximal to distal manner with mature canalized proximal segments existing concurrently with solid undifferentiated distal tips (9). Coincident with epithelial differentiation, the U G M condenses to form smooth muscle sheaths immediately surrounding the epithelium, with looser connective tissue between individual ducts. The smooth muscles differentiate in a proximal to distal manner along the prostatic ducts, from the urethra towards the tips (10). By 20 days postnatal the epithelium and mesenchyme are differentiated. During puberty, the prostate undergoes significant proliferation with wet weight and D N A content increasing more rapidly than during the early postnatal period (11). Prostate secretory proteins are first detectable just prior to puberty (about postnatal day 20) and greatly increase in abundance  as serum testosterone levels rise (7). Although  accumulating evidence supports the opinion that neuroendocrine (NE) cells share a common origin with exocrine epithelial cells (12), their exact origin is still unclear.  2  1.1.3  Roles of Hormones in Prostate Development  Development, growth and function of the prostate are androgen-dependent. The synthesis of testosterone by the fetal testes begins at E13.5 in male mice (13). The UGS in both male and female embryos are morphologically indistinguishable until the onset of prostate development in males at E17.5. In females, due to lack of androgen, the UGS forms the lower portion of the vagina as well as the urethra (7). The development of prostatic tissue is not determined by fetal genetic sex, but rather by exposure to androgen. Application of androgen to cultured female UGS during the appropriate developmental period (i.e. E13.5 to P5) can induce development of the prostate (14), while removal of androgen from male UGS at E16.5 can totally abolish prostate development (15). The cellular response to systemic androgens is mediated by nuclear androgen receptor (AR) (7). Inactivating mutations of the A R result in the absence of prostate in both human and mouse (16-19). In the UGS testosterone is converted to dihydrotestosterone (DHT), which has a 10-fold greater affinity for A R (20), by the locally produced enzyme type II 5-a reductase (21). Mutation of the gene encoding type II 5-a reductase causes reduced prostate growth and development (22, 23), indicating DHT is a crucial local mediator of prostate development. At the initiation of prostate development (El7.5) A R is only detected in the U G M which is the actual target, and mediator, of the morphogenetic effects of androgens on the epithelium (24, 25). The expression of A R in the U G E gradually increases after birth, suggesting that androgens may not be directly involved in epithelial cytodifferentiation (9, 10). Tissue recombination and grafting experiments utilizing the UGS from testicular  3  feminized (Tfm) mice which have an inactivated A R due to a frame-shift mutation, showed that mesenchymal A R is required for establishing prostate identity. Although androgens are necessary to induce prostate budding, the process of ductal growth and branching displays less critical androgen dependency. Postnatal castration retards, but does not arrest, growth and branching of the ductal network (11). Indeed, ductal growth and branching in vivo are maximal at the lowest physiologic levels of androgen and diminish as the levels begin to rise at puberty (5). In the mature prostate androgens are believed to act on prostatic smooth muscle to maintain a fully differentiated growthquiescent epithelium via paracrine stromal-epifhelial interactions. The role of the epithelial A R seems to be in regulating secretory functions (26, 27). In addition to the androgens, the developing prostate is also sensitive to estrogens. Estrogen can affect prostate development both directly and indirectly. Exposure to high doses of estrogen (8-fold above normal levels) during fetal life inhibits prostate development (28). Estrogen-treated prostate explants are smaller with a decreased proportion of epithelial cells and lumen and an increased stroma. This is concurrent with a disturbance in the branching pattern and a reduction in ductal canalization (29). In contrast, low concentrations of estrogens (50% increase above normal levels) result in permanent increases in prostatic A R levels and prostate size (30). However, mice with inactivating mutations in the known nuclear estrogen receptors (ER), E R a or ERp, as well as E R a and ERp compound mutant mice, have normal prostatic development, suggesting that prostatic development is sensitive to the overall endocrine environment and that local actions of known ERs are not required for prostatic development (31).  4  Other circulating hormones also affect prostate development. Prolactin (PRL) stimulates the proliferation and differentiation of prostate cells both directly and indirectly (32). Hypertrophy of different lobes of the prostate was found in transgenic (Tg) mice overexpressing rat Prl (33). Prl knock out mice develop a small ventral prostate (VP), consistent with the ability of PRL to potentiate the effects of androgens on the prostate (34). Growth hormone (GH) is also essential for prostate development. G H can either potentiate the action of androgens on the prostate or act directly on this gland by a mechanism that does not depend on testicular androgens (35). G H antagonist Tg mice have significantly impaired prostate development with regard to the number of terminal duct tips in the dorsal prostate (DP) and V P (36).  1.1.4  Mesenchymal-Epithelial (M/E) Interactions in Prostate Development  M/E interactions between the endoderm-derived epithelium and mesoderm-derived mesenchyme are required for organogenesis of numerous organs including lung (37), pancreas (38), tooth (39) and kidney (40). Although the molecular mechanisms involved in M / E interactions within the various organs are unclear, it is believed that growth factors produced in the mesenchyme are required for epithelial differentiation and vice versa. The importance of M / E interactions in controlling prostate development has been well established. Both morphogenesis and cellular differentiation of the prostate depend on M / E interactions. U G M grafts grown in vivo alone have few differentiated smooth muscle cells and U G E grafts alone are maintained as undifferentiated epithelium (1). Tissue recombination experiments have also demonstrated an inductive role of U G M in prostate epithelial branching and differentiation. First, the induction by the U G M shows  5  regional specificity. Ventral U G M induces V P specific secretory protein expression in dorsal U G E (41), while complete U G M can induce the ventral epithelium of the adult prostate to express secretory proteins characteristic of all lobes (42). Second, other epithelial tissues derived from the cloaca, i.e. bladder, can be induced to differentiate into prostatic glandular epithelium after association with U G M (1). Finally, the instructive role of U G M seems to be conserved across species. Recombinations of mouse U G M and human fetal bladder epithelium are able to form glandular structures. The morphogenetic process is similar to that normally observed during human prostatic development (43). On the other hand, rat U G M recombined with human U G E differentiates into stroma with characteristics of the human but not the rat prostate, indicating that epithelium can affect mesenchymal differentiation and that the M / E interactions are reciprocal (44).  1.1.5  Molecular Mechanisms Regulating Prostate Development  Although substantial progress has been made in understanding prostate development in recent years, the molecular mechanisms regulating this complex biological process are largely unknown. Mesenchymal growth factors and their receptors are likely to be involved in M / E interactions. For example, two fibroblast growth factor (FGF) family members, FGF7 and FGF 10, have been implicated in prostate development. FGF7 is a potent proliferative inducer and its transcript is expressed in the mesenchyme of the developing prostate in neonatal mice (45-47). Tg mice overexpressing FGF7 develop prostate hyperplasia (48). However, loss of FGF7 function does not cause morphological abnormalities in the reproductive tract of adult male mice (49), probably due to the functional redundancy within the FGF family. The transcript encoding FGF 10 is also  6  expressed in the U G M and FGF 10 has been shown to stimulate growth and branching of the neonatal (P0) V P in vitro in the absence of testosterone (50). FgflO-l- mice have either prostate agenesis or rudimentary prostate buds (51), suggesting it is required for prostate development. Another protein implicated in prostate development is transforming growth factor pi (TGFpi) which is localized in the mesenchyme surrounding ductules in the fetal and neonatal prostate (52, 53). Growth of the developing prostate can be inhibited by addition of exogenous TGFpi to in vitro organ cultures (54). Prostate phenotypes of TgfBl Tg and K O mice have not been described (55, 56). Tg mice carrying the C3(l)-TGFp type II dominant negative receptor exhibit an abnormal morphology with multiple layers of epithelial cells lining the proximal ducts of the prostate (57). Although epidermal growth factor (EGF) protein, as well as transforming growth factor a (Tgfa) transcripts, are also expressed in mouse prostate (58, 59), there is no evidence that either of these factors are required for prostate development. Insulin like growth factor 1 (Igfl)-/- mice have a small prostate with reduced branching (36, 60). However, since these mice have dramatically reduced circulating androgen levels, the prostate phenotype might be indirect (7). Other genes involved in development of numerous organs have also been implicated in prostate development, including bone morphogenetic protein 4 (Bmp4), Sonic hedgehog (Shh) and Notchl. Bmp4, encoding a TGFp superfamily member, is highly expressed in the male U G M from E14 through birth. Bmp4+I- mice exhibit an increased number of duct tips in both the V P and coagulating gland (CG), suggesting it restricts prostate ductal budding and branching morphogenesis (61). Shh is expressed in the U G E and the time course of its expression coincides with the formation of the main prostatic  7  ducts (62). Transcripts encoding the SHH receptor, Patched 1 (PTC1), are detected in the U G M , as are the transcripts encoding their downstream transcriptional regulators GLI1, GLI2 and GLI3 (63). Organ cultures treated with SHH show decreased cell proliferation and increased differentiation of luminal epithelial cells (64). S H H signaling is not essential for prostatic induction, i.e. bud formation, although Shh knockout fetuses display morphological defects such as an increase in ductal tip number (65, 66). This effect is at least in part mediated through suppression of FgflO expression, in addition to the upregulation of expression of Bmp4, activin A and Tgffll, which are expressed in the stromal cells and their proteins have been shown to inhibit prostatic epithelial branching (64, 67-70). Notchl is expressed at high levels in the epithelial cells of the developing prostate and downregulated in the adult. It is involved in cell differentiation and cell fate determination  (71).  Elimination  of Notch 1 -expressing  cells inhibits branching  morphogenesis, growth, and differentiation of early postnatal prostate in culture, suggesting that Notch 1 -expressing cells define the progenitor cells in the prostatic epithelial cell lineage, which are indispensable for prostatic development and re-growth (72). Other molecules involved in prostate development include activin A and its antagonist follistatin. Activin A is expressed in both mesenchyme and epithelium while its receptors (i.e. ActRIA, ActRIIA, and ActRIIB) are expressed in the epithelium. Follistatin is also expressed in the epithelium. Application of activin A or follistatin to in vitro organ cultures indicates that they have opposite effects: activin A limits epithelial growth and morphogenesis, while follistatin promotes this process (70). Hyaluronan, a polysaccharide component of the extracellular matrix, and its receptor CD44, are also involved in prostate morphogenesis. Treatment of anterior prostate (AP) organ cultures  8  with hyaluronan/CD44 antagonists impairs branching morphogenesis (73). Inhibition of fucosyltransferase  1, a transmembrane enzyme, reduces epithelial proliferation of  cultured rat AP (74). The interaction between urokinase plasminogen activator (uPA) and its receptor (uPAR) in prostate development has also been studied. Disruption of this interaction inhibits growth and differentiation in V P cultures with associated increases in apoptosis (75). Transcription factors have also been extensively studied in prostate development. The Hox gene family plays important roles in development and levels ofHox gene expression correlate with morphogenetic activity in prostate ductal development (76). The Hoxl3 paralogs are especially important to prostate development (77). Hoxal3 and Hoxdl3 are expressed in the U G M and U G E from E l 5 and gradually diminish with age (78). Gene targeting to delete one allele of Hoxal3 results in a prostate phenotype characterized by decreased size and branching of the dorsal-lateral prostate (DLP) and V P (78). Homozygous deletion of Hoxal3 disrupts genitourinary development, resulting in severe hypoplasia of the UGS (79). Hoxdl3 deficient mice have a similar, but milder, phenotype (80). Compound Hoxal3+/-, Hoxdl3-/-  mutant mice exhibit a more severe prostate  phenotype, including absence of the A P (79). The similarity in expression patterns of Hoxal3  and Hoxdl3  during prostate development and the overlap between the  phenotypes resulting from their inactivations is consistent with their additive function, and partial functional redundancy. Hoxbl3 is expressed in both developing and adult prostate (81, 82). Mice with a disruption in Hoxbl3 exhibit absence of VP-specific secretory proteins and defects in the cellular morphology of the V P epithelium, indicating that HOXB13 is specifically required for the proper differentiation of the V P epithelium  9  in the adult. Double homozygous mutants for Hoxbl3 and Hoxdl3 showed a 50% reduction in the number of duct tips, providing evidence for redundant functions of HOXB13 and HOXD13 in V P development (77). HoxalO mutants showed reduced A P branching and partial A P to D L P transformation, while the D L P and V P were not affected (83). Nkx3.1, encoding a non-Hox homeobox protein, is required for the proper development of the prostate and displays androgen-dependent expression (84, 85). Nkx3.1 homozygous mutants show reductions in duct tips and impaired secretory function (86). In addition, both Nkx3.1 heterozygous and homozygous mutants display a high incidence of prostatic intraepithelial neoplasias (PIN) (87). Loss of function of NKX3.1 in mice cooperates with loss of function of the phosphatase and tensin homolog (PTEN) tumor suppressor gene in prostate cancer (PCa) progression (88). Tumor protein 63 (p63), a homologue of p53, is expressed in the majority of U G E cells during prostate development and gradually becomes restricted to basal cells in the mature prostate (8). p63-l- mice die at birth and no prostatic epithelial buds are present (89). p63-l- male UGS engrafted under the renal capsule develops into prostate tissue with N E and luminal cells, but the basal cells are absent indicating this molecule is essential for differentiation of basal cells (90). Despite this wealth of knowledge regarding the growth and transcription factors regulating prostate development, many questions remain. For example, although all evidence indicates an important role for M / E interactions in prostate development, the mesenchymal factors that mediate the effects of androgen on the developing U G E , as well as the epithelial factors that direct mesenchymal differentiation remain to be  10  identified. Moreover, the mechanisms responsible for developmental heterogeneity, i.e. development of region-specific identity and differences of cellular proliferation, secretory activity and cellular architecture along the proximal-distal axis of the prostate ducts, are still to be understood (7).  1.2 The Adult Mouse Prostate: Function, Anatomy and Histology The prostate is an exocrine gland found in all male mammals. It expels a complex proteolytic solution to liquify the semen after ejaculation thus facilitating transportation of sperm (3). Interest in understanding prostate biology was stimulated mostly by the high incidence of human prostatic proliferative diseases, i.e. benign prostatic hyperplasia (BPH) and prostate adenocarcinoma. The mouse prostate is located beneath the bladder, surrounding the urethra. It is composed of four pairs of distinct lobes: A P or C G , VP, DP and LP. The DP and LP are often grouped together as D L P . Each lobe has distinct histological features, with extensive epithelial-infolding in the AP, significant but less extensive epithelial-infolding in the DLP, and minimal epithelial-infolding in the V P (7). In addition, these lobes also show specific biochemical features as was shown by two-dimensional electrophoresis (2DE) of the protein profiles (91). The mature mouse prostate is composed of epithelial ducts surrounded by a stromal layer. There are three types of cells lining the epithelial ducts: luminal cells (also called tall columnar or secretory cells), basal cells and a minor population of N E cells. The luminal cells produce secretory proteins which comprise part of the semen. Basal cells, may be reserve or stem cells capable of differentiating into columnar secretory cells (8,  11  92). The function of the N E cells is still not clear. The stromal component of the prostate encompasses all cellular and extracellular elements outside of the epithelial basal lamina. The stromal cells adjacent to the epithelial cells are a thin layer of smooth muscle cells. Other stromal cells include: fibroblasts, blood vessels and their associated pericytes, wandering connective tissue cells, nerve terminals, and lymphatic cells, all of which are embedded in a loose collagenous extracellular matrix (93).  1.3 Prostate Cancer (PCa) 1.3.1  Brief Overview of PCa  PCa constitutes a major heath problem in Western countries. It is the most frequently diagnosed cancer in Canadian men and the third leading cause of male cancer deaths. In 2004, it was estimated that 20,100 men would be diagnosed with PCa and 4,200 would die of it (http://www.cancer.ca/ccs/internet/standard/0,3182,3172_12851  langld-en,00.  html). Prostate specific antigen (PSA) is a sensitive serum marker for the early diagnosis of PCa, but it is not specific as its level is also frequently elevated in patients with B P H and prostatitis. Patients with localized PCa can be treated by surgery or radiation therapy. For patients with advanced disease, hormonal ablation therapy is initially successful as most of the tumors are androgen-dependent at the time of diagnosis (94). However, the disease will eventually progress to androgen-independence with metastases, most commonly to the bone (95). At present, there is no effective therapy for PCa at this stage. Thus development of new diagnostic markers and therapeutic targets is urgently required.  12  1.3.2  Genetic and Epigenetic Factors in PCa  The risk of developing PCa is strongly influenced by familial history. Familial PCa accounts for about 40% of patients who present younger than 55 years and for up to 9% of those presenting at 85 years or older (96). Genome wide linkage analysis and loss of heterozygosity (LOH) studies identified several potential high-penetrant loci, such as HPC2/ELAC2 on chromosome 17pll, R N A S E L on chromosome lq24-25, and MSR1 on chromosome 8p22, that are likely to be involved in familial PCa (97, 98). However, none of the loci have indisputably been verified by independent studies, confirming the tremendous heterogeneity in the predisposition to PCa (99). In addition, common low penetrant loci might also contribute to the development of PCa. A number of genes containing polymorphisms have been studied extensively for association with PCa including: AR, PSA, type II 5-alpha reductase and cytochrome P450 (Cyp) isoforms (100). Epigenetic changes also happen in some PCa patients. Hypermethylation of the promoter region of glutathione S-transferase class n gene (Gstpl) is the most frequent somatic change associated with PCa (101).  1.3.3  Disregulation of Oncogenes and Tumor Suppressor Genes (TSGs)  Oncogenes and TSGs are important factors in PCa. The role of many oncogenes has been extensively studied. R A S is activated by virtually all of the growth factors upregulated in PCa, yet Ras mutations are infrequent in PCa and do not yield clinically useful prognostic information (102). A number of studies have shown that c-Myc mRNA is elevated in PCa (l03), as well as in B P H patients (104). C-ERB-B2 (HER2) has been studied in PCa and its clinical utility as a tumor marker or therapeutic target remains  13  unclear. Amplification and overexpression of HER2 is not common in early stage PCa. Instead, overexpression of HER2 is more common in androgen-independent PCa (60%) and correlates with shortened survival (105). Enhancer of zeste homologue 2 (EZH2), a repressor of transcription, is increased significantly in PCa at both m R N A and protein levels and repression of Ezh2 transcripts by siRNA decreases PCa cell proliferation (106). TSGs such as retinoblastoma (Rb), Nkx3.1, p53, Pten andp27, have been implicated in PCa (107). Reduced levels, or absence, of RB protein have been found in some PCa patients (108). Wild type (WT) Rb cDNA, introduced into the Rb-mutated PCa cell line DU145, results in a 15 fold decrease in tumorigenicity of cancer cells in nude mice (109). Transfection of WT p53 suppresses growth of human PCa cells containing mutant p53 alleles (110). Deletion or mutation of Pten is reported in a significant number of cancers, including PCa. The cyclin-dependent kinase inhibitory protein, P27, inhibits cell proliferation and induces arrest in the G l phase of the cell cycle. Decreased expression of P27 is seen in PIN and its expression is also related to the pathological stage and grade of the tumor (107). Although alleles of many TSGs are deleted in PCa patients, the frequency of mutations in TSGs is low in primary PCa, but becomes much higher in advanced PCa (111). For example, p53 mutations are rare in early, localized PCa, whereas they are found in about 40% of all advanced PCa (99). Further, evidence that haploinsufficiency of Nkx3.1 and Pten may be involved in PCa comes from analysis of mouse models (86, 112). Although oncogenes and TSGs have been extensively studied in PCa, none of them has been linked conclusively with the initiation or early progression of PCa. This is  14  probably due to small patient cohorts, methodological variations (98) and the heterogeneity of the disease.  1.3.4  Signaling Pathways Involved in PCa  It has become evident that development of cancer often involves inappropriate reactivation of pathways that are active during normal development. For example, aberrant proliferation and branching morphogenesis can be seen in PCa. Furthermore, abnormal outgrowth of the stem cells may also play a key role in PCa development (113). Thus, the study of prostate development may provide insights into the molecular mechanisms important in prostate neoplasia as well as reveal new markers for its detection (114). Many signaling pathways are involved in PCa. Development of PCa is androgendependent (115) and reactivation of the androgen signaling pathway is an important factor in development of androgen-independent PCa after androgen depletion therapy (116). A R mutations are not common in primary PCa but are frequently found in advanced and metastatic disease (117). A R amplification is detected in about 30% of androgen-independent  PCa after  androgen withdrawal (118). Ligand-independent  induction of A R transactivation is promoted by growth factors such as EGF, IGF1 and FGF7, as well as cytokines (e.g. Interleukin 6) (119) through activation of proliferation and differentiation pathways such as the protein kinase B (120) and mitogen-activated protein kinase pathways (121). In advanced PCa EGF, TGFa, FGF7 and IGF1, as well as their receptors, have been reported to be overexpressed (122). Perturbations in the rate of apoptosis are also important contributors in PCa. BCL2, an apoptosis suppressing  15  oncoprotein, is highly expressed in advanced PCa (123) and overexpression of Bcl2 protects PCa cells from apoptosis and confers resistance to androgen depletion (124).  1.3.5  Changes in Stromal-Epithelial Interactions in PCa  In the adult prostate reciprocal interactions between the stroma and the epithelium maintain prostatic functional differentiation and growth-quiescence. The deregulation of stromal-epithelial interactions and the continued interactions of carcinoma cells with the stromal microenvironment contribute to both initiation and progression of carcinogenesis. Transformed epithelium is incapable of maintaining normal differentiation of adjacent smooth muscle in PCa (125); thus the stroma undergoes progressive loss of smooth muscle with the appearance of carcinoma-associated fibroblasts. In turn, the altered stroma can promote carcinogenesis of nontumorigenic prostate epithelial cells, while normal stroma has no such effect (126). Increases in autocrine and paracrine growth factor loops appear to correlate with PCa progression from a localized and androgendependent disease to the androgen-independent state. Conversion from paracrine to autocrine mechanisms of growth factors support, and may also be involved in, malignant transformation of prostatic epithelial cells (127, 128).  1.3.6  Modelling PCa  Human PCa is a heterogeneous disease. There is a long latency period between the appearance of PIN and clinically detectable carcinomas. This makes understanding the molecular mechanisms involved in PCa initiation and the early steps of progression nearly impossible (88). Much of our knowledge of the molecules involved in PCa comes  16  from analysis of mouse models which provide significant advantages for demonstrating the molecular mechanisms involved in PCa initiation and progression.  1.3.7  Xenograft Models  Several mouse PCa models have been described (http://emice.nci.nih.gov/emice/ mouse_models/organ_models/prostate_models/models_overview).  Among  them  are  xenograft models. Human PCa cells or tissue fragments are injected or grafted into immunodeficient mice such as nude mice or Severe Combined Immunodeficient (SCID) mice. There are 3 major graft sites in xenografting studies: sub-cutaneous, sub-renal capsule, and orthotopic. Each of these sites has its own advantages and disadvantages. The traditional sub-cutaneous engraftment is easily accessible and has a high graftcarrying capacity. However, the site is poorly vascularized and the success rate is still very low, especially for primary tumors. Orthotopic grafting has the perceived advantage that the site is representative of the environment in which the tumor originated. However, it is technically demanding and has a limited capacity. In addition, due to the effect of androgen ablation on the host prostate, it is impractical for castration related studies. Approximately 25 xenograft models of human PCa have been established using the subcutaneous and orthotopic grafts (129). These models were derived from either primary tumors or distant site metastases, thus representing the whole history of clinical behavior of human PCa. However, it is important to note that the success rate of these xenografts is still very low, though increased significantly from approximately 5% in early 80s to 38% in early 90s (129). Wang et al showed a take rate of 50% and 70% using sub-cutaneous and orthotopic xenograft sites respectively. They have also developed the sub-renal  17  capsule xenograft model, which was shown to be the most efficient in terms of take rate for PCa tissues amongst the three sites (90%). The sub-renal capsule and the orthotopic grafts exhibit better histopathologic differentiation than the sub-cutaneous  grafts,  although the surgery is more technically demanding and the xenograft carrying capacity is lower than sub-cutaneous grafting (130). The major advantages of using xenograft models for PCa pre-clinical testing are that human cells or tissues are used and changes in PSA levels, which mark PCa progression, can be easily monitored. Major inherent limitations of this model are that the absence of an immune system means it cannot be used to study PCa immunity and that the xenograft tumor cells rarely develop metastases (129, 130).  1.3.8  Knockout (KO) and Tg Mouse Models of PCa  K O mouse models have provided key insights into the molecular mechanisms involved in PCa. However, relatively few K O mice display prostate phenotypes. One possible explanation is that these mutant mice die of unrelated developmental defects or diseases in other tissues before developing PCa; an alternative possibility is that the prostate phenotypes have simply not been examined. Two K O mouse models with prostate phenotypes have been demonstrated. Nkx3.1 heterozygous and homozygous mice display an extensively hyperplastic epithelium with focal areas of severe dysplasia (86, 88) thus providing a model for studying mechanisms of PCa initiation. Pten heterozygous mice develop cancers of multiple tissues including the prostate (131, 132). As indicated above, conventional K O manipulations of various genes may cause embryonic lethality, thus prostate specific ablation of such genes has been used to  18  examine their roles in prostate development and PCa. For example, ablation of the retinoid X receptor a specifically in the prostate leads to the development of preneoplastic lesions (133). Similarly, prostate-specific inactivation of P T E N leads to the development of invasive adenocarcinoma with metastases (134). The resulting phenotypes of most single-gene Tg or K O mice consist primarily of hyperplasias that often develop in animals of advanced age, suggesting that expression or loss of a single gene is not sufficient to lead to carcinoma. Two studies have investigated the effect of cooperative loss-of-function of 2 genes. Compound mutant mice with homozygous Nkx3.1 deletions and heterozygous Pten deletion (88) display an increased incidence of high grade PIN (HGPIN)/early carcinoma lesions, which resemble the early stages of human prostate carcinogenesis. Likewise, concomitant inactivation of one Pten allele and one or both p27 alleles accelerates spontaneous neoplastic transformation and the incidence of prostate tumors (135). Most Tg mouse models have been generated using promoters directing expression of the transgene to prostate epithelial cells. The most commonly used include the probasin (136), C3(l) (137) and Psp94 promoters (138). There are two general classes of Tg models of PCa. The first class utilizes the promoters mentioned above to express various molecules ranging from growth factors to regulators of the cell cycle and apoptosis (139) including c-Myc (140, 141), Igfl (142), AR (143), Bcl2 (144, 145), Fg/7 (146), Fgf8b (147) and Tgfp (57). It is interesting to note that most of these Tg lines are prone to develop PIN but not adenocarcinoma and often these phenotypes do not arise until the mice are of advanced age (>6 months).  19  The second class consists of models resulting from expression of simian virus 40 (SV40) early genes. SV40 is a virus of monkey origin and the most potent transforming virus frequently found in human tumors. Its early genes, including large T antigen (TAg) and small t antigen (tAg), encode regulatory proteins required for viral replication. SV40 contributes to transformation by perturbing several intracellular pathways. T A g simultaneously disables both the RB and the P53 tumor-suppressor pathways, whereas tAg perturbs protein phosphatase 2A (PP2A) (148). Only a limited number of promoters have been used to drive the expression of SV40 early Ags. Among them probasin promoter (136) is by far the most frequently used. Models of this type include the TRansgenic Adenocarcinoma of the Mouse Prostate (TRAMP) model that utilizes the minimal rat probasin promoter to express both TAg and tAg (136), as well as a number of Tg lines using the long probasin promoter to express only large TAg, collectively termed the " L A D Y " models (137). Disease progression in the " L A D Y " model is less aggressive than in T R A M P mice, perhaps because this transgene lacks the small tAg. The " L A D Y " mice consistently develop multifocal low grade PIN (LGPIN) that progresses to HGPIN and early invasive carcinoma, but generally fail to metastasize (149). Also in this class of models are the C3(l)-Tag (both T A g and tAg) mice (137) which display a reproducible progression from LGPIN to HGPIN to invasive carcinoma. However, the utility of these lines is limited due to the onset of tumors in multiple tissue sites such as bulbourethral gland and the urethra (150). The recently developed Psp94-Tag (138) mice have SV40 Tag (both TAg and tAg) driven by the Psp94 promoter. PSP94 is a prostate specific secretory protein. These mice develop prostate hyperplasia as early as 10 weeks of age, with subsequent emergence of PIN and eventually metastatic disease at about 20  20  weeks of age. Other models in this class include Cryptdin-2-1Ag  (151) and Gg-SV40  TAg models (152). However, transgene expression is not prostate specific in these models and the mice develop N E tumors of a highly aggressive nature which do not appear to progress through an androgen-dependent stage. Thus the utility of these models for studying PCa is limited.  1.3.9  TRansgenic Adenocarcinoma of the Mouse Prostate (TRAMP)  The T R A M P model was generated on a C57B1/6 background utilizing the prostate specific rat probasin promoter (-426/+28) to drive expression of both SV40 TAg and tAg, as mentioned above. Because probasin expression is prostate specific and regulated by androgen (153), the transgene is detected specifically in the epithelium of the prostate from as early as 4 weeks of age and expression is maximal by the time the mice are approximately 12 weeks of age  (http://thegreenberglab.fhcrc.org/research/research_  tramp_uses.html). The transgene is expressed primarily in the D L P and V P , with relatively minor expression in the A P and seminal vesicles (SV) (153). Histological changes, such as changes in nuclear size, may be observed within scattered regions of prostatic acini as early as 4-6 weeks of age, which mimics human LGPIN. By the time the T R A M P mice are 10-16 weeks old, HGPIN is generally present throughout the majority of the D L P (154). It should be noted that the V P usually has a less severe phenotype. At 24 weeks of age, approximately 100% of males develop poorly differentiated and invasive adenocarcinomas (137). Metastasis can be seen at about 18-24 weeks of age and is commonly detected in the lymph nodes and lungs and less frequently in the bone, kidney and adrenal glands (139). The development of PCa in T R A M P mice  21  is initially androgen-dependent, as castration of the mice at 12 weeks of age causes an initial regression of the prostate. However, the remainder of the cells eventually progress to poorly differentiated androgen-insensitive cancer (139). Despite the many parallels between T R A M P and clinical PCa, there are concerns that invasive N E carcinoma has been observed in some T R A M P mice, which may limit the use of this model (155).  1.3.10 Differences Between Mouse PCa Models and Human PCa Although mouse models are widely used to study human PCa, it should be noted that there are some significant differences between the human and mouse prostate. The adult human prostate is not divided into discreet lobular structures as the mouse is. The mouse DLP has been stated to be the most homologous to the human peripheral zone since they originate from the same area in the UGS during embryonic development (156). In mouse individual ductules are surrounded by only a few layers of smooth muscle cells, unlike the abundant intervening dense fibromuscular stroma surrounding the epithelium in human prostate. The human prostate has a continuous basal cell layer while the basal cells in the mouse prostate are discontinuous. Most importantly, mice do not spontaneously develop PCa while it is the most commonly diagnosed cancer in men (155). The pathology of genetically manipulated mice is also different from that of humans. For example, the PIN lesion in human is different from that in mouse. Human PIN lesions can be graded as LGPIN and HGPIN, with HGPIN considered having true potential for progression to invasion. In mouse models, some develop only mouse PIN (mPIN) without the capacity for invasion or metastasis. The histological features of PIN  22  in human and mouse also differ. Compared with human HGPIN, the atypical nuclei seen in SV40-based mPIN lesions appear to be more elongated, more hyperchromatic, and have a greater mitotic and apoptotic rate. There are also obvious histological differences in mPIN lesions in the various mouse models and these differences do not seem to relate to the variations in the potential for development of invasive carcinoma (155). The ideal mouse model of PCa would be one that accurately mimics the heterogeneity, androgenindependent growth and bone metastasis of human PCa. However, there has not yet been a mouse model developed which contains all the features mentioned above, suggesting more work is required in this area.  1.4 Gene Expression Profiling 1.4.1  Summary of Techniques  Nearly all biological events are associated with extensive changes in gene expression. Thus, by comparing gene expression profiles at different stages of development or disease progression, genes that play important roles in these processes can be identified (157). Many techniques to identify differentially expressed genes have been developed and used widely because of their high efficiency and the tremendous amount of information  they  provide. These  technologies  rely on sequencing,  differential  hybridization or combinations thereof and include: differential display, subtractive hybridization, microarray analysis, expressed sequence tags (ESTs) and serial analysis of gene expression (SAGE). Differential display is a commonly used technique to identify genes differentially expressed in different cells or tissues. The 3' terminal portions of mRNAs are amplified  23  by reverse transcription using anchored primers designed to bind to the 5' boundary of the poly-A tails. This is followed by polymerase chain reaction (PCR) amplification with additional upstream primers of arbitrary sequences. Thus certain mRNA sub-populations are amplified and then visualized by running the PCR products on a polyacrylamide gel. This allows direct side-by-side comparison to identify genes differentially expressed in 2 or more samples. The differentially expressed bands can be cut from the gel and sequenced to allow identification. This method has been widely applied in developmental biology, cancer research, neuroscience, pathology, endocrinology, plant physiology, and many other fields (158). Subtractive hybridization is also used to identify differentially expressed genes in two samples. Briefly, cDNA from one sample (called "driver") is hybridized with c D N A from another sample (called "tracer") so that sequences common to both samples are "subtracted", leaving a population of cDNA species enriched for sequences preferentially expressed in one of the samples (159). These differentially expressed cDNAs are then sequenced to allow identification. Recently much attention has focused on the use of microarrays for determination of the mRNA fingerprint of a cell type or tissue. Thousands of cDNAs or oligonucleotides (called "probes") are printed on slides. Fluorescently labeled mRNA from the tissue of interest (called "target") is added and hybridization between probe and target provides a quantitative measure of the abundance of a particular sequence in the target population. Alternatively, two samples to be compared can be labeled with different dyes and used to hybridize to the same slide. After stringent post-hybridization washes, the slide is examined to determine the pattern of binding of the sample (or samples) to the probes.  24  Comparison of hybridization patterns enables the identification of mRNAs that differ in abundance in two or more target samples (160, 161). ESTs are small pieces of cDNA (usually 200 to 500 nucleotides long) generated by random cloning of the reverse-transcribed cDNA representing mRNA species expressed by a cell type or tissue. Ends of the clone (one or both) are sequenced and these "tags" are used  to  identify  the  gene  by  matching  base pairs  to  chromosomal D N A  (http://www.ncbi.nlm.nih.gov/About/primer/est.html). The relative abundance of each EST is assessed via the number of clones representative of each sequence (159). SAGE is a powerful gene expression profiling method originally developed by Velculescu et al (162). Historically, when the method was developed, it was believed that: 1) a short sequence tag (14 base pairs) contains sufficient information to uniquely identify a transcript; 2) the concatenation of tags allows for an increased efficiency in a sequence-based analysis. In the original S A G E protocol, total R N A was isolated from the tissue of interest using standard R N A extraction methods. c D N A was synthesized from poly(A) R N A using biotinylated oligo (dT) primers bound to streptavidin-coated beads. +  The cDNA was then cleaved with an anchoring enzyme (i.e. Nlalll) and the 3'-terminal c D N A fragments bound to the beads were isolated and separated into 2 aliquots. Two different oligonucleotide adapters, both containing recognition sites for a tagging enzyme (i.e. BsmFl), were linked separately to the bound cDNA fragments. The tagging enzyme is a class II restriction endonuclease that cleaves the D N A at a certain number of bases 3' to the recognition site, resulting in the release of a short tag plus the linker from the beads. The 3'-ends of the released tags plus linkers from each sample were then blunted and ligated to each other to form ditags. After P C R amplification of the ditags using  25  primers whose sequences were derived from the linker sequences, the linkers were released from the ditags by digestion with the anchoring enzyme. The ditags were then concatenated and cloned into a sequencing vector. Sequencing enables individual tags to be identified and the abundance of the transcripts for a given cell line or tissue to be determined. Comparing the sequence information from the tags with GenBank and other databases provides qualitative information about the transcribed genes. The frequency of a specific tag within the S A G E tag population correlates with its relative abundance in the cell and gives quantitative information about expressed genes. The S A G E method therefore allows for comparison of expressed genes under various physiological and pathological conditions (162). Since over 30 tags can be read serially in a single sequencing reaction, it is at least 30-fold more efficient (163) than conventional EST analysis. In theory, S A G E is sensitive enough to detect low-abundance transcripts (164). For example, a 14 base pair tag SAGE library can potentially identify 4  10  (1,048,576)  unique transcripts, a number sufficient to identify virtually all transcripts in the whole genome since the human genome was predicted to contain only about 30,000-40,000 genes (165). A major difference between SAGE and D N A microarrays is that S A G E does not require prior knowledge of the sequence of the transcripts. This makes it a potent tool for the analysis of organisms whose genomes are largely unknown to identify novel genes as well as alternatively processed transcripts (166). Furthermore, S A G E data are digital, easily exchanged between laboratories for comparison and can be added to by scientists for years to come. As with all other techniques, SAGE has its own drawbacks. First, the expense of generating and sequencing a SAGE library could limit the sequencing depth thus  26  reducing the ability to detect low abundance transcripts. Similarly, the expense makes it impractical to repeat the libraries and thus information on biological variation and experimental precision is not available. Second, depending on the anchoring enzyme (most often M a l l l ) used some mRNA species will be missed i f they do not contain the recognition site for this enzyme. It has been shown that about 2-3% of the annotated transcriptome in Drosophila and C. elegans would be missed due to lack of Nlalll sites (167). Third, errors observed in SAGE data are inherent to the experimental process, including sampling, P C R and sequencing errors, as well as the nonuniqueness and nonrandomness of tag sequences (168). Sequencing artifacts have the potential to affect novel gene discovery. Similarly, tag-to-gene mapping is also hampered by ambiguous tags, sequencing errors and the incomplete genome information (169). Finally, SAGE tags need to be mapped to the genes they represent using databases which may sometimes contain incomplete and nonvalidated data and this may affect the accuracy of SAGE quantification. Based on comparisons of advantages vs. disadvantages, S A G E was used for our studies and thus will be discussed in more detail below.  1.4.2  Technical Improvement of SAGE  One disadvantage of the original SAGE protocol was the need for a relatively large amount of total R N A (~5u.g of poly(A) R N A ; approximately 5 1 0 cells), thus limiting x  6  its utility for analysis of small biological samples such as tissue biopsies or microdissected material (170). Numerous attempts have been made to reduce the quantity of starting material. Two methods, SADE (171) and MiniSAGE (172), diminish the loss  27  of material throughout the procedure thus permitting the amount of input R N A to be lowered. For example, an mRNA capture kit was used to allow the first five steps to occur within one tube thus significantly reducing the loss of material between successive steps. Similar methods were applied to MicroSAGE (170). In addition, MicroSAGE added a limited number of additional PCR cycles to generate sufficient ditags. Some other PCR-amplification-based methods were also developed. P C R - S A G E (173) and SAGE-Lite (174) both rely upon the inherent poly(C) terminal transferase activity of reverse transcriptase  to switch templates  during D N A polymerization (Clontech  Switching Mechanism At 5' end of R N A Template (SMART) system). This allows generation of PCR-amplified cDNA prior to the SAGE procedure. Similarly, single cell SAGE (scSAGE) (175) has recently been demonstrated, although there was no convincing validation of the results. Since the additional PCR steps are likely to introduce biases into the SAGE libraries, small amplified R N A - S A G E (SAR-SAGE) (176), an alternative method to generate sufficient transcripts, was devised in which a loop of linear amplification of R N A was included before cDNA synthesis. Various methods have also been used to increase the efficiency of SAGE. A modified protocol using biotinylated primers to prepare P C R reactions of S A G E ditags helped to remove contaminating linker sequence thus increasing the yield by 43% (i.e. from -22 tags to about 35 tags per concatemer) (177). A heating step introduced to disrupt small concatemer aggregates increased this to an average of 67 tags per concatemer (178). Adding a single purification step before M a l l l digestion of the ditags increased the yield of digested ditags by removing soluble contaminants (179). Introducing low-cycle P C R amplification of the 3' cDNA before the BsmFl digestion and gel purification of the  28  5sraFI-released tag fragments before ditag formation provides a large quantity of initial 3' cDNAs and high quality tags and ditags for the construction of S A G E libraries (180). The original S A G E protocol used 7VMII as the anchoring enzyme and BsmFl as the tagging enzyme to yield a tag of 14 base pairs. Alternative enzymes have also been used. Sau3Al, an enzyme that recognizes a 4 base pair cutting site, has been used as the anchoring enzyme to generate 14 base pair tags at different 3' sites of genes (171). Although a 14 base pair tag SAGE library was believed to be able to identify virtually all transcripts in the whole genome (165), in reality, 14 base pair tags are not absolutely specific and ambiguous tags exist. It was reported that about 8% of the D. melanogaster and 15% of the C. elegans genes can not be detected unambiguously (167). This rate may be even higher in species with increased gene numbers such as human and mouse, which makes data analysis more difficult. Several groups have tried to solve this problem by generating longer S A G E tags. The M A G E method generated S A G E tags of 18 base pairs using Rsal as the anchoring enzyme (169, 181). By using Mmel as the tagging enzyme, the LongSAGE method generates 21 base pair tags and thus is able to distinguish 4  17  different tags (182). Comparison of the short S A G E and LongSAGE methods showed that 4-5% of the short tags corresponded to more than one confident long tag. The tradeoff when using LongSAGE is that the average number of 21 base pair tags sequenced from a library is only 50% to 60% of what could be achieved with the same budget when sequencing a short library (183). Robust long S A G E (184) has substantially improved the LongSAGE method by reducing the amount of mRNA needed, enhancing ditag formation, and reducing ditag PCR reactions and improving cloning efficiency.  29  The application of SAGE to identify the 5' ends of transcripts has been described by 2 groups (185, 186). These methods are based on the 5'-end capping method (187) (188). The 5' cap of the mRNA is removed and two different oligo adapters containing both a type lis restriction enzyme site (Mmel) and another restriction enzyme site, such as Xhol (185) or BamHl (186), are added to the 5'-end of the two aliquots of mRNAs. The remaining steps are basically the same as in the original SAGE procedure. Combination of data from both 5' SAGE and 3' SAGE allows the precise localization of the start and ending point of the transcripts (186).  1.4.3  Application of SAGE  Since it was first reported, SAGE has been used extensively in two types of applications: 1.) Comparison of samples in different states to identify differentially expressed transcripts; 2.) Novel gene discovery. SAGE has been widely used for expression analysis in different organisms including loblolly pine (189), rice (190), yeast, fungus, C. elegans, Drosophila, rat, mouse and human (167). The common biological phenomena being analyzed using SAGE are human diseases, among which the analysis of various types of cancer is probably the most common. By comparing the gene expression profiles derived from cancers and normal tissues, genes involved in tumorigenesis and cancer progression can be identified. SAGE has been used in studies of gastrointestinal cancer, lung cancer, thyroid cancer, breast and ovarian cancer, neuroblastoma and glioblastoma, renal cancer and PCa (191). In these studies, SAGE was used for the identification of specific cancer markers that can help in early detection of minimal disease (192), genes and pathways involved in the  30  regulation of cell senescence (193), comparison of a severe phenotype to a mild one to identify genes involved in disease progression (194), evaluation of downstream gene expression changes affected by a key tumor suppressor gene (195), genes related to cancer progression or metastasis (196), and genes involved in drug resistance or sensitivity (197). In addition, SAGE has been used in a variety of other human diseases such as immune diseases (169) and cardiovascular diseases (198). The interest in understanding stem cell biology has also brought S A G E analysis to this field (199, 200). SAGE libraries have also been generated to study embryonic development in Drosophila as well as in mouse. SAGE has been applied to detect genomic responses of the Drosophila embryo to J N K signaling (201). It was used to identify genes involved in determination of mouse limb identity (202), mouse neocortex development (203), and mouse fetal thymocyte differentiation (204). Validation of the differentially expressed genes identified through analysis of the SAGE libraries can be achieved by numerous techniques including Quantitative Real Time P C R (RT-qPCR) (192, 205) and Northern blot analysis (206). It is conservatively estimated that there are more than 15,000 previously unpredicted genes in the human genome (182). S A G E can efficiently identify novel transcripts or novel genes in the genome. The "Generation of Longer cDNA fragments from SAGE tags for Gene Identification (GLGI)" method verified more than 1000 tags from the public SAGE database by converting SAGE tags into longer 3' complementary cDNA fragments (207). A reverse transcription (RT)-PCR based method, Rapid analysis of unknown SAGE tags (Rast-PCR), was used to identify potentially novel genes from tags that do not match to known genes (208). By using a reverse primer derived from the M l 3  31  sequence added to the 3' end of the transcripts during the cDNA synthesis step and the SAGE tag as the forward primer, they successfully amplified the 3' end of all 14 selected transcripts. SAGE R A C E (SARA)-PCR (209) generated the 3' end of the transcript using a similar method.  1.4.4  Bioinformatic Analysis of SAGE Data  The SAGE technique produces a digital output which necessitates efficient computational tools for data management and analysis. The main functions of SAGE software are: 1) extraction of tag sequences and counts from raw sequence files, 2) matching tags to reference sequences in other databases (a process termed "tag mapping") and 3) comparison of tag abundance between different S A G E libraries (210). A variety of S A G E databases are publicly available. SAGEnet (www.sagenet.org) contains S A G E libraries from colon cancer, pancreatic cancer and corresponding normal tissue, as well as some mouse and yeast SAGE libraries. This site is known for its relatively comprehensive  introduction to SAGE techniques and references to its  application in the literature. Most of current SAGE libraries are shared resources available  in the  Gene  Expression  www.ncbi.nlm.nih.gov/geo/)  and  Omnibus  NCBI  (GEO) public  SAGEmap  repository  (http://  (http://www.ncbi.nlm.nih.gov/  projects/SAGE/) websites. There are now -500 SAGE libraries in SAGEmap which can be easily accessed by scientists from all over the world (http://www.ncbi.nlm.nih.gov/ SAGE/index.cgi?cmd=printstats). In addition, SAGEmap (http://www.ncbi.nlm.nih.gov/ projects/SAGE/) (211) also serves as a tool for the assignment of tags to transcriptional clusters in the UniGene database. To enhance the utility of the S A G E data, the Cancer  32  Genome Anatomy Project (CGAP) created SAGE Genie (http://cgap.nci.nih.gov/SAGE), a website for the analysis and presentation of SAGE data. This database contains data from both normal and cancer tissues of human and mouse and provides a more userfriendly interface and an easy means to view the expression of any gene (212). However, there are 2 caveats with the use of these databases: 1) depth of sequencing of libraries from different sources; 2) lack of validation. These may affect the accuracy of tag mapping and library comparisons. Thus use of these databases to identify genes of interest must be followed by studies to confirm the analyses. Tag mapping is defined as the unambiguous determination of the identity of the gene represented by a SAGE tag (167). This is achieved by matching the real experimental SAGE tags to a pool of digital SAGE tags which have known annotations. These digital SAGE tags are generated by extracting 14 base pair sequences from the 3' most 7VMII site of transcripts in a database such as UniGene (211). The mapping can be problematic since cDNA sequences are not available for all genes in any organism. Many SAGE tags will not be found in these databases and tags could be misassigned. In a method described by Pleasance et al (167), a conceptual transcriptome was generated using genomic sequence and annotation by extending predicted coding regions to include UTRs on the basis of EST and cDNA alignments, U T R length distributions, and polyadenylation signals thus increasing the accuracy of tag mappings. Another program called SAGEScreen (213) was developed to evaluate and correct the PCR errors and sequencing errors. It was shown to eliminate 78% of error tags and the occurrence of singleton tags. SAGE is usually applied to identify expression changes in different samples, e.g. normal and diseased tissue (214). Since each SAGE library represents only one  33  measurement, it is possible that many pair-wise differences between the libraries are simply the result of random sampling from two populations that do not differ (215). Different statistical methods have been used to evaluate the significance of tag count differences observed in comparisons between two SAGE libraries (216-220). Each has their advantages and drawbacks (214). A comparison amongst different hypothesistesting based methods (i.e. Chi-square, Fisher's exact test, and Audic and Claverie's Bayesian method) was carried out to assess the specificity, power and robustness of these statistical models. A Monte Carlo approach to simulate S A G E experiments suggested that the Chi-square test has the best power and robustness (214). Another comparison that also included SAGE300 and Madden's method showed that for S A G E libraries of equal as well as different size, all the tests give essentially the same results except the Madden test, which can only be used for libraries of similar size (215).  1.4.5  Gene Expression Profiling of the Prostate  Gene profiling techniques have been extensively used in PCa studies (221). Identification of genetic changes and specific target genes in PCa is quite difficult as PCa is a very heterogeneous disease. The use of novel profiling technologies offers the possibility to investigate broad gene expression profiles and thus improve our understanding of genetic alterations in PCa (94). Comparisons have been made between normal and tumor cells, androgen-dependent and independent cells, and cells before and after a specific treatment (222) to identify sensitive and specific markers of PCa for diagnosis and therapy.  34  Prostate specific EST libraries have been generated. A web-based prostate expression database (PEDB) was constructed. ESTs and full-length c D N A sequences derived from more than 40 human prostate cDNA libraries are maintained and represent a wide spectrum of normal and pathological conditions (223, 224). The differential display technology was used in the characterization of genes critical in early prostate carcinogenesis (225), genes differentially  expressed in androgen-dependent  and  androgen-independent PCa (226), as well as genes involved in tumor metastasis (227). SAGE has been used to identify androgen regulated genes in LNCaP cells (228, 229), genes differentially expressed between normal and PCa tissues (230) and genes and pathways involved in the regulation of senescence in prostate epithelial cells (PrECs) (193). Subtractive hybridization was used to identify PCa-related genes (231, 232). Microarray comparisons have been used to compare transcript profiles between normal and tumor cells (233), B P H and PCa tissues (234), hormone-refractory PCa xenografts and xenografts of the parental, hormone-sensitive cells (235), or before and after treatment (236). Combinations of two techniques, such as subtractive hybridization and microarray, were used to more effectively identify candidate genes involved in PIN and PCa development (237, 238). No gene expression profiling studies have been carried out during human prostate development and few have been done in mouse. Mouse EST c D N A libraries containing a total of 62,046 ESTs were constructed at various stages of mouse prostate development: E16.5 UGS, P2, P10, P20, P35 (puberty), 3 month old normal adult, and 14 month old aged adult animals (239). However, with so few tags sequenced, this study has limited scope and depth. The application of gene expression profiling methods in prostate  35  development is still in its infancy and such studies will certainly provide insights into the molecular mechanisms regulating this process.  1.5 Wingless-Type Mouse Mammary Tumor Virus Integration Site Genes (Wnt) and Secreted Frizzled-Related Proteins (Sfrps) 1.5.1  The WNT Signaling Pathway  The WNTs are a family of secreted factors that regulate a variety of developmental processes such as cell differentiation, polarity, migration and proliferation (240). There are at least 19 mouse Wnt genes (241). The WNT ligands bind to seven-transmembrane spanning receptors called Frizzled (FZD) and 10 FZD receptors have been identified in mouse (242). What determines the specificity of WNT-receptor interactions is still not clear (243). There are three main branches of the W N T pathway: the canonical W N T pathway, the planar cell polarity (PCP) pathway and the W N T / C a PCP pathway and the W N T / C a  2+  2+  pathway (244). The  pathway are also collectively called the non-canonical  WNT pathway. Some W N T ligands, such as WNT1, WNT3a and WNT8 are involved in the canonical W N T pathway, while others such as WNT4, WNT5a and WNT 11 are believed to be involved in the non-canonical WNT pathway. However, several WNTs appear to function in both pathways, e.g. WNT5a (245). The non-canonical WNTs have been shown to antagonise the canonical pathway (245). The canonical W N T pathway is by far the most extensively studied. It is characterized by stabilization of the p-catenin protein and its translocation to the nucleus. P-catenin also functions as a component of the cadherin complex, which controls cell-cell adhesion and influences cell migration. In the absence of W N T signals, cytoplasmic P-  36  catenin is phosphorylated by a multiprotein complex for subsequent ubiquitination and degradation. This complex consists of the adenomatous polyposis coli (APC) tumor suppressor protein, axin, and the glycogen synthase kinase, GSK3p. On activation, the WNT ligand binds to the cysteine-rich domain (CRD) of the F Z D receptor (242). Two co-receptors, low density lipoprotein (LDL)-receptor-related protein 5 (LRP5) and LRP6, are required in canonical W N T signaling. It is not clear i f the non-canonical pathway also requires LRP5/LRP6 (245). The FZD receptor may transduce its signal through activation of dishevelled (DVL), leading to dissociation of the P-catenin-phosphorylation complex. How D V L interacts with the FZD receptor or the complex is unclear. The dissociation of the complex releases P-catenin and thus results in accumulation of cytoplasmic P-catenin, which goes into the nucleus and binds to the T-cell factor (TCF)/lymphoid enhancer binding factor (LEF) family of transcription factors and induces target gene expression (246). The non-canonical W N T pathway is independent of P-catenin. The PCP pathway is defined by a set of genes that, when mutated, result in defects in the polarity of cells in a planar tissue. This pathway involves R H O A and Jun kinase (JNK) and controls cell cytoskeletal rearrangements that regulate the polarity of the cells. Activation of the WNT/Ca  2+  pathway involves an increase in intracellular C a  2+  and activation of C a 2+  sensitive signaling components, such as calmodulin-dependent kinase (CaMK) (244). Unlike the canonical WNT signaling pathway, studies of the noncanonical WNT pathway in development have been hampered by the lack of a robust, specific, and facile functional assay. Assays for the activities of the multiple downstream targets of the noncanonical WNT pathways, including JNK, C a  37  2+  flux, P K C , and CaMKII, are also  technically challenging. In addition, none of the above targets is specific to any of the noncanonical WNT pathways (247). WNTs are critical mediators of cell-to-cell signaling during development and regulate such intriguing processes as the generation of cell polarity, and the specification of pattern formation and cell fate (246). Activation of the W N T pathway controls a wide variety of developmental processes including early gastrulation and axis formation (248), cardiogenesis (249), adipogenesis (250), nephrogenesis (251), as well as development of the mammary gland (252), lung (253), central nervous system (254), limb (255) and placenta (256). It also functions in the differentiation or maintenance of stem cells (257). For example, WNT/p-catenin signaling is required in the entire central nervous system for expanding the progenitor cell population by simultaneously promoting cell proliferation and blocking apoptosis and differentiation (258). The expression of WNT pathway members in the developing prostate has not been well characterized. WNT signaling has also been associated with the development of a multitude of different tumor types. Individuals with Axin2 mutations display a predisposition to colon cancer (259). Mutations in B-catenin and Ape have also been found in sporadic colon cancers and a large variety of other tumor types such as gastric, liver and ovarian cancers (260). The involvement of the WNT pathway in PCa has also been described. B-catenin mutations have been found in some PCa patients (261, 262). Treatment of LNCaP cells with WNT3A significantly enhances cell growth in the absence of androgens (263). High levels of Wntl and B-catenin expression are associated with advanced, metastatic, hormone-refractory PCa (264). WNT inhibitory factor 1 (WIF1), a WNT pathway antagonist, is downregulated in various cancers including PCa (265). The interaction  38  between the canonical WNT pathway and the A R has also been addressed. P-catenin has a strong interaction with the A R but not with other steroid hormone receptors (266). Pcatenin significantly enhanced androgen-stimulated transcriptional activity by the A R and overexpression of P-catenin enhanced androgen-mediated transcription (267, 268). In turn, the androgen receptor can promote P-catenin nuclear translocation (269). However, P-catenin/TCF-related transcription in PCa cell lines is inhibited by androgen treatment, possibly due to competition between A R and TCF/LEF for P-catenin (270). A direct interaction between A R and TCF4 was also reported and this may contribute to selective gene regulation in normal and neoplastic prostate growth (271).  1.5.2  The Secreted Frizzled-Related Proteins (SFRPs)  The WNT pathway is regulated by three types of extracellular antagonists. The SFRP family and WIF1 bind to WNT proteins to antagonize their binding with the F Z D receptors. In contrast, the Dickkopf (DKK) family binds to the LRP5/LRP6 component of the WNT-receptor complex and thus may only antagonise the canonical W N T pathway (245). There are 5 members of the SFRP family in mammals (SFRP1 to SFRP5) (245). SFRPs have an N-terminal C R D domain and a C-terminal netrin like (NTR) domain (272). The C R D domain which is both necessary and sufficient for W N T ligand binding (242) shares 30-50% sequence similarity with those of the FZD receptors (245). Thus, it is believed that SFRPs inhibit W N T signaling by interacting with the W N T ligands through the C R D domain, preventing them from binding to the F Z D receptors (273). However, the CRD of SFRP1 appears to interact with itself as well as with FZD (274).  39  Therefore, SFRPs may block WNT signaling by forming a non-functional complex with FZD (245). In addition, there is evidence showing that SFRPs antagonize W N T through the C-terminal N T R domain (275). The N T R domain shares weak sequence similarity with the netrin protein. This domain has also been found in tissue inhibitors of metalloproteases (TIMPs) and some complement proteins (276) and its function is still not clear. The functions of the various SFRPs are still unknown. In some cases, their expression patterns complement those of specific Wnts, which may support the idea that they antagonise W N T function. It has been suggested that SFRPs facilitate boundary definition in the developing organism by limiting the range of W N T activity. However, overlapping expression may simply reflect the regulation of SFRPs by WNTs. For example, Sfrp2 transcripts were shown to be induced by WNT4 in kidney development (277). To further confound the issue, SFRPs do not always function as W N T antagonists. SFRP1 potentiates the activity of wingless (WG), the Drosophila homolog of WNT, at low concentrations (275). A n alternative hypothesis is that SFRPs can sequester and transport WNTs to cellular sites that have a high concentration of receptors, where they can be released as active ligands (245, 272). In addition, SFRP members may play different roles and even antagonise each other. For example, SFRP1 and SFRP2 elicit opposite responses in MCF-7 breast cancer cells (278) and SFRP2 antagonises the effects of SFRP 1 in metanephric kidney development (279). SFRPs are also implicated in development of diseases. Elevated transcription of Sfrp is found in diseases associated with increased apoptosis. For example, Sfrp3 and Sfrp4  40  mRNA are upregulated in failing ventricles (280). SFRPs also regulate cell growth and thus may be involved in tumorigenesis. Given the oncogenic potential of constitutive WNT signaling, SFRPs have been postulated to act as tumor suppressor proteins. Sfrpl mRNA is downregulated in cervical cancer (281), breast cancer (282) and ovary and kidney cancer (283). Sfrp4 mRNA is overexpressed in PCa. Membranous expression of SFRP4 predicts for good prognosis in localized PCa and inhibits PC3 cellular proliferation in vitro (284). However, SFRPs may also play a positive role in cell growth. For example, Sfrp4 mRNA has been shown to be upregulated in the stroma of endometrial and breast cancers (285). The precise mechanism by which SFRPs regulate WNT signaling and cell growth remains poorly understood. The Sfrp2 gene is located on Chromosome 3 in the mouse and encodes a protein of approximately 30 kDa in size (286). R N A encoding SFRP2 is expressed highly in mouse brain, heart, kidney, lung and thymus, but not in the liver or spleen (287). High levels of Sfrpl transcripts are also found in the retina relative to other tissues examined (286). During development Sfrp2 transcripts are expressed in the eye, brain, neural tube, craniofacial mesenchyme, joints, testis, pancreas and below the epithelia of oesophagus, aorta and ureter where smooth muscles develop (288). It is also expressed in the mesenchyme of developing teeth (289). The expression of Sfrp2 transcripts in lung and kidney has been studied more extensively. From E10.5 to E14.5, Sfrpl transcripts are highly expressed in the mesenchyme adjacent to the epithelium and at low levels in the epithelium of the developing lung. SFRP2 protein is present at higher levels in the epithelium between E l 1.5 and E14.5 (290). In the developing kidney Sfrpl was shown to be induced by  41  WNT4  (277).  Its  expression  localizes predominantly in the  condensing and  epithelializing metanephrogenic mesenchyme while no expression was found in the ureter and its branches. After E15.5 Sfrp2 is confined to the glomerular precursors of the outer cortical region but missing in the inner part of the developing kidney where tubuli and mature glomeruli are located, suggesting that SFRP2 is involved in early differentiation of the nephric glomeruli (277, 288). In addition, SFRP2 was shown to antagonise the morphogenic blocking effects of SFRPT thus promoting tubule formation in metanephri (279). The involvement of SFRP2 in disease has not been extensively studied and is controversial. Sfrp2 transcripts are upregulated in retinitis pigmentosa, a degenerative human retina disease in which the photoreceptor neurons undergo apoptosis (291). However, breast cancer cells transfected with Sfrp2 became more resistant to different proapoptotic stimuli (278). Involvement of SFRP2 in cancer has also been demonstrated. Sfrp2 transcripts are produced by malignant glioma cell lines and transfection of Sfrp2 promotes glioma tumor growth in nude mice (292). SFRP2 is also detected in primary cultures of canine mammary gland tumors but not in normal mammary tissues (293, 294) and was found to induce tumorous transformation in normal mammary epithelial cells and inhibit apoptosis (295).  1.6 Hypotheses and Objectives Understanding the molecular mechanisms regulating mouse development requires an intimate knowledge of the molecules involved. Transcriptome analyses of numerous mouse tissues at critical stages in their development will help us to identify genes that  42  are involved in these processes. The objective of the Mouse Atlas project was to define the normal state for many tissues by determining, in a comprehensive and quantitative fashion, the number and identity of genes expressed throughout  development  (http://www.mouseatlas.org/about). SAGE is a powerful gene profiling technique and thus it has been chosen for this work. By the time of completion, approximately 200 SAGE libraries will have been constructed and made publicly accessible. M / E interactions have been shown to be critical to prostate development. It is proposed that epithelial differentiation is regulated by growth factors and secreted extracellular molecules produced by the surrounding mesenchyme. The role of some growth factors, e.g. FGF, and their receptors have been studied. However, other molecules involved in this process are still to be identified. The hypotheses underlying this thesis were: 1) gene expression profiling using SAGE on prostate tissue at critical stages during development will help us to identify molecules and pathways whose expression in prostate development have not been demonstrated before; 2) secreted extracellular molecules likely to be involved in M / E interactions critical for prostate development will be differentially expressed during prostate development and in the male and female U G S ; 3) since cancer is believed to be a process that recapitulates development, these molecules may be re-expressed and/or play critical roles in PCa development and progression. The objectives of this study were: 1) to validate the U G S and prostate SAGE libraries; 2) to identify genes differentially expressed during prostate development, as well as in male and female UGS, using bioinformatic analyses; 3) to identify secreted extracellular proteins not previously described in prostate development since these  43  proteins may be involved in the M / E interactions which are critical for prostate development; 4) to validate expression of these candidate genes identified in prostate development; and 5) to examine expression of these candidate genes in PCa.  44  CHAPTER 2. MATERIALS AND METHODS  2.1 Mouse Maintenance and Mating Strategy C57B1/6J mice (originally obtained from the Jackson Laboratories) were used for all tissue dissections, SAGE library construction and validation studies. A l l mice were bred and maintained at the British Columbia Cancer Research Centre animal facility. They were housed in microisolator units and provided with sterilized food, water, and bedding. Matings were set up in the afternoon by putting 3 female mice into a cage with one male mouse. Vaginal plugs detected the following morning were designated embryonic day 0.5. For the PCa studies, T R A M P mice, originally obtained from Dr. Norman Greenberg, were used and maintained as indicated above. The breeding strategy was to mate female T R A M P mice to WT C57B1/6J males. The offspring were genotyped for SV40 Tag expression using previously described techniques (296). The PCR-based screening assay was performed  using a forward primer from the rat probasin promoter  (5'-  C C G G T C G A C C G G A A G C T T C C A C A A G T G A A T T T A - 3 ' ) and a reverse primer from SV40 genomic sequence ( 5 ' - A G G C A T T C C A C C A C T G C T C C C A T T C A T C - 3 ' ) to amplify a 1250 base pair fragment.  2.2 Tissue Collection The following tissues were collected for construction of the 6 S A G E libraries: E16.5 male UGS, E16.5 female UGS, E16.5 male U G E , E16.5 male U G M , P0 prostate and the DLP of 12 week old adult mice. A l l adult mice were sacrificed by CO2 asphyxiation,  45  while the El6.5 embryos and PO mice were decapitated. El6.5 embryos were harvested from timed-pregnant female mice and whole mouse embryos were collected in ice-cold lx Phosphate Buffered Saline (PBS, 32g NaCl, 0.8g KC1, 4.6g N a H P 0 , 0.8g K H P 0 2  4  2  4  dissolve in 4 litres distilled water, pH7.2) for further dissection. Embryos were staged according to Theiler criteria (http://genex.hgu.mrc.ac.uk/Databases/Anatomy/new/). Dissection of both the male and female UGS was carried out following a previously described protocol (2). Specifically, the entire urogenital system consisting of the bladder, UGS, urethra and two Wolffian ducts (male embryos) or Mullarin ducts (female embryos) was removed from the El6.5 embryos using a dissection microscope (Leica) and put in ice- cold C a - and Mg -free Hank's Balanced Salt Solution (HBSS, StemCell 2+  2+  Technologies Inc.). The urogenital sinuses were subsequently isolated by removing the bladder, the urethra and the Wolffian or Mullarin ducts and transferred to RNALater (Qiagen) for R N A isolation (see below) (Figure 2.1 A , B). Separation of the mesenchyme from the epithelium of the intact E16.5 male UGS was performed with modifications of the above mentioned protocol (2). The E16.5 male UGS was digested with Trypsin (Sigma) in HBSS (1% w/v) for 1.5 hours at 4°C in the fridge. The digestion was stopped by pipetting out the trypsin and adding Fetal Bovine Serum (FBS, StemCell Technologies Inc.) in HBSS (20% v/v). The tissues were then transferred to fresh HBSS on ice to separate the mesenchyme and epithelium. To prevent tissue stickiness induced by trypsin-mediated D N A release, a few crystals of DNase I (Sigma) were added before separation. The U G M and U G E were then separated manually by carefully peeling the mesenchyme away using two 1ml syringes with 30G needles with the aid of the dissection microscope (Figure 2.1C, D). The intact PO prostate and the DLP  46  from 12 week old adult mice were dissected in a similar manner by first removing the entire urinary tract and then isolating the PO prostate or the 12 week D L P (Figure 2.1 E H). To study the expression patterns of candidate genes during PCa development, the DLP of 2 week, 4 week, 12 week and 26 week old T R A M P mice and their WT littermates were dissected for performing RT-qPCR. At each stage the D L P of 2 mice were pooled for R N A isolation. 4 week and 8 week old T R A M P mice and their W T littermates were also used for immunofluorescence staining (n=2).  2.3 RNA Isolation Dissected tissues for SAGE library construction were removed from the RNALater (Qiagen) and transferred to Trizol (Invitrogen) for homogenization. Tissues were homogenized using a PowerGen 125 homogenizer (Fisher Scientific). Following tissue disruption total R N A was isolated according to the Trizol R N A isolation protocol provided by Invitrogen. The protocol for R N A isolation from the DLP of the adult mice was changed due to a problem with poor R N A quality. Total R N A was isolated using the RNeasy Lipid Tissue Mini Kit (Qiagen) to remove possible contaminants and thus increase R N A quality. The R N A amounts isolated and used for S A G E library construction are listed in Table 2.1. For the validation studies different homogenization methods were performed. The E16.5 U G S and PO prostate were homogenized by passing the tissues through 1ml syringes with a series of different sized needles (21.5G, 26.5G and 30.5G). To disrupt the 12 week old adult DLP, dissected tissues were first transferred from RNALater to liquid  47  Figure 2.1 Tissues dissected for SAGE library construction. The entire urogenital system of E l 6.5 male (A) and female (B) embryos was removed for isolation of the UGS. The bladder (B), urethra (U), testes (T), and ovaries (O) are indicated, as is the UGS area that was isolated for library construction. The El6.5 U G M (C) and U G E (D) were isolated as described in the methods section. The PO prostate used for library construction is shown in dorsal (E) and ventral views (F) with the bladder (B), urethra (U), testes (T), dorsal-lateral prostate (DLP), ventral prostate (VP) and seminal vesicle (SV) indicated. The 12 week old adult prostate is shown in dorsal (G) and ventral (H) views with the lateral prostate (L), ventral prostate (V), bladder (B), urethra (U), seminal vesicle (SV), dorsal prostate (DP) and coagulating gland (CG) indicated.  48  49  Table 2.1: Summary of samples collected for SAGE library construction.  Tissue  Number of Embryos/Mice Used  Total RNA Isolated  RNA used for Library Construction  SM032 El6.5 Male UGS  15  25ug  5lig  SM035 El6.5 Male UGM  44  25.2ng  SM065 El6.5 Male UGE  19  10.7ug  10.7ug  SM054 El6.5 Female UGS  11  22.4ng  10ug  SM025 El6.5 PO Prostate  7  18ug  18ug  SM026 E16.512W DLP  3  15ug  15^g  50  nitrogen and then ground to powder using a mortar and pestle (VWR). The powder was transferred to Qiazol reagent (Qiagen) for R N A isolation. R N A was stored at -80°C prior to library construction or validation studies.  2.4 SAGE Library Construction A l l 6 S A G E libraries were constructed by technicians at the British Columbia Genome Sciences Centre (BCGSC) using the I-SAGE kit (Invitrogen) adapted for LongSAGE according to the LongSAGE protocol (version 1.0a) from the SAGEnet website (http://www.sagenet.org/resources). Briefly, isolated R N A was treated with DNase I (lU/u.1, Invitrogen) and then assayed for quality and quantified using an Agilent 2100 Bioanalyzer (Agilent Technologies) and the R N A 6000 Nano LabChip kit (Caliper Technologies). Libraries were constructed using 5-18jag of DNase I treated total RNA. The mRNAs were bound to oligo (dT) magnetic beads (Invitrogen) and c D N A was synthesized using Superscript™ II reverse transcriptase and E. coli D N A polymerase (Invitrogen). The bead-bound cDNAs were then cut with the Nlalll  (Invitrogen)  anchoring enzyme which recognizes the sequence 5'-CATG-3'. The magnetic beads were then collected in a magnetic stand. Only the most 3' end of the fragments attached to the beads were retained while the other fragments were washed away. The beads were then recovered and divided equally into 2 new tubes. A pair of 40 base pair adapters was ligated to the 2 cDNA pools using T4 D N A ligase. Two different pairs of adapters were used during library construction due to an adapter contamination problem. The adapter sequences for the El6.5 male UGS, U G M , PO and adult prostate libraries were: Adapter A :  51  5' -TTTGG A T T T G C T G G T G C A G T A C A A C T A G G C T T A A T A T C C G A C A T G - 3 ' 3'-amino(C7) C C T A A A C G A C C A C G T C A T G T T G A T C C G A A T T A T A G G C T P 0 - 5 ' 4  Adapter B: 5' - T T T C T G C T C G A A T T C A A G C T T C T A A C G A T G T A C G T C C G A C A T G - 3 ' 3'-amino(C7) G A C G A G C T T A A G T T C G A A G A T T G C T A C A T G C A G G C T P 0 - 5 ' 4  The adapter sequences for the U G E and the El6.5 female UGS libraries were: Adapter A : 5' -TTTGG A T T T G C T G G T G C A A C G T C A C T A G G C T TA A T A T C C G A C A T G - 3 ' 3'-amino(C7) C C T A A A C G A C C A C G T T G C A G T G A T C C G A A T T A T A G G C T P 0 - 5 ' 4  Adapter B: 5' -TTTCTGCTCG A A T T C A A G A C A G A A A C G A T G T A C G T C C G A C A T G - 3 ' 3'-amino(C7) G A C G A G C T T A A G T T C T G T C T T T G C T A C A T G C A G G C T - P 0 - 5 ' 4  The adapters contained cohesive 4 base pair overhangs complementary to the M a l l l digested cDNA, a Type IIS restriction enzyme (tagging enzyme) recognition site at the 3' end, and priming sites for P C R amplification. The 3' end of the adapters was modified with an amino group to prevent self-ligation. Cleavage of the ligation products with the tagging enzyme, Mmel, resulted in release of the adapter with a short tag of the cDNA from the beads. Mmel recognizes the site 5'-TCCRAC-3' and cuts 21 base pairs downstream from the adapter, releasing a -60 base pair fragment with a 2 base pair overhang. The fragment consisted o f - 4 0 base pairs of adapter sequence and -21 base pairs of unique sequence (tag) from a single transcript. After Mmel digestion, the tags generated in pool A were ligated with tags generated in pool B to form -130 base pair ditags using T4 D N A ligase (Invitrogen).  52  The ditags were amplified by RT-PCR using the primer sequences from the adapters (see above) to produce sufficient ditags for subsequent generation of concatemers. The scale-up P C R was performed for 25 cycles using 1/80 dilution of template and two 96well plates with 50ul reactions per well. The LongSAGE adapter molecules and scale-up PCR oligonucleotide primers were purchased from Invitrogen. The 130 base pair ditags produced after scale-up P C R were then purified from a 12% polyacrylamide gel in order to separate the 130 base pair ditags from the 100 base pair adapter contaminants. The ditags were excised, eluted from the gel and purified using spin columns (Invitrogen). Digesting the 130 base pair ditags with Nlalll produced 34 base pair ditags without adapter sequence which were gel-purified and ligated to form concatemers. Concatemer bands of various sizes were separated on a polyacrylamide gel and the high molecular weight bands were excised and purified. Concatemers were cloned into the pZErO-1 vector (Invitrogen) and transformations achieved with One Shot TOP 10 Electrocomp E. coli (Invitrogen). Following screening of transformants by colony PCR, concatemer size fractionation was carried out to select clones for sequencing (i.e. 1000-1500 base pairs with an average true insert size of 800 base pairs, or 45 tags per clone). Colony picking was performed using a Q-Pix robot (Genetix) and inoculations made into 2X Y T media with 50|a,g/ml Zeocin and 7.5% glycerol. Following overnight culture, these glycerol stocks were used to inoculate larger volume cultures for plasmid preparation using a standard alkaline lysis procedure adapted for high-throughput processing with microtitre plates. D N A sequencing was performed with BigDye version 3 dye terminator cycle sequencing reactions run on Tetrad thermal cyclers (MJ Research). Sequencing reaction products were purified by isopropanol  53  precipitation and then run on model 3700 and 3730x1 capillary D N A sequencers (Applied Biosystems). Resulting sequence data was collected automatically using custom D N A sequencing Laboratory Information Management System (LIMS) software and processed by trimming of reads for sequence quality, and removal of non-recombinant clones, linker-derived tags and duplicate ditags. A schematic diagram of S A G E library construction is shown in Figure 2.2.  2.5 Bioinformatic Analysis of the SAGE Libraries A l l data analyses, including tag-to-gene mapping and comparisons amongst the various libraries, were done using DiscoverySpace (http://www.bcgsc.ca/bioinfo/software /discoveryspace/), a S A G E data analysis program developed by the bioinformatic group at the BCGSC,  with the CMOST  plug-in  (http://www.bcgsc.ca/bioinfo/software/  discoveryspace/CMOST_ plugin_docs). Libraries were analyzed using a 99.9% cutoff for singletons and 95% cutoff for overall sequence quality to ensure only high quality tags were  included for  bioinformatic  analysis.  Duplicate ditags  were  masked.  In  DiscoverySpace the comparison of a tag in 2 libraries is tested for significance using the Audic and Claverie approximation to the Poisson difference, which is analogous to the ttest for the difference of two means for normally distributed variables (218). Pair-wise comparisons between libraries used a cutoff of 99% (confidence intervals) (218). A l l 21 base pair SAGE tags were annotated using CMOST which draws tag data from 8 different sources: Reference Sequences (RefSeq), Mammalian Gene Collection (MGC); Ensembl transcripts; Ensembl EST transcripts; Transcription units (based on Ensembl  54  Figure 2.2 Schematic diagram of SAGE library construction. Isolated R N A was bound to oligo (dT) magnetic beads on which double-stranded c D N A was synthesized. The resulting cDNA was then cleaved by the anchoring enzyme, Nlalll, which recognizes and cleaves D N A immediately 3' of the sequence C A T G . These cDNAs were divided into two halves, then ligated to linkers (or adapters) A and B , respectively. The linkers contain an M a l l l 4-nucleotide cohesive overhang, a type IIS recognition sequence and a priming site for PCR amplification. Type IIS restriction enzymes (Mmel) cleave the D N A at a defined distance 21 nucleotides 3' of its recognition sequence, releasing the linkeradapted SAGE tag from each cDNA. The cDNAs from each pool were mixed together and then ligated using T4 D N A ligase. These linker-adapted ditags were amplified by PCR, digested with Nlalll to release the primer-adapters and the SAGE ditags were purified. The isolated ditags were then ligated to form concatemers and cloned into a high-copy plasmid vector for sequencing.  55  Total R N A  1 cDNA synthesis  -AAAAA ^ TTTTT-U  |  O Malll  O  GTAC-  O  CATG-  CATG-  Mme\  Mme\ CATG-  B  O  CATG  -o CATG-  •O a  01V0 P C R amplification a  Malll CATG-  CATG-  -CATG-OIVO-  =01V0 -CATG-91V3"  -CATG-91V0-  -91V0  pZErO-1  Sequencing CATGTTCATTATAATCTCAAA CATGATCAACACCGCAACCTT CATGTGGCTCGGTCACTTGGG CATGAAGAGGCAAGACGAAAA  56  141 84 20 119  genes); Golden path; Mitochondria (GenBank) and Non-protein coding genes (GenBank). Tags were also blasted to the mouse genome using Ensembl B L A S T (Basic Local Alignment and Search Tool, http://www.ensembl.org/Multi/blastview?species=Mus_ musculus). The molecular function, biological process  and cellular component  classifications for the various genes were obtained using the Gene Ontology (GO) Browser of DiscoverySpace (http://www.bcgsc.ca/bioinfo/software/discoveryspace/GO primer). Digital Northern analysis was carried out by determining tag counts for the gene of interest in 262 human and 149 mouse SAGE libraries present in the C G A P S A G E Genie website  using the  online tool at:  http://cgap.nci.nih.gov/SAGE/mSAGE_Digital_  Northern.  2.6 Reverse Transcription (RT) To perform RT-qPCR, total R N A , isolated as indicated above, was treated with DNase I (lU/ul, Invitrogen) prior to cDNA synthesis. The c D N A reaction mix ( l u l random primers (Invitrogen, 150ng/ul), l u g total R N A , ljil  lOmM dNTP mix  (Invitrogen), dF^O to 12ui) was heated to 65°C for 5 minutes and incubated on ice for at least 1 minute, after which the following reagents were added: 4ul 5X First-Strand Buffer (Invitrogen), l u l 0.1M DTT (Invitrogen), l u l RNaseOUT Recombinant RNase Inhibitor (40units/ul, Invitrogen), l u l of Superscript III Reverse Transcriptase (200units/ul, Invitrogen). The reaction was incubated at 25°C for 5 minutes, 50°C for 60 minutes and then inactivated by heating at 70°C for 15 minutes. After the reaction, DNase-, RNasefree dF^O was added to make a lOOul final volume per ug total R N A . c D N A quality was  57  confirmed by the lack of detectable genomic D N A using primers spanning intronic regions of fi-actin (forward primer, 5 ' - A T G G A T G A C G A T A T C G C T G - 3 ' ,  reverse  primer, 5' - A T G A G G T A G T C T G T C A G G T - 3 ' ) .  2.7 Quantitative Real Time PCR (RT-qPCR) All  RT-qPCR  primers  (Table  2.2)  were  designed  (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) primers  was  confirmed  by  blasting  their  using  Primer  3  and the specificity of the  sequences  using  NCBI  BLAST  (http://www.ncbi.nlm.nih.gov/BLAST/). RT-qPCR was performed using the iCycler iQ Real Time PCR (BioRad). Primer efficiency was tested by performing RT-qPCR on serial dilutions (1/2, 1/4, 1/20, 1/40) of 1 (0.1 (0.01 jag) reverse-transcribed c D N A with a nocDNA reaction as a negative control. A standard curve was generated and the slope of the standard curve was used to determine the efficiency (Eff) of the P C R reaction using the following equation: Eff=10 " (  1/slope)  -l. Primers with an efficiency of 1±0.2 (0.8-1.2) were  chosen for RT-qPCR validation of the SAGE tags. The generation of specific PCR products was confirmed by melting curve analysis. The reaction conditions for RT-qPCR were as follows: 10 minutes at 95 °C followed by 45 cycles (95 °C for 15 seconds, 60 °C for 1 minute). Ubiquitin C (Ubc) was used as an internal control for the 3 prostate developmental libraries (El6.5 male UGS, P0 prostate and the adult prostate) because its expression was similar in all 3 libraries. Glyceraldehyde-3-phosphate dehydrogenase (Gapdh) was used as an internal control when comparing female U G S to male U G S and P-2-microglobulin (B2m) was chosen as the internal control for comparison of the U G M and U G E for the same reason. ft-actin was chosen as the internal control for comparing  58  Table 2.2: Primer sequences used to amplify the WNT family members and housekeeping genes selected for RT-qPCR. Gene  Forward Primer  Reverse Primer  Product Size (base pair)  Gapdh  AACTTTGGCATTGTGGAAGG  ATGCAGGGATGATGTTCTGG  130  Ubc  AGGTCAAACAGGAAGACAGACGTA  TCACACCCAAGAACAAGCACA  80  (3-actin  GCTCTTTTCCAGCCTTCCTT  CGGATGTCAACGTCACACTT  109  B2m  GAGCCCAAGACCGTCTACTG  TGGATTTGTAATTAAGCAGGTTCA  91  Wntl 1  CAGGATCCCAAGCCAATAAA  CCCCATGGCATTTACACTTC  99  Wnt3a  CCATCTTTGGCCCTGTTCT  TCACTGCGAAAGCTACTCCA  87  Wnt8a  TCGTGGACAGTTTGGAGAAA  AGATGCCATGACACTTGCAG  120  Wnt8b  TTCCAAGCAGTTTGTGGATG  TTACACGTGCGTTTCATGGT  117  Wnt4  CCGGGCACTCATGAATCTT  ACGTCTTTACCTCGCAGGAG  113  Fzdl  CTCTTCACGGTGCTCACGTA  GTAACAG CCGGACAGGAAAA  84  Fzd7  CCAGGTGGATGGTGACCTAC  GACGTCCCGATGAAGAGGTA  114  Lrp5  CTGTGGCTGTGCTTCACACT  T G G C T G A A C A G C A A G A A G GT  81  Axin  CCGGAGCTATTCCGAGAAC  TTGGCATTCTTCCCAGATTC  120  Naked 1  CTGGCGGGGATAGAGAACTA  AGTGGGGTTGGAGATTCGAG  96  P-catenin  TGGCCATCTTTAAGTCTGGTG  TCCTGATGGAGCAGGAGATT  115  Lefl  TGAGTGCACGCTAAAGGAGA  AGTATTTGGCCTGCTCTTCC  89  RhoA  AAGGACCAGTTCCCAGAGGT  TGTCCAGCTGTGTCCCATAA  110  Sfrp2  AGGACAACGACCTCTGCATC  TTCTTGGTTTTGCAGGCTTC  97  59  expression levels of genes of interest between T R A M P and their WT littermates as it showed the best consistency (i.e. most comparable levels) amongst the three housekeeping genes tested (Gapdh, Ubc and P-actin). The formula used for calculation of gene expression fold changes was: Ratio = ( l + E a r g e t ) t  A(contro  '  _treated  V  (l+E ference) re  A(contro1  "  treated)  2.8 Tissue Preparation for Immunofluorescence and in situ Hybridization A l l tissues were dissected as indicated above and fixed immediately in 4% paraformaldehyde (PFA, Sigma) in PBS (w/v) at 4°C overnight followed by infiltration with 20% sucrose (Gibco BRL) in PBS (w/v) at 4°C overnight. A l l W T and T R A M P adult prostates were embedded in OCT (Optimal Cutting Temperature compound, Tissue-Tek) the following morning and stored at -80°C. 5 urn-thick sections were cut using a microtome cryostat (MICROM International GmbH) and stored at -80°C until use. The El6.5 embryos and PO prostates were embedded in OCT and cut into 5urn sections by Wax-it Histology at the University of British Columbia.  2.9 In situ Hybridization The Sfrpl cDNA sequence was obtained from GenBank (http://www.ncbi.nlm.nih. gov/entrez/viewer.fcgi?db=nucleotide&val= 6677894) and the primers to amplify the probe were designed using Primer 3. The primer sequences were: forward primer, 5'A C G A G A C C A T G A A G G A G G T G - 3 ' ; reverse primer, 5' - G G A G A T G C G C T T G A A C T C TC-3'. The specificity of the primers was confirmed using NCBI B L A S T . A 665 base pair fragment corresponding to the mouse Sfrpl transcript sequence between +196nt to  60  +861nt was amplified from E16.5 male UGS cDNA (See above, Reverse Transcription) with the above primers. PCR was performed using Platinum® Taq D N A Polymerase High Fidelity (5U/ul, Invitrogen), at 95°C 30 seconds, 60°C 30 seconds, 72°C 40 seconds for 35 cycles. The PCR product was purified from a 1% agarose gel using the MinElute Gel Extraction Kit (Qiagen), cloned into the TOPO pCRII® dual promoter vector (Invitrogen), and transformed into TOPO 10 competent cells (Invitrogen) according to the TOPO TA® cloning kit (Invitrogen). Transformed cells were spread on a LuriaBertani (LB) agar plate [Tryptone (Becton, Dickinson and Company, BD) lOg, Yeast extract (BD) 2.5g, NaCl ( E M science) 5g, Agar (BD) 7.5g, add distilled water to 500ml] with Ampicillin (Sigma, 100ug/ml) and incubated at 37°C overnight. Colonies were picked the next day and inoculated into 3ml L B broth (Tryptone lOg, Yeast extract 2.5g, NaCl 5g, add distilled water to 500ml) with Ampicillin (lOOug/ml) and grown with shaking at 37°C overnight. The plasmid was purified the following morning using QIAprep Spin Miniprep Kit (Qiagen). The orientation of the insert was determined using PCR with 4 different primer combinations: both M l 3 primers, both gene specific primers, M l 3 left primer with right gene specific primer, and the right primer of M l 3 with left, gene specific primer. The product was then digested with either EcoKV or Hindlll (Invitrogen) to generate a linearized vector containing the Sfrp2 sequence on one end with either an SP6 or a T7 promoter upstream (a schematic diagram is shown in Figure 2.3). In our case, Hindlll was used and the antisense probes were transcribed and labelled with digoxigenin (DIG) using T7 R N A polymerase (Roche) at 37°C for 2 hours according to the manufacturer recommendations. The probes were purified using  61  Figure 2.3 Schematic diagram of Sfrp2 probe synthesis for in situ hybridization by in vitro transcription. Sfrp2 cDNA amplified by RT-PCR was cloned into pCRII TOPO Vector (Invitrogen) and the direction of the insert was determined using P C R with different primer combinations. The plasmid was then linearized by a restriction enzyme (Hindlll) to leave the insert downstream of the T7 R N A polymerase site. The linearized sequence was used as the template for in vitro transcription according to the protocol (Roche).  Cut with restriction enzyme {Hind\\\)  SP6  Antisense probe  62  ProbeQuant™ G-50 Micro Columns (Amersham Pharmacia Biotech Inc.). The integrity of the probes was confirmed on a 1% agarose gel. Probes were stored at -20°C until use. A l l remaining steps were done at room temperature unless otherwise specified. To perform in situ hybridization, cryosections were air-dried for an hour and then fixed in 4% PFA for 10 minutes. Sections were then washed in PBS (pH 7.4) 3x for 3 minutes each on a rocking shaker and then treated with proteinase K (Invitrogen, 0.3ug/ml in 50mMTris pH 7.5, 5mM EDTA) for 10 minutes. Sections were re-fixed in 4% P F A for 5 minutes and washed in PBS 3x for 3 minutes as mentioned above. The slides were then pre-hybridized with hybridization buffer [50% Formamide (Invitrogen), 5><SSC pH 4.5 (20xSSC: 175.3g NaCl ( E M science), 88.2g Sodium Citrate (Fisher Scientific), adjust p H and add distilled H 2 O to 1L), 50ug/ml yeast tRNA (Invitrogen), 1% Sodium Dodecyl Sulfate (SDS, Sigma), 50(ig/ml heparin (Sigma)] in a humidified chamber for 2 hours at 55°C. Probes were diluted 1:20 in the hybridization buffer, denatured by heating for 5 minutes at 80°C and cooled to room temperature for 5 minutes. 200ul of probe was added to each slide which was then covered with a HybriWell™ (Grace Bio-Labs) coverslip and incubated in a moist chamber at 65°C overnight. On day 2 the HybriWell™ coverslips were removed and the slides transferred to prewarmed (65°C) 5*SSC (pH 7) and incubated on a rocker for 30 minutes at room temperature, after which the slides were incubated with prewarmed (70°C) 0.2xSSC (pH 7) for 3 hours at 70°C, 0.2xSSC for 5 minutes and l x M A B (pH7.5, maleic acid 11.6g, NaCl 8.7g in 1L dH20) for 5 minutes. The samples were blocked in Powerblock™ (InnoGenex) for 1 hour and then incubated with anti-DIG-Alkaline Phosphatase (AP) antibody (Roche) diluted at 1:1000 in the Antibody Diluent (BioGenex) at 4°C overnight. On day 3, the slides were washed 3x in  63  l x M A B with Tween-20 (0.1% v/v) for 15 minutes followed by a wash in dFJ^O-Tween20 (0.1% v/v) for 20 minutes on a rocking shaker. B M purple (Roche) was added and the slides covered with foil to allow adequate color development, (i.e. 4 hours to 4 days). Relative expression levels were estimated based on subjective comparisons of the staining intensity.  2.10 All  Immunofluorescence steps were performed at room temperature unless otherwise specified.  Cryosections were air-dried for one hour, fixed in 4% P F A for 10 minutes and then washed in PBS (pH 7.4) 3x for 3 minutes each. Samples were blocked using Blocking Reagent solution (ImmunoVision Technologies) for 1 hour. 20-50ul (depending on the size of the sample) anti-SFRP2 rabbit IgG or anti-P-catenin rabbit IgG (200u_/ml, Santa Cruz Technologies) (1:200 in Blocking Reagent) were added and slides were incubated in a moist chamber at 4°C overnight. On day 2, slides were washed with PBS-Tween-20 (0.1% v/v) 5x 10 minutes each and incubated with 20-50ul Alexa Fluor® 488 donkey anti-rabbit secondary antibody (2mg/ml, Molecular Probes) (1:200 in Blocking Reagent) for 1 hour followed by washing in PBS-Tween-20 (0.1% v/v) 5x 10 minutes each. Slides were mounted with GelTol Aqueous Mounting Medium (Immunon) and pictures were taken using an Axioplan 2 imaging fluorescent microscope (Carl Zeiss). Relative expression levels were estimated based on subjective comparisons of the staining intensity.  64  CHAPTER 3. RESULTS  3.1 Analysis of the Prostate SAGE Libraries 3.1.1  Rationale for Library Selection  The importance of androgen and M / E interactions in prostate development has been well characterized through tissue recombination experiments. Numerous growth factors and transcription factors involved in prostate development have been identified. However, the molecular mechanisms underlying this process are still unclear. In order to further our understanding of the molecular mechanisms regulating prostate development, we constructed a total of six LongSAGE libraries from tissues collected at three critical time points in prostate development. A library was created from the male UGS at El6.5, the time immediately prior to initiation of prostate bud formation (7). A female UGS library was made at this same stage to identify genes expressed at higher levels in the male UGS and thus likely to be involved in prostate development. To facilitate identification of genes and pathways involved in the critical M / E interactions, we further dissected the intact El6.5 male UGS into its separate components - the U G M and the U G E from which 2 additional SAGE libraries were constructed. A postnatal day 0 (PO) prostate library was made since extensive epithelial branching occurs at this time (7), while a 12 week old adult DLP library was created to represent a time of relative growth "quiescence". The dissected tissues used for library construction are shown in Figure 2.1. We selected the DLP of the adult prostate for library construction as it is believed to be homologous to the human prostate peripheral zone, the site where PCa commonly commences (297).  65  The basic properties of the 6 libraries analyzed in this study are summarized in Table 3.1. "Total counts" refers to the total number of tags obtained from processing of the raw sequencing data. It includes duplicate ditags and tags with poor base-calls. The useful tags, those not containing poor base-calls (N), comprised approximately 99% of the total tag counts in the libraries. Duplicate ditags accounted for 5 % - l l % of the total tags in each library except in the adult D L P library, which contained 21.3% duplicate ditags. This difference likely arose due to the abundant transcripts of prostatic secretory proteins in the adult library. To reduce possible P C R bias and avoid the effect of sequencing errors during data analysis, duplicate ditags were masked and a 99.9% quality cutoff was used for singletons and 95% quality cutoff for all other tags during data analysis. The quality cutoff was converted from the Phred score and represented the estimated error probability of sequencing. Total tag counts after quality cutoffs represent 70% to 85% of the useful tags. This is also shown in Table 3.1. Tag distributions in the 6 libraries were compared and found to be very similar (Figure 3.1 and Table 3.2). The most abundant tags were singletons, accounting for about 60%) of the total tag types. Tags above 5 counts made up about 15% of the total tag types while tags with counts between 2 and 5 made up the remaining 25%. Similarly, G O analyses based on tag types revealed no obvious differences in GO components, i.e. molecular function, biological process, and cellular component, amongst the libraries (data not shown). One goal of making SAGE libraries at different stages of prostate development was to identify differentially expressed genes at the various stages. Thus pair-wise comparisons of tag abundance were carried out amongst the constructed libraries to  66  Table 3.1: SAGE library summary.  Library  Total Counts  Useful Tags  Tags from Duplicate Ditags  Tag Counts *  Tag Types *  E16.5 Male UGM  126,274  125,023 (99%)  12,600(10%)  89,056  27,663  El6.5 Male UGE  144,668  143,780 (99.4%)  16,068(11.1%)  107,087  25,066  El6.5 Male UGS  118,194  117,363 (99.3%)  9,616(8.1%)  86,751  23,980  El6.5 Female UGS  114,754  114,361 (99.7%)  7,686 (6.7%)  88,956  25,285  PO Prostate  99,274  97,788 (98.5%)  5,157 (5.2%)  65,455  20,306  12W DLP  120,944  120,115 (99.3%)  25,402 (21%)  70,352  19,225  *99.9% singleton quality cutoff; 95% overall quality cutoff.  67  Figure 3.1 Tag distributions in the 6 libraries. The y-axis shows the number of tag types.  '%  x  \  y  Table 3.2: Tag distributions in the 6 libraries. counts>5  counts 2-5  Singletons  El6.5 male U G M  3,178 (11.5%)*  6,405 (23.2%)  18,080 (65.4%)  El6.5 male U G E  4,215 (16.8%)  6,551 (26.1%)  14,300 (57%)  El6.5 male UGS  3,425 (14.3%)  6,395 (26.7%)  14,160 (59%)  El6.5 female UGS  3,475 (13.7%)  6,127 (24.2%)  15,683 (62%)  PO prostate  2,564(12.6%)  5,789 (28.5%)  11,953 (58.9%)  Adult prostate DLP  2,404(12.5%)  5,068 (26.4%)  11,753 (61.1%)  * Percentage in brackets indicates proportion in relation to the total tag types.  68  evaluate differences between various pairs of libraries (Figure 3.2 and Table 3.3). Figure 3.2 shows the tag abundance in one library versus another, while Table 3.3 summarizes the data presented in Figure 3.2. We found that most of the tags (97-99% of the total tag types) in each pair-wise comparison were expressed at a similar level in both libraries. These comparisons also indicate that the El6.5 male U G S library and the PO prostate library are more similar to each other (99.4%) than to the adult prostate library (~97%), consistent with the fact that they represent developmentally similar stages (Figure 3.2A, B and C, Table 3.3). Comparison between the male and the female U G S library also revealed a high degree of similarity (99.24%) (Figure 3.2D, Table 3.3), suggesting that the initiation of prostate differentiation in the male UGS is regulated either by only a small group of transcripts or by low abundance genes within the singleton pool. Comparison of the U G M and U G E libraries showed that 98% of the tags were similarly expressed (Figure 3.2E, Table 3.3). To investigate the biological significance of genes represented by tags showing differential expression in the pair-wise comparisons, further analyses were carried out. 54 tags were expressed at higher levels in the El6.5 male UGS library than in PO prostate library (obtained from Figure 3.2A) as well as than in the adult D L P library (obtained from Figure 3.2C). Annotation of these tags showed that many were ribosomal proteins, perhaps suggesting a higher requirement for protein biosynthesis at this stage. Similarly, 54 tags were expressed at higher levels in the PO prostate library than either the El6.5 male UGS library (obtained from Figure 3.2A) or the adult DLP library (obtained from Figure 3.2B), including genes involved in cytoskeleton formation such as collagens and a-actin, perhaps correlating with branching morphogenesis and tissue remodeling  69  Figure 3.2 Pair-wise comparisons of the S A G E libraries. Comparisons between (A) El6.5 male UGS and PO prostate, (B) adult DLP and PO prostate, (C) El6.5 male UGS and adult DLP, (D) E16.5 female UGS and male UGS, (E) E16.5 U G E and U G M , are shown. Each dot in the diagram represents a unique tag. The blue dots in the funnel area represent tags that were expressed at similar levels in the two libraries. The green dots outside the funnel represent tags that were differentially expressed in the two libraries. Lines of the funnel represent different cutoffs (confidence intervals). A 99% cutoff was used in data analysis. A  B Adult D L P v s PO  E 1 6 . 5 M a l e U G S v s PO  00  1«0I  10  w  Adult D L P  E16.5 Male U G S  D E16.5 Female U G S vs E16.5 Male U G S  E 1 6 . 5 M a l e U G S v s Adult D L P  I  4 SO  1«01  40  l«07  SO  1*01  E16.5 Female U G S  E16.5 Male U G S  70  Table 3.3: Pair-wise comparisons of the SAGE libraries. Numbers and percentages of the tag types differentially expressed (based on 99% confidence intervals) are listed.  E16.5 Male UGS vs. PO E16.5 male UGS vs. Adult PO vs. Adult El6.5 Male vs. Female UGS E16.5 Male UGM vs. UGE  Similarly Expressed Tags  t inE16.5 Male UGS  34,933 (99.4%) 34,521 (97%) 31,642 (97.6%) 38,639 (99.24%) 42,000 (98%)  83 (0.24%) 516 (1.45%)  t in P0  f in Adult  t inE16.5 Female UGS  t inE16.5 Male UGM  T in E16.5 Male UGE  408 (0.95%)  447 (1.04%)  118 (0.34%)  414 (1.28%) 147 t (0.38%)  538 (1.51%) 371 (1.14%) 148 (0.38%)  71  required at this time point. In the adult D L P library there were 325 tags more highly expressed compared to the other 2 libraries, many of which are prostate secretory enzymes only expressed in the mature prostate. A list of the top 10 tags highly expressed in each of these 3 libraries is shown in Tables 3.4, 3.5 and 3.6. Comparison between the male and female UGS libraries showed that most tags more highly expressed in the female UGS library were ribosomal proteins, as was observed in the male UGS library. The majority of the tags differentially highly expressed in the U G M and U G E libraries were also ribosomal proteins which may be involved in differential protein syntheses (Data not shown).  3.2 Validation of the Prostate SAGE Libraries To confirm SAGE data obtained for the 5 prostate libraries (except the female UGS library), bioinformatic validation was carried out by searching the libraries for genes known to be expressed in the developing and adult prostate (Table 3.7). The Refseq sequences corresponding to these genes were retrieved using DiscoverySpace and SAGE tags were generated. The most 3' end tags (1 position tags) were used to query our st  libraries for their expression. Tags from other positions were also searched i f the 1  st  position tag was ambiguous or was not found in any of the libraries. 18/27 genes were expressed as expected. Probasin (153) and prostate P-defensin (239) are prostate specific secreted proteins whose transcripts were expected, and detected, at high levels in the adult prostate S A G E library. A gene encoding PSP94, a secreted protein unique to the DLP, was also expressed specifically in the adult library (297, 298). Important developmental genes such as Shh (62) and Bmp4 (61) had the highest tag counts in the  72  Table 3.4: Top 10 tags more highly expressed in the E16.5 male U G S than in the PO and adult prostate libraries. (99% cutoff)  Tag  Tag Counts (Tags per 100,000)  Gene *  Male UGS  P0  Adult  ATACTGACATTTTGTAG  945  814  583  NC 005089.lt  CCCTTCTTCTCTCCCTT  679  130  28  hemoglobin alpha, adult chain 1  TG G ATCCTG AG A ACTTC  610  159  28  hemoglobin, 3 adult minor chain  AACAATTTGGGCTCTTT  331,  196  135  ribosomal protein L9  ATGACTGATAGCAAGTC  325  153  182  NC 005089.lt  TTCATTATAATCTCAAA  262  138  73  Prothymosin alpha  ATCAACACCGCAACCTT  241  177  97  GNAS  GAGCGTTTTGGGTCCAG  208  142  82  peptidylprolyl isomerase A  TTTTATGTTTAAATAAA  199  105  132  ribosomal protein L32  AACAGGTTCAATCAGCT  188  133  107  ribosomal protein S25  * Tag to gene mapping was carried out using DiscoverySpace with the CMOST plug-in. t These tags are located on the mitochondria genome which has only one annotation for all genes.  Table 3.5: Top 10 tags more highly expressed in the P0 library than in the E16.5 male UGS and adult prostate libraries. (99% cutoff)  Tag  Tag Counts (Tags per 100,000)  Gene *  Male UGS  P0  adult  GCTGCCCTCCACCATAT  316  648  469  NC 005089.lt  GGCAAGCCCCAGCGTCT  225  373  220  RpllOa protein  ACTTATTATGCAAGCTG  100  237  40  decorin precursor  ATACTGAAGCCCCACTT  122  182  104  ribosomal protein L13  GGAGTAAGAACACAGCT GACCACCTCTGTTTTAT  45 31  125 116  33 0  H3 histone, family 3B microfibril-associated glycoprotein 2  GTTCCAAAGAAGTCTTG  46  , .116 i  0  collagen alpha 2 (I) chain precursor  GGCTTCCGCGAGGGTAC GAACATTGCACCACACG  48 23  0 86 ' 84^ • 9  CCCAATGGCCCAATAAA  40  77 ,  4  midkine secreted acidic cysteine rich glycoprotein procollagen, type VI, alpha 2  * Tag to gene mapping was carried out using DiscoverySpace with the CMOST plug-in. t These tags are located on the mitochondria genome which has only one annotation for all genes.  73  Table 3.6: Top 10 tags more highly expressed in the adult library than in the E16.5 male UGS and PO prostate libraries. (99% cutoff)  Tag  Tag Counts (Ta gs per 100,000)  Gene *  Male UGS  PO  Adult  TTCTAATCGGTATTAGT  2  0  ,2264.  probasin  GTACCTGTGAGAATGGA GGAGGTAGACCCTTTTT  1 0  0 0  2196 1366  prostate P-defensin 1 hypothetical protein 9530008N10  TGCCAACTGATCAGTCA CTCACTTAGTGTAAGCT CTATATGTATACTTCTG  0 0 0  0 '< 1195 0 900 631; o 1  TAACTGACAATAAAAGC  10  15  Transcript Derived From ESTs palmitoyl-protein thioesterase-like protein 28537044 similar to BOP-1 protein  623  metallothionein 2  TGACAAAACGTCAATCA  0  0  593  p-microseminoprotein  AAAAAGTACCAGAGCTG  35  64  522  BC033600  ACTAGACTCAGATTGTC  0  0  516  CAAA01003848  * Tag to gene mapping was carried out using DiscoverySpace with the C M O S T plug-in.  74  Table 3.7: Known genes used for bioinformatic validation of the E16.5 male UGS, PO prostate and the adult prostate SAGE libraries. Tag counts (Tags per 100,000)  P value Male UGS vs. Adult  P value P0 vs Adult  Expected Expression  1  0.507f  0.413  All  0  7  0.173  0.039  All  0  0  30  4.217E-8  1.039E-6  All  Prostate stem cell antigen  16  14  1  0.002  0.008  All  Tgfpl  0  .0  0  N/A*  N/A  All  Type II 5 alpha reductase  0  0  0  N/A  N/A  All  ERl(a)  0  0  0  N/A  N/A  All  ER2((3)  0  6  1  0.365  0.187  All  Probasin  2  0  2264  0.000  0.000  Adult  Prostate p-defensin 1  1  0  2196  0.000  0.000  Adult  Psp94  0  0  593  2.867E-146  7.964E-120  Adult  Notch 1  1  0  0  0.507  N/A  El6.5 and/or P0  Hoxal3  0  0  0  N/A  N/A  El6.5 and/or P0  HoxalO  8  14  9  0.894  0.371  El6.5 and/or P0  Hoxdl3  0  0  0  N/A  N/A  El6.5 and/or P0  Shh  9  8  0  0.010  0.025  El6.5 and/or P0  Ptc  0  0  0  N/A  N/A  El6.5 and/or P0  Gene  Male UGS  P0  Adult  Androgen receptor  1  2  p63  2  Nkx3.1  Glil  0  0  0  N/A  N/A  El6.5 and/or P0  Gli2  0  0  0  N/A  N/A  El6.5 and/or P0  Gli3  0  0  0  N/A  N/A  El6.5 and/or P0  Bmp4  13  12  0  0.002  0.003  El6.5 and/or P0  Bmp receptor, type 1A  8  6  4  0.390  0.649  El6.5 and/or P0  Bmp receptor, type 1B  0  0  23  2.342E-6  2.786E-5  El6.5 and/or P0  Tgfp2  2  3  4  0.179  0.664  El6.5 and/or P0  Tgfl33  0  3  0  N/A  0.223  El6.5 and/or P0  FgflO  0  0  3  0.179  0.275  El6.5 and/or P0  Fgf7  1  0  0  0.507  N/A  El6.5 and/or P0  f p values were rounded off to 3 decimal places. * N/A: p value is not available since statistics can not be performed when both counts were "0".  75  male UGS. As expected for such genes, expression levels were decreased at PO and tags were not detected in the adult prostate library, which correlates with previously reported data (61, 62). However, some genes known to be expressed in the prostate were either not detected (9 of 27 tested, 33%), or were present at very low levels (less than 5 counts) (7 of 27 tested, 26%) in the libraries. For example, AR, which is expressed in the U G M from E l 5.5 and upregulated in the adult prostate, was found only as a singleton in each of the 3 prostate libraries (E16.5 male UGS, PO and adult). It was not detected in the female UGS library, or in the male U G M and U G E libraries. NkxS.l, the earliest known prostate specific marker which is expressed from El5.5, was only detected in the adult prostate library (84, 299). Steroid 5 a-reductases are known to be expressed in the U G S of both male and female (300). Unfortunately these genes were not detected in any of the 6 libraries. Transcripts for growth factors such as TGFp\ FGF7, FGF 10 and their receptors which have been reported to be expressed in the developing prostate (45-47, 52-54), were either present in the libraries at very low levels or not detected. Additional validation of the U G E and U G M libraries was performed by looking for markers known to be specifically expressed in each tissue. Keratins are a family of proteins expressed specifically in the epithelium. Among them, keratins 5, 7, 8, 14, 18 and 19 have been shown to be expressed in the developing rat prostate epithelium (9). As expected, transcripts for all these keratins were expressed highly in the U G E library (Table 3.8). Keratin 7 and keratin 18 were also present in the U G M library at low abundance, probably due to slight contamination of the library by epithelial tissue during separation. Shh expression was detected only in the U G E library as expected (62). Validation of mesenchymal specific markers is also shown in Table 3.8. Vimentin, the  76  Table 3.8: Expression of known mesenchymal or epithelial specific genes in U G M and UGE SAGE libraries. Mesenchyme  Epithelium Gene  UGE  UGM  Vimentin  6  16  a-actin  3  34  Desmin  2  4  Bmp4  0  8  UGE  UGM  Keratin 5  45*  0  Keratin 7  50  4  Keratin 8  39  0  Keratin 14  35  0  Keratin 18  71  4  Keratin 19  63  0  Shh  41  0  Tag counts were normalized to tags per 100,000.  77  first marker detected throughout the mesenchyme (10), is downregulated with prostate development (125). Desmin is upregulated during prostate development and higher in the adult than the developing prostate, which correlates with the literature (10). Lower levels of vimentin and a-actin were also detected in the U G E library, suggesting trace contamination during separation. Bmp4 was detected in the U G M library, correlating well with previously reports (61).  3.3 Identification oiSfrp2 As A Candidate Gene As outlined in the introduction, M / E interactions play a critical role in prostate development. Tissue recombination experiments reveal that secreted molecules produced by the mesenchyme are required for epithelial differentiation. To identify factors that may be involved in prostate development, additional comparisons were carried out amongst the 6 constructed SAGE libraries. We hypothesized that genes critically important in early prostate development would be differentially expressed in the male and female UGS. We further hypothesized that they would be more highly expressed at early stages of prostate development and downregulated in the adult prostate. Therefore, we first compared the male and female UGS libraries to identify tags upregulated in the male UGS, and 147 such tags, identified (Figure 3.3; left panel). Next we compared each of the developmental libraries (the El6.5 male UGS library and the PO prostate library) to the adult prostate library to identify genes more highly expressed at early stages of prostate development. 242 tags were identified that were present at higher levels in both the El6.5 male UGS and PO prostate libraries as compared with the adult prostate library (Figure 3.3; middle and right panels). Combining the results of these comparisons produced a list  78  38,639  Female UGS  34,521 538  147  Male UGS  Male UGS  Adult DLP  Adult DLP  414 tags  516 tags  147 tags  PO prostate  242 tags  34 tags  Figure 3.3 Identification of Sfrp2 through bioinformatic comparisons of the S A G E libraries. The male and female UGS SAGE libraries were first compared to identify tags more highly expressed in the male UGS library (left panel; 147 tags identified). Next, the male UGS library was compared with the adult DLP library to identify tags more highly expressed in the UGS library (middle panel). Finally, the PO prostate library was compared with the adult D L P library to identify tags more highly expressed in the PO library (right panel). Commonly expressed, upregulated tags from the middle and right panels (242 tags) were then compared with the tags upregulated in the male versus female UGS to identify a total of 34 tags that were more highly expressed in the male UGS and in the developing prostate libraries.  79  of 34 tags that fulfilled both criteria (Table 3.9). These tags were annotated using CMOST and were then analyzed using GO browser. Although growth factors and their receptors are proposed to be involved in M / E interactions during prostate development none were detected as part of these 34 tags. However GO analysis identified 4 genes encoding extracellular space molecules including: Decorin precursor, Sfrp2, Glypican 3, and Insulin-like growth factor binding protein 2 (Igfbpl). Decorin, a small proteoglycan, is a component of the extracellular matrix that plays a role in matrix assembly (301). Glypican-3, a cell-surface heparan sulfate proteoglycan, plays a negative role in growth control (302). Glypican-3 also modulates B M P - and FGF-mediated effects during renal branching morphogenesis (303). IGFBP2 is a growth factor binding protein that has been shown to be overexpressed in prostatic intraepithelial neoplasia (PIN) and invasive cancer (304, 305). IGFBP2 also has a potent stimulatory effect on growth of PCa cells (306). SFRP2, as introduced in Chapter 1, is a member of the SFRP family that are believed to be antagonists of the WNT pathway (272).  3.4 Expression of WNT Family Members during Prostate Development The identification ofSfrp2 as one of the developmentally upregulated genes raised the interesting possibility that the WNT pathway may be involved in prostate development. The expression and function of WNT pathway members in prostate development have not been previously described. Mining of the literature produced a list of family members involved in the pathway, either directly or as regulators. The Refseq sequences corresponding to these genes were retrieved using DiscoverySpace and SAGE tags were  80  Table 3.9: Sequence and identity of tags expressed at higher levels in the developing prostate libraries. P value}  Sequence  Gene identified  P value*  P valuef  AACAATTTGGGCTCTTT  ribosomal protein L9  1.014E-11A  8.904E-16  0.006  2.141E-4  0.008  AAGAAAATAGAGGACAA  ribosomal protein L23a  1.320E-5  AAGAGGCAAGACGAAAA  ribosomal protein S15a  1.039E-5  1.999E-11  6.805E-5  AAGGAAGAGATGGCTCG  vimentin  0.008  1.581E-6  1.055E-10  AATTTCAAAACACCACG  ribosomal protein S17  2.761E-7  7.392E-11  1.005E-6  ACAAAGGTTAAAAAAAA  Lmo4 protein  5.182E-5  7.672E-6  0.001  ACATCATAGATGACATC  ribosomal protein L12  2.210E-4  3.713E-6  1.438E-4  ACTTATTATGCAAGCTG  decorin  0.008  6.601E-6  8.370E-25  ACTTCAGCCAGATTAGC  t-complex testis expressed 1  2.493E-4  1.295E-4  0.003  AGCAGTCCCCTCCCTAG  Mtco2  2.078E-9  1.221E-7  4.040E-5  AGGAGGACTTAACCAAA  Mtnd2  5.912E-16  4.976E-5  6.289E-4  AGGCAGACAGTTGCTGT  eukaryotic translation elongation factor 1 alpha 1  5.278E-5  8.462E-5  4.310E-5  Mtco3  6.U2E-24  3.078E-16  3.120E-7  1.025E-5  3.495 E-12  5.424E-5  5.087E-5  1.117E-8  9.150E-7  ATACTGACATTTTGTAG ATCAACACCGCAACCTT ATGCCCTCAAATAAAAA  guanine nucleotide binding protein a stimulating isoform a secreted frizzled-related sequence protein 2  CAAACTCTCACAGCAAT  sparc/osteonectin  0.005  2.425E-8  1.022E-31  CACAGAACCAGCAGTCT  chaperonin subunit 3 (gamma)  0.002  6.584E-5  0.003  CATTGCGTGGTTGTAAT  Wbscrl protein  2.010E-7  9.746E-4  0.002  CCACAGAGCTGTCAGAA  Ena-vasodilator stimulated phosphoprotein  0.004  4.129E-4  8.015E-5  CCCCAGCCAGTGCCTAC  ribosomal protein S3  2.389E-4  0.002  0.002  CCTGATCTTTACTTCTA  laminin receptor 1  7.798E-9  4.102E-8  3.190E-4  GATGTGGCTGCTTTTAA  eukaryotic translation elongation factor I beta 2  1.162E-4  0.008  6.289E-4  GATTTCTTTGACAAAAA  glypican 3  0.003  4.23 7E-6  7.300E-5  GCAGGCACTCAATAAAT  tubulin, beta 5  4.255E-6  2.907E-21  1.165E-18  GGCTTAAGTAGGAAAGT  Anxal protein  0.001  1.294E-4  0.004  GGGAGCGAAAACGTTAA  inhibitor of DNA binding 2  0.002  0.003  0.002  0.003  2.799E-7  2.532E-6  1.058E-5  1.007E-8  5.239E-6  GTCTGCTGATGGCCAGA TATTTATATTTGGAAAG  guanine nucleotide binding protein, beta 2, related sequence 1 insulin-like growth factor binding protein 2  TCTCTGACTTGATAAGC  marcks-related protein  0.007  7.757E-14  3.939E-6  TGAAATAAACTCAGTAT  nucleophosmin 1  2.086E-4  3.219E-6  0.009  TGATATGAGCTCACAGA  lactate dehydrogenase 2, B chain  2.181E-4  2.312E-6  0.005  TGTTCATCTTGTTTTAA  procollagen, type III, alpha 1  0.003  2.058E-26  5.196E-36  TTCATTATAATCTCAAA  prothymosin alpha  8.345E-5  1.819E-20  1.991E-4  TTTTTAATGTTGTCAGT  H3 histone, family 3A  3.416E-5  2.571E-16  7.168E-9  * p value of the comparison between the female and male UGS libraries, t p value of the comparison between the male UGS and adult D L P libraries, f p value of the comparison between the P0 prostate and adult DLP libraries. A p values were rounded off to 3 decimal places.  81  generated. The most 3' end tags (1 position tag) were used to search our libraries for st  their expression. Tags from other positions were also searched i f the 1 position tag was st  ambiguous or was not found in any of the libraries (Table 3.10). We observed that many WNT pathway members (25 out of 57 searched, 44%) were not detected and some members (17 out of 57 searched, 30%) were detected at very low levels (less than 5 counts). P-catenin, a downstream member of the canonical W N T pathway, had abundant tags in our libraries. O f particular interest were the W N T pathway members that showed a similar trend in expression to that of Sfrp2, i.e. higher tag counts in the male UGS library, or with decreased expression during development (p<0.05). This suggests that they may also be regulated developmentally. Included in this list were: Fzdl, P-catenin, Dapper homolog 1, and Dkk3. Members occupying various positions in the WNT signaling cascade were selected to validate their expression in the developing and adult prostate by RT-qPCR. Some tags that were represented at higher levels than other WNT members in the libraries, including Wntll, Fzdl, Fzd7, Axin, Nakedl and RhoA were also selected. Another 2 genes, Lrp5 and Lefl, are required components for the canonical WNT pathway and thus were also selected. Results revealed developmentally differential expression of all the selected members (Figure 3.4). With the exception of Lrp5 and RhoA, all of the transcripts were downregulated during development (Figure 3.4). RT-qPCR also revealed differential expression of Nakedl, Lefl, Lrp5, Fzdl and Fzdl in the male and female El6.5 UGS (Figure 3.5), with Nakedl, Lefl, LrpS highly expressed in the female UGS and Fzdl and Fzd7 highly expressed in the male UGS. Taken together, these data showed that a number of WNT pathway members were differentially expressed in the developing prostate,  82  Table 3.10: Expression of W N T pathway members in the 6 S A G E libraries analyzed. Gene  El6.5 Male UGM  El6.5 Male UGE  El6.5 Female UGS  E16.5 Male UGS  PO Prostate  Adult DLP  Wntl  0*  0  0  0  0  0  Wnt 10a  0  3  0  0  0  0  Wnt 10b  0  0  0  0  0  0  Wntll  3  2  3  0  5  0  Wntl 6  0  0  0  0  0  0  Wnt2  0  0  0  0  0  0  Wnt2b  0  0  0  0  0  0  Wnt3  0  0  0  0  0  0  Wnt3a  0  0  0  0  0  0  Wnt4  0  0  0  1  0  0  Wnt5a  0  0  0  0  0  0  Wnt5b  0  0  0  0  0  0  Wnt6  0  0  0  1  0  0  Wnt7a  0  3  6A  0  0  0  Wnt7b  0  1  1  1  0  3  Wnt8a  0  0  0  0  0  0  Wnt8b  0  0  0  0  0  0  Wnt9a  0  0  0  0  0  0  Wnt9b  0  0  0  0  0  0  Frizzled 1  17§  2  7  7t  9t  0  Frizzled 10  0  0  0  0  0  0  Frizzled 2  3  2  3  3  2  3  Frizzled 3  0  0  0  0  0  0  Frizzled 4  0  0  0  0  0  0  Frizzled 5  0  0  0  0  0  0  Frizzled 6  2  9  4  0  3  3  Frizzled 7  12  8  16A  2  15  6  Frizzled 8  0  0  0  0  0  0  Sfrpl  1  0  0  5  2  0  Sfrp2  16  12  8  36Af  29J  0  Sfrp4  0  0  0  2  0  0  Sfrp5  0  0  1  0  0  0  83  Table 3.10: Expression of W N T pathway members in the 6 S A G E libraries analyzed (Continued). Gene  El6.5 Male UGM  El6.5 Male UGE  El6.5 Female UGS  E16.5 Male UGS  P0 Prostate  Adult DLP  Lrp5  0  1  1  0  0  1  Lrp6  0  0  0  0  0  0  Axin  1  0  2  2  5  0  Axin2  1  2  2  2  3  4  Ape  0  0  0  0  0  0  CK1, alpha 1  13  13  9  10  18  7  Lcfl  0  0  3  0  0  0  P-catenin  30  69  34  60Af  47  37  Dapper 1  1  3  9  12t  3  1  Dvl2  0  0  6  2  5  0  Dvl3  0  0  0  0  0  0  Dvll  0  0  0  0  0  0  GSK3P  0  0  0  1  0  0  Naked1  11  4  3  6  5  1  Naked2  2  0  1  0  0  0  Dickkopfl  0  0  0  0  0  0  Dickkopf2  0  0  4  1  0  1  DickkopD  2  0  1  7t  3  0  Dickkopf4  0  0  0  0  0  0  TcB  0  3  1  2  3  0  Tcf7  0  1  0  0  0  0  RhoA  11  10  15  20  29  21  Wifl  0  1  0  0  2  17ft  * Tag counts were normalized to tags per 100,000. Sfrp2 and P-catenin are highlighted. Members showing the same expression trend as Sfrp2 are shown in pink. § Indicates counts that are statistically "upregulated" at p<0.05 in a comparison of the El6.5 male U G M and U G E libraries. A Indicates counts that are statistically "upregulated" at p<0.05 in a comparison of the El6.5 female and male UGS libraries. t Indicates counts that are statistically "downregulated" at p<0.05 in a comparison of the El6.5 male UGS and the adult DLP libraries. J Indicates counts that are statistically "downregulated" at p<0.05 in a comparison of the P0 and adult DLP libraries.  84  Figure 3.4 RT-qPCR analysis of WNT family member expression during prostate development. Expression levels of the indicated transcripts were determined using RTqPCR in the El6.5 male UGS, the PO prostate, and the adult DLP. Each bar represents the mean ± SD of 3 technical replicates using a single sample consisting of tissues pooled from at least 2 mice. The y-axis indicates the fold change relative to Ubc (See Chapter 2, Materials and Methods). Statistical analyses were carried out using the Student's t-test where * indicates p<0.05 compared to the adult prostate.  Lrp5  *  210-  i UGS  85  PO  • Adult  I  Figure 3.5 RT-qPCR analysis of WNT family member expression in E16.5 male and female UGS. Expression levels of the indicated transcripts were determined using RTqPCR in the El6.5 male and female UGS. Each bar represents the mean ± SD of 3 technical replicates using a single sample consisting of tissues pooled from at least 2 mice. The y-axis indicates the fold change relative to Gapdh (See Chapter 2, Materials and Methods). Statistical analyses were carried out using the Student's t-test where * indicates p<0.05 compared to the adult prostate.  Frizzled7  Frizzledl *  21 0 -i  i  — i ••• female  male  *  10-  ,  J  0female  86  \mm\ male  ,  suggesting that they may play a role in prostate organogenesis.  3.5 Assessment of the Expression Pattern of Sfrp2 by RT-qPCR and Digital Northern Confirmation of the regulated expression of other members of the W N T signaling pathway in the developing prostate suggested that further analysis of Sfrp2 expression was warranted. Since Sfrp2 encodes a secreted extracellular molecule differentially expressed in the male and female UGS, as well as in the developing prostate, it is possible that it may be involved in M / E interactions. To determine if Sfrp2 is expressed in other tissues, we performed Digital Northern analysis. As shown in Table 3.11, Sfrp2 was detected at a higher level in the male U G S library than in all other 149 mouse SAGE libraries in the database with the exception of the retina library. The high levels of Sfrp2 transcripts in the retina confirmed previous work (286, 307) (Table 3.11). In addition to the UGS and prostate libraries, Sfrp2 was also present in the telencephalon at E9.5 (TS 15), correlating with previous reports (308). Digital Northern analysis of Sfrp2 expression in human SAGE libraries was somewhat different from our mouse S A G E libraries (Table 3.12). In a human normal prostate library 30 tags per 200,000 were detected while in mouse there were only 2 tags per 200,000. Due to the lack of developmental human SAGE libraries, the expression pattern of Sfrp2 during human development could not be analyzed. RT-qPCR was performed to confirm the expression pattern and levels of Sfrp2 suggested by the S A G E data (Figure 3.6). The expression pattern of Sfrp2 during prostate development showed the same trend as in the S A G E libraries. It was highly expressed at  87  Table 3.11: Digital Northern analysis ofSfrp2 expression in mouse SAGE libraries. Library Retina normal TS20| Urogenital_sinus_male_normal_TS24  Total Tags 52319 107747  Tags per 200,000 179 63  Prostate_normal_Od  92631  45  Retina_normal_TS22 Urogenital_sinus_male_mesenchyme_normal_TS24  54733 112423  40 33  Bladder normal_TS22  135961  29  Optic_cup_normal_TS 17 Urogenital_sinus_female_normal_TS24 Urogenital_sinus_male_epithelium_normal_TS24  127386 114350 143679  26 19 19  Bladder_normal_TS24 Telencephalon_ventral_normal_TS 17  73830 104750  18 17  Mammary_gland_normal_35d  135062  16  Telencephalon_normal_TS 15  117444  15  Lung_normal_TS 19  116547  10  Prostate dorsal lobe normal 12w  94713  2  * To search all mouse SAGE libraries for expression of Sfrp2, the 21 base pair tag (CAT G A T G C C C T C A A A T A A A A A ) was truncated to a 14 base pair tag ( C A T G A T G C C C T C AA). | TS: Theiler stage.  88  Table 3.12: Digital Northern analysis of Sfrp2 expression in human SAGE libraries.  Library  Total Tags  Tags per 200000  Breast carcinoma associated myofibroblast  63813  366  Breast_carcinoma_associated_myofibroblast  65091  362  Breast_carcinoma_associated_myofibroblast  74281  207  Brain_meningioma_grade_II  44857  142  Breast carcinoma_associated_stroma  68024  129  Brain meningioma gradell  48711  102  Breast carcinoma _associated_myoepithelium  81452  98  Brain meningioma_grade_I  54647  87  Breast_normal_stroma  79152  65  Peritoneum_normal  53527  56  Breast normal myoepithelium  69006  46  Breast carcinoma_associated_stroma  57049  45  Breast_carcinoma_associated_myoepithelium  37435  42  Brain astrocytoma_grade_III  109886  41  Pancreas adenocarcinoma  33582  41  White Blood Cells breast carcinoma_associated  34399  40  Carti lage_chondrosarcoma_grade_2  93846  38  Brain_meningioma_grade_I  69253  34  Brain_astrocytoma_grade_III  99939  32  Pancreas_adenocarcinoma  33213  30  Prostate_normal  59277  30  Prostate adenocarcinoma  64951  6  * To search all human SAGE libraries for expression of Sfrpl, the 21 base pair tag (CAT G A T T T C T T C A A G T A A A A G ) was truncated to a 14 base pair tag ( C A T G A T T T C T T C A A).  89  Figure 3.6 Comparison of Sfrp2 expression levels using S A G E (right panels) and R T - q P C R (left panels). Sfrp2 transcript levels were determined in the male UGS, PO prostate, and adult D L P using RT-qPCR and SAGE (A, B respectively). Statistical analyses of RT-qPCR were carried out using the Student's t-test and S A G E statistical analyses were performed as described in Material and Methods. * indicates p<0.05 compared to the adult prostate. Sfrp2 transcript levels were also determined in the male versus female UGS using both RT-qPCR and SAGE (C, D respectively). * indicates p<0.05 compared to the female UGS. Finally, Sfrp2 transcript levels were determined in the male U G E versus the male U G M using both RT-qPCR and S A G E (E, F respectively). * indicates p<0.05 compared to the U G M . In all cases, the RT-qPCR data represents the mean ± SD of 3 biological replicates. Each replicate consisted of tissues pooled from 2 mice. SAGE  RT-qPCR 200  A  *  150 100 50  ~j~  0  r  UGS  | J- | P0  I I  i I  Adult  90  E l 6.5 and downregulated with maturation of the prostate (Figure 3.6A, B). Comparison of Sfrp2 expression levels in the male and female UGS also confirmed the S A G E data, with higher expression in the male U G S (Figure 3.6C, D). The RT-qPCR result comparing Sfrp2 expression in the U G M and the U G E also correlated well with the SAGE data, with Sfrp2 detected in both the mesenchyme and the epithelium. Expression in the U G M was higher than in the U G E (Figure 3.6E, F). Taken together, both RT-qPCR analyses and the S A G E data showed differential expression of Sfrp2 during prostate development, suggesting that its role in this process is worthy of further study.  3.6 RT-qPCR Analysis of Wnt4 and p-catenin Expression Levels SFRPs are believed to act on the WNT pathway by binding to the W N T ligands, thus preventing them from binding to their receptors. WNT3a (309), Xenopus WNT8 (XWNT8) (310) and WNT4 (277) have all been reported to interact with SFRP2. In order to investigate whether any of these WNT ligands may interact with SFRP2 in prostate development, the expression levels of Wnt3a, Wnt8a and Wnt8b (orthologs of Xwnt8) (311), and Wnt4 in the El6.5 and PO prostate were examined by RT-qPCR. No expression of Wnt3a, Wnt8a or Wnt8b was detected in the developing prostate (Figure 3.7 and data not shown). Wnt4 was detected in all tissues examined (Figure 3.8). During development, Wnt4 was highly expressed in the E16.5 male UGS and PO prostate with its expression downregulated in the adult (Figure 3.8A and B). It was expressed at slightly higher levels in the male UGS than the female U G S , with expression mainly in the epithelium of the male UGS. Thus, of the 4 WNT ligands known to interact with SFRP2, only transcripts encoding WNT4 were detected and developmentally regulated.  91  Figure 3.7 Wnt3a, 8a and 8b are not expressed in the PO prostate. RT-qPCR was carried out on duplicate samples using primers specific for each transcript. R l ES cell cDNA was used as a positive control. The band arising in the Wnt3a RT-qPCR in the PO sample is primer dimers (arrow). Similar results were obtained with the El6.5 male UGS (not shown here).  PO  PO  ES  Wnt3a  ES  PO  PO  ES  Wnt8a  92  ES  PO  PO  E S - ES  Wnt8b  Figure 3.8 Comparison of Wnt4 expression levels using S A G E (right panels) and R T - q P C R (left panels). Wnt4 transcript levels were determined in the male UGS, PO prostate, and adult D L P using RT-qPCR and SAGE (A, B respectively). Statistical analyses of RT-qPCR were carried out using the Student's t-test and S A G E statistical analyses were performed as described in Material and Methods. * indicates p<0.05 compared to the adult prostate. Wnt4 transcript levels were also determined in the male versus female UGS using both RT-qPCR and SAGE (C, D respectively). Finally, Wnt4 transcript levels were determined in the U G E versus the U G M using both RT-qPCR and SAGE (E, F respectively). * indicates p<0.05 compared to the U G M . In all cases, the RTqPCR data represents the mean ± SD of 3 biological replicates. Each replicate consisted of tissues pooled from 2 mice. RT-qPCR  SAGE  F  1 0.8 0.6  0.4 0.2 n u  93  i 1  UGE  1 1  UGM  The expression of B-catenin, a downstream effector of the canonical pathway, was also validated by RT-qPCR (Figure 3.9). As with  Wnt4, its expression was  downregulated during prostate development. It was expressed at slightly higher levels in the male UGS compared to the female UGS and in the U G E versus the U G M .  3.7 In situ hybridization of Sfrp2 The RT-qPCR, as well as the Digital Northern analyses, suggested that Sfrp2 expression is regulated during prostate development. Since Sfrp2 was expressed at higher levels in the male U G M while Wnt4 and B-catenin were shown to be predominantly present in the epithelium, it raises the interesting possibility that the WNT pathway may be involved in M / E interactions. We next determined the spatial and temporal expression pattern of Sfrp2 mRNA during prostate development using in situ hybridization. As described previously, we detected Sfrp2 expression in the nasal cavity (288) and telencephalon (308) (Data not shown). Consistent with the SAGE and RT-qPCR data, the probe for Sfrp2 weakly stained the male El6.5 U G E , with more predominant expression in the U G M (Figure 3.10 A , B). The mesenchymal staining pattern with the Sfrp2 probe was intriguing. In the lower UGS and urethra, Sfrp2 was detected in the U G M immediately adjacent to the U G E . In contrast, in the upper U G S , Sfrp2 levels were significantly lower in the adjacent mesenchyme suggesting that it might be excluded from this region. A similar expression pattern was also seen in the PO prostate (Figure 3.10C, D). Of note, in the PO prostate, Sfrp2 was also detected in the nascent prostate buds (Figure 3.10D, indicated by arrow). Sfrp2 staining could not be detected in the adult prostate (Figure 3.10E, F).  94  Figure 3.9 Comparison of P-catenin expression levels using S A G E (right panels) and R T - q P C R (left panels). P-catenin transcript levels were determined in the male UGS, PO prostate, and adult D L P using RT-qPCR and SAGE (A, B respectively). Statistical analyses of RT-qPCR were carried out using the Student's t-test and SAGE statistical analyses were performed as described in Material and Methods. * indicates p<0.05 compared to the adult prostate. P-catenin transcript levels were also determined in the male versus female UGS using both RT-qPCR and SAGE (C, D respectively). Finally, /?catenin transcript levels were determined in the U G E versus the U G M using both RTqPCR and SAGE (E, F respectively). * indicates p<0.05 compared to the U G M . In all cases, the RT-qPCR data represents the mean ± SD of 3 biological replicates. Each replicate consisted of tissues pooled from 2 mice.  95  Figure 3.10 In situ hybridization of Sfrp2 during prostate development. Expression of Sfrp2 in the El6.5 male urogenital system is shown at low (2.5x; A ) and high magnification (40x, B) with the UGS, bladder (B), urogenital sinus epithelium (E), and urogenital sinus mesenchyme (M) indicated. Expression of Sfrp2 in the PO prostate is shown at low (lOx; C) and high (40x; D) magnification. Arrow shows the nascent epithelial buds. Expression of Sfrp2 in the adult D L P is shown at low (lOx; E) and high (40x; F) magnification. In all cases the magnified area is indicated by the outline in the lower magnification photo. A  B  M  *  JB B 40x  2.5x D  M  10x  40x  10x  40x  96  3.8 Immunofluorescence to Detect SFRP2 and p-catenin Protein During Prostate Development Most factors involved in M / E interactions are secreted molecules that function in a paracrine manner. For example, WNT proteins have been shown to move a short distance to function in other cells (312). SFRP2 is also a secreted protein. Previous work indicates differential expression of the Sfrp2 transcript and protein in developing lung (290). To determine i f expression of the SFRP2 protein during prostate development was consistent with the R N A result, immunofluorescent antibody staining was performed on prostate tissues at various stages of development. In keeping with its role as a secreted protein, SFRP2 was localised to the cell membrane and/or extracellular space in the tissues examined (Figure 3.11). Detection of the SFRP2 protein correlated with its R N A expression pattern, i.e. it was present in both U G M and U G E at E16.5 and in both the urethral epithelium and nascent prostate buds at PO. Similarly SFRP2 was barely detected in the adult DLP (Figure 3.11). SFRP2 has been implicated in regulating p-catenin protein expression via the WNT signaling pathway (278). To further investigate the possibility that SFRP2 may interact with the canonical W N T pathway, P-catenin immunoreactivity was investigated. Pcatenin was detected in the epithelium of the developing prostate as well as in the adult DLP  (Figure  3.12).  The  antibody  used  for  this  analysis  detects  both  cytoplasmic and nuclear P-catenin. Although nuclear P-catenin is an indicator of active WNT signaling, we could not detect nuclear staining in any of the sections.  97  Figure 3.11 Immunofluorescent detection of SFRP2 in the developing and adult prostate. Immunofluorescent antibody staining was carried out to detect SFRP2 protein in the El6.5 male UGS (A), the PO prostate (C), and the 12-week adult D L P (E). Corresponding controls in which the 1 antibody was omitted are also shown (B, D, and F). Final magnification is 20x in all cases. st  SFRP2 antibody  R  98  No 1 antibody control  Figure 3.12 Immunofluorescent detection of p-catenin in the developing and adult prostate. Immunofluorescent antibody staining was carried out to detect p-catenin protein in the El6.5 male UGS (A, 40*), the PO prostate (C, 20*), and the 12-week adult DLP (E, 20x). Corresponding controls in which the 1 antibody was omitted are also shown (B, D, and F). st  P-catenin antibody  No 1 antibody control st  F •  .. •>• M 40x D  99  3.9 RT-qPCR analysis of Sfrp2, p-catenin and Wnt4 in the TRAMP Mouse Model of PCa In many cases, cancer seems to be caused by the deregulation of pathways that are normally involved in development and thus many developmental genes have been implicated in cancer initiation and progression. We have detected expression of Sfrp2 and WNT pathway members in the developing prostate, indicating that they may play a role in prostate development. As outlined in the introduction, evidence suggestive of a role for the W N T pathway in PCa is just beginning to emerge. For example, direct activating mutations in P-catenin, as well as inactivating mutations of its regulators such as A P C have been discovered in multiple human cancers including PCa (261, 313, 314). WIF1, a WNT pathway antagonist, is downregulated in many cancers including PCa (265). Sfrp2 mRNA has been reported to be upregulated in mammary cancer (293, 294) and was found to induce tumorous transformation in normal mammary epithelial cells and to inhibit apoptosis (295). The role of SFRP2 in PCa has not been previously described. Therefore, RT-qPCR and immunofluorescent staining were carried out to investigate expression of Sfrp2 transcript and its protein in T R A M P mice. Expression levels of Sfrp2,  P-catenin  and Wnt4 transcripts were determined at 4  different stages (2 week, 4 week, 12 week and 26 week) during PCa development and progression in T R A M P Tg mice compared to their WT littermates (Figure 3.13). RTqPCR results showed that Sfrp2 was upregulated in the Tg mice at 4 weeks of age compared to their WT littermates. At 12 weeks of age Sfrp2 transcript levels in the prostate were back to WT levels and stayed at that level at 26 weeks (Figure 3.13A). /?-  catenin  was similarly expressed in WT and Tg mice at all stages of PCa development  100  Figure 3.13 RT-qPCR analysis of Wnt4, P-catenin and Sfrp2 expression in the DLP of TRAMP mice during PCa development. Expression levels of Sfrp2 (A), P-catenin (B), and Wnt4 (C) were determined in T R A M P (Tg) as well as wild type (WT) mice of the indicated ages (x-axis). The y-axis indicates fold change relative to P-actin. Data are normalized according to the 2 week old mice. Statistical analyses were carried out using the Student's t-test where * indicates p<0.05 compared to age-matched W T mice. In all cases, the RT-qPCR data represents the mean ± SD of 3 technical replicates. The replicates consisted of tissues pooled from 2 mice. A : Sfrpl 1.2  n  1 0.8  • WT  0.6  Tg  0.4 0.2 0 2  4 Age  1.2  12  28  (weeks)  B: P-Catenin  1 0.8  - WT  0.6  Tg  0.4 0.2 0 4 Age  14  12  28  (weeks)  C: Wnt4  12 10 8  • WT  6  Tg  4 2 0 4 Age  12  28  (weeks)  101  (Figure 3.13B). Wnt4 expression was normal in Tg mice at 2 and 4 weeks of age, but was upregulated in 12 and 26 week old T R A M P mice (Figure 3.13C).  3.10  Immunofluorescence to Detect SFRP2 and p-catenin Protein in the Prostate  of TRAMP mice Since transcript levels do not always correlate with protein levels, we performed immunofluorescence analyses to detect SFRP2 and P-catenin protein in both 4 week old and 8 week old T R A M P mice. Expression of SFRP2 was barely detectable in these samples and there was no detectable difference in the expression pattern between the WT and T R A M P mice (Figure 3.14). Cytoplasmic P-catenin was detected in the prostate epithelium of both 4 and 8 week old T R A M P mice as well as the WT controls of the same.age (Figure 3.15). There were no obvious differences in expression patterns and no nuclear staining was observed in WT or T R A M P mice.  3.11  Summary In this study, we constructed six SAGE libraries at critical stages during mouse  prostate development. Bioinformatic analyses were carried out and Sfrp2, a gene encoding a W N T pathway member, was identified as a candidate gene that may be involved in mouse prostate development. In addition, we detected other members of the WNT pathway in our SAGE libraries and confirmed differential expression of some members during prostate development as well as in male and female UGS by RT-qPCR. The expression of Sfrp2 was validated by RT-qPCR. We observed that Sfrp2 is highly  102  Figure 3.14 Immunofluorescent detection of SFRP2 protein in T R A M P mice and W T littermates. SFRP2 protein levels were localized in the prostates of 4-week old WT (A) and T R A M P (Tg; B), as well as 8-week old WT (C) and Tg (D), mice. The control (E) contained no primary antibody. Final magnification is 20x in all cases (n=2). A  B  No Ab control  103  Figure 3.15 Immunofluorescent detection of p-catenin protein in T R A M P mice and W T Iittermates. P-catenin protein levels were localized in the prostates of 4-week old WT (A) and T R A M P (Tg; B), as well as 8-week old WT (C) and Tg (D), mice. The control (E) contained no primary antibody. Final magnification is 20x in all cases (n=2). A  B  No Ab control  104  expressed during prostate development and is downregulated in the adult prostate. The expression levels of WNT  ligands reported to interact with SFRP2 were also examined  by RT-qPCR and only Wnt4 was detected. It was expressed in the epithelium of the developing prostate. We have also confirmed expression of P-catenin, which encodes a downstream molecule of the canonical WNT pathway, by RT-qPCR. The localization of Sfrpl  transcript and its protein was determined by  in situ  hybridization and  immunofluorescence, respectively. These results showed that both Sfrpl transcript and its protein product were highly expressed in the U G M and present at low levels in the UGE. However, immunofluorescence to detect active (i.e. nuclear) P-catenin protein failed. Expression of Sfrpl, Wnt4 and  P-catenin  during PCa initiation and progression was  also evaluated using the T R A M P mouse model. We found that Sfrpl and Wnt4 transcripts were upregulated relative to WT mice at 4 and 8 weeks of age, respectively. However, /?-  catenin  expression was not changed during disease progression. In addition, we did not  detect changes in the levels or expression patterns of SFRP2 and P-catenin proteins.  105  CHAPTER 4. DISCUSSION  4.1  SAGE Library Analyses SAGE is a powerful gene expression profiling tool that allows digital analysis of  overall gene expression patterns. It has been widely used to study various organisms such as yeast, human and mouse (167). A majority of these studies have been focused on human diseases and their mouse models and relative few were on development. This study is the first comprehensive analyses of mouse prostate development using this powerful technique. It contributes to the establishment of a mouse prostate transcriptome database at critical stages during prostate development, which is an important step toward understanding the molecular mechanisms regulating this process. The libraries were all sequenced to be a similar size (-100,000 total tags) to facilitate data analysis and make the comparisons more accurate since library size has been shown to affect the reliability of comparisons (169). Data analysis of the constructed SAGE libraries revealed that they were similar in terms of tag qualities, numbers and distributions, indicating the libraries are consistent and comparable. Approximately 60% of the total tag types in each library were singletons suggesting that our sequencing depth was not sufficient to reach saturation. A study analyzing a SAGE library constructed with mouse embryonic stem cell R N A showed that new transcript discovery reaches a plateau when 300,000 simulated tags have been sequenced (213). In human colorectal cancer cell lines, unique tag discovery approached zero at -650,000 tags collected (315). This indicates that somewhere between 300,000 and 600,000 tags must be sequenced to detect the whole transcriptome of a tissue. Our  106  libraries were not sequenced to a sufficient depth to achieve this goal. Moreover, it is this lack of depth that may account for differences in quantitative gene expression profiling using SAGE and RT-qPCR (see later). In pair-wise comparisons of the 6 SAGE libraries, we found that most of the tags (9799% of the total tags) in the libraries compared were expressed at a similar level (singletons included). This is also due to the existence of large proportion of low abundance tags (tag counts lower than 5) in these libraries. As indicated by Audic et al, variations in counts such as 7 -^0, or 2 ->12 are significant (P < 0.01) evidence of differential gene expression (218). Since most of tags in our S A G E libraries were lower than 5 counts, comparison of these low abundance tags between two libraries did not reach sufficient statistical significance thus giving the overall profiles of the compared libraries a high degree of similarity. It has been shown that prostatic tissue can be induced from El6.5 female UGS by androgen application (14). Since neither the E16.5 male or female U G S are fully committed to their final fates it is not surprising that these libraries were highly similar (99.24%o) in their gene expressions profiles. Comparison of the El6.5 male U G M and U G E libraries showed that less tags were similarly expressed (98%) than in the comparison between E16.5 male U G M and UGS libraries (98.66%, data not shown), as well as in the comparison between El6.5 male U G E and UGS libraries (98.41%, data not shown). The difference between the U G E and U G M may be caused by tissue heterogeneity, i.e. distinct differentiation pathways between the U G M and U G E , or molecules specific to the U G M or U G E that might be involved in M / E interactions. Bioinformatic validation of these SAGE libraries detected some genes known to be  107  expressed at each stage, confirming the libraries reflect the biological characteristic they were meant to represent. However, there were also some genes known to be expressed in the prostate that were not detected. There are numerous possible reasons. First, as indicated above, it is likely that our sequencing depth was not sufficient to detect the entire transcriptome of the tissues examined. It therefore follows that detection of specific genes is apt to be dependent on their expression levels. For example, although Nkx3.1 is known to be expressed in the developing prostate as early as El5.5 and its expression level is upregulated in the adult, it was only detected in the adult library. A second possibility is that tags for some of the genes used for searching the libraries were generated from a 3' U T R which has not been annotated, thus genes associated with these tags would not be detected by our program. In addition, experimental errors may also affect the results. For example, tags mutated by PCR and/or sequencing errors would not be annotated with their corresponding genes. Due to the expense associated with construction and sequencing of S A G E libraries, biological replicates are not routinely carried out. Thus information on biological variation and experimental precision is not available. The expression patterns of candidate genes selected from SAGE libraries need to be further validated. We used RTqPCR to validate gene expression changes detected in our SAGE data. RT-qPCR results of some genes with high tag counts in the SAGE libraries, such as Sfrp2, correlated well with the S A G E data. However, we also observed discrepancies between these two sets of results when the SAGE tag counts were low, indicating that S A G E can not accurately quantify genes detected at low levels. For example, although Wnt4 was detected by SAGE in the male UGS library, it appeared as a singleton and its S A G E counts did not  108  correlate with RT-qPCR. In addition, experimental errors, i.e. sequencing and RT-PCR errors, may also affect quantification of the tag. SAGE libraries can be used for identification of novel, as well as differentially expressed, genes in various tissues. We compared the 3 prostate S A G E libraries (El6.5 male UGS, PO prostate and adult DLP) to 86 other mouse SAGE libraries to identify putative novel "prostate specific" tags. The tags identified were further investigated using Rapid Amplification of cDNA Ends (RACE)-PCR. As summarized in Appendix 1, no novel tags were identified. In a second study, RT-qPCR validation of the highest PO specific transcription factors which may be involved in branching morphogenesis at this stage, as well as known transcription factors involved in branching morphogenesis of other tissues, was carried out and their expression in PCa development investigated by RT-qPCR. These analyses are described in Appendix 2. The importance of M / E interactions in prostate development has been well established through numerous tissue recombination experiments (1, 41-44). However, the molecular mechanisms involved in these M / E interactions remain unclear. There is abundant evidence that secreted extracellular molecules functioning through paracrine mechanisms play an important role. Members of the F G F (45, 48-51), TGFp (52-54), BMP4 (61) and SHH (62, 63, 65, 66) signaling pathways have all been implicated in the M / E interactions involved in prostate development. We thus searched our prostate libraries, as well as 79 other mouse LongSAGE libraries (www.mouseatlas.org), for expression of various members of these pathways. Components of all of them were detected in the prostate libraries (data not shown). These analyses also revealed that the  109  components of these pathways were relatively evenly distributed throughout the libraries, highlighting their rather ubiquitous nature throughout development.  4.2  SFRP2 and the WNT Pathway in Prostate Development The W N T pathway has been implicated in M / E interactions during uterus  development (316) as well as in branching morphogenesis of the mammary gland (317), kidney (251), and lung (318). Intriguingly, members of the WNT signaling pathway were also expressed in the prostate libraries, including Sfrp2, Wnt4, Wnt6, Wntll, Fzdl, Fzd7, and DkkS, etc. Surprisingly, although this pathway has been implicated in PCa (261-265), it has not, to our knowledge, been studied in prostate development, despite the fact that we find ample evidence of the expression of numerous components of this pathway in our libraries. Some W N T pathway members detected in these SAGE libraries were further validated using RT-qPCR. Successful amplification of these W N T pathway members confirmed that they are differentially expressed during prostate development. Although Lrp5 and Axin, genes encoding two essential members of the canonical pathway, were not detected in the analyzed libraries, they were detected using RT-qPCR. Not all members of the WNT pathway examined showed the same expression pattern, i.e. highly expressed during early prostate development and downregulated in the adult. A possible reason is that RT-qPCR levels do not always reflect corresponding protein levels because post-transcriptional regulation may be involved, thus even i f no changes in R N A levels were detected, changes in protein levels might occur.  110  It should be noted that both SAGE and RT-qPCR were performed with bulk tissues which contain different cell populations, including both epithelial cells such as columnar and basal cells as well as stromal cells like fibroblasts and smooth muscle cells. The proportion of these cell populations may change during development. Thus differential expression of candidate genes during development may simply be caused by changes in the proportion of cells expressing them in the tissue. We observed expression of transcripts of P-catenin, the downstream member of the canonical WNT pathway, as well as of Wntl 1 and RhoA, members of the noncanonical WNT pathway, in our SAGE libraries and their expression was confirmed by RT-qPCR, suggesting that both the canonical and non-canonical WNT pathways may be involved in prostate development. In addition, we performed immunofluorescent staining to detect Pcatenin protein in the developing prostate. P-catenin is not only a molecule of the canonical WNT pathway, but also a component of the cadherin complex, which controls cell-cell adhesion and influences cell migration (319). Although P-catenin was detected in the cytoplasm attached to membrane (non-active P-catenin), we could not detect nuclear P-catenin (active P-catenin) staining in the developing prostate. This can be explained as follows: first, the canonical WNT pathway is not active in prostate development. This could be caused by the antagonizing effect of SFRP2. Secondly, the canonical WNT pathway does function in the developing prostate, but the level of nuclear P-catenin may be too low to detect. Sfrp2 was differentially expressed in the male and female UGS libraries as well as in the developing prostate libraries (i.e. El6.5 male UGS and PO prostate) and the adult DLP library, suggesting that it may be involved in prostate development. We further  111  investigated its expression in other SAGE libraries using Digital Northern. Analysis of the mouse SAGE libraries revealed high levels of Sfrp2 transcripts in the retina and telencephalon, confirming previous work (286, 307, 308). The relatively high levels of Sfrpl in the male UGS library compared to all other 149 mouse S A G E libraries in the database suggest that it may play a role in prostate development. The Digital Northern analysis of Sfrpl expression in the human SAGE libraries was somewhat different from the mouse SAGE libraries. In the human normal prostate sample 30 tags per 200,000 were detected while in mouse there were only 2 tags per 200,000. This could be due to species differences. For example, human prostate contains more stroma than mouse. If Sfrpl is expressed at higher levels in the mesenchyme/stroma than in the epithelium as was shown in the mouse UGS, it might explain why Sfrpl is relatively more highly expressed in the human prostate. In addition, we only dissected the D L P from the mouse. It is possible that other lobes may have a higher expression i.e., more similar to the human. Finally, our mouse SAGE libraries were constructed from pooled samples which minimize individual variation. The samples used for human S A G E libraries, however, are not likely pooled. Thus the Digital Northern results of human might reflect individual variation. In addition, no validation has been done on human S A G E libraries. The expression pattern of candidate genes in the human S A G E library needs to be further validated through experiments such as RT-qPCR. Both Sfrpl mRNA and its protein were expressed at higher levels in the El6.5 male UGS and P0 prostate than in the adult prostate, suggesting it may play an important role in the early stages of prostate development. To date the functions of the various SFRPs are not fully understood. In situ hybridization and immunofluorescent staining results  112  showed that Sfrp2 transcript and its protein were expressed in both mesenchyme and epithelium, correlating with the results of RT-qPCR and SAGE. The mesenchymal Sfrp2 expression pattern was particularly intriguing. In the urethra, Sfrp2 mRNA was expressed predominantly in the mesenchyme immediately surrounding the U G E while in the upper UGS its expression in the mesenchyme adjacent to the epithelium was significantly reduced. Immunostaining showed a similar localization of SFRP2 protein. It is possible to speculate on several reasons for, and consequences of, this expression pattern. First, SFRP2 may be involved in tissue remodeling during budding and branching morphogenesis of the developing prostate. SFRP2 contains a netrin-like domain which is also present in the TIMPs. In the TIMPs, the netrin domain appears to function in the inhibitory activity of these molecules against extracellular matrix metalloproteinases (MMPs). Thus the coupling of the F Z D and netrin-like domains in the SFRP family raises the possibility that their interactions with WNTs and the WNT signaling network may involve matrix-stabilizing activities (275). In fact, Sfrp2 overexpression was shown to inhibit glioma cell motility, and is associated with downregulation of M M P - 2 activity (286). MMPs are involved in cell invasion and in embryonic development and organogenesis (320). During rat prostate development M M P - 2 expression is upregulated with the branching morphogenesis process (321). Thus, SFRP2 expression may prohibit epithelial budding in the urethra while its expression in the upper U G S may be downregulated to allow the prostate bud to grow out. Secondly, SFRP2 was also expressed in the U G E and nascent epithelial buds at lower levels than in the U G M . If SFRP2 inhibits branching morphogenesis of the prostate, it is difficult to explain its expression in the epithelium. One possibility is that epithelial and  113  mesenchymal SFRP2 play different roles in the developing prostate. For example, SFRP2 rescues the branching-blocking effects of SFRP1 in the metanephros and may promote tubule formation by permitting WNT4 signaling in the developing kidney (279). This suggests that epithelial SFRP2 may in fact promote epithelial budding and branching. SFRP2 may also play a role in inducing epithelial differentiation. It has been shown to promote stem cell differentiation into neural progenitors while WNT1 inhibits this process (322), raising the possibility that expression of SFRP2 in the developing prostate epithelium may promote the differentiation of epithelial cells while W N T may inhibit differentiation. SFRPs have been proposed to antagonize the WNT pathway (273). WNT3a, WNT4 and WNT8 have been shown to interact with SFRP2 (277, 309, 310). To investigate if these ligands interact with SFRP2 during prostate development, we performed RT-qPCR to detect expression of these ligands. Only Wnt4 transcripts were detected in the developing prostate. The expression of Sfrp2 mRNA has been shown to be regulated by WNT4 (277). In turn, SFRP2 modulates WNT4 signaling in renal organogenesis (277). However, the expression pattern of Wnt4 did not correlate with that of Sfrp2 during prostate development. This could be caused by differences in tissue-specificity of the WNT signaling pathway. Thus, Sfrp2 may also be regulated by other molecules or pathways in prostate development. It is known that WNTs also function as long-range morphogenetic signals that can act on distant neighbors (323) and are involved in M / E interactions (316, 318). It is therefore, possible that WNT4 acts on the surrounding mesenchymal cells to induce differentiation of the U G M . In this model SFRP2 might act to limit the spread of WNT4's influence.  114  Examination of the expression pattern of the F Z D receptors (i.e. FZD3 (324), FZD6 (325)) that bind WNT4 will clarify this possibility. Wnt4 Tg and K O mice have been generated. Overexpression of human Wnt4 in mice disrupts testicular vasculature and testosterone synthesis (326). The morphology of the SV was altered in these mice but a prostate phenotype was not described. Sexual development in males with a mutation in Wnt4 appears to be normal (327). Thus the prostate phenotype was either not noticed or these mice are normal and redundant WNT pathway members may play a compensatory role.  4.3  SFRP2 and the WNT Pathway in PCa The WNT pathway has been identified as one of the key signaling pathways in  cancer and it is involved in various cancers such as gastric, liver, ovarian and other cancers (260). The role of SFRP2 in cancer development is still not clear. SFRP2 has been shown to promote tumor growth (292), induce tumorous transformation and inhibit apoptosis (295). However, given the oncogenic potential of constitutive WNT signaling, SFRPs have also been postulated to act as tumor suppressor genes. Our Digital Northern analysis of Sfrp2 expression in the human SAGE libraries suggested that Sfrp2 is downregulated in PCa compared to normal prostate (Table 3.10). In contrast, using the T R A M P mouse model of PCa, we showed that Sfrp2 mRNA levels were increased in the 4 week old T R A M P mice compared to their WT littermates. The discrepancy between our T R A M P data and the human S A G E data could be caused by the difference between human and mouse PCa, the heterogeneity of the disease, or differences in the stages at which samples were collected. Since total prostate tissues were used in our study while  115  the human SAGE library used micro-dissected epithelial cells, it is possible tissue heterogeneity may also be a reason for the difference. Since no validation of human SAGE libraries is available at present, we can not rule out that the Digital Northern data of human Sfrp2 expression may reflect individual variations. Although T R A M P is a PCa model mimicking human cancer, there are differences between T R A M P and human PCa. There has been a concern about the high frequency of N E characteristics in the T R A M P model, which is relatively infrequent in humans (only about 10% of cases) (328). We did not detect differences in SFRP2 protein expression patterns in T R A M P and WT mice. Discrepancies between R N A and protein levels have been demonstrated before. In a study to identify androgen regulated genes using both SAGE and proteomics, the change in intensity for most of the affected proteins identified could not be predicted based on the level of their corresponding R N A (229). The upregulation of Sfrp2 mRNA, as well as other candidate genes tested, therefore, may simply be at transcriptional levels. An alternative explanation is that, although the protein level of SFRP2 in Tg mice did not appear to change, subtle differences might not be effectively detected and quantified by immunofluorescence staining. Although WNTs have been implicated in multiple cancers, the role of WNT4 in cancer has not been elucidated and data are only descriptive. For example, amplification of the Wnt4 gene was found in squamous non-small cell lung carcinoma (329). It was also abnormally expressed in human breast cancer cell lines (330). Some transcripts of the W N T ligands have been shown to be overexpressed during PCa development, e.g. Wntl (264), Wnt2 (331) and Wnt11 (332). WNT3a was shown to induce AR-mediated transcription and cell growth in a ligand-independent manner (263). We showed, for the  116  first time, that Wnt4 transcripts are upregulated in the DLP of 12 week old T R A M P mice. Due to the lack of a reliable antibody to detect endogenous WNT4, protein levels could not be determined. A role for WNT4 and SFRP2 in PCa development in the T R A M P model can be inferred by their expression patterns. It is possible that at early stages of PCa, SFRP2 is upregulated as a negative feedback regulator to inhibit cell proliferation caused by transgene expression, thus inhibiting WNT4 upregulation at this stage. At a later stage, e.g. 12 weeks, SFRP2 was downregulated thus allowing progression to occur. A n alternative explanation is that SFRP2 and WNT4 play synergistic roles in PCa development, i.e. upregulated SFRP2 facilitates the transformation of the epithelial cells and WNT4 promotes cancer progression. However, it should also be noted that the differential expression of transcripts could be caused by tissue heterogeneity between the WT and Tg mice, i.e. higher level of expression in Tg mice was due to increase in the proportion of epithelial cells expressing candidate genes instead of upregulation of the transcripts in each cell. This may explain why we did not detect changes in SFRP2 protein using immunofluorescent staining. We did not detect changes in either R N A or protein levels of P-catenin in T R A M P mice as old as 26 week of age compared to their WT littermates, indicating either that Pcatenin levels did not change or the changes were too subtle to be detected. A recently published paper showed a significant decrease in P-catenin protein levels in 28 week old T R A M P mice (333). One possible reason for this discrepancy could be due to the difference in experimental methods used: immunoblotting vs. immunofluorescence  117  staining. Immunofluorescence is not quantitative, while immunoblotting can be. Thus changes in fluorescence intensity may not be easily differentiated. WNT4 is involved in both the canonical (334) and non-canonical W N T pathways (325, 335). SFRPs have been shown to regulate non-canonical WNT signaling in cancer development (336). Since no changes in P-catenin transcript or protein levels were detected during PCa development in the T R A M P mice, it is likely that WNT4 and SFRP2 signal through the non-canonical W N T pathway. In addition, the involvement of other signaling molecules can not be ruled out. For example, Sfrp2 was shown to be upregulated by SHH (337) which is an important regulator of prostate development (65, 66). Overexpression of SHH dramatically accelerates prostate tumor growth (338). Although a role for SHH in the T R A M P mouse model of PCa has not been described, it would be interesting to investigate if SHH is expressed and regulates Sfrp2 expression. Taken together, we have provided preliminary evidence that transcripts of the WNT pathway members including SFRP2 are differentially expressed during prostate development and in a mouse model of PCa. This raises the possibility that they may play important roles in these processes. Further functional studies may provide new insights into the molecular mechanisms regulating these processes and help to identify new cancer markers and therapeutic targets.  4.4  Future Directions We have used SAGE to study mouse prostate development and identified WNT  pathway members in the developing prostate. The expression pattern of Sfrp2 and transcripts of other selected WNT pathway members have been determined. However,  118  these are only the first step and many questions remain. 1) What is the role of WNT4 and SFRP2 in mouse prostate development? Although Wnt4 Tg and K O mice have been generated, there is no report of a prostate phenotype. A detailed analysis of prostate tissue in these mice would be necessary. Similar functional studies, such as overexpression or K O , are required to determine the role of SFRP2 in prostate development. 2) SFRP2 was predominantly expressed in the developing mesenchyme. Thus, it will be interesting to see  if  inhibition  of  SFRP2  in  the  mesenchyme  alters  epithelial  cell  proliferation/differentiation. Tissue recombination and xenograft experiments can be used to test this hypothesis through combining Sfrp2 null mesenchyme with WT epithelium. 3) Does SFRP2 interact with WNT4 in prostate development? This question could be answered using coimmunoprecipitation techniques provided appropriate antibodies were developed. 4) What downstream genes are regulated by SFRP2 and WNT4? Further studies combining Sfrp2 or Wnt4 Tg or K O models with other gene expression profiling methods, e.g. microarray, and protein profiling methods, e.g. 2-DE and mass spectrometric (MS) analysis, can be carried out to identify these genes and pathways. 5) Does S H H regulate Sfrp2 expression in prostate development? How are SFRP2 and WNT4 linked to other factors involved in prostate development such as FGF10? To answer these questions, expression of SFRP2 and WNT4 in Shh or FgflO Tg or K O mice can be determined by RT-qPCR. 6) Are there other SFRPs involved in prostate development? SFRP1 was shown to antagonize SFRP2 in MCF-7 breast cancer cells and in kidney development (278, 279). We have also observed Sfrpl and Sfrp4 tags in the El6.5 male UGS SAGE library. Their expression can be validated by RT-qPCR and their function could also be determined. 7) There are 19 W N T ligands in mouse and we  119  examined only transcripts of 5 of them, including W n t l l and other 4 W N T ligands previously shown to interact with SFRP2. The expression patterns of other W N T members are still to be examined. 9) What are the specific FZD receptors interacting with SFRP2 and/or WNT4 in prostate development? Are they mesenchymal or epithelial? RTqPCR, in situ hybridization and immunofluorescence staining will inform us of their presence and localization. Coimmunoprecipitation can be carried out on isolated U G E or U G M to detect the interaction between SFRP2 and/or WNT4 and these FZD receptors i f antibodies were available. 10) Which WNT pathway functions in prostate developmentcanonical or noncanonical-or are both involved in this process? To answer this question, blocking of the canonical WNT pathway using methods such as R N A i targeting various components of the pathway, or expressing an excess of the ligand binding domain of the FZD receptor, will give us some hint (http://www.stanford.edu/~rnusse/assays/inhib. htm). However, to date an effective way to block the noncanonical WNT pathway has not been described. We have determined the expression of Sfrp2, B-catenin and Wnt4 in the prostate of T R A M P mice. The functions of these molecules, as well as other W N T pathway members, in PCa development require further studies. Since we did not detect changes in SFRP2 or P-catenin using immunofluorescent staining, a more quantitative technique, such as immunoblotting, must be used to quantify their protein levels. Questions which need to be answered are as follows: 1) Are Sfrp2, Wnt4 and B-catenin expression patterns the same in other PCa models, or are they T R A M P specific? A different cancer model more closely mimicking human PCa (i.e. Pten+I- mice), or human PCa samples can be used to elucidate the expression patterns of these molecules. 2) Do SFRP2 and WNT4  120  affect PCa growth? This can be tested using cell culture in vitro or xenograft models in vivo. 3) Are SFRP2 or WNT4 oncogenic or do they function as tumor suppressor proteins? Will overexpression or K O of Sfrp2 or Wnt4 cause PCa? To answer these questions, prostate specific Tg or K O Sfrp2 or Wnt4 mice can be generated to determine if they have prostate phenotypes. 4) Do changes in SFRP2 and/or WNT4 expression affect A R expression and its transcriptional activity? A R is known to play a critical role in PCa progression and crosstalk between some members of the W N T pathway and the A R has been demonstrated in PCa cells (266-271). The effect of SFRP2 or WNT4 on A R transcriptional activity will complement knowledge of the crosstalk between these two pathways. Taken together, such functional studies will not only help us to further our understanding of prostate development, but may also facilitate generation of novel PCa models and lead to identification of putative therapeutic targets.  121  REFERENCES 1.  2. 3.  4. 5. 6. 7. 8. 9.  10.  11. 12.  13.  14.  15. 16.  Cunha, G. R., H . Fujii, B . L . Neubauer, J. M . Shannon, L . Sawyer, and B. A . Reese. 1983. Epithelial-mesenchymal interactions in prostatic development. I. morphological observations of prostatic induction by urogenital sinus mesenchyme in epithelium of the adult rodent urinary bladder. J Cell Biol 96:1662. Staack, A . , A . A . Donjacour, J. Brody, G. R. Cunha, and P. Carroll. 2003. Mouse urogenital development: a practical approach. Differentiation 71:402. Cunha, G. R., A . A . Donjacour, P. S. Cooke, S. Mee, R. M . Bigsby, S. J. Higgins, and Y . Sugimura. 1987. The endocrinology and developmental biology of the prostate. Endocr Rev 8:338. Timms, B . G., T. J. Mohs, and L . J. Didio. 1994. Ductal budding and branching patterns in the developing prostate. J Urol 151:1427. Sugimura, Y . , G. R. Cunha, and A . A . Donjacour. 1986. Morphogenesis of ductal networks in the mouse prostate. Biol Reprod 34:961. Aumuller, G. 1991. Postnatal development of the prostate. Bull Assoc Anat (Nancy) 75:39. Marker, P. C , A . A . Donjacour, R. Dahiya, and G. R. Cunha. 2003. Hormonal, cellular, and molecular control of prostatic development. Dev Biol 253:165. Wang, Y . , S. Hayward, M . Cao, K . Thayer, and G. Cunha. 2001. Cell differentiation lineage in the prostate. Differentiation 68:270. Hayward, S. W., L . S. Baskin, P. C. Haughney, A . R. Cunha, B . A . Foster, R. Dahiya, G. S. Prins, and G. R. Cunha. 1996. Epithelial development in the rat ventral prostate, anterior prostate and seminal vesicle. Acta Anat (Basel) 155:81. Hayward, S. W., L . S. Baskin, P. C. Haughney, B . A . Foster, A . R. Cunha, R. Dahiya, G. S. Prins, and G. R. Cunha. 1996. Stromal development in the ventral prostate, anterior prostate and seminal vesicle of the rat. Acta Anat (Basel) 155:94. Donjacour, A . A., and G. R. Cunha. 1988. The effect of androgen deprivation on branching morphogenesis in the mouse prostate. Dev Biol 128:1. Xue, Y . , F. Smedts, A . Verhofstad, F. Debruyne, J. de la Rosette, and J. Schalken. 1998. Cell kinetics of prostate exocrine and neuroendocrine epithelium and their differential interrelationship: new perspectives. Prostate Suppl 8:62. Pointis, G., M . T. Latreille, T. M . Mignot, Y . Janssens, and L . Cedard. 1979. Regulation of testosterone synthesis in the fetal mouse testis. J Steroid Biochem 11:1609. Cunha, G. R. 1975. Age-dependent loss of sensitivity of female urogenital sinus to androgenic conditions as a function of the epithelia-stromal interaction in mice. Endocrinology 97:665. Takeda, H., I. Lasnitzki, and T. Mizuno. 1986. Analysis of prostatic bud induction by brief androgen treatment in the fetal rat urogenital sinus. J Endocrinol 110:467. Brown, T. R., D. B. Lubahn, E. M . Wilson, D. R. Joseph, F. S. French, and C. J. Migeon. 1988. Deletion of the steroid-binding domain of the human androgen receptor gene in one family with complete androgen insensitivity syndrome:  122  17.  18.  19.  20.  21. 22.  23.  24.  25. 26.  27. 28. 29.  30.  evidence for further genetic heterogeneity in this syndrome. Proc Natl Acad Sci U S A 85:8151. Charest, N . J., Z. X . Zhou, D. B. Lubahn, K . L . Olsen, E. M . Wilson, and F. S. French. 1991. A frameshift mutation destabilizes androgen receptor messenger R N A in the Tfm mouse. Mol Endocrinol 5:573. Gaspar, M . L . , T. Meo, P. Bourgarel, J. L . Guenet, and M . Tosi. 1991. A single base deletion in the Tfm androgen receptor gene creates a short-lived messenger R N A that directs internal translation initiation. Proc Natl Acad Sci U S A 88:8606. He, W. W., M . V . Kumar, and D. J. Tindall. 1991. A frame-shift mutation in the androgen receptor gene causes complete androgen insensitivity in the testicularfeminized mouse. Nucleic Acids Res 19:2373. Deslypere, J. P., M . Young, J. D. Wilson, and M . J. McPhaul. 1992. Testosterone and 5 alpha-dihydrotestosterone interact differently with the androgen receptor to enhance transcription of the M M T V - C A T reporter gene. M o l Cell Endocrinol 88:15. Russell, D. W., and J. D. Wilson. 1994. Steroid 5 alpha-reductase: two genes/two enzymes. Annu Rev Biochem 63:25. Andersson, S., D. M . Berman, E. P. Jenkins, and D. W. Russell. 1991. Deletion of steroid 5 alpha-reductase 2 gene in male pseudohermaphroditism. Nature 354:159. Mahendroo, M . S., K . M . Cala, D. L . Hess, and D. W. Russell. 2001. Unexpected virilization in male mice lacking steroid 5 alpha-reductase enzymes. Endocrinology 142:4652. Cunha, G. R., J. M . Shannon, B . L . Neubauer, L . M . Sawyer, H . Fujii, O. Taguchi, and L . W. Chung. 1981. Mesenchymal-epithelial interactions in sex differentiation. Hum Genet 58:68. Cooke, P. S., P. Young, and G. R. Cunha. 1991. Androgen receptor expression in developing male reproductive organs. Endocrinology 128:2867. Donjacour, A . A . , and G. R. Cunha. 1993. Assessment of prostatic protein secretion in tissue recombinants made of urogenital sinus mesenchyme and urothelium from normal or androgen-insensitive mice. Endocrinology 132:2342. Hayward, S. W., and G. R. Cunha. 2000. The prostate: development and physiology. Radiol Clin North A m 38:1. Prins, G. S. 1992. Neonatal estrogen exposure induces lobe-specific alterations in adult rat prostate androgen receptor expression. Endocrinology 130:3703. Jarred, R. A . , B. Cancilla, G. S. Prins, K . A . Thayer, G . R. Cunha, and G . P. Risbridger. 2000. Evidence that estrogens directly alter androgen-regulated prostate development. Endocrinology 141:3471. . vom Saal, F. S., B . G. Timms, M . M . Montano, P. Palanza, K . A . Thayer, S. C. Nagel, M . D. Dhar, V . K . Ganjam, S. Parmigiani, and W. V . Welshons. 1997. Prostate enlargement in mice due to fetal exposure to low doses of estradiol or diethylstilbestrol and opposite effects at high doses. Proc Natl Acad Sci U S A 94:2056.  123  31.  32.  33.  34.  35.  36.  37.  38.  39. 40.  41.  42.  43.  44.  Jarred, R. A . , S. J. McPherson, J. J. Bianco, J. F. Couse, K . S. Korach, and G. P. Risbridger. 2002. Prostate phenotypes in estrogen-modulated transgenic mice. Trends Endocrinol Metab 13:163. Reiter, E., B . Hennuy, M . Bruyninx, A . Cornet, M . Klug, M . McNamara, J. Closset, and G. Hennen. 1999. Effects of pituitary hormones on the prostate. Prostate 38:159. Wennbo, H., J. Kindblom, O. G. Isaksson, and J. Tornell. 1997. Transgenic mice overexpressing the prolactin gene develop dramatic enlargement of the prostate gland. Endocrinology 138:4410. Steger, R. W., V . Chandrashekar, W. Zhao, A . Bartke, and N . D . Horseman. 1998. Neuroendocrine and reproductive functions in male mice with targeted disruption of the prolactin gene. Endocrinology 139:3691. Reiter, E., O. Kecha, B . Hennuy, S. Lardinois, M . Klug, M . Bruyninx, J. Closset, and G. Hennen. 1995. Growth hormone directly affects the function of the different lobes of the rat prostate. Endocrinology 136:3338. Ruan, W., L . Powell-Braxton, J. J. Kopchick, and D. L . Kleinberg. 1999. Evidence that insulin-like growth factor I and growth hormone are required for prostate gland development. Endocrinology 140:1984. Demayo, F., P. Minoo, C. G. Plopper, L . Schuger, J. Shannon, and J. S. Torday. 2002. Mesenchymal-epithelial interactions in lung development and repair: are modeling and remodeling the same process? A m J Physiol Lung Cell Mol Physiol 283:L510. Madsen, O. D., J. Jensen, N . Blume, H . V . Petersen, K . Lund, C. Karlsen, F. G . Andersen, P. B . Jensen, L. I. Larsson, and P. Serup. 1996. Pancreatic development and maturation of the islet B cell. Studies of pluripotent islet cultures. Eur J Biochem 242:435. Tucker, A. S., and P. T. Sharpe. 1999. Molecular genetics of tooth morphogenesis and patterning: the right shape in the right place. J Dent Res 78:826. Kanwar, Y . S., J. Wada, S. Lin, F. R. Danesh, S. S. Chugh, Q. Yang, T. Banerjee, and J. W. Lomasney. 2004. Update of extracellular matrix, its receptors, and cell adhesion molecules in mammalian nephrogenesis. A m J Physiol Renal Physiol 286:F202. Timms, B . G., C. W. Lee, G. Aumuller, and J. Seitz. 1995. Instructive induction of prostate growth and differentiation by a defined urogenital sinus mesenchyme. Microsc Res Tech 30:319. Hayashi, N . , G. R. Cunha, and M . Parker. 1993. Permissive and instructive induction of adult rodent prostatic epithelium by heterotypic urogenital sinus mesenchyme. Epithelial Cell Biol 2:66. Cunha, G. R., M . Sekkingstad, and B. A . Meloy. 1983. Heterospecific induction of prostatic development in tissue recombinants prepared with mouse, rat, rabbit and human tissues. Differentiation 24:174. Hayward, S. W., P. C. Haughney, M . A . Rosen, K . M . Greulich, H . U . Weier, R. Dahiya, and G. R. Cunha. 1998. Interactions between adult human prostatic epithelium and rat urogenital sinus mesenchyme in a tissue recombination model. Differentiation 63:131.  124  45.  46.  47. 48.  49. 50. 51. 52.  53.  54.  55.  56.  57.  58.  Sugimura, Y . , B. A . Foster, Y . K . Horn, J. H . Lipschutz, J. S. Rubin, P. W. Finch, S. A . Aaronson, N . Hayashi, J. Kawamura, and G. R. Cunha. 1996. Keratinocyte growth factor (KGF) can replace testosterone in the ductal branching morphogenesis of the rat ventral prostate. Int J Dev Biol 40:941. Thomson, A . A., B. A . Foster, and G. R. Cunha. 1997. Analysis of growth factor and receptor mRNA levels during development of the rat seminal vesicle and prostate. Development 124:2431. Thomson, A . A . 2001. Role of androgens and fibroblast growth factors in prostatic development. Reproduction 121:187. Kitsberg, D. I., and P. Leder. 1996. Keratinocyte growth factor induces mammary and prostatic hyperplasia and mammary adenocarcinoma in transgenic mice. Oncogene 13:2507. Guo, L . , L . Degenstein, and E. Fuchs. 1996. Keratinocyte growth factor is required for hair development but not for wound healing. Genes Dev 10:165. Thomson, A . A . , and G. R. Cunha. 1999. Prostatic growth and development are regulated by FGF10. Development 126:3693. Donjacour, A . A . , A . A . Thomson, and G . R. Cunha. 2003. FGF-10 plays an essential role in the growth of the fetal prostate. Dev Biol 261:39. Timme, T. L . , G. Yang, L . D. Truong, D. Kadmon, S. H . Park, and T. C. Thompson. 1995. Transforming growth factor-beta localization during mouse prostate morphogenesis and in prostatic growth abnormalities. World J Urol 13:324. Timme, T. L., L . D. Truong, V . W. Merz, T. Krebs, D. Kadmon, K . C. Flanders, S. H . Park, and T. C. Thompson. 1994. Mesenchymal-epithelial interactions and transforming growth factor-beta expression during mouse prostate morphogenesis. Endocrinology 134:1039. Itoh, N . , U . Patel, A . S. Cupp, and M . K . Skinner. 1998. Developmental and hormonal regulation of transforming growth factor-beta 1 (TGFbetal), -2, and -3 gene expression in isolated prostatic epithelial and stromal cells: epidermal growth factor and TGFbeta interactions. Endocrinology 139:1378. Shull, M . M . , I. Ormsby, A . B. Kier, S. Pawlowski, R. J. Diebold, M . Yin, R. Allen, C. Sidman, G. Proetzel, D. Calvin, and et al. 1992. Targeted disruption of the mouse transforming growth factor-beta 1 gene results in multifocal inflammatory disease. Nature 359:693. Pierce, D. F., Jr., M . D. Johnson, Y . Matsui, S. D. Robinson, L . I. Gold, A . F. Purchio, C. W. Daniel, B. L . Hogan, and H . L . Moses. 1993. Inhibition of mammary duct development but not alveolar outgrowth during pregnancy in transgenic mice expressing active TGF-beta 1. Genes Dev 7:2308. Kundu, S. D., I. Y . Kim, T. Yang, L . Doglio, S. Lang, X . Zhang, R. Buttyan, S. J. Kim, J. Chang, X . Cai, Z. Wang, and C. Lee. 2000. Absence of proximal duct apoptosis in the ventral prostate of transgenic mice carrying the C3(l)-TGF-beta type II dominant negative receptor. Prostate 43:118. Itoh, N . , U . Patel, and M . K . Skinner. 1998. Developmental and hormonal regulation of transforming growth factor-alpha and epidermal growth factor receptor gene expression in isolated prostatic epithelial and stromal cells. Endocrinology 139:1369.  125  59.  60.  61.  62.  63.  64.  65.  66.  67.  68. 69.  70.  71.  72.  Haughney, P. C , S. W. Hayward, R. Dahiya, and G. R. Cunha. 1998. Speciesspecific detection of growth factor gene expression in developing murine prostatic tissue. Biol Reprod 59:93. Baker, J., M . P. Hardy, J. Zhou, C. Bondy, F. Lupu, A . R. Bellve, and A . Efstratiadis. 1996. Effects of an Igfl gene null mutation on mouse reproduction. Mol Endocrinol 10:903. Lamm, M . L., C. A . Podlasek, D. H . Barnett, J. Lee, J. Q. Clemens, C. M . Hebner, and W. Bushman. 2001. Mesenchymal factor bone morphogenetic protein 4 restricts ductal budding and branching morphogenesis in the developing prostate. Dev Biol 232:301. Podlasek, C. A., D. H . Barnett, J. Q. Clemens, P. M . Bak, and W. Bushman. 1999. Prostate development requires Sonic hedgehog expressed by the urogenital sinus epithelium. Dev Biol 209:28. Lamm, M . L . , W. S. Catbagan, R. J. Laciak, D. H . Barnett, C. M . Hebner, W. Gaffield, D. Walterhouse, P. Iannaccone, and W. Bushman. 2002. Sonic hedgehog activates mesenchymal GUI expression during prostate ductal bud formation. Dev Biol 249:349. Wang, B . E., J. Shou, S. Ross, H . Koeppen, F. J. De Sauvage, and W. Q. Gao. 2003. Inhibition of epithelial ductal branching in the prostate by sonic hedgehog is indirectly mediated by stromal cells. J Biol Chem 278:18506. Berman, D. M . , N . Desai, X . Wang, S. S. Karhadkar, M . Reynon, C. Abate-Shen, P. A . Beachy, and M . M . Shen. 2004. Roles for Hedgehog signaling in androgen production and prostate ductal morphogenesis. Dev Biol 267:387. Freestone, S. H . , P. Marker, O. C. Grace, D. C. Tomlinson, G. R. Cunha, P. Harnden, and A . A . Thomson. 2003. Sonic hedgehog regulates prostatic growth and epithelial differentiation. Dev Biol 264:352. Pu, Y . , L . Huang, and G. S. Prins. 2004. Sonic hedgehog-patched Gli signaling in the developing rat prostate gland: lobe-specific suppression by neonatal estrogens reduces ductal growth and branching. Dev Biol 273:257. Ball, E. M . , and G. P. Risbridger. 2001. Activins as regulators of branching morphogenesis. Dev Biol 238:1. Zhao, H . , A . Patra, Y . Tanaka, L . C. L i , and R. Dahiya. 2002. Transforming growth factor-beta(s) and their receptors in aging rat prostate. Biochem Biophys Res Commun 294:464. Cancilla, B., R. A . Jarred, H . Wang, S. L . Mellor, G. R. Cunha, and G. P. Risbridger. 2001. Regulation of prostate branching morphogenesis by activin A and follistatin. Dev Biol 237:145. Shou, J., S. Ross, H . Koeppen, F. J. de Sauvage, and W. Q. Gao. 2001. Dynamics of notch expression during murine prostate development and tumorigenesis. Cancer Res 61:7291. Wang, X . D., J. Shou, P. Wong, D. M . French, and W. Q. Gao. 2004. Notchlexpressing cells are indispensable for prostatic branching morphogenesis during development and re-growth following castration and androgen replacement. J Biol Chem 279:24733.  126  73.  74.  75.  76.  77.  78. 79.  80. 81. 82. 83.  84.  85.  86.  87.  88.  Gakunga, P., G. Frost, S. Shuster, G. Cunha, B . Formby, and R. Stern. 1997. Hyaluronan is a prerequisite for ductal branching morphogenesis. Development 124:3987. Marker, P. C , J. P. Stephan, J. Lee, L. Bald, J. P. Mather, and G. R. Cunha. 2001. fucosyltransferasel and H-type complex carbohydrates modulate epithelial cell proliferation during prostatic branching morphogenesis. Dev Biol 233:95. Elfman, F., R. Bok, M . Conn, M . Shuman, and G. Cunha. 2001. Urokinase plasminogen activator amino-terminal peptides inhibit development of the rat ventral prostate. Differentiation 69:108. Oefelein, M . , C. Chin-Chance, and W. Bushman. 1996. Expression of the homeotic gene F£ox-dl3 in the developing and adult mouse prostate. J Urol 155:342. Economides, K . D., and M . R. Capecchi. 2003. Hoxbl3 is required for normal differentiation and secretory function of the ventral prostate. Development 130:2061. Podlasek, C. A., J. Q. Clemens, and W. Bushman. 1999. Hoxa-13 gene mutation results in abnormal seminal vesicle and prostate development. J Urol 161:1655. Warot, X . , C. Fromental-Ramain, V . Fraulob, P. Chambon, and P. Dolle. 1997. Gene dosage-dependent effects of the Hoxa-13 and Hoxd-13 mutations on morphogenesis of the terminal parts of the digestive and urogenital tracts. Development 124:4781. Podlasek, C. A., D. Duboule, and W. Bushman. 1997. Male accessory sex organ morphogenesis is altered by loss of function of Hoxd-13. Dev Dyn 208:454. Zeltser, L . , C. Desplan, and N . Heintz. 1996. Hoxb-13: a new Hox gene in a distant region of the H O X B cluster maintains colinearity. Development 122:2475. Sreenath, T., A . Orosz, K . Fujita, and C. J. Bieberich. 1999. Androgenindependent expression of hoxb-13 in the mouse prostate. Prostate 41:203. Podlasek, C. A., R. M . Seo, J. Q. Clemens, L . M a , R. L . Maas, and W. Bushman. 1999. Hoxa-10 deficient male mice exhibit abnormal development of the accessory sex organs. Dev Dyn 214:1. Bieberich, C. J., K . Fujita, W. W. He, and G. Jay. 1996. Prostate-specific and androgen-dependent expression of a novel homeobox gene. J Biol Chem 271:31779. Sciavolino, P. J., E. W. Abrams, L . Yang, L . P. Austenberg, M . M . Shen, and C. Abate-Shen. 1997. Tissue-specific expression of murine Nkx3.1 in the male urogenital system. Dev Dyn 209:127. Bhatia-Gaur, R., A . A . Donjacour, P. J. Sciavolino, M . Kim, N . Desai, P. Young, C. R. Norton, T. Gridley, R. D. Cardiff, G. R. Cunha, C. Abate-Shen, and M . M . Shen. 1999. Roles for Nkx3.1 in prostate development and cancer. Genes Dev 13:966. Abdulkadir, S. A . , J. A . Magee, T. J. Peters, Z. Kaleem, C. K . Naughton, P. A . Humphrey, and J. Milbrandt. 2002. Conditional loss of Nkx3.1 in adult mice induces prostatic intraepithelial neoplasia. Mol Cell Biol 22:1495. Kim, M . J., R. D. Cardiff, N . Desai, W. A . Banach-Petrosky, R. Parsons, M . M . Shen, and C. Abate-Shen. 2002. Cooperativity of Nkx3.1 and Pten loss of  127  89.  90. 91.  92.  93. 94. 95. 96.  97. 98. 99. 100.  101.  102. 103.  104.  function in a mouse model of prostate carcinogenesis. Proc Natl Acad Sci U S A 99:2884. Signoretti, S., D. Waltregny, J. Dilks, B . Isaac, D. Lin, L . Garraway, A . Yang, R. Montironi, F. McKeon, and M . Loda. 2000. p63 is a prostate basal cell marker and is required for prostate development. A m J Pathol 157:1769. Kurita, T., R. T. Medina, A . A . Mills, and G. R. Cunha. 2004. Role of p63 and basal cells in the prostate. Development 131:4955. Lee, C , Y . Tsai, H . H . Harrison, and J. Sensibar. 1985. Proteins of the rat prostate: I. Preliminary characterization by two-dimensional electrophoresis. Prostate 7:171. Collins, A . T., F. K . Habib, N . J. Maitland, and D. E. Neal. 2001. Identification and isolation of human prostate epithelial stem cells based on alpha(2)beta(l)integrin expression. J Cell Sci 114:3865. Aumuller, G. 1983. Morphologic and endocrine aspects of prostatic function. Prostate 4:195. Eder, I. E., J. Bektic, P. Haag, G. Bartsch, and H . Klocker. 2004. Genes differentially expressed in prostate cancer. B J U Int 93:1151. Cher, M . L . 2001. Mechanisms governing bone metastasis in prostate cancer. Curr OpinUrol 11:483. Carter, B . S., G. S. Bova, T. H . Beaty, G. D. Steinberg, B . Childs, W. B . Isaacs, and P. C. Walsh. 1993. Hereditary prostate cancer: epidemiologic and clinical features. J Urol 150:797. Dong, J. T. 2001. Chromosomal deletions and rumor suppressor genes in prostate cancer. Cancer Metastasis Rev 20:173. Karayi, M . K . , and A . F. Markham. 2004. Molecular biology of prostate cancer. Prostate Cancer Prostatic Dis 7:6. Porkka, K . P., and T. Visakorpi. 2004. Molecular mechanisms of prostate cancer. Eur Urol 45:683. Nwosu, V . , J. Carpten, J. M . Trent, and R. Sheridan. 2001. Heterogeneity of genetic alterations in prostate cancer: evidence of the complex nature of the disease. Hum Mol Genet 10:2313. Nakayama, M . , C. J. Bennett, J. L . Hicks, J. I. Epstein, E. A . Platz, W. G. Nelson, and A . M . De Marzo. 2003. Hypermethylation of the human glutathione Stransferase-pi gene (GSTP1) CpG island is present in a subset of proliferative inflammatory atrophy lesions but not in normal or hyperplastic epithelium of the prostate: a detailed study using laser-capture microdissection. A m J Pathol 163:923. Carter, B . S., J. I. Epstein, and W. B . Isaacs. 1990. ras gene mutations in human prostate cancer. Cancer Res 50:6830. Buttyan, R., I. S. Sawczuk, M . C. Benson, J. D. Siegal, and C. A . Olsson. 1987. Enhanced expression of the c-myc protooncogene in high-grade human prostate cancers. Prostate 11:327. Fox, S. B., R. A . Persad, J. Royds, R. N . Kore, P. B . Silcocks, and C. C. Collins. 1993. p53 and c-myc expression in stage A l prostatic adenocarcinoma: useful prognostic determinants? J Urol 150:490.  128  105.  106.  107.  108.  109.  110.  111.  112.  113.  114.  115.  116.  117.  118.  Jorda, M . , A . Morales, Z. Ghorab, G. Fernandez, M . Nadji, and N . Block. 2002. Her2 expression in prostatic cancer: a comparison with mammary carcinoma. J Urol 168:1412. Varambally, S., S. M . Dhanasekaran, M . Zhou, T. R. Barrette, C. Kumar-Sinha, M . G. Sanda, D. Ghosh, K . J. Pienta, R. G. Sewalt, A . P. Otte, M . A . Rubin, and A. M . Chinnaiyan. 2002. The polycomb group protein EZH2 is involved in progression of prostate cancer. Nature 419:624. Deutsch, E., L . Maggiorella, P. Eschwege, J. Bourhis, J. C. Soria, and B . Abdulkarim. 2004. Environmental, genetic, and molecular features of prostate cancer. Lancet Oncol 5:303. Bookstein, R., P. Rio, S. A . Madreperla, F. Hong, C. Allred, W. E. Grizzle, and W. H . Lee. 1990. Promoter deletion and loss of retinoblastoma gene expression in human prostate carcinoma. Proc Natl Acad Sci U S A 87:7762. Bookstein, R., J. Y . Shew, P. L . Chen, P. Scully, and W. H . Lee. 1990. Suppression of tumorigenicity of human prostate carcinoma cells by replacing a mutated RB gene. Science 247:712. Isaacs, W. B., B. S. Carter, and C. M . Ewing. 1991. Wild-type p53 suppresses growth of human prostate cancer cells containing mutant p53 alleles. Cancer Res 51:4716. Cairns, P., K . Okami, S. Halachmi, N . Halachmi, M . Esteller, J. G. Herman, J. Jen, W. B . Isaacs, G. S. Bova, and D. Sidransky. 1997. Frequent inactivation of P T E N / M M A C 1 in primary prostate cancer. Cancer Res 57:4997. Kwabi-Addo, B., D. Giri, K. Schmidt, K. Podsypanina, R. Parsons, N . Greenberg, and M . Ittmann. 2001. Haploinsufficiency of the Pten tumor suppressor gene promotes prostate cancer progression. Proc Natl Acad Sci U S A 98:11563. Bruchovsky, N . , P. S. Rennie, A . J. Coldman, S. L . Goldenberg, M . To, and D. Lawson. 1990. Effects of androgen withdrawal on the stem cell composition of the Shionogi carcinoma. Cancer Res 50:2275. Pisters, L . L . , X . H . Wang, L . W. Chung, and J. T. Hsieh. 2001. Differential expression of c-erb B2/neu, epidermal growth factor receptor, cytokeratin 8, and the prostatic steroid-binding protein gene in rat ventral prostate during postnatal development. Prostate 47:164. Huggins, C , and C. V . Hodges. 1972. Studies on prostatic cancer. I. The effect of castration, of estrogen and androgen injection on serum phosphatases in metastatic carcinoma of the prostate. C A Cancer J Clin 22:232. Amler, L. C , D. B. Agus, C. LeDuc, M . L . Sapinoso, W. D. Fox, S. Kern, D. Lee, V . Wang, M . Leysens, B . Higgins, J. Martin, W. Gerald, N . Dracopoli, C. Cordon-Cardo, H . I. Scher, and G. M . Hampton. 2000. Dysregulated expression of androgen-responsive and nonresponsive genes in the androgen-independent prostate cancer xenograft model CWR22-R1. Cancer Res 60:6134. Marcelli, M . , M . Ittmann, S. Mariani, R. Sutherland, R. Nigam, L . Murthy, Y . Zhao, D. DiConcini, E. Puxeddu, A . Esen, J. Eastham, N . L . Weigel, and D. J. Lamb. 2000. Androgen receptor mutations in prostate cancer. Cancer Res 60:944. Abate-Shen, C , and M . M . Shen. 2000. Molecular genetics of prostate cancer. Genes Dev 14:2410.  129  119.  120. 121.  122. 123.  124.  125. 126. 127.  128.  129. 130.  131.  132. 133.  134.  Sadar, M . D., M . Hussain, and N . Bruchovsky. 1999. Prostate cancer: molecular biology of early progression to androgen independence. Endocr Relat Cancer 6:487. Ghosh, P. M . , S. Malik, R. Bedolla, and J. I. Kreisberg. 2003. Akt in prostate cancer: possible role in androgen-independence. Curr Drug Metab 4:487. Zhu, X., and J. P. Liu. 1997. Steroid-independent activation of androgen receptor in androgen-independent prostate cancer: a possible role for the M A P kinase signal transduction pathway? Mol Cell Endocrinol 134:9. Gioeli, D. 2004. Signal transduction in prostrate cancer progression. Clin Sci (Lond). 108:293 McDonnell, T. J., P. Troncoso, S. M . Brisbay, C. Logothetis, L . W. Chung, J. T. Hsieh, S. M . Tu, and M . L . Campbell. 1992. Expression of the protooncogene bcl2 in the prostate and its association with emergence of androgen-independent prostate cancer. Cancer Res 52:6940. Raffo, A . J., H . Perlman, M . W. Chen, M . L . Day, J. S. Streitman, and R. Buttyan. 1995. Overexpression of bcl-2 protects prostate cancer cells from apoptosis in vitro and confers resistance to androgen depletion in vivo. Cancer Res 55:4438. Wong, Y . C , and N . N . Tam. 2002. Dedifferentiation of stromal smooth muscle as a factor in prostate carcinogenesis. Differentiation 70:633. Cunha, G. R., S. W. Hayward, and Y . Z. Wang. 2002. Role of stroma in carcinogenesis of the prostate. Differentiation 70:473. Gao, J., J. T. Arnold, and J. T. Isaacs. 2001. Conversion from a paracrine to an autocrine mechanism of androgen-stimulated growth during malignant transformation of prostatic epithelial cells. Cancer Res 61:5038. Wang, Y . Z., and Y . C. Wong. 1998. Sex hormone-induced prostatic carcinogenesis in the noble rat: the role of insulin-like growth factor-I (IGF-I) and vascular endothelial growth factor (VEGF) in the development of prostate cancer. Prostate 35:165. van Weerden, W. M . , and J. C. Romijn. 2000. Use of nude mouse xenograft models in prostate cancer research. Prostate 43:263. Wang, Y . , M . P. Revelo, D. Sudilovsky, M . Cao, W. G. Chen, L . Goetz, H . Xue, M . Sadar, S. B. Shappell, G. R. Cunha, and S. W. Hayward. 2005. Development and characterization of efficient xenograft models for benign and malignant human prostate tissue. Prostate. 64:149 Podsypanina, K., L. H . Ellenson, A. Nemes, J. Gu, M . Tamura, K . M . Yamada, C. Cordon-Cardo, G. Catoretti, P. E. Fisher, and R. Parsons. 1999. Mutation of Pten/Mmacl in mice causes neoplasia in multiple organ systems. Proc Natl Acad Sci U S A 96:1563. D i Cristofano, A . , B . Pesce, C. Cordon-Cardo, and P. P. Pandolfi. 1998. Pten is essential for embryonic development and tumour suppression. Nat Genet 19:348. Huang, J., W. C. Powell, A. C. Khodavirdi, J. Wu, T. Makita, R. D. Cardiff, M . B . Cohen, H . M . Sucov, and P. Roy-Burman. 2002. Prostatic intraepithelial neoplasia in mice with conditional disruption of the retinoid X receptor alpha allele in the prostate epithelium. Cancer Res 62:4812. Wang, S., J. Gao, Q. Lei, N . Rozengurt, C. Pritchard, J. Jiao, G. V . Thomas, G. L i , P. Roy-Burman, P. S. Nelson, X . L i u , and H . Wu. 2003. Prostate-specific  130  135.  136.  137. 138.  139. 140.  141.  142.  143.  144.  145.  146.  147.  148.  deletion of the murine Pten tumor suppressor gene leads to metastatic prostate cancer. Cancer Cell 4:209. D i Cristofano, A . , M . De Acetis, A . Koff, C. Cordon-Cardo, and P. P. Pandolfi. 2001. Pten and p27KIPl cooperate in prostate cancer tumor suppression in the mouse. Nat Genet 27:222. Greenberg, N . M . , F. DeMayo, M . J. Finegold, D. Medina, W. D. Tilley, J. O. Aspinall, G. R. Cunha, A . A . Donjacour, R. J. Matusik, and J. M . Rosen. 1995. Prostate cancer in a transgenic mouse. Proc Natl Acad Sci U S A 92:3439. Winter, S. F., A . B. Cooper, and N . M . Greenberg. 2003. Models of metastatic prostate cancer: a transgenic perspective. Prostate Cancer Prostatic Dis 6:204. Gabril, M . Y . , T. Onita, P. G. Ji, H . Sakai, F. L . Chan, J. Koropatnick, J. L . Chin, M . Moussa, and J. W. Xuan. 2002. Prostate targeting: PSP94 gene promoter/enhancer region directed prostate tissue-specific expression in a transgenic mouse prostate cancer model. Gene Ther 9:1589. Huss, W. J., L . A . Maddison, and N . M . Greenberg. 2001. Autochthonous mouse models for prostate cancer: past, present and future. Semin Cancer Biol 11:245. Ellwood-Yen, K., T. G. Graeber, J. Wongvipat, M . L . Iruela-Arispe, J. Zhang, R. Matusik, G. V . Thomas, and C. L . Sawyers. 2003. Myc-driven murine prostate cancer shares molecular features with human prostate tumors. Cancer Cell 4:223. Zhang, X . , C. Lee, P. Y . Ng, M . Rubin, A . Shabsigh, and R. Buttyan. 2000. Prostatic neoplasia in transgenic mice with prostate-directed overexpression of the c-myc oncoprotein. Prostate 43:278. DiGiovanni, J., K . Kiguchi, A . Frijhoff, E. Wilker, D. K . Bol, L . Beltran, S. Moats, A . Ramirez, J. Jorcano, and C. Conti. 2000. Deregulated expression of insulin-like growth factor 1 in prostate epithelium leads to neoplasia in transgenic mice. Proc Natl Acad Sci U S A 97:3455. Stanbrough, M . , I. Leav, P. W. Kwan, G. J. Bubley, and S. P. Balk. 2001. Prostatic intraepithelial neoplasia in mice expressing an androgen receptor transgene in prostate epithelium. Proc Natl Acad Sci U S A 98:10823. Bruckheimer, E. M . , S. Cho, S. Brisbay, D. J. Johnson, J. R. Gingrich, N . Greenberg, and T. J. McDonnell. 2000. The impact of bcl-2 expression and bax deficiency on prostate homeostasis in vivo. Oncogene 19:2404. Zhang, X . , M . W. Chen, A. Ng, P. Y . Ng, C. Lee, M . Rubin, C. A . Olsson, and R. Buttyan. 1997. Abnormal prostate development in C3(l)-bcl-2 transgenic mice. Prostate 32:16. Foster, B . A., A . Evangelou, J. R. Gingrich, P. J. Kaplan, F. DeMayo, and N . M . Greenberg. 2002. Enforced expression of FGF-7 promotes epithelial hyperplasia whereas a dominant negative FGFR2iiib promotes the emergence of neuroendocrine phenotype in prostate glands of transgenic mice. Differentiation 70:624. Song, Z., X . Wu, W. C. Powell, R. D. Cardiff, M . B. Cohen, R. T. Tin, R. J. Matusik, G. J. Miller, and P. Roy-Burman. 2002. Fibroblast growth factor 8 isoform B overexpression in prostate epithelium: a new mouse model for prostatic intraepithelial neoplasia. Cancer Res 62:5096. Gazdar, A . F., J. S. Butel, and M . Carbone. 2002. SV40 and human tumours: myth, association or causality? Nat Rev Cancer 2:957.  131  149.  150.  151.  152.  153.  154.  155.  156. 157. 158. 159.  160.  161. 162.  Kasper, S., P. C. Sheppard, Y . Yan, N . Pettigrew, A . D. Borowsky, G. S. Prins, J. G. Dodd, M . L . Duckworth, and R. J. Matusik. 1998. Development, progression, and androgen-dependence of prostate tumors in probasin-large T antigen transgenic mice: a model for prostate cancer. Lab Invest 78:i. Shibata, M . A., C. L . Jorcyk, D. E. Devor, K . Yoshidome, S. Rulong, J. Resau, N . Roche, A . B. Roberts, J. M . Ward, and J. E. Green. 1998. Altered expression of transforming growth factor betas during urethral and bulbourethral gland tumor progression in transgenic mice carrying the androgen-responsive C3(l) 5' flanking region fused to SV40 large T antigen. Carcinogenesis 19:195. Garabedian, E. M . , P. A . Humphrey, and J. I. Gordon. 1998. A transgenic mouse model of metastatic prostate cancer originating from neuroendocrine cells. Proc Natl Acad S c i U S A 95:15382. Perez-Stable, C , N . H . Altman, J. Brown, M . Harbison, C. Cray, and B. A . Roos. 1996. Prostate, adrenocortical, and brown adipose tumors in fetal globin/T antigen transgenic mice. Lab Invest 74:363. Greenberg, N . M . , F. J. DeMayo, P. C. Sheppard, R. Barrios, R. Lebovitz, M . Finegold, R. Angelopoulou, J. G. Dodd, M . L . Duckworth, J. M . Rosen, and et al. 1994. The rat probasin gene promoter directs hormonally and developmentally regulated expression of a heterologous gene specifically to the prostate in transgenic mice. Mol Endocrinol 8:230. Gingrich, J. R., R. J. Barrios, B . A . Foster, and N . M . Greenberg. 1999. Pathologic progression of autochthonous prostate cancer in the T R A M P model. Prostate Cancer Prostatic Dis 2:70. Shappell, S. B., G. V . Thomas, R. L . Roberts, R. Herbert, M . M . Ittmann, M . A . Rubin, P. A . Humphrey, J. P. Sundberg, N . Rozengurt, R. Barrios, J. M . Ward, and R. D. Cardiff. 2004. Prostate pathology of genetically engineered mice: definitions and classification. The consensus report from the Bar Harbor meeting of the Mouse Models of Human Cancer Consortium Prostate Pathology Committee. Cancer Res 64:2270. Price, D. 1963. Comparative Aspects of Development and Structure in the Prostate. Natl Cancer Inst Monogr 12:1. Tuteja, R., and N . Tuteja. 2004. Serial Analysis of Gene Expression: Applications in Human Studies. J Biomed Biotechnol 2004:113. Stein, J., and P. Liang. 2002. Differential display technology: a general guide. Cell Mol Life Sci 59:1235. Byers, R. J., J. A . Hoyland, J. Dixon, and A . J. Freemont. 2000. Subtractive hybridization—genetic takeaways and the search for meaning. Int J Exp Pathol 81:391. Kurella, M . , L . L . Hsiao, T. Yoshida, J. D. Randall, G . Chow, S. S. Sarang, R. V . Jensen, and S. R. Gullans. 2001. D N A microarray analysis of complex biologic processes. J A m Soc Nephrol 12:1072. Murphy, D. 2002. Gene expression studies using microarrays: principles, problems, and prospects. Adv Physiol Educ 26:256. Velculescu, V . E., L . Zhang, B . Vogelstein, and K . W. Kinzler. 1995. Serial analysis of gene expression. Science 270:484.  132  163. 164.  165. 166. 167. 168. 169. 170.  171. 172.  173.  174.  175. 176.  177. 178.  179.  180.  Adams, M . D. 1996. Serial analysis of gene expression: ESTs get smaller. Bioessays 18:261. Sun, M . , G. Zhou, S. Lee, J. Chen, R. Z. Shi, and S. M . Wang. 2004. SAGE is far more sensitive than EST for detecting low-abundance transcripts. B M C Genomics 5:1. Boheler, K . R., and M . D. Stern. 2003. The new role of S A G E in gene discovery. Trends Biotechnol 21:55. Ye, S. Q., T. Lavoie, D. C. Usher, and L. Q. Zhang. 2002. Microarray, SAGE and their applications to cardiovascular diseases. Cell Res 12:105. Pleasance, E. D., M . A . Marra, and S. J. Jones. 2003. Assessment of SAGE in transcript identification. Genome Res 13:1203. Stollberg, J., J. Urschitz, Z. Urban, and C. D. Boyd. 2000. A quantitative evaluation of SAGE. Genome Res 10:1241. Yamamoto, M . , T. Wakatsuki, A . Hada, and A . Ryo. 2001. Use of serial analysis of gene expression (SAGE) technology. J Immunol Methods 250:45. Datson, N . A., J. van der Perk-de Jong, M . P. van den Berg, E. R. de Kloet, and E. Vreugdenhil. 1999. MicroSAGE: a modified procedure for serial analysis of gene expression in limited amounts of tissue. Nucleic Acids Res 27:1300. Virion, B., L. Cheval, J. M . Buhler, E. Billon, A. Doucet, and J. M . Elalouf. 1999. Serial microanalysis of renal transcriptomes. Proc Natl Acad Sci U S A 96:15286. Ye, S. Q., L . Q. Zhang, F. Zheng, D. Virgil, and P. O. Kwiterovich. 2000. miniSAGE: gene expression profiling using serial analysis of gene expression from 1 microg total R N A . Anal Biochem 287:144. Neilson, L., A . Andalibi, D. Kang, C. Coutifaris, J. F. Strauss, 3rd, J. A . Stanton, and D. P. Green. 2000. Molecular phenotype of the human oocyte by PCR-SAGE. Genomics 63:13. Peters, D. G., A . B. Kassam, H . Yonas, E. H . O'Ffare, R. E. Ferrell, and A . M . Brufsky. 1999. Comprehensive transcript analysis in small quantities of mRNA by SAGE-lite. Nucleic Acids Res 27:e39. Schober, M . S., Y . N . Min, and Y . Q. Chen. 2001. Serial analysis of gene expression in a single cell. Biotechniques 31:1240. Vilain, C , F. Libert, D. Venet, S. Costagliola, and G. Vassart. 2003. Small amplified R N A - S A G E : an alternative approach to study transcriptome from limiting amount of mRNA. Nucleic Acids Res 31 :e24. Powell, J. 1998. Enhanced concatemer cloning-a modification to the SAGE (Serial Analysis of Gene Expression) technique. Nucleic Acids Res 26:3445. Kenzelmann, M . , and K . Muhlemann. 1999. Substantially enhanced cloning efficiency of SAGE (Serial Analysis of Gene Expression) by adding a heating step to the original protocol. Nucleic Acids Res 27:917. Angelastro, J. M . , L . P. Klimaschewski, and O. V . Vitolo. 2000. Improved N l a l l l digestion of PAGE-purified 102 bp ditags by addition of a single purification step in both the SAGE and microSAGE protocols. Nucleic Acids Res 28:E62. Lee, S., J. Chen, G. Zhou, and S. M . Wang. 2001. Generation of high-quantity and quality tag/ditag cDNAs for SAGE analysis. Biotechniques 31:348.  133  181.  182.  183.  184.  185.  186.  187. 188. 189.  190.  191. 192.  193.  194.  Ryo, A., N . Kondoh, T. Wakatsuki, A . Hada, N . Yamamoto, and M . Yamamoto. 2000. A modified serial analysis of gene expression that generates longer sequence tags by nonpalindromic cohesive linker ligation. Anal Biochem 277:160. Saha, S., A . B. Sparks, C. Rago, V . Akmaev, C. J. Wang, B . Vogelstein, K . W. Kinzler, and V . E. Velculescu. 2002. Using the transcriptome to annotate the genome. Nat Biotechnol 20:508. Lu, J., A . Lai, B. Merriman, S. Nelson, and G. Riggins. 2004. A comparison of gene expression profiles produced by SAGE, long SAGE, and oligonucleotide chips. Genomics 84:631. Gowda, M . , C. Jantasuriyarat, R. A . Dean, and G. L . Wang. 2004. RobustLongSAGE (RL-SAGE): a substantially improved LongSAGE method for gene discovery and transcriptome analysis. Plant Physiol 134:890. Hashimoto, S., Y . Suzuki, Y . Kasai, K . Morohoshi, T. Yamada, J. Sese, S. Morishita, S. Sugano, and K . Matsushima. 2004. 5'-end S A G E for the analysis of transcriptional start sites. Nat Biotechnol 22:1146. Wei, C. L., P. Ng, K. P. Chiu, C. H . Wong, C. C. Ang, L . Lipovich, E. T. Liu, and Y. Ruan. 2004. 5' Long serial analysis of gene expression (LongSAGE) and 3' LongSAGE for transcriptome characterization and genome annotation. Proc Natl Acad Sci U S A 101:11701. Maruyama, K . , and S. Sugano. 1994. Oligo-capping: a simple method to replace the cap structure of eukaryotic mRNAs with oligoribonucleotides. Gene 138:171. Carninci, P., and Y . Hayashizaki. 1999. High-efficiency full-length cDNA cloning. Methods Enzymol 303:19. Lorenz, W. W., and J. F. Dean. 2002. SAGE profiling and demonstration of differential gene expression along the axial developmental gradient of lignifying xylem in loblolly pine (Pinus taeda). Tree Physiol 22:301. Bahl, A., B. Brunk, R. L . Coppel, J. Crabtree, S. J. Diskin, M . J. Fraunholz, G. R. Grant, D. Gupta, R. L . Huestis, J. C. Kissinger, P. Labo, L . L i , S. K . McWeeney, A. J. Milgram, D. S. Roos, J. Schug, and C. J. Stoeckert, Jr. 2002. PlasmoDB: the Plasmodium genome resource. A n integrated database providing tools for accessing, analyzing and mapping expression and sequence data (both finished and unfinished). Nucleic Acids Res 30:87. Tuteja, R., and N . Tuteja. 2004. Serial analysis of gene expression (SAGE): application in cancer research. Med Sci Monit 10:RA132. Bosma, A . J., B. Weigelt, A . C. Lambrechts, O. J. Verhagen, R. Pruntel, A . A . Hart, S. Rodenhuis, and L . J. van't Veer. 2002. Detection of circulating breast tumor cells by differential expression of marker genes. Clin Cancer Res 8:1871. Untergasser, G., H . B. Koch, A . Menssen, and H . Hermeking. 2002. Characterization of epithelial senescence by serial analysis of gene expression: identification of genes potentially involved in prostate cancer. Cancer Res 62:6255. Hunter, S., A . Young, J. Olson, D. J. Brat, G . Bowers, J. N . Wilcox, D . Jaye, S. Mendrinos, and A . Neish. 2002. Differential expression between pilocytic and anaplastic astrocytomas: identification of apolipoprotein D as a marker for low-  134  195.  196.  197.  198.  199.  200.  201.  202.  203.  204.  205.  206.  207.  208.  grade, non-infiltrating primary CNS neoplasms. J Neuropathol Exp Neurol 61:275. Caldwell, M . C , C. Hough, S. Furer, W. M . Linehan, P. J. Morin, and M . Gorospe. 2002. Serial analysis of gene expression in renal carcinoma cells reveals VHL-dependent sensitivity to TNFalpha cytotoxicity. Oncogene 21:929. Parle-McDermott, A . , P. McWilliam, O. Tighe, D. Dunican, and D. T. Croke. 2000. Serial analysis of gene expression identifies putative metastasis-associated transcripts in colon tumour cell lines. Br J Cancer 83:725. Gu, Z., R. Y . Lee, T. C. Skaar, K . B . Bouker, J. N . Welch, J. Lu, A . Liu, Y . Zhu, N . Davis, F. Leonessa, N . Brunner, Y . Wang, and R. Clarke. 2002. Association of interferon regulatory factor-1, nucleophosmin, nuclear factor-kappaB, and cyclic A M P response element binding with acquired resistance to Faslodex (ICI 182,780). Cancer Res 62:3428. Patino, W. D., O. Y . Mian, Y . Shizukuda, and P. M . Hwang. 2003. Current and future applications of SAGE to cardiovascular medicine. Trends Cardiovasc Med 13:163. Richards, M . , S. P. Tan, J. H . Tan, W. K . Chan, and A . Bongso. 2004. The transcriptome profile of human embryonic stem cells as defined by SAGE. Stem Cells 22:51. Anisimov, S. V., K . V . Tarasov, D. Tweedie, M . D. Stern, A . M . Wobus, and K . R. Boheler. 2002. S A G E identification of gene transcripts with profiles unique to pluripotent mouse R l embryonic stem cells. Genomics 79:169. Jasper, H . , V . Benes, C. Schwager, S. Sauer, S. Clauder-Munster, W. Ansorge, and D. Bohmann. 2001. The genomic response of the Drosophila embryo to JNK signaling. Dev Cell 1:579. Margulies, E. H . , S. L . Kardia, and J. W. Innis. 2001. A comparative molecular analysis of developing mouse forelimbs and hindlimbs using serial analysis of gene expression (SAGE). Genome Res 11:1686. Gunnersen, J. M . , C. Augustine, V . Spirkoska, M . Kim, M . Brown, and S. S. Tan. 2002. Global analysis of gene expression patterns in developing mouse neocortex using serial analysis of gene expression. Mol Cell Neurosci 19:560. Zhao, F. Q., Z. M . Sheng, M . M . Tsai, A . E. Hubbs, R. Wang, T. J. O'Leary, D. J. Izon, and J. K . Taubenberger. 2002. Serial analysis of gene expression in murine fetal thymocyte cell lines. Int Immunol 14:1383. El-Rifai, W., C. A . Moskaluk, M . K . Abdrabbo, J. Harper, C. Yoshida, G . J. Riggins, H . F. Frierson, Jr., and S. M . Powell. 2002. Gastric cancers overexpress S100A calcium-binding proteins. Cancer Res 62:6823. de Jonge, R. R., I. N . van Schaik, J. P. Vreijling, D. Troost, and F. Baas. 2004. Expression of complement components in the peripheral nervous system. Hum Mol Genet 13:295. Chen, J., M . Sun, S. Lee, G. Zhou, J. D. Rowley, and S. M . Wang. 2002. Identifying novel transcripts and novel genes in the human genome by using novel SAGE tags. Proc Natl Acad Sci U S A 99:12257. van den Berg, A., J. van der Leij, and S. Poppema. 1999. Serial analysis of gene expression: rapid RT-PCR analysis of unknown S A G E tags. Nucleic Acids Res 27:el7.  135  209. 210. 211.  212.  213. 214. 215. 216.  217.  218. 219.  220.  221. 222. 223.  224.  225.  Kannbley, U . , K . Kapinya, U . Dirnagl, and G. Trendelenburg. 2003. Improved protocol for SAGE tag-to-gene allocation. Biotechniques 34:1212. Tuteja, R., and N . Tuteja. 2004. Serial analysis of gene expression (SAGE): unraveling the bioinformatics tools. Bioessays 26:916. Lash, A . E., C. M . Tolstoshev, L . Wagner, G. D. Schuler, R. L . Strausberg, G. J. Riggins, and S. F. Altschul. 2000. SAGEmap: a public gene expression resource. Genome Res 10:1051. Boon, K., E. C. Osorio, S. F. Greenhut, C. F. Schaefer, J. Shoemaker, K . Polyak, P. J. Morin, K . H . Buetow, R. L. Strausberg, S. J. De Souza, and G. J. Riggins. 2002. A n anatomy of normal and malignant gene expression. Proc Natl Acad Sci USA99:11287. Akmaev, V . R., and C. J. Wang. 2004. Correction of sequence-based artifacts in serial analysis of gene expression. Bioinformatics 20:1254. Man, M . Z., X . Wang, and Y . Wang. 2000. POWER_SAGE: comparing statistical tests for SAGE experiments. Bioinformatics 16:953. Ruijter, J. M . , A . H . Van Kampen, and F. Baas. 2002. Statistical evaluation of SAGE libraries: consequences for experimental design. Physiol Genomics 11:37. Zhang, L., W. Zhou, V . E. Velculescu, S. E. Kern, R. H . Hruban, S. R. Hamilton, B. Vogelstein, and K . W. Kinzler. 1997. Gene expression profiles in normal and cancer cells. Science 276:1268. Michiels, E. M . , E. Oussoren, M . Van Groenigen, E. Pauws, P. M . Bossuyt, P. A . Voute, and F. Baas. 1999. Genes differentially expressed in medulloblastoma and fetal brain. Physiol Genomics 1:83. Audic, S., and J. M . Claverie. 1997. The significance of digital gene expression profiles. Genome Res 7:986. Madden, S. L . , E. A . Galella, J. Zhu, A . H . Bertelsen, and G. A . Beaudry. 1997. SAGE transcript profiles for p53-dependent growth regulation. Oncogene 15:1079. Kal, A . J., A . J. van Zonneveld, V. Benes, M . van den Berg, M . G. Koerkamp, K . Albermann, N . Strack, J. M . Ruijter, A . Richter, B . Dujon, W. Ansorge, and H. F. Tabak. 1999. Dynamics of gene expression revealed by comparison of serial analysis of gene expression transcript profiles from yeast grown on two different carbon sources. Mol Biol Cell 10:1859. Feroze-Merzoug, F., M . S. Schober, and Y . Q. Chen. 2001. Molecular profiling in prostate cancer. Cancer Metastasis Rev 20:165. Liang, P. 2002. A decade of differential display. Biotechniques 33:338. Nelson, P. S., W. L . Ng, M . Schummer, L . D. True, A . Y . Liu, R. E. Bumgarner, C. Ferguson, A . Dimak, and L. Hood. 1998. A n expressed-sequence-tag database of the human prostate: sequence analysis of 1168 c D N A clones. Genomics 47:12. Hawkins, V . , D. Doll, R. Bumgarner, T. Smith, C. Abajian, L . Hood, and P. S. Nelson. 1999. PEDB: the Prostate Expression Database. Nucleic Acids Res 27:204. Magi-Galluzzi, C , and M . Loda. 1996. Molecular events in the early phases of prostate carcinogenesis. Eur Urol 30:167.  136  226.  227. 228.  229.  230.  231. 232.  233.  234.  235.  236.  237.  238.  239.  Blok, L . J., M . V . Kumar, and D. J. Tindall. 1995. Isolation of cDNAs that are differentially expressed between androgen-dependent and androgen-independent prostate carcinoma cells using differential display PCR. Prostate 26:213. Bangma, C. H . , Y . Nasu, C. Ren, and T. C. Thompson. 1999. Metastasis-related genes in prostate cancer. Semin Oncol 26:422. X u , L . L . , N . Shanmugam, T. Segawa, I. A . Sesterhenn, D. G. McLeod, J. W. Moul, and S. Srivastava. 2000. A novel androgen-regulated gene, P M E P A 1 , located on chromosome 20ql3 exhibits high level expression in prostate. Genomics 66:257. Waghray, A., F. Feroze, M . S. Schober, F. Yao, C. Wood, E. Puravs, M . Krause, S. Hanash, and Y . Q. Chen. 2001. Identification of androgen-regulated genes in the prostate cancer cell line LNCaP by serial analysis of gene expression and proteomic analysis. Proteomics 1:1327. Waghray, A . , M . Schober, F. Feroze, F. Yao, J. Virgin, and Y . Q. Chen. 2001. Identification of differentially expressed genes by serial analysis of gene expression in human prostate cancer. Cancer Res 61:4283. Jiang, F., and Z. Wang. 2004. Identification and characterization of PLZF as a prostatic androgen-responsive gene. Prostate 59:426. Vaarala, M . H . , K . S. Porvari, A . P. Kyllonen, M . V . Mustonen, O. Lukkarinen, and P. T. Vihko. 1998. Several genes encoding ribosomal proteins are overexpressed in prostate-cancer cell lines: confirmation of L7a and L37 overexpression in prostate-cancer tissue samples. Int J Cancer 78:27. Chetcuti, A., S. Margan, S. Mann, P. Russell, D. Handelsman, J. Rogers, and Q. Dong. 2001. Identification of differentially expressed genes in organ-confined prostate cancer by gene expression array. Prostate 47:132. Luo, J., D. J. Duggan, Y . Chen, J. Sauvageot, C. M . Ewing, M . L . Bittner, J. M . Trent, and W. B. Isaacs. 2001. Human prostate cancer and benign prostatic hyperplasia: molecular dissection by gene expression profiling. Cancer Res 61:4683. Bubendorf, L . , M . Kolmer, J. Kononen, P. Koivisto, S. Mousses, Y . Chen, E. Mahlamaki, P. Schraml, H . Moch, N . Willi, A . G. Elkahloun, T. G. Pretlow, T. C. Gasser, M . J. Mihatsch, G. Sauter, and O. P. Kallioniemi. 1999. Hormone therapy failure in human prostate cancer: analysis by complementary D N A and tissue microarrays. J Natl Cancer Inst 91:1758. L i , Y . , X . L i , M . Hussain, and F. H . Sarkar. 2004. Regulation of microtubule, apoptosis, and cell cycle-related genes by taxotere in prostate cancer cells analyzed by microarray. Neoplasia 6:158. Ying, S. Y . , and S. L . Lin. 2004. Gene expression in precursor cells of prostate cancer associated with activin by combination of subtractive hybridization and microarray technologies. Biochem Biophys Res Commun 313:104. Porkka, K . P., and T. Visakorpi. 2001. Detection of differentially expressed genes in prostate cancer by combining suppression subtractive hybridization and c D N A library array. J Pathol 193:73. Abbott, D. E., C. Pritchard, N . J. Clegg, C. Ferguson, R. Dumpit, R. A . Sikes, and P. S. Nelson. 2003. Expressed sequence tag profiling identifies developmental and  137  240.  241. 242. 243. 244. 245. 246. 247. 248. 249.  250.  251.  252.  253. 254.  255.  256.  257.  anatomic partitioning of gene expression in the mouse prostate. Genome Biol 4:R79. Kuhl, M . , L . C. Sheldahl, M. Park, J. R. Miller, and R. T. Moon. 2000. The Wnt/Ca2+ pathway: a new vertebrate Wnt signaling pathway takes shape. Trends Genet 16:279. Katoh, M . 2002. WNT and FGF gene clusters (review). Int J Oncol 21:1269. Huang, H . C., and P. S. Klein. 2004. The Frizzled family: receptors for multiple signal transduction pathways. Genome Biol 5:234. Hsieh, J. C. 2004. Specificity of WNT-receptor interactions. Front Biosci 9:1333. Lustig, B., and J. Behrens. 2003. The Wnt signaling pathway and its role in tumor development. J Cancer Res Clin Oncol 129:199. Kawano, Y . , and R. Kypta. 2003. Secreted antagonists of the Wnt signalling pathway. J Cell Sci 116:2627. Cadigan, K . M . , and R. Nusse. 1997. Wnt signaling: a common theme in animal development. Genes Dev 11:3286. Yang, Y . 2003. Wnts and wing: Wnt signaling in vertebrate limb development and musculoskeletal morphogenesis. Birth Defects Res C Embryo Today 69:305. Liu, P., M . Wakamiya, M . J. Shea, U . Albrecht, R. R. Behringer, and A. Bradley. 1999. Requirement for Wnt3 in vertebrate axis formation. Nat Genet 22:361. Pandur, P., M . Lasche, L. M . Eisenberg, and M . Kuhl. 2002. Wnt-11 activation of a non-canonical Wnt signalling pathway is required for cardiogenesis. Nature 418:636. Ross, S. E., N . Hemati, K . A . Longo, C. N . Bennett, P. C. Lucas, R. L . Erickson, and O. A . MacDougald. 2000. Inhibition of adipogenesis by Wnt signaling. Science 289:950. Majumdar, A . , S. Vainio, A . Kispert, J. McMahon, and A . P. McMahon. 2003. W n t l l and Ret/Gdnf pathways cooperate in regulating ureteric branching during metanephric kidney development. Development 130:3175. Brisken, C , A . Heineman, T. Chavarria, B. Elenbaas, J. Tan, S. K . Dey, J. A . McMahon, A . P. McMahon, and R. A . Weinberg. 2000. Essential function of Wnt-4 in mammary gland development downstream of progesterone signaling. Genes Dev 14:650. L i , C , J. Xiao, K . Hormi, Z. Borok, and P. Minoo. 2002. Wnt5a participates in distal lung morphogenesis. Dev Biol 248:68. McMahon, A . P., A . L . Joyner, A . Bradley, and J. A . McMahon. 1992. The midbrain-hindbrain phenotype of Wnt-1-/Wnt-1- mice results from stepwise deletion of engrailed-expressing cells by 9.5 days postcoitum. Cell 69:581. Yamaguchi, T. P., A . Bradley, A . P. McMahon, and S. Jones. 1999. A Wnt5a pathway underlies outgrowth of multiple structures in the vertebrate embryo. Development 126:1211. Monkley, S. J., S. J. Delaney, D. J. Pennisi, J. H . Christiansen, and B . J. Wainwright. 1996. Targeted disruption of the Wnt2 gene results in placentation defects. Development 122:3343. Wang, J., and A . Wynshaw-Boris. 2004. The canonical Wnt pathway in early mammalian embryogenesis and stem cell maintenance/differentiation. Curr Opin Genet Dev 14:533.  138  258.  259.  260. 261. 262.  263.  264.  265.  266. 267. 268.  269.  270.  271.  272.  Zechner, D., Y . Fujita, J. Hulsken, T. Muller, I. Walther, M . M . Taketo, E. B. Crenshaw, 3rd, W. Birchmeier, and C. Birchmeier. 2003. beta-Catenin signals regulate cell growth and the balance between progenitor cell expansion and differentiation in the nervous system. Dev Biol 258:406. Lammi, L . , S. Arte, M . Somer, H . Jarvinen, P. Lahermo, I. Thesleff, S. Pirinen, and P. Nieminen. 2004. Mutations in AXIN2 cause familial tooth agenesis and predispose to colorectal cancer. A m J Hum Genet 74:1043. Giles, R. H., J. H . van Es, and H. Clevers. 2003. Caught up in a Wnt storm: Wnt signaling in cancer. Biochim Biophys Acta 1653:1. Voeller, H . J., C. I. Truica, and E. P. Gelmann. 1998. Beta-catenin mutations in human prostate cancer. Cancer Res 58:2520. Chesire, D. R., C. M . Ewing, J. Sauvageot, G. S. Bova, and W. B. Isaacs. 2000. Detection and analysis of beta-catenin mutations in prostate cancer. Prostate 45:323. Verras, M . , J. Brown, X . L i , R. Nusse, and Z. Sun. 2004. Wnt3a growth factor induces androgen receptor-mediated transcription and enhances cell growth in human prostate cancer cells. Cancer Res 64:8860. Chen, G., N . Shukeir, A . Potti, K . Sircar, A . Aprikian, D. Goltzman, and S. A . Rabbani. 2004. Up-regulation of Wnt-1 and beta-catenin production in patients with advanced metastatic prostate carcinoma: potential pathogenetic and prognostic implications. Cancer 101:1345. Wissmann, C , P. J. Wild, S. Kaiser, S. Roepcke, R. Stoehr, M . Woenckhaus, G. Kristiansen, J. C. Hsieh, F. Hofstaedter, A. Hartmann, R. Knuechel, A. Rosenthal, and C. Pilarsky. 2003. WIF1, a component of the Wnt pathway, is down-regulated in prostate, breast, lung, and bladder cancer. J Pathol 201:204. Yang, F., X . L i , M . Sharma, C. Y . Sasaki, D. L . Longo, B . Lim, and Z. Sun. 2002. Linking beta-catenin to androgen-signaling pathway. J Biol Chem 277:11336. Truica, C. I., S. Byers, and E. P. Gelmann. 2000. Beta-catenin affects androgen receptor transcriptional activity and ligand specificity. Cancer Res 60:4709. Chesire, D. R., C. M . Ewing, W. R. Gage, and W. B . Isaacs. 2002. In vitro evidence for complex modes of nuclear beta-catenin signaling during prostate growth and tumorigenesis. Oncogene 21:2679. Mulholland, D. J., H . Cheng, K . Reid, P. S. Rennie, and C. C. Nelson. 2002. The androgen receptor can promote beta-catenin nuclear translocation independently of adenomatous polyposis coli. J Biol Chem 277:17933. Mulholland, D. J., J. T. Read, P. S. Rennie, M . E. Cox, and C. C. Nelson. 2003. Functional localization and competition between the androgen receptor and T-cell factor for nuclear beta-catenin: a means for inhibition of the Tcf signaling axis. Oncogene 22:5602. Amir, A. L., M . Barua, N . C. McKnight, S. Cheng, X . Yuan, and S. P. Balk. 2003. A direct beta-catenin-independent interaction between androgen receptor and T cell factor 4. J Biol Chem 278:30828. Jones, S. E., and C. Jomary. 2002. Secreted Frizzled-related proteins: searching for relationships and patterns. Bioessays 24:811.  139  273.  274.  275.  276.  277. 278.  279.  280.  281.  282.  283.  284.  285.  Lin, K . , S. Wang, M . A . Julius, J. Kitajewski, M . Moos, Jr., and F. P. Luyten. 1997. The cysteine-rich frizzled domain of Frzb-1 is required and sufficient for modulation of Wnt signaling. Proc Natl Acad Sci U S A 94:11196. Bafico, A . , A . Gazit, T. Pramila, P. W. Finch, A . Yaniv, and S. A . Aaronson. 1999. Interaction of frizzled related protein (FRP) with Wnt ligands and the frizzled receptor suggests alternative mechanisms for FRP inhibition of Wnt signaling. J Biol Chem 274:16180. Uren, A . , F. Reichsman, V . Anest, W. G. Taylor, K . Muraiso, D. P. Bottaro, S. Cumberledge, and J. S. Rubin. 2000. Secreted frizzled-related protein-1 binds directly to Wingless and is a biphasic modulator of Wnt signaling. J Biol Chem 275:4374. Banyai, L . , and L . Patthy. 1999. The N T R module: domains of netrins, secreted frizzled related proteins, and type I procollagen C-proteinase enhancer protein are homologous with tissue inhibitors of metalloproteases. Protein Sci 8:1636. Lescher, B., B. Haenig, and A . Kispert. 1998. sFRP-2 is a target of the Wnt-4 signaling pathway in the developing metanephric kidney. Dev Dyn 213:440. Melkonyan, H . S., W. C. Chang, J. P. Shapiro, M . Mahadevappa, P. A . Fitzpatrick, M . C. Kiefer, L . D. Tomei, and S. R. Umansky. 1997. SARPs: a family of secreted apoptosis-related proteins. Proc Natl Acad Sci U S A 94:13636. Yoshino, K., J. S. Rubin, K . G. Higinbotham, A . Uren, V . Anest, S. Y . Plisov, and A. O. Perantoni. 2001. Secreted Frizzled-related proteins can regulate metanephric development. Mech Dev 102:45. Schumann, FL, J. Holtz, H . R. Zerkowski, and M . Hatzfeld. 2000. Expression of secreted frizzled related proteins 3 and 4 in human ventricular myocardium correlates with apoptosis related gene expression. Cardiovasc Res 45:720. Ko, J., K. S. Ryu, Y . H . Lee, D. S. Na, Y . S. Kim, Y . M . Oh, I. S. Kim, and J. W. Kim. 2002. Human secreted frizzled-related protein is down-regulated and induces apoptosis in human cervical cancer. Exp Cell Res 280:280. Ugolini, F., E. Charafe-Jauffret, V . J. Bardou, J. Geneix, J. Adelaide, F. LabatMoleur, F. Penault-Llorca, M . Longy, J. Jacquemier, D. Birnbaum, and M . J. Pebusque. 2001. WNT pathway and mammary carcinogenesis: loss of expression of candidate tumor suppressor gene SFRP1 in most invasive carcinomas except of the medullary type. Oncogene 20:5810. Zhou, Z., J. Wang, X . Han, J. Zhou, and S. Linder. 1998. Up-regulation of human secreted frizzled homolog in apoptosis and its down-regulation in breast tumors. Int J Cancer 78:95. Horvath, L . G., S. M . Henshall, J. G . Kench, D. N . Saunders, C. S. Lee, D. Golovsky, P. C. Brenner, G. F. O'Neill, R. Kooner, P. D. Strieker, J. J. Grygiel, and R. L . Sutherland. 2004. Membranous expression of secreted frizzled-related protein 4 predicts for good prognosis in localized prostate cancer and inhibits PC3 cellular proliferation in vitro. Clin Cancer Res 10:615. Abu-Jawdeh, G., N . Cornelia, Y . Tomita, L . F. Brown, K . Tognazzi, S. Y . Sokol, and O. Kocher. 1999. Differential expression of frpHE: a novel human stromal protein of the secreted frizzled gene family, during the endometrial cycle and malignancy. Lab Invest 79:439.  140  286.  287.  288.  289. 290.  291.  292.  293.  294.  295.  296.  297.  298.  299.  Rattner, A . , J. C. Hsieh, P. M . Smallwood, D. J. Gilbert, N . G. Copeland, N . A . Jenkins, and J. Nathans. 1997. A family of secreted proteins contains homology to the cysteine-rich ligand-binding domain of frizzled receptors. Proc Natl Acad Sci U S A 94:2859. Shirozu, M . , H . Tada, K . Tashiro, T. Nakamura, N . D. Lopez, M . Nazarea, T. Hamada, T. Sato, T. Nakano, and T. Honjo. 1996. Characterization of novel secreted and membrane proteins isolated by the signal sequence trap method. Genomics 37:273. Leimeister, C , A . Bach, and M . Gessler. 1998. Developmental expression patterns of mouse sFRP genes encoding members of the secreted frizzled related protein family. Mech Dev 75:29. Sarkar, L., and P. T. Sharpe. 1999. Expression of Wnt signalling pathway genes during tooth development. Mech Dev 85:197. Tebar, M . , O. Destree, W. J. de Vree, and A . A . Ten Have-Opbroek. 2001. Expression of Tcf/Lef and sFrp and localization of beta-catenin in the developing mouse lung. Mech Dev 109:437. Jones, S. E., C. Jomary, J. Grist, H . J. Stewart, and M . J. Neal. 2000. Altered expression of secreted frizzled-related protein-2 in retinitis pigmentosa retinas. Invest Ophthalmol Vis Sci 41:1297. Roth, W., C. Wild-Bode, M . Platten, C. Grimmel, H . S. Melkonyan, J. Dichgans, and M . Weller. 2000. Secreted Frizzled-related proteins inhibit motility and promote growth of human malignant glioma cells. Oncogene 19:4210. Lee, J. L., C. J. Chang, L . L . Chueh, and C. T. Lin. 2003. Expression of secreted frizzled-related protein 2 in a primary canine mammary tumor cell line: a candidate tumor marker for mammary tumor cells. In Vitro Cell Dev Biol Anim 39:221. Lee, J. L . , C. J. Chang, S. Y . Wu, D. R. Sargan, and C. T. Lin. 2004. Secreted frizzled-related protein 2 (SFRP2) is highly expressed in canine mammary gland tumors but not in normal mammary glands. Breast Cancer Res Treat 84:139. Lee, J. L . , C. T. Lin, L . L. Chueh, and C. J. Chang. 2004. Autocrine/paracrine secreted Frizzled-related protein 2 induces cellular resistance to apoptosis: a possible mechanism of mammary tumorigenesis. J Biol Chem 279:14602. Gingrich, J. R., R. J. Barrios, R. A . Morton, B . F. Boyce, F. J. DeMayo, M . J. Finegold, R. Angelopoulou, J. M . Rosen, and N . M . Greenberg. 1996. Metastatic prostate cancer in a transgenic mouse. Cancer Res 56:4096. Imasato, Y . , T. Onita, M . Moussa, H . Sakai, F. L . Chan, J. Koropatnick, J. L . Chin, and J. W. Xuan. 2001. Rodent PSP94 gene expression is more specific to the dorsolateral prostate and less sensitive to androgen ablation than probasin. Endocrinology 142:2138. Kwong, J., J. W. Xuan, P. S. Chan, S. M . Ho, and F. L . Chan. 2000. A comparative study of hormonal regulation of three secretory proteins (prostatic secretory protein-PSP94, probasin, and seminal vesicle secretion II) in rat lateral prostate. Endocrinology 141:4543. Sciavolino, P. J., E. W. Abrams, L . Yang, L . P. Austenberg, M . M . Shen, and C. Abate-Shen. 1997. Tissue-specific expression of murine Nkx3.1 in the male urogenital system. Developmental Dynamics 209:127.  141  300.  301.  302.  303.  304.  305.  306.  307.  308.  309.  310.  311.  312. 313.  Berman, D. M . , H . Tian, and D. W. Russell. 1995. Expression and regulation of steroid 5 alpha-reductase in the urogenital tract of the fetal rat. Molecular Endocrinology 9:1561. Dyne, K . M . , M . Valli, A . Forlino, M . Mottes, H . Kresse, and G. Cetta. 1996. Deficient expression of the small proteoglycan decorin in a case of severe/lethal osteogenesis imperfecta. A m J Med Genet 63:161. Chiao, E., P. Fisher, L . Crisponi, M . Deiana, I. Dragatsis, D . Schlessinger, G. Pilia, and A . Efstratiadis. 2002. Overgrowth of a mouse model of the SimpsonGolabi-Behmel syndrome is independent of IGF signaling. Dev Biol 243:185. Grisaru, S., D. Cano-Gauci, J. Tee, J. Filmus, and N . D. Rosenblum. 2001. Glypican-3 modulates B M P - and FGF-mediated effects during renal branching morphogenesis. Dev Biol 231:31. Kanety, H . , Y . Madjar, Y . Dagan, J. Levi, M . Z. Papa, C. Pariente, B. Goldwasser, and A . Karasik. 1993. Serum insulin-like growth factor-binding protein-2 (IGFBP-2) is increased and IGFBP-3 is decreased in patients with prostate cancer: correlation with serum prostate-specific antigen. J Clin Endocrinol Metab 77:229. Y u , H . , M . R. Nicar, R. Shi, H . J. Berkel, R. Nam, J. Trachtenberg, and E. P. Diamandis. 2001. Levels of insulin-like growth factor I (IGF-I) and IGF binding proteins 2 and 3 in serial postoperative serum samples and risk of prostate cancer recurrence. Urology 57:471. Moore, M . G., L . A . Wetterau, M . J. Francis, D. M . Peehl, and P. Cohen. 2003. Novel stimulatory role for insulin-like growth factor binding protein-2 in prostate cancer cells. Int J Cancer 105:14. Chang, J. T., N . Esumi, K . Moore, Y . L i , S. Zhang, C. Chew, B . Goodman, A . Rattner, S. Moody, G. Stetten, P. A . Campochiaro, and D. J. Zack. 1999. Cloning and characterization of a secreted frizzled-related protein that is expressed by the retinal pigment epithelium. Hum Mol Genet 8:575. K i m , A . S., S. A . Anderson, J. L . Rubenstein, D. H . Lowenstein, and S. J. Pleasure. 2001. Pax-6 regulates expression of SFRP-2 and Wnt-7b in the developing CNS. J Neurosci 2LRC132. Barnes, T., L . M . Galli, A . Anglade, P. Francis-West, K . Willert, R. Nusse, P. Bovolenta, L.W. Burrus. 2004. Analysis of the Role of Sfrps in Modulating Wnt3a Activity. Dev Biol 271:605. Ladher, R. K . , V . L . Church, S. Allen, L . Robson, A . Abdelfattah, N . A . Brown, G. Hattersley, V . Rosen, F. P. Luyten, L . Dale, and P. H . Francis-West. 2000. Cloning and expression of the Wnt antagonists Sfrp-2 and Frzb during chick development. Dev Biol 218:183. Saitoh, T., T. Mine, and M . Katoh. 2002. Expression and regulation of WNT8A and WNT8B mRNAs in human tumor cell lines: up-regulation of WNT8B mRNA by beta-estradiol in MCF-7 cells, and down-regulation of W N T 8 A and WNT8B mRNAs by retinoic acid in NT2 cells. Int J Oncol 20:999. Zecca, M . , K . Basler, and G. Struhl. 1996. Direct and long-range action of a ^ wingless morphogen gradient. Cell 87:833. Gerstein, A . V . , T. A . Almeida, G. Zhao, E. Chess, M . Shih Ie, K . Buhler, K . Pienta, M . A . Rubin, R. Vessella, and N . Papadopoulos. 2002. APC/CTNNB1  142  314. 315.  316. 317.  318. 319. 320.  321.  322.  323. 324.  325.  326.  327. 328.  (beta-catenin) pathway alterations in human prostate cancers. Genes Chromosomes Cancer 34:9. Chesire, D. R., and W. B . Isaacs. 2002. Ligand-dependent inhibition of betacatenin/TCF signaling by androgen receptor. Oncogene 21:8453. Velculescu, V . E., S. L . Madden, L . Zhang, A . E. Lash, J. Y u , C. Rago, A . Lai, C. J. Wang, G. A . Beaudry, K . M . Ciriello, B. P. Cook, M . R. Dufault, A . T. Ferguson, Y . Gao, T. C. He, H . Hermeking, S. K . Hiraldo, P. M . Hwang, M . A . Lopez, H . F. Luderer, B . Mathews, J. M . Petroziello, K . Polyak, L . Zawel, K . W. Kinzler, and et al. 1999. Analysis of human transcriptomes. Nat Genet 23:387. Mericskay, M . , J. Kitajewski, and D. Sassoon. 2004. Wnt5a is required for proper epifhelial-mesenchymal interactions in the uterus. Development 131:2061. Uyttendaele, H . , J. V . Soriano, R. Montesano, and J. Kitajewski. 1998. Notch4 and Wnt-1 proteins function to regulate branching morphogenesis of mammary epithelial cells in an opposing fashion. Dev Biol 196:204. Shannon, J. M . , and B. A . Hyatt. 2004. Epithelial-mesenchymal interactions in the developing lung. Annu Rev Physiol 66:625. Nelson, W. J., and R. Nusse. 2004. Convergence of Wnt, beta-catenin, and cadherin pathways. Science 303:1483. Lelongt, B., B. Legallicier, R. Piedagnel, and P. M . Ronco. 2001. Do matrix metalloproteinases MMP-2 and MMP-9 (gelatinases) play a role in renal development, physiology and glomerular diseases? Curr Opin Nephrol Hypertens 10:7. Wilson, M . J., J. D. Kirby, Y . Zhao, A . A . Sinha, and P. S. Cooke. 1997. Neonatal hypothyroidism alters the pattern of prostatic growth and differentiation, as well as plasminogen activator and metalloprotease expression, in the rat. Biol Reprod 56:475. Aubert, J., H . Dunstan, I. Chambers, and A . Smith. 2002. Functional gene screening in embryonic stem cells implicates Wnt antagonism in neural differentiation. Nat Biotechnol 20:1240. Logan, C. Y . , and R. Nusse. 2004. The Wnt signaling pathway in development and disease. Annu Rev Cell Dev Biol 20:781. Lyuksyutova, A . I., C. C. Lu, N . Milanesio, L . A . King, N . Guo, Y . Wang, J. Nathans, M . Tessier-Lavigne, and Y . Zou. 2003. Anterior-posterior guidance of commissural axons by Wnt-frizzled signaling. Science 302:1984. Lyons, J. P., U . W. Mueller, H . Ji, C. Everett, X . Fang, J. C. Hsieh, A . M . Barth, and P. D. McCrea. 2004. Wnt-4 activates the canonical beta-catenin-mediated Wnt pathway and binds Frizzled-6 CRD: functional implications of Wnt/betacatenin activity in kidney epithelial cells. Exp Cell Res 298:369. Jordan, B . K . , J. H . Shen, R. Olaso, H . A . Ingraham, and E. Vilain. 2003. Wnt4 overexpression disrupts normal testicular vasculature and inhibits testosterone synthesis by repressing steroidogenic factor 1/beta-catenin synergy. Proc Natl Acad Sci U S A 100:10866. Vainio, S., M . Heikkila, A . Kispert, N . Chin, and A . P. McMahon. 1999. Female development in mammals is regulated by Wnt-4 signalling. Nature 397:405. Abate-Shen, C , and M . M . Shen. 2002. Mouse models of prostate carcinogenesis. Trends Genet 18:S1.  143  329.  330.  331. 332.  333.  334.  335. 336.  337.  338.  Garnis, C., J. Campbell, J. J. Davies, C. Macaulay, S. Lam, and W. L . Lam. 2005. Involvement of multiple developmental genes on chromosome lp in lung tumorigenesis. Hum Mol Genet 14:475. Huguet, E. L., J. A . McMahon, A . P. McMahon, R. Bicknell, and A . L . Harris. 1994. Differential expression of human Wnt genes 2, 3, 4, and 7B in human breast cell lines and normal and disease states of human breast tissue. Cancer Res 54:2615. Katoh, M . 2001. Frequent up-regulation of WNT2 in primary gastric cancer and colorectal cancer. Int J Oncol 19:1003. Zhu, H . , M . Mazor, Y . Kawano, M . M . Walker, H . Y . Leung, K . Armstrong, J. Waxman, and R. M . Kypta. 2004. Analysis of Wnt gene expression in prostate cancer: mutual inhibition by WNT 11 and the androgen receptor. Cancer Res 64:7918. Gupta, S., N . Ahmad, S. R. Marengo, G. T. MacLennan, N . M . Greenberg, and H . Mukhtar. 2000. Chemoprevention of prostate carcinogenesis by alphadifluoromethylornithine in T R A M P mice. Cancer Res 60:5125. Guo, X . , T. F. Day, X . Jiang, L . Garrett-Beal, L . Topol, and Y . Yang. 2004. Wnt/beta-catenin signaling is sufficient and necessary for synovial joint formation. Genes Dev 18:2404. Moon, R. T., J. D. Brown, and M . Torres. 1997. WNTs modulate cell fate and behavior during vertebrate development. Trends Genet 13:157. Caldwell, G. M . , C. Jones, K . Gensberg, S. Jan, R. G. Hardy, P. Byrd, S. Chughtai, Y . Wallis, G. M . Matthews, and D. G. Morton. 2004. The Wnt antagonist sFRPl in colorectal tumorigenesis. Cancer Res 64:883. Lee, C. S., L . A . Buttitta, N . R. May, A . Kispert, and C. M . Fan. 2000. SHH-N upregulates Sfrp2 to mediate its competitive interaction with WNT1 and WNT4 in the somitic mesoderm. Development 127:109. Fan, L . , C. V . Pepicelli, C. C. Dibble, W. Catbagan, J. L . Zarycki, R. Laciak, J. Gipp, A . Shaw, M . L . Lamm, A . Munoz, R. Lipinski, J. B . Thrasher, and W. Bushman. 2004. Hedgehog signaling promotes prostate xenograft tumor growth. Endocrinology 145:3961.  144  Appendix 1. Identification of prostate specific genes  Introduction The hypothesis underlying this work was that prostate specific genes (i.e. those expressed only in the prostate) are likely to play important roles in prostate development and its normal physiological functions. Such genes may also be dysregulated during prostate cancer (PCa) development and thus be novel targets for therapy. For example, Nkx3.1, which encoding a homeobox transcription factor expressed relatively specifically in the prostate, is both a critical developmental gene and a candidate tumor suppressor (1, 2). The identification of prostate specific genes may also help us to monitor the progression of PCa and its metastasis. In humans, the level of Prostate Specific Antigen (PSA), a prostate specific protein, is used as a biomarker for screening PCa as well as evaluating its progression, recurrence and prognosis clinically (3).  Materials and Methods Rapid Amplification of c D N A Ends (RACE)-PCR The dissection of, and R N A isolation from, the dorsal-lateral lobe of 12 week old adult C57B1/6J mice was carried out as described in Chapter 2, Materials and Methods. The isolated total R N A was then split into 2 aliquots for both 5 ' - R A C E P C R and 3'R A C E PCR. The R A C E - P C R was performed using the GeneRacer™ Kit (Invitrogen) following manufacturer's instructions. Unless otherwise specified, all reagents used in the following steps were included in the kit.  145  For 5'-RACE, total R N A was treated with calf intestinal phosphatase (CIP, lOU/uf) to remove the 5' phosphate. This eliminates truncated mRNA and non-mRNA from subsequent ligation with the GeneRacer R N A Oligo. The CIP treated R N A was then treated with tobacco acid pyrophosphatase (TAP, 0.5U/ul) to remove the 5' cap structure from the intact, full-length mRNA. The 5'-end of the mRNA was ligated with the GeneRacer R N A Oligo using T4 ligase (5U/|ai). The Oligo sequence is: 5'CG A C U G G A G C A C G A G G A C A C U G A C A U G G A C U G A A G G A G U A G A A A - 3 ' .  This  oligo provides a known priming site for GeneRacer PCR primers in the later steps. The mRNA was then converted to cDNA using Superscript II RT (200U/)^1) following the standard RT method (see Chapter 2, Materials and Methods). To perform the 5' R A C E PCR, forward and reverse primers were designed and purchased from Invitrogen. The forward primer for the 5' R A C E PCR was modified from the original sequence provided in the kit to reduce the Tm. The sequence is:  5'-TGGAGCACGAGGACACTGA-3'  (150ng/uf). The reverse primers were generated to be complementary to the tag sequences of interest. Selected tags are listed in Table A L L The following P C R was carried out using a Gradient PCR PTC-200 (MJ Research) under the following conditions: samples were denatured at 95°C for 30 seconds, annealed (temperature varies from 50°C to 70°C) for 30 seconds, and extended at 72°C for 2 minutes for 35 cycles. Different annealing temperatures were tested until the expected band appeared. For the 3' R A C E , cDNA synthesis was carried out by standard RT using Superscript II RT (200U/ul) following the Invitrogen protocol. Tags of interest were used as forward primers (Table A l . l ) . The reverse primer for the 3' R A C E was modified from that  146  provided in the kit to reduce the Tm. The sequence is: 5 ' - C A A C G A T A C G C T A C G T A A C G - 3 ' . The PCR was performed under the same conditions as in the 5' R A C E . Table A 1.1 Primer sequences of selected prostate specific tags for RACE-PCR. Tag AATTGCAAACTAGAATA GACTCCAGGAATGCCTC GCAAGGAAGGCAGGTTG GGGAGAAGTCCTGTGTG TGCCAGCTGATCAGTCA ATAAAATCAACTCACAC GCAAGGCTGGACATCAG  Primer for 5' RACE TATTCTAGTTTGCAATTCATG GAGGCATTCCTGGAGTCCATG CAACCTGCCTTCCTTGCCATG CACACAGGACTTCTCCCCATG TGACTGATCAGCTGGCACATG GTGTGAGTTGATTTTATCATG CTGATGTCCAGCCTTGCCATG  Primer for 3' RACE CATGAATTGCAAACTAGAATA CATGGACTCCAGGAATGCCTC CATGGCAAGGAAGGCAGGTTG CATGGGGAGAAGTCCTGTGTG CATGTGCCAGCTGATCAGTCA CATGATAAAATCAACTCACAC CATGGCAAGGCTGGACATCAG  The successfully amplified PCR products were purified from a 1% gel using the MinElute Gel Extraction Kit (Qiagen), cloned into the TOPO pCR2.1® vector (Invitrogen) and transformed into TOPO 10 competent cells according to the TOPO TA® cloning kit (Invitrogen). Transformation and inoculation of the cells are described in Chapter 2, Materials and Methods. The plasmid was then purified using a QIAprep Spin Miniprep Kit (Qiagen). PCR was performed using M13 primers provided in the TOPO TA® cloning kit (Invitrogen) and the PCR products were run on a gel to confirm successful insertion of the R A C E PCR product. Bands of the correct size were excised, purified with QIAprep Spin Miniprep K i t (Qiagen) and sent for sequencing (UBC NAPS). To identify the amplicons, sequencing results were analyzed by combining the 5' and 3' R A C E results. The combined sequences were then blasted using Ensembl B L A S T (http://www.ensembl.org/Multi/blastview?species=Mus_musculus).  147  Results The comparison strategy to identify tags of interest is shown in Figure A L L Tags commonly expressed amongst the 3 prostate libraries (5,618 in total) were compared to 86 other SAGE libraries (including 41 G E O libraries from NCBI). Only 4 tags were identified  as  "prostate-specific" using this  comparison strategy.  They were:  933011 lN05Rik; ATPase, Na+/K+ transporting, alpha 1 polypeptide (Atplal); sine oculis-related homeobox 1 homolog (Sixl) and keratin 15 (Krtl-15). Searching for the expression of these genes in the UniGene database showed these 4 genes are expressed in many tissues and thus are not "prostate specific".  A  B  Figure A1.1. Strategy to search for prostate specific tags. 5,618 tags that were commonly expressed in the 3 prostate libraries (A) were identified and compared to 86 mouse SAGE libraries. Only 4 tags were left after the comparison (B).  A second comparison was therefore carried out using an alternate strategy. In this case, comparison of all 45,244 tag types present in the 3 prostate libraries with 86 other SAGE libraries identified 6,265 tags as "prostate specific" (Figure A1.2). Among them, 62 tags present at above 5 counts were chosen for further analysis (Table A1.2). Of these, most of the tags (58/62 tags) were only present in the adult prostate library. 55/62 tags were unambiguously mapped to known genes by CMOST (Chapter 2, Materials and Methods), including transcripts of some prostate specific secreted proteins such as  148  probasin and beta-defensin, and were thus not of interest. The 7 tags that could not be matched to any database were chosen as potential novel genes for subsequent analysis by RACE-PCR. Both 3' and 5' R A C E were carried out for all 7 candidate genes resulting in 8 products from 4 tags that were sent for sequencing analysis. Surprisingly, 3/7 gave no product in either 3' or 5' RACE-PCR, suggesting the tags might have been generated through sequencing errors. Comparisons of the sequencing results to the Ensemble genome sequence using Ensembl B L A S T revealed that 2/7 were previously identified genes likely representing tags that span splice sites. One of these represents a novel exon. The additional sequence information obtained for 2/7 tags suggested the possibility that they represent novel transcripts (Table A1.3). Primers were designed against sequences from both the 5' and 3' ends in order to attempt to amplify these products from R N A isolated from 12 week adult prostate dorsal lobe for further validation. RT-PCR amplification of the potential novel gene products was unsuccessful.  Figure A 1.2. Strategy of the second search for prostate specific tags. Tags that were expressed in the 3 prostate libraries (A) were compared to 86 other mouse SAGE libraries (B).  149  Table A 1.2 Prostate specific tags (2  n  search.)  Tag  Counts in Adult library  GACACGGACAACTGTCC AGTTGAGCTGATTGTGA  40 32  GTGGTCCCTCGGTCAAC  31  GGTACACTACTTTGGTA TGCCCAGGTTCTCTGGC GACTCCAGGAATGCCTC TGTGAGCGACTGTCTCT TCTGTATATGTTGTACC CGCCTGCGCGTGCCCGC TCAGAGCTTGCAGAAAG TGGAAGACACCAAAGCC TATGACTATGTTGACAG GTTGTAGGATCCATATA AATTGCAAACTAGAATA GCTGAGCTGCGCAGCAG TCCTAATCGGTATTAGT CTACTCAGCTTTCATCT CTCTAATCGGTATTAGT AACTTTTCATTTGGAGT GATTCTATCAAGTCCAG GTGCGTCGAATGAGAGC GCAAGGCTGGACATCAG GGTGGAATGAAGAGAGA GTACCCGTGAGAATGGA GTACCTGCGAGAATGGA GTACCTGTGGGAATGGA GTGATAAAATTACATAG TGCATCCTCACTCAGGC AGAGCCAGAGCGACCCT TTCTAGTCGGTATTAGT CACCGGGCCTCTGCGTC CTCTGTAACCTGTAGAA TTTGTCTTCCCACAGTG CATTCCATTGACAGTCA CCAGGAGCTAGCACAAA GCCAGCTCAGAATGTAC GGCACCTGTGTGCACGT GGTGATCTCTGTAATAA TAGCTTTTACCACAAAT TGCCAACCGATCAGTCA TTCTAACCGGTATTAGT  30 26 24 23 22 19 19 19 18 15 14 14 14 12 12 11 11 11 10 10 10 10 10 10 10 9 9 8 8 8 7 7 7 7 7 7 7 7 6 6 6 6 6 6 6 6 5 5 5 5 5 5 5 5 5 5  CATCCTCCCAAGAGCTC CGACAAAACGTCAATCA CTTGGACAGCCTCAACC GACAGCGGGCATTTCCC GATTTGGCAGTGTGAAA GCAAGGAAGGCAGGTTG GGGAGAAGTCCTGTGTG TTCCAATCGGTATTAGT ATAAAATCAACTCACAC TGGTAGCTCGTGGTCTC CAAGGGATTGTTCAAAT CACTAGTAACCAAACCA CCACTGGAGCAGGTCAT GAAGTCAAGTGACAGAT GCACCTGTGAGAATGGA GCTGAAATGGGCTGTCA GGGGGTAGACCCTTTTT TGACGAAACGTCAATCA  Annotation Probasin Homeobox B13 ATPase, H+ transporting, V1 subunit C, isoform2 RIKEN full-length enriched library, clone:A630095E13 Hypothetical protein 9530008N106  1  Probasin 5430419D17Rik Mus musculus similar to A430096B05Rik protein RIKEN full-length enriched library, clone:A630095E13 Experimental Autoimmune prostatitis antigen 1 ATPase, H+/K+ transporting, non-gastric, alpha polypeptide Mus musculus similar to A430096B05Rik protein RIKEN cDNA 9530053A07 gene Probasin spermine binding protein; mature glycoprotein Probasin Membrane-spanning 4-domains, subfamily A, member5r hypothetical protein 9530008N10 ENSMUSESTT00000003793 Transcript Derived From ESTs Ambiguous Defensin Defensin Defensin Defensin beta 1 cis-AB transferase Mus musculus similar to A430096B05Rik protein Probasin Spermine binding protein Membrane-spanning 4-domains, subfamily A, member 5r Experimental autoimmune prostatitis antigen 1 Spermine binding protein; mature glycoprotein RIKEN cDNA A030004J04 Procollagen, type IV, alpha 6 RIKEN cDNA 9530053A07 gene Defensin beta 1 Experimental prostatitis antigen 2 ENSMUSESTT00000017891 Transcript Derived From ESTs Probasin precursor CEA-related cell adhesion molecule 1 PSP94 uroplakin 1A ENSMUSG00000029855 1700015L13Rik  Probasin precursor RIKEN cDNA 6430501H15 Claudin-8 RIKEN full-length enriched library, clone:A430006D18 RIKEN full-length enriched library, clone:A630095E13 Defensin beta 2 Defensin Experimental autoimmune prostatitis antigen 2 Experimental autoimmune prostatitis antigen 1 PSP94  150  Table A 1.2 Prostate specific tags (2  n  Tag TGCCAGCTGATCAGTCA TGCCGACTGATCAGTCA TTCTAATCAGTATTAGT  search.) (Continued)  Counts in Adult library  Annotation  5 5  ambiguous Probasin precursor  Table A 1.3 RACE-PCR results of selected prostate specific tags. Tag counts Tag AATTGCAAACTAGAATA GACTCCAGGAATGCCTC GCAAGGAAGGCAGGTTG GGGAGAAGTCCTGTGTG TGCCAGCTGATCAGTCA ATAAAATCAACTCACAC GCAAGGCTGGACATCAG  (in adult only) 14 24 6 6 5 5 10  Gene Amplified Not successful Eapal Potential novel gene * Not successful IgG Fc binding protein Not successful Potential novel gene *  Table A 1.3 * These 2 tags were amplified successfully by 5' RACE and 3' RACE but were not successfully amplified using new primers designed from both ends of the transcripts.  Discussion To date there have been few prostate specific genes identified. We successfully detected known prostate specific genes using our comparison strategy, suggesting that this was a valid approach. In the 1st comparison, a very stringent set of criteria were used (i.e. tags commonly expressed in all 3 libraries). This may explain why only 4 genes were found after the comparison and the expression of each of them was very low in the 3 prostate libraries (tag counts less than 5). The second comparison used less stringent criteria and identified a large number of candidates. It is interesting that most of the prostate "specific" tags were found only in the adult library. There are two explanations for this. First, developing prostates (El6.5 male U G S and PO prostate) are not well differentiated and they have not acquired the characteristics of the prostate, thus the prostate "specific" genes are not expressed at these stages. Second, prostate "specific"  151  genes maybe expressed at very low levels during developmental stages so that they are not detectable with the current sequencing depth of these S A G E libraries. In the analysis of prostate specific genes we observed multiple tags matching one gene (Table A1.2). Some of these tags were derived from different positions within the transcript, which are likely to be caused by splicing variants or incomplete enzyme digestion during S A G E library construction. Alternatively, they could be misassigned tags that were generated by sequencing or P C R errors or even from unknown genes. Others were from the same tag but with a single base difference. These tags are likely to be generated from sequencing or PCR errors. However, it is also possible that these tags may be from unknown genes that can not be mapped by CMOST since the databases we used are still incomplete. During the tag mapping process, CMOST can take unmapped tags as sequencing errors and will correct them by single base permutation, insertion and deletion. This function increases tag mapping greatly. However, at the same time, it may also create some tag mapping errors. Thus these tags could be either sequencing errors corrected by the CMOST, or were novel tags that were incorrectly mapped by it. We successfully amplified 2 splice variants of known prostate specific genes using RACE-PCR, indicating that it is a useful technique to validate S A G E tags of interest. However, we were unsuccessful in identifying novel prostate specific genes. The attempt to amplify the 2 tags identified as novel genes was not successful, indicating that either these 2 tags were generated by sequencing error or R A C E - P C R amplified the wrong products due to technical problems such as PCR error. There are several possible reasons for our failure to identify novel prostate specific genes. First, except for known secreted proteins of the prostate, "prostate specific genes" may be expressed highly in the prostate  152  but also be present at very low levels in other tissues. For example, transcripts of probasin and PSP94, two prostate secretory proteins, are also expressed in the spleen library. Since we only looked for tags "specific" to the prostate, we may have excluded some of these genes during our comparisons. The other explanation is that some prostate specific genes may be expressed at very low levels and the current sequencing depth of our SAGE libraries is insufficient to detect them.  References 1.  2. 3.  Bhatia-Gaur, R., A . A . Donjacour, P. J. Sciavolino, M . K i m , N . Desai, P. Young, C. R. Norton, T. Gridley, R. D. Cardiff, G. R. Cunha, C. Abate-Shen, and M . M . Shen. 1999. Roles for Nkx3.1 in prostate development and cancer. Genes Dev 13:966. Shen, M . M . , and C. Abate-Shen. 2003. Roles of the Nkx3.1 homeobox gene in prostate organogenesis and carcinogenesis. Dev Dyn 228:767. So, A . , L . Goldenberg, and M . E. Gleave. 2003. Prostate specific antigen: an updated review. Can J Urol 10:2040.  153  Appendix 2. Transcriptional regulation of branching morphogenesis in prostate development and cancer  Introduction Branching morphogenesis is the formation of a tree-like structure via the ramification of epithelial tubules during embryogenesis. This process is fundamental to the development of a number of organ systems that share similar tissue structure, such as the kidney, lung, breast and prostate (1). In the prostate, postnatal day 0 (PO) is the time when extensive branching starts. There is some evidence to suggest that transcription factors involved in branching morphogenesis may also be involved in cancer since branching morphogenesis involves epithelial "invasion" of the mesenchymal tissue, a process akin to invasion by cancer cells (2). These previous studies led to the hypotheses driving this work: (a) transcription factors regulating the initiation and progression of prostate branching morphogenesis are likely to be more highly expressed in the PO SAGE library (relative to the UGS or adult prostate library), (b) Selected transcription factors involved in prostate branching may play a role in development and/or progression of PCa; (c) candidate genes will be differentially expressed at various stages of tumor development.  Materials and methods Bioinformatic analyses techniques and the procedure for quantitative real time PCR (RT-qPCR) are described in Chapter 2, Materials and Methods. Primers for RT-qPCR are listed in Table A2.1.  154  Table A2.1 Primer sequences for the transcription factors selected for RT-qPCR validation.  Gene  Left primer  Right primer  Product length (base pair)  Ddefl  CACAGATTTGCCCACATCAC  GGGTGCCTAGAGGGAGTGTT  104  HoxDll  CTCGGATGCTCAACCTCACT  CAGACGGTCCCTGTTCAGTT  89  JunB  ATCCCTATCGGGGTCTCAAG  CCTGACCCGAAAAGTAGCTG  84  Ncoa6  GGGCCACAGAGTTTACATCC  CCTTGCATGAAGTTGGGATT  95  Sncaip  CTGAACATTGTCAACGAAGGA  TTGTTTCCGTTCTCATCGTG  89  teashirt3  TGCCGTCGTATCATTCATGT  TTTTGACGTGTGGCTCTCTG  100  Podl  CGCTCACTTAAGGCAGATCC  TCAGGTCATTCTCTGGTTTGC  110  Erm  CGAGTTGTCGTCCTGTAGCC  TAATGGCTTGAACCCAGAGG  112  HoxB5  AGGGGCAGACTCCACAGATA  CCAGGGTCTGGTAGCGAGTA  114  FoxA2  CATCCGACTGGAGCAGCTA  TGTGTTCATGCCATTCATCC  92  Pbxi  ATGCAGCTGAAACAGAGCAC  CTGTGGCTTGCTTGTTGAAA  100  WT1  GGAGAGCCAGCCTACCATC  GAAGGAATGGTTGGGGAACT  115  Liml  CCAAAGAGAACAGCCTCCAC  TGAGACGTTGGCACTTTCAG  113  Er81  GGGAAGTGCTGGGCAATAAT  ATTTCCTCGGGATCTGGACT  110  Etsl  GCCATCAAGCAAGAGGTGTT  GAGGCGGTCACAACTATCGTA  120  Pea3  TACCACCATGGAGAGCAGTG  GAGGCTCTGCTGCTGTTCTG  120  Idl  TGAGGTCCGAGGCAGAGTAT  ATGCGCCTGAAAAGTAAGGA  91  Slug  ACATTGCCTTGTGTCTGCAA  CAGTGAGGGCAAGAGAAAGG  110  Figure A2.1. The strategy for identifying P0 specific genes.  155  Results The strategy for identifying genes of interest in the PO library is shown in Figure A2.1. Comparison of the PO library to the UGS plus the adult prostate libraries identified 9,669 tags as PO specific. GO analysis of these tags indicated that 141 tags were transcription factors (Table A2.2). Among them, most (122/141) tag were expressed at low levels (tag, count =1). No transcription factors known to be involved in prostate development were identified in this group, indicating that our sequencing depth was not sufficient to identify these genes which are likely to be expressed at very low levels. Multiple tags for some genes were again observed (as discussed in Appendix 1). The 6 most highly expressed transcription factors were chosen for RT-qPCR analysis. They were: HoxDll,  JunB, Development and differentiation enhancing (Ddefl), nuclear  receptor coactivator 6 (Ncoa6), Synuclein alpha interacting protein (Synphilin) and Teashirt3 (Tsh3). None of the above transcription factors has previously been implicated in branching morphogenesis except HoxDll,  which was shown to be critical for  branching of the developing kidney (3). Because of the inconsistent levels of Gapdh in our developing prostate SAGE libraries, Ubiquitin C (Ubc) was used as a housekeeping control for RT-qPCR experiments (Figure A2.2A). Two additional genes, HoxAlO and GUI, were used as controls. HoxAlO was chosen because its expression levels in the 3 SAGE libraries remained constant, while the expression pattern of GUI during prostate development has been reported previously (4). RT-qPCR results showed that the expression pattern of HoxAlO correlates well with the SAGE data. GUI, although not detected by SAGE, was expressed at higher levels during development and decreased in the adult prostate, displaying the expression pattern previously reported (Figure A2.2A).  156  RT-qPCR analysis of the 5 transcription factors selected as "PO specific" showed that all of them were more widely expressed (FigureA2.2B). The Teashirt3 primers did not yield a product and thus analysis of this gene was excluded from the experiment. Among the 5 transcription factors, Synphilin, Ddefl and HoxDll  were most highly expressed in the  El6.5 male UGS and were downregulated with development. Surprisingly, Noca6 and JunB were expressed at their lowest level in the PO sample (FigureA2.2B).  Table A2.2. 141 Transcription factor tags specifically expressed in the PO prostate library. Sequence  PO counts  CMOST Mappings - CMOST Methodology - Best Mouse Mappings  CATCTAGACGCTTACCC  6  development and differentiation enhancing  GGGCAGCGACTTTCTAA  5  Synuclien, alpha interacting protein (synphilin)  TATTGAAAGGCGCTAAT  5  nuclear receptor coactivator 6  AGTCACTGTCTTTTTGT  4  Hox-Dll  ATGTGACTCTTCATATA  4  teashirt 3  GCCCCCTTCCAGCGTAT  4  Jun-B oncogene  ACTGCTTTTTAAATTTT  3  Unknown (protein for MGC:70076)  TTTTGTAAAACTTGTGT  3  AW538212 protein  AAAAGTGGCAAGTCTCC  2  ISL1 transcription factor, LIM/homeodomain(islet 1)  AAACCTTCTGTAATCGC  2  nuclear factor I/A  AAATATTACAAGATACA  2  Hsf2 protein  AGAAATTTAGTAAAATA  2  nuclear receptor subfamily 2, group F, member 1  AGCTCATAAAAGTTTTT  2  interferon regulatory factor 2  GACACGCAAGAACGCAT  2  jun D proto-oncogene  GACAGTAATGTCGGCAC  2  centaurin, beta 5  TACCATCTATCCTAAGG  2  Msx-2  TTAACCCTGACTTTATG  2  nuclear receptor subfamily 1, group D, member 2  TTATCAGCCAGACAGGA  2  Pbxl protein  TTTAATGTGACTTCCCC  2  homeo box B5  AAAAGTGGCAAGTCTCT  1  ISL1 transcription factor, LIM/homeodomain(islet 1)  AAATCAAAGGTGATTCT  1  feminization 1 homolog b nuclear factor of activated T-cells, cytoplasmic, calcineurin-dependent 1  AACCATTCTTAGTAGAC  1  AACTCAAATATACCAAA  1  myotrophin  AAGAATCTTTTAATGTC  1  nuclear respiratory factor 1  AATTTCAGAAAGCCTAC  1  homeobox protein A10  ACACTTAATGTTAATAT  1  myotrophin lung carcinoma myc related oncogene 1  ACCAACGGTCCTTGCCC  1  ACCCCGGTGCCCAAGTG  1  twist homolog 2  ACTCAATAAACCATTAC  1  paired box gene 8  ACTGCTCTTTAAAATTT  1  Unknown (protein for MGC70076)  AGAAATTTAATAAAAAA  1  nuclear receptor subfamily 2, group F, member 1  157  Table A2.2. 141 Transcription factor tags specifically expressed in the PO prostate library (Continued). Sequence  PO counts  CMOST Mappings - CMOST Methodology - Best Mouse Mappings  AGCCAGATTTTCAGGTT  1  transcription elongation regulator 1 (CA150)  AGCTTAGGGTGGCTTCT  1  Sox 13 protein  AGGACCAGGAAGTGCCC  1  AGGGAGTCAGGCTGCTG  ,  forkhead box M 1 nuclear factor of kappa light polypeptide gene enhancer in B-cells 2, 49/pl00  AGGTGCACACGTTCGTG  1  hairy and enhancer of split 6  ATGATCATACCGGTAGC  1  glutaryl-Coenzyme A dehydrogenase  ATGCACAGAAACCAGTG  1  RIKEN cDNA 1110054N06  CAACACACAAGCACTAT  1  androgen receptor  CACAGATAGGCTTGTCT  1  Testis specific high mobility group protein  CACAGGGCTTGTCATCT  1  Ankrd25 protein  CAGCATCAGTCTCCTGA  1  general transcription factor II H, polypeptidel  CAGCCTCCTACCCCCCA  1  upstream binding transcription factor, RNA polymerase I  CAGTACCACTGACTTGG  1  forkhead box M1  CCAAAACAGTATAAAAT  1  histone deacetylase 2  CCAACATCAGTTAAAGC  1  AW538212 protein  CCACACACACTGGAATG  1  F830020C16Rik protein  CCACACCCAGGTCTCGG  1  ankyrin repeat and SOCS box-containing protein6  CCATCTTGACTGTAGGG  1  thyroid hormone receptor alpha  CCCACCCGGAAGCACTT  1  Nr2f6 protein  CCCCCCACACAGTCCCC  1  transforming growth factor beta 1 induced transcript 4, isoform 1  CCCTAAATCAAGCAAAA  1  WD40 protein Ciaol  CCTAAAGCAGCTGGTAC  1  centaurin, beta 5  CCTGACCAATGGGCACA  1  longevity assurance homolog 4  CCTTAAGCCAATTTATA  1  E26 avian leukemia oncogene 2, 3' domain  CGCAGATGTGAGGGACA  1  early growth response 1  CGCGGACTAGCTGCGGC  1  jun D proto-oncogene insulin related protein 2 (islet 2)  P  CGCTTTGGGGGTTATGT  1  CGGGCCCGCTCACGAGC  1  transcription factor 21  CTACAGCTGCAACACGA  1  T-cell leukemia, homeobox 3  CTCCCTTTTCGAAAAGG  1  Tcfe2a protein  CTCTCTCTCTTCTGAAG  1  Pbx/Knotted 1 homeobox 2  CTGAGTGCCAGTGGCAC  1  ankyrin repeat and F Y V E domain containing 1  CTGCTGTGCTTCCTCTG  1  ankyrin repeat and F Y V E domain containing 1 BCL-6 interacting corepressor isoform a  CTTTGAAATGTATGTTC  1  CTTTGTTCCCAGAATAA  1  Znrdl protein  GAAAGCAAACAAATGGA  1  retinoid X receptor gamma  GAAGATGAGGATGACAC  .1  MAX-interacting protein  GACCAATCTTTTTTTTG  1  transcription elongation factor A (SII) 1  GACCATCAGCAAAGCCG  1  Nrfl protein  GACTGTATTTATTTTCA  1  histone deacetylase 2  GAGAATGGGCTTAACCA  1  E2F transcription factor 2  GAGCGTAGCTGGGAAAG  1  transcription factor 3  GAGTCCTCCACCTTCGC  1  ativating transcription factor 5  GATAATCTAGATATATG  1  ELK3, member of ETS oncogene family, isoform b  GATGAATGACAGCTGTA  1  trans-acting transcription factor 1  158  Table A2.2. 141 Transcription factor tags specifically expressed in the PO prostate library (Continued). Sequence  PO counts  CMOST Mappings - CMOST Methodology - Best Mouse Mappings  GATTCCAATGTTCTTCT  1  Transcription factor 21  GCAACAAACTCTAAGTT  1  F-box and WD-40 domain protein 7, archipelago homolog  GCACATTTTGAATCTGC  1  androgen receptor  GCACCTGTGGGACACCT  1  GATA binding protein 2  GCAGACAACTATGTGGT  1  TATA box binding protein  GCAGGAATGCCGTCCCC  1  SRY (sex determining region Y)-box 5  GCAGGCTTCACAGGGAC  1  Mrg2 protein  GCATTGTTTATTTGAAG  1  Zfp95 protein  GCCAGGGCCATAGCTTT  1  MYC-associated zinc finger protein  GCTATCGCCCTGAACAG  ,  MAD homolog 3  GGACTACTGGAGAGGTA  1  sine oculis-related homeobox 5 homolog  GGCTACATTACACAATA  1  activating transcription factor 1  GGCTGGGGTTGACGCCA  1  metastais-associated gene family, member 2;  GGCTGGGTCGGAAAGGC  1  Nr2f6 protein  GGGACTCTCCTCAGTTA  1  Sox 10 protein  GGGTGGCTGTCCCCAGG  1  serum response factor  GGTCCCCTTAGATATTC  • i  RIKEN cDNA A630035I11 gene  GGTCTGCTGGCCAAGGA  1  heart and neural crest derivatives expressed transcript 2  GGTTTCATCAGCAAGGA  1  signal transducer and activator of transcription3  GTAGCCGGAGGGGACCA  1  Znrdl protein  GTCACATCTCCTGTACA  1  serum response factor  GTCATTTTCTAGTCTTT  1  zinc finger homeobox la  GTCCTGGCCTTAACTTT  1  Ets2 repressor factor  GTGAGAAGGCCTCCGCA  1  Iroquois related homeobox 1  GTGCTGGAAATGGTTGG  1  nuclear factor I/A  GTGTGAAAGGCTGAGTA  1  Thra protein  GTGTGCTGCTCAGACTC  1  POU domain, class 6, transcription factor 1  GTTTGTTTGGTTTTTTA  1  RIKEN cDNA 9430065N20 gene  TAAAAATCCAAAATACA  1  heat shock factor 1  TAAACCTCGAGGTGGAA  1  Jun-B oncogene  TAACAAAGACTACGGTG  1  Nfib protein  TAACTGCAATTGGGCTG  1  transcription factor A, mitochondrial  TACACAATAATTTTTTT  1  GATA binding protein 6  TACCTCACCCGGGACCG  1  homeobox protein A9  TACGGCTCGCCCGGGGA  1  homeobox C9  TACTAAGGCCGCAGCCC  1  peroxisome proliferator activator receptor delta  TACTACTTTGACTTTTC  1  sterol regulatory element binding factor 1  TAGCAATTGCACCGTGC  1  FBJ osteosarcoma oncogene B  TATATAGCATTACTTCT  1  Hhex protein  TATCAGTTTTCCCCTAC  1  mesenchyme homeobox 1  ' TATTCATAGTGTTTTAG  1  MAD homolog 3  TCAGAGAGAAGACTGCT  1  RIKEN cDNA 1110054N06  TCATTAGAAGTTACCTC  1  UDP-glucose pyrophosphorylase 2  TCATTTGACCTTTTTTT  1  ISL1 transcription factor, LIM/homeodomain(islet 1)  TCCAGCTCAGAAGAGGA  1  sine oculis homeobox homolog 1  TCCTTTGACTTTTTTTT  1  ISL1 transcription factor, LIM/homeodomain(islet 1)  159  Table A2.2. 141 Transcription factor tags specifically expressed in the PO prostate library (Continued). CMOST Mappings - CMOST Methodology - Best Mouse Mappings  Sequence  PO counts  TCCTTTTTTATAGCCCT  1  forkhead box 11  TCGAAAGACCTCAGGGT  1  FBJ osteosarcoma oncogene  TCGAGAGACCTCAGGGT  1  FBJ osteosarcoma oncogene  TCTAAATAGAATACTTT  1  Ankra2 protein  TCTCCCAGGTAGCTGCT  1  F830020C16Rik protein  TGACAACTTCAAAATGC  1  myotrophin  TGCGGGAGTCGTGGTCA  1  general transcription factor II H, polypeptide 1  TGGAGCCAGTGCAGAGG  1  transcription factor-like 1  TGGCCAAGTGGCAGAGT  1  Sox 18 protein  TGGGCGGCAGCTGGGGG  1  ets variant gene 4 (E1A enhancer binding protein, E1AF)  TGGTGGCTCACAACCAC  1  RIKEN cDNA 1810037003  TGTACACTGAAAAATAA  1  BCL-6 interacting corepressor isoform a  TGTGCGCATTGGGGTGG  1  transcription factor 1  TTGTATATGAAGGAGAA  1  transcription factor 3  TTGTATATGAAGGAGAT  1  transcription factor 3  TTGTTAGGAAATACCGG  1  Satbl protein  TTTAGAGAGGTGGAGGG  1  transcription factor 21  TTTTGATAAAGAATGAA  1  GA repeat binding protein, beta 1  TTTTTGGACAAAAACTT  1  catenin beta  Since PO is the time point when branching morphogenesis of the prostate commences, it is very likely that transcription factors known to be involved in branching will also be highly expressed in our PO prostate library. We selected 12 transcription factors previously shown to be involved in branching morphogenesis of kidney, lung and mammary gland for RT-qPCR analysis (TableA2.3) (5-13). RT-qPCR analysis revealed two expression patterns for these genes. Liml, Idl, Pea3, Etsl, HoxB5 and Wtl were highly expressed in the El6.5 male UGS and progressively downregulated with development (Figure A2.3A). Slug, Pbxl, Er81, Podl, Erm and FoxA2 were expressed highly in both El6.5 male UGS and PO prostate and downregulated in the adult (FigureA2.3J3). The expression patterns of these known branching genes suggest that they may also play important roles in development and branching morphogenesis of the prostate.  160  A  CD D> C CO .c o  PO  > o m  Adult  ;a  03 CO  a  DC  O Q.  Gli  o  lip  HoxalO  I  PO  m  •12 •8 4  o o c 3  U G S PO  Adult  Figure A2.2 RT-qPCR results for selected genes. (A) Housekeeping gene and controls. Ubc was used as the housekeeping gene and GUI and HoxAW as experimental controls (B) PO specific genes. Note: Bars represent RT-qPCR results; lines represent SAGE tag counts. Each bar represents the mean ± SD of 3 technical replicates using a single sample consisting of tissues pooled from at least 2 mice. UGS, El6.5 male urogenital sinus; PO, postnatal day 0 prostate; Adult, 12 week prostate dorsal-lateral lobe. TableA2.3 Transcription factors implicated in branching were selected from the literature for RT-qPCR validation. Annotation transcription factor 21 Erm Hoxb5 Forkhead box A2 pre B-cell leukemia transcription factor 1 Wilms tumor homolog LIM homeobox protein 1 Er81 E26 avian leukemia oncogene 1 Pea3 Idl slug  UGS* 38 1 0 1 2 2 1 2 0 0 4 3  Tag counts Adult PO 24 1 2 0 2 0 0 0 2 1 5 0 0 0 2 0 0 0 1 0 14 1 3 0  function lung and kidney branching lung branching lung branching lung branching kidney branching ureteric bud outgrowth ureteric bud outgrowth Expressed in branching organs expressed in branching organs mammary gland branching blood vessel branching trachea branching  Table A2.3 Tag counts of the selected transcription factors in the 3 prostate SAGE libraries is shown. * UGS, El6.5 urogenital sinus; PO, postnatal day 0 prostate; Adult, 12 week adult prostate dorsal-lateral lobe.  161  A  Pea3  Liml  0) O)  c 03 .c o  UGS PO  UGS PO  03 >  UGS PO  m  Ets1  ig or o  HoxB5  ST ca  Wt1  o o c 3  0.5  CL i  o  UGS  PO  UGS  Adult  PO Adult  Er81  I UGS  > O m  PO  CO  o o c  Erm Hill— UGS  i—r-f -  PO  t  i |  Adult  Figure A2.3 RT-qPCR results for selected transcription factors during prostate development. (A) Transcription factors highly expressed at El6.5 and downregulated with development; (B) Transcription factors expressed highly in both El6.5 male UGS and PO prostate but downregulated in the adult prostate. Note: Bars represent RT-qPCR results; lines represent SAGE tag counts. Each bar represents the mean ± SD of 3 technical replicates using a single sample consisting of tissues pooled from at least 2 mice. UGS, El6.5 male urogenital sinus; PO, postnatal day 0 prostate; Adult, 12 week prostate dorsal-lateral lobe.  Branching morphogenesis is also believed to play an important role in PCa (14). Many transcription factors involved in branching morphogenesis have been shown to play roles in cancer development, e.g. Pea3 in breast cancer (15). To determine the expression pattern of selected branching genes in PCa, we chose 4 time points representing various stages of PCa progression in the T R A M P mice (refer to Chapter 1 Introduction for a description of the model). Six candidate transcription factors showing relatively more  162  significant changes (-10 fold) in expression levels in the PO prostate compared to the adult prostate were selected for the T R A M P study including GUI, Slug, Er81, Pea3, Podl and Sncaip (Synphilin). Among them, Er81, Slug, GUI and Pea3 have previously been shown to be upregulated in several cancer tissues including breast cancer and PCa (15-17). Podl and Synphilin have no known involvement in cancer development. RTqPCR analysis showed that all the 6 genes were upregulated at 4 weeks of age in T R A M P mice compared to their wild type controls (Figure A2.4).  12W  12W  26W  26W  Synphilin 12  0.6 0.4 •  Q2 12W  12W  26W  26W  2W  4W  12W  26W  WT (C57 BL/6J) Tg (TRAMP) Figure A2.4 RT-qPCR results of candidate gene expression patterns at different time points during PCa progression in TRAMP mice. In all cases, the RT-qPCR data represents the mean ± SD of 3 technical replicates. The replicates consisted of tissues pooled from 2 mice. * P<0.05  Discussion We selected "PO specific" transcription factors from our S A G E libraries for RTqPCR validation. However, the expression levels revealed by RT-qPCR did not agree with the SAGE data, nor were any of the transcription factors specifically expressed in the PO prostate. Since these transcription factors were expressed at low levels in the SAGE libraries (<10 tag counts), the quantification by SAGE may not be accurate. Three  163  of them, Ddefl, HoxDll  and Synphilin, were highly expressed and downregulated with  development, indicating that they are likely to be involved in initiation of budding in prostate development. Detection of known branching genes in the PO prostate library suggests that there may be a common genetic program that regulates branching morphogenesis in various tissues, although the specific growth factors involved are likely to differ from one tissue to another. Since many of the known branching genes are highly expressed in both the El6.5 and/or PO libraries, this suggests that the processes of prostate budding and branching morphogenesis may share similar regulatory mechanisms. Liml was only highly expressed at E16.5, suggesting that it may play an important role in the initiation of prostate development. The results suggest that selected genes are upregulated in T R A M P mice at an early stage of cancer development. This observation requires additional validation to determine if the increase reflects an altered tissue composition or if they are specifically increased in the cell types of interest. This result may also provide further evidence that the processes of development and cancer share some common patterns of gene expression. It is interesting that these genes went up at 4 weeks of age and dropped to a normal level afterwards. This may indicate that these genes were involved in initiation of the branching process in PCa. It will also be important to determine i f the genes of interest are differentially expressed in other models of PCa that more closely mimic human disease.  164  References 1. 2.  3.  4.  5.  6.  7.  8.  9.  10.  11.  12.  13.  Shah, M . M . , R. V . Sampogna, H . Sakurai, K . T. Bush, and S. K . Nigam. 2004. Branching morphogenesis and kidney disease. Development 131:1449. Fata, J. E., Z. Werb, and M . J. Bissell. 2004. Regulation of mammary gland branching morphogenesis by the extracellular matrix and its remodeling enzymes. Breast Cancer Res 6:1. Patterson, L . T., M . Pembaur, and S. S. Potter. 2001. H o x a l l and H o x d l l regulate branching morphogenesis of the ureteric bud in the developing kidney. Development 128:2153. Lamm, M . L . , W. S. Catbagan, R. J. Laciak, D. H . Barnett, C. M . Hebner, W. Gaffield, D. Walterhouse, P. Iannaccone, and W. Bushman. 2002. Sonic hedgehog activates mesenchymal G l i l expression during prostate ductal bud formation. Dev Biol 249:349. Quaggin, S. E., L . Schwartz, S. Cui, P. Igarashi, J. Deimling, M . Post, and J. Rossant. 1999. The basic-helix-loop-helix protein podl is critically important for kidney and lung organogenesis. Development 126:5771. Samakovlis, C , G. Manning, P. Steneberg, N . Hacohen, R. Cantera, and M . A . Krasnow. 1996. Genetic control of epithelial tube fusion during Drosophila tracheal development. Development 122:3531. Chotteau-Lelievre, A . , R. Montesano, J. Soriano, P. Soulie, X . Desbiens, and Y . de Launoit. 2003. PEA3 transcription factors are expressed in tissues undergoing branching morphogenesis and promote formation of duct-like structures by mammary epithelial cells in vitro. Dev Biol 259:241. Volpe, M . V . , L . Pham, M . Lessin, S. J. Ralston, I. Bhan, E. Cutz, and H . C. Nielsen. 2003. Expression of Hoxb-5 during human lung development and in congenital lung malformations. Birth Defects Res A Clin M o l Teratol 67:550. Zhou, L . , C. R. Dey, S. E. Wert, C. Yan, R. H . Costa, and J. A . Whitsett. 1997. Hepatocyte nuclear factor-3beta limits cellular diversity in the developing respiratory epithelium and alters lung morphogenesis in vivo. Dev Dyn 210:305. Schnabel, C. A . , R. E. Godin, and M . L . Cleary. 2003. Pbxl regulates nephrogenesis and ureteric branching in the developing kidney. Dev Biol 254:262. Lee, S. B . , K . Huang, R. Palmer, V . B . Truong, D. Herzlinger, K . A . Kolquist, J. Wong, C. Paulding, S. K . Yoon, W. Gerald, J. D. Oliner, and D. A . Haber. 1999. The Wilms tumor suppressor WT1 encodes a transcriptional activator of amphiregulin. Cell 98:663. Kola, I., S. Brookes, A . R. Green, R. Garber, M . Tymms, T. S. Papas, and A . Seth. 1993. The Etsl transcription factor is widely expressed during murine embryo development and is associated with mesodermal cells involved in morphogenetic processes such as organ formation. Proc Natl Acad Sci U S A 90:7588. Lyden, D., A. Z. Young, D. Zagzag, W. Yan, W. Gerald, R. O'Reilly, B . L . Bader, R. O. Hynes, Y . Zhuang, K . Manova, and R. Benezra. 1999. Idl and Id3 are required for neurogenesis, angiogenesis and vascularization of tumour xenografts. Nature 401:670.  165  14.  15.  16. 17.  Cunha, G. R., A . A . Donjacour, P. S. Cooke, S. Mee, R. M . Bigsby, S. J. Higgins, and Y . Sugimura. 1987. The endocrinology and developmental biology of the prostate. Endocr Rev 8:338. Kurpios, N . A., N . A . Sabolic, T. G. Shepherd, G. M . Fidalgo, and J. A . Hassell. 2003. Function of PEA3 Ets transcription factors in mammary gland development and oncogenesis. J Mammary Gland Biol Neoplasia 8:177. Hajra, K . M . , D. Y . Chen, and E. R. Fearon. 2002. The S L U G zinc-finger protein represses E-cadherin in breast cancer. Cancer Res 62:1613. Fan, L . , C. V . Pepicelli, C. C. Dibble, W. Catbagan, J. L . Zarycki, R. Laciak, J. Gipp, A . Shaw, M . L. Lamm, A . Munoz, R. Lipinski, J. B. Thrasher, and W. Bushman. 2004. Hedgehog signaling promotes prostate xenograft tumor growth. Endocrinology 145:3961.  166  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0092285/manifest

Comment

Related Items